AU703874B2 - Method and system for obtaining information from a plurality of remote stations - Google Patents

Description

-r la METHOD AND SYSTEM FOR OBTAINING INFORMATION FROM A PLURALITY OF REMOTE STAT2IONS 9 11 12 13 14 16 17 i 19 21 22 23 24 26 27 28 29 S* 30 31 32 33 34 36 37 A N3 8 FIELD OF THE INVENTION This invention relates generally to a method and system for obtaining information from a plurality of remote stations.

BACKGROUND OF THE INVENTION It is a common requirement to target and possibly identify quickly one or several out of a plurality of participants according to specific selection criteria.

It is frequently required to select one or more participants according to a "priority" or "characteristic value" based on specified selection criteria for the purpose of allocating a particular task to the participant or participants having the highest priority.

In dispatching systems, for example, for dispatching a taxi or messenger to a customer at a specified location, it is desirable that a suitable (and preferably the most suitable) taxi or messenger be sent to a particular customer. Generally the nearest, unoccupied taxi which has sufficient accommodation should be dispatched to the customer. Furthermore, it is desirable that the allocation be accomplished in the minimum possible time.

Typical existing dispatching systems include a central dispatch station having a transmitter and receiver or a transceiver in each of the participating vehicles for communicating with the central dispatch station. Typically, a voice request is transmitted by a dispatcher to each of the participating vehicles, and the dispatcher decides which of the vehicles is most suited to the task in hand based on the replies from the vehicles.

Such a system would be capable of simple SUBSTITUTE SHEET (RULE 26) I- I M 0e WO 95/27963 PCT/EP95/01330 2 1 implementation if the selection criteria related to static 2 variables only. Thus, if the only selection criterion were 3 a taxi's current distance from the customer and each taxi 4 were stationary, it would merely be necessary to extract the taxis' locations once, after which it would be simple to 6 determine which taxi were nearest to the customer's 7 location. However, in practice, the selection criteria 8 relate to dynamic variables which, by definition, are 9 changing constantly and therefore it is necessary continuously to update each taxi's distance from the 11 customer's location (and/or other information required to 12 choose a taxi for the given task) or at least to do so each 13 time a taxi is to be dispatched.

14 In some typical prior art systems, this is done by providing the dispatcher with a periodically updated map 16 t.hat shows the respective location of each of the taxis.

17 This updating is accomplished by the periodic transmission 18 of a location message by each of the taxis via a 19 communication channel. In order to ensure that the transmitted data can be received quickly and without 21 corruption, the total spectrum width of the communication 22 system must be very large.

23 In a system described in EP 0389488 job requests are 24 dispatched by a controller to mobile vehicles which messages include information about the location of a job. Each 26 vehicle has a receiver, transmitter and circuitry to compare 27 the requirements of the job with the status of the vehicle.

28 If the results of the comparison is that the vehicle is 29 suitable fo- the job, then it transmits a message back to the controller volunteering itself for the job.

31 It should also be noted that, even in the specific 32 case of a taxi or messenger service, distance from the 33 customer location is by no means the only criterion 34 according to which a task may be allocated. Thus, it may well be that the nearest messenger or taxi is already 36 occupied and is therefore not available for performing the 37 task. Alternatively, the nearest available taxi may not 38 have sufficient room for carrying all the passengers to whom SUBSTITUTE SHEET (RULE 26)

LI

WO 95/27963 PCT/EP95/01330 3 1 a taxi must be sent; or perhaps a particularly bulky load 2 must be carried and the nearest, available taxi or messenger 3 is inadequate for the task.

4 Yet a further consideration is that it is often preferred to dispatch to a customer an idle taxi waiting at 6 the taxi rank rather than go through the process of 7 transmitting a voice message and awaiting responses from 8 taxis in the field prior to allocating the task to one of 9 them. In the event that several idle taxis are waiting at the taxi rank, or where several taxis are reasonably close 11 to the customer, it is often preferable that the taxi which 12 has been idle for the longest period of time be selected.

13 Furthermore, it may not always be desirable to 14 dispatch the nearest available taxi to a particular customer location if other customers, albeit further away, have made 16 prior requests which have not yet been serviced.

17 Even apart from some of the basic limitations of prior 18 art systems described above, it is often desirable to target 19 and possibly to identify participants according to several selection criteria. This is somewhat analogous to performing 21 a database search by means of key words which can be 22 combined according to the rules of Boolean or other logic 23 systems. However, database records are generally static and 24 are stored at a single location. In contrast to this, the attributes of the participants that are the subject of the 26 present invention are dynamic and constantly changing, and 27 cannot be characterized by static data which can be stored 28 at a single site. Thus, if the dynamic data characterizing 29 such participants are to be searched at a single site, then the data must first be downloaded to the site where the 31 search is to be performed. During the time that such data 32 are downloaded, they may well change, thereby compromising 33 the accuracy of the search which is subsequently performed.

34 Another application which requires the receiving and processing of information from a large number of sources is 36 IVHS. In this application, for example, information on 37 position and speed from a large number of vehicles is 38 processed in order to obtain information on road delays.

SUBSTITUTE SHEET (RULE 26) 4 Again, the sending of large amounts of information required substantial bandwidths, even though the vehicles themselves need not be identified.

Another previously unsolved problem is the tracking or mapping of the position of large numbers of vehicles. Prior solutions to this problem required the broadcasting by each vehicle of an information bearing signal including at least its position. When large numbers of vehicles are to be tracked, the amount of information to be transmitted (and the communication overhead associated with the transmission) is very large and the available time/bandwidth necessary is either unavailable or if it is available, such broadband systems are expensive. The alternative of trading bandwidth for time, results in a 15 system which is too slow for many uses.

SUMMARY OF THE INVENTION Accordingly, the invention provides a method of tracking changes at a plurality of remote stations each 20 having a varying attribute affecting a characteristic value computed according to a predetermined procedure, involving: assigning at least one transmission slot to each of the remote stations, wherein the presence or absence of energy in a particular slot defines a single bit 25 of information; determining, by the respective stations, of their characteristic values relative to a previously determined characteristic value; and transmitting, by the respective stations, of the difference between their determined characteristic values and a previously determined characteristic value that was transmitted in at least one previous transmission slot, in said at least one assigned transmission slot.

The invention also provides a method of mapping a plurality of remote stations each having a varying attribute affecting a characteristic value computed in accordance with a predetermined procedure, involving: H-\Cgo.ty\Keep\Nick\22162,95.doc 22/01/99 ii 5 assigning at least one transmission slot to each of the remote stations, wherein the presence or absence of energy in a particular slot defines a single bit of information; determining at the respective stations, of their characteristic values; initially transmitting by the respective stations, of a signal responsive to their determined characteristic values in said at least one assigned slot, said transmitted characteristic value having a first characteristic value resolution; and si sequently transmitting, by the stations, of a signal related to their respective characteristic values in said at least one assigned transmission slot, said subsequent transmission having a finer characteristic value resolution relative to said previously transmitted characteristic value.

As used herein the term "priority" or "characteristic value" means, in addition to its normal

S*

•o 9. 9 e a o e It, 11 J-M, ha ,Cy-ty Kwt Or.,'Jl I -K 211- I q q, L Is 6 meaning, a characterization according to a protocol which takes into consideration one or more elements associated with a person or object being characterized.

BRIEF DESCRIPTION OF THE DRAWINGS In order to better understand the invention and to see how the same may be carried out in practice, nonlimiting preferred embodiments of the invention will now be described with reference to the accompanying drawings, in which: Fig. 1 shows schematically the principal components of a preferred system for carrying out a dispatching function in accordance with one preferred embodiment of the invention; Fig. 2 is a flow diagram showing the principal steps associated with a preferred method of carrying out a dispatching function in accordance with the preferred e* embodiment of the invention; Fig. 3 shows schematically how vehicles are 20 targeted during an initial phase of targeting based on distance from the customer location in accordance with a preferred embodiment of the invention for carrying out a S"dispatching function; ~Fig. 4 shows schematically how vehicles are 25 targeted during a second iteration of targeting based on distance; Figs. 5A and 5B are two portions of a state diagram showing various options associated with a first targeting phase according to a preferred embodiment of the invention for carrying out a dispatching function; Figs. 6A and 6B are two portions of a state diagram showing various options associated with a second identification phase according to a preferred embodiment of the invention for carrying out a dispatching function; Figs. 7A and 7B show timing diagrams relating to the H:\Cgowty\reep\ Nck\\21#2.95.doc 22/0t199

I

-7 THIS PAGE IS INTENTIONALLY BLANK H:\Cgo.ty\Keep\N2ck\22162.95.doc 22/01199 -8 THIS PAGE IS INTENTIONALLY BLANK H:\Cgo-.ty\Keep\t4.ck\22162.95.doc 22/01/99 9- THIS PAGE IS INTENTIONALjLY BLANK H:\Cgo-ty\Keep',ltck\22162.95.doC 22/01/99 10 THIS PAGE IS INTENTIONALLY

BLANK

H;\Cgowty\Keep~tNick\2216 2 '.95,doc 22/01/99 11 THIS PAGE IS INTENTION~ALLY BLANK

I

.7' H:\CgCwty\KeepqtJc 2 2 6 2 .95 .doc 2210119') 12 a.

THIS PAGE IS INTENTIONALLY BLANK

I

H: gowyN~ep~li- ,.2lh.95doc22/01/99 13 THIS PAGE IS INTENTIONALLY BLANK H:\Cgowty\Ke\N1ck\22162.95.doc 22/01/99 ~s I WO 95/27963 PCTEP95/01330 14 1 targeting and identification phases r ectively of a 2 priority discrimination according to a preferred embodiment 3 of the invention for carrying out a dispatching function; 4 Fig. 8 is a block diagram showing the principal components in a control center according to a preferred 6 embodiment of the invention for carrying out a dispatching 7 function; 8 Fig. 9 is a block diagram showing the principal 9 components of a control unit in respect of each of the remote units in accordance with a preferred embodiment of 11 the invention for carrying out a dispatching function; 12 Fig. 10 shows an initial map generated in an IVHS 13 system in accordance with a preferred embodiment of the 14 invention; Fig. 11 shows a second, more detailed map, generated 16 during a second iteration in an IVHS application in 17 accordance with a preferred embodiment of the invention; 18 Fig. 12 shows a graph of additional information which 19 is generated in an IVHS application in accordance with a preferred embodiment of the inveiition; 21 Fig. 13 shows a graph of further additional 22 information which is generated in an IVHS application in 23 accordance with a preferred embodiment of the invention; 24 Fig. 14 is a general block diagram of a transmitter for an IVHS system in accordance with a preferred embodiment 26 of the invention; 27 Fig. 15 is a block diagram of a receiver for useful 28 for both IVHS and dispatching systems in accordance with a 29 preferred embodiment of the invention; and Figs 16A-16C shows a scheme for slot distribution for 31 tracking of individual vehicles.

32 33 34 36 37 38 SUBSTITUTE SHEET (RULE 26) I WO 95/27963 PCT/EP95/01330 15 1 DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT 2 Fig. 1 shows a typical scenario of a dispatch 3 situation in connection with which the invention may be 4 employed. In this scenario a geographical area 10 is defined by a boundary 11 within which a system according to 6 the invention is operational.

7 An identification system according to the invention 8 includes a controller such as a control center (base 9 station) 12 and, optionally, a plurality of taxi stands 13, 14 and 15 which constitute sub-control units each of which 11 serves a respective region within area 10 and which may 12 forward a customer request to control center 12.

13 Associated with each of stands 13, 14 and 15 are 14 respective groups of participants taxis) of which two groups 18 and 19 associated with stands 13 and 16 respectively, are shown in Fig. 1. The groups of 17 participants 18 and 19 generally comprise some taxis which 18 are stationary proximate their respective stands awaiting 19 instructions therefrom and other taxis such as 20, 21, 22, 23 and 24 which are circulating within area 10 and are 21 either available for performing a task on receiving 22 instructions or, alternatively, are occupied and therefore 23 unavailable.

24 A customer 25 located somewhere within area 10 relays a request for service to control center 12 telephonically 26 via a Public Switched Telephone Network (PSTN). Control 27 center 12, in turn, broadcasts an invitation message to all 28 of the participants in area 10 either directly or via a 29 remote station 30 which is typically located so as to cover all of area 10. Control center 12 can also receive messages 31 via station 30. Alternatively, control center 12 broadcasts 32 and receives messages directly.

33 Remote station 30 may be located inside area 10, or 34 alternatively if the area is built up with tall buildings, outside the boundary of the built-up area, if this siting 36 reduces the blocking of signals between the taxis and the 37 repeater station by tall building and the like or for other 38 reasons.

SUBSTITUTE SHEET (RULE 26) "1 WO 95/27963 PCT/EP95/01330 16 1 Sometimes, customer 25 telephones a particular taxi 2 stand since this is the nearest stand to the customer's 3 location. In this case it is generally preferable that one 4 of the taxis associated with the taxi stand be dispatched to the customer unless, of course, all of the taxis associated 6 with the stand are currently occu 'ed (or otherwise 7 unsuitable), in which case one of the taxis associated with 8 another of the stands will be allocated for the task.

9 In this case, the association of a taxi with a particular stand may constitute at least one of the criteria 11 involved in choosing the taxi to be allocated to customer 12 25. Such a selection criterion is a static variable and, 13 once fixed, never varies because a taxi is always associated 14 with one stand. However, the actual priority assigned to each of the taxis is also a function of several independent 16 dynamic factors which are subject to constant fluctuation.

17 Of these, the taxi's distance from the customer is the most 18 important example. However, other dynamic conditions 19 pertaining to a taxi's instantaneous status also affect the respective priority of the taxi so that, for example, a taxi 21 which is currently occupied or one which has insufficient 22 occupancy for the number of passengers to be collected would 23 not participate In the selection process and a taxi which is 24 waiting at a stand would get priority. The idle time of the taxi can also be an important criteria.

26 It will be appreciated that in general there are many 27 different contributory factors, or selection criteria, which 28 influence the priority assigned to each individual taxi 29 within area 10. Moreover, it is generally the case that each selection criterion has a different "weight" associated 31 therewith so that the final magnitude of the priority 32 associated with each respective taxi is built up from many 33 different selection criteria each of which exert a different 34 influence on the actual priority assigned.

For example, in the simplest case, it may be that only 36 distance from the taxi to customer 25 is of concern. Such a 37 simple case would not take into consideration the fact that 38 other customers may already have requested service and may SUBSTITUTE SHEET (RULE 26)

I

IC- 1 WO 95/27963 PCT/EP95/01330 17 1 not yet have been processed. Thus, another customer near 2 customer 25 but still somewhat further away from the nearest 3 available taxi in area 10, may have a prior claim for 4 service. However, in the simplest of systems where only distance from the taxi to the customer is important, such a 6 prior claim would not be recognized.

7 In a preferred system where many factors are taken 8 into account the priority assigned to each taxi may often be 9 viewed as a multi-dimensional vector which is the vector sum of component priority vectors each relating to a different 11 selection criterion.

12 A preferred method for allocating a task to one of the 13 taxis in response to a request by customer 25 will now be 14 explained with reference to Figs. 2 to 7. For the sake of simplicity of presentation only, it will initially be 16 assumed that the only selection criterion of interest is the 17 distance of a taxi from customer 18 Fig. 2, shows a flow diagram of the operation of a 19 preferred system of the invention. The left portion of the flow diagram comprises operation of a first, targeting, 21 phase and the right portion of the flow diagram comprises 22 operation of a second, identification phase. The targeting 23 phase starts with the broadcasting of a call message to all 24 of the participating taxis informing them that a priority must be determined for responding to a pending request for 26 service. The criteria for determining the priority may be 27 sent together with the call or, alternatively, may be part 28 of a preset protocol used for all such determinations.

29 Alternatively, there may be several such protocols one of which the call identifies by a code. In the very simple case 31 of a dispatching system wherein distance of a taxi from a 32 specified location is the only selection criterion, there is 33 no need to inform the taxi of the selection criterion each 34 time an invitation message is transmitted.

Responsive to the call, each of the taxis uses the 36 selection criteria to determine its own priority in 37 accordance with the protocol. The protocol also includes a 38 plurality of ranges of priority values and a communication SUBSTITUTE SHEET (RULE 26) I I 1 WO 95/27963 PCTIEP95/01330 18 1 protocol which subdivides a time period and/or a frequency 2 range into a plurality of time or time/frequency slots each 3 of which is associated with one of the ranges of priorities.

4 Each of the participants who is not immediately eliminated from further participation owing to gross 6 unsuitability they are already occupied or is already 7 responding to a call) responds to the call message by 8 transmitting an indication signal in the appropriate time 9 or time/frequency slot in accordance with his respective priority. The indication time slots all start at a time 11 relative to a time base common to all of the participants.

12 It is important to note that for this embodiment of the 13 invention all those responding participants having a 14 priority within the same range respond at the same time and frequency. As a result, substantially simultaneous 16 indication signals are received by the control center from 17 those participants having the highest priority as determined 18 at the current priority resolution according to the 19 protocol. The indication signals, which are preferably pulsed CW they are pulsed signals at a particular 21 frequency having no information content other than that 22 given by the time at which the transmission occurs and the 23 frequency of the transmission), are sufficiently non- 24 destructive with respect to other simultaneously transmitted indication signals and have at least sufficient pulse width 26 so as to permit an indication that at least one of the taxis 27 has responded in a given time slot. In certain cases it may 28 be necessary to add some dithering or other variations to 29 the signals so as to avoid destructive interference between the signals.

31 Since the indication signals may, and typically do, 32 overlap, even a fairly narrow bandwidth broadcast channel 33 may be employed, there being no requirement to discriminate 34 between different indication signals in the same priority slot. Fu2thermore, intermodulation effects between the 36 indication signals in a given slot are not important, since 37 only the presence of at least one indication is required, 38 and the intermodulation does not affect this determination.

SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCT/E P95/01330 19 1 The control center monitors any response and 2 determines which slots have a true indicator signal (as 3 opposed to noise or other transients). Preferably, the time 4 slots are arranged in descending order of priorities, such that the control center may ignore all slots after the first 6 one (or some other small number of) "occupied" slot.

7 The control center targets all those responding taxis 8 having the highest priority range in respect of which a 9 valid indication signal has been received. Except as will be described below, taxis having a lower priority are 11 excluded from further consideration.

12 If a predetermined criteria for stopping the targeting 13 phase has not been reached, then an additional call is 14 broadcast requesting all those taxis which have a priority within the highest priority range to respond. The response 16 of the taxis is similar to that sent in the previous step 17 except that the time or time/frequency slots now represent 18 sub-ranges within the highest indicated priority range or 19 ranges. In general the call will include this range and may include an indication of the protocol for dividing the slots 21 among the priorities.

22 It should be understood that, for the more general 2D case of multiple criteria, the priority vector may be a 24 function of the iteration number and/or the priority range.

Thus for example, the first iteration may be used to elim- 26 inate taxis which are far away from the destination without 27 giving much weight to the idle (waiting) time. The second 28 iteration may give a greater weight to the idle time or to 29 other factors. In general, taxis which have moved closer to the destination since the last call and have an increased 31 priority may participate in the second iteration even if 32 they did not have the highest priority in the previous 33 iteration or were not detected as having this priority.

34 Furthermore, a special slot (hereinafter also referred to as a "miss" trap) may be provided for taxis whose priority is 36 now higher than the highest range detected in the previous 37 iteration. These taxis would take precedence over the other 38 taxis by using the special slot.

SUBSTITUTE SHEET (RULE 26)

II

P~ C WO 95/27963 'PCT/EP95101330 1 This iterative reduction of the number of participants 2 continues until a predetermined criteria is reached. This 3 criteria may include consideration of the priority 4 resolution achieved. The criteria may include a statistical estimate of the number of vehicles which have not been 6 eliminated. For example, if in a given iteration in which a 7 substantial number of sub-slots have been allocated, only 8 one or a few sub-slots contain a response, it is then fairly 9 certain that the number of taxis left in the system is small (or even one) and the iteration process (and the targeting 11 phase) is ended.

12 Another iterative approach which may sometimes be used 13 is to restrict the responders in the first phase to a single 14 range or a limited number of ranges. Assuming that the range of interest is between 0-10 km from the customer. A first 16 call would only ask for responses from those taxis which are 17 closer than 5 km from the customer. This distance could be 18 divided into ranges or a single range could be used. If 19 there were no responses, then the call would request responses from those taxis in the range of 5-10 km. If there 21 were a response, then further delineation of the range would 22 successively narrow the range of distances. In this method 23 of restriction all positive responses to the query are 24 preferably broadcast in the same time/frequency slot.

Preferably, in the targeting phase no participants are 26 actually identified, and therefore it is not yet possible 27 for the control center to dispatch a particular taxi to the 28 customer. Before this can be done, it is first necessary to 29 complete a second, identification, phase wherein one of the taxis targeted at the end of the first phase is uniquely 31 identified.

32 An additional call is broadcast or otherwise 33 transmitted to the participants indicating that an 34 identification phase is to begin. All of the targeted participants remaining at the end of the first phase are 36 invited to broadcast or otherwise transmit their 37 identification codes in one of a number of identification 38 slots (which may be time or time/frequency slots, DS-CDMA or SUBSTITUTE SHEET (RULE 26) I I CIILU WO 95/27963 PCTLEP9501330 21 1 FDMA slots). These slots have a duration (or information 2 bandwidth) commensurate with the information to be 3 transmitted by the taxis. The number of identification time 4 slots is determined in accordance with the protocol and is application-dependent, and may be based on the number of 6 participants which are believed to be (or estimated to be) 7 left.

8 For example, in a dispatching system, the priority 9 scale may extend from a distance of 10 km from the customer location and the initial priority resolution (so far as the 11 distance criterion is concerned) may be 1 km which is 12 reduced during two successive iterations to 100 m and 13 finally to 1 m. At such a fine priority resolution it is not 14 to be expected that more than a small number of taxis will be targeted so a fast converging identification phase having 16 only a few time slots ought to be sufficient for identifying 17 one of the targeted participants. It is not suggested that 18 a 1 m distance is significant in determining priorities for 19 taxis, however use of such fine distinctions aids in reducing the number of taxis which participate in the 21 identification phase. However, as will be explained below, 22 the protocol has built therein sufficient discrimination to 23 allow for possible errors in the number of identification 24 time slots allocated and to compensate for such errors as required.

26 The identification slots are preferably not assigned 27 in any way, and the taxis choose their slots in sore random 28 way. It can be expected that at for least some of the slots 29 more than one taxi will broadcast its identification information. Such broadcasts probably can not be read by the 31 control center which will choose the first taxi which it can 32 identify. If multiple dispatches are re lired to the same 33 destination, as for example where there are too many 34 passengers for one taxi, the second phase may have to be repeated several times until the required number of taxis 36 are dispatched. Furthermore, in extreme cases, it may be 37 necessary to call for identification from taxis having a 38 lower priority.

SUBSTTITITE SHEET (RULE 26) II I -III r ol-. C- IL WO 95/27963 PCTIEP95101330 22 1 As in the targeting phase, a slot may be provided in 2 the identification phase for taxis having a higher priority 3 than the call. These taxis may have moved closer to the 4 destination or their signal may not have been received by the receiver due to interference or blockage.

6 Alternatively, the stations may be identified in an 7 alternative embodiment of the identification phase in which 8 a slot is assigned to each of the remote stations (including 9 those which are not targeted, since the system has no indication of those which are left). All of the stations 11 which are targeted are invited to broadcast in their 12 identification slot. One or more of these responding 13 stations is then chosen. Using this system of identification 14 of the remote stations frees them of the need to transmit any information bearing signal, simplifying the system.

16 Having described the overall method for iteratively 17 targeting, during a first phase, successively fewer 18 participants and then, during a second phase, identifying a 19 desired number of the targeted participants, there will now be described a specific application thereof to the scenario 21 depicted in Fig. 1 and with reference to Figs. 3 to 7 of the 22 drawings.

23 Referring then to Fig. 3, a customer 25 has requested 24 a taxi. Shown within a circular target area 40 centered about customer 25 and having a boundary 35 at different 26 distances from customer 25 are four ta:i vehicles designated 27 as Va, Vb, Vc, and Vd. Vehicles outside boundary 35 are 28 excluded from consideration.

29 Target area 40 is split into a plurality of concentric sectors of which only the outermost sectors 42, 43, 44 and 31 45 are shown each having a width AR and being radially 32 disposed with respect to the customer. Adjacent sectors 42 33 and 43 or 43 and 44 or 44 and 45 are contiguous although for 34 the sake of clarity and explanation they are shown in Fig. 3 as separated from each other.

36 It will be noted that vehicles Vb and Vc are within 37 the first (innermost) sector 42, vehicle Vd is within the 38 middle sector 44 and vehicle Va is in the last (outermost) SUBSTITUTE SHEET (RULE 26)

I

WO 95127963 PCT/EP95/01330 -23- 1 sector 45. Since it is desired to allocate the task of 2 servicing the customer to the vehicle which is closest to 3 him, it is clear that one of the two vehicles Vb and Vc in 4 the innermost sector 42 must be identified as the most suitable for the task. It will also be apparent that the 6 number of vehicles which exist in any particular sector is a 7 function of the width of the sector. Thus, if the width of 8 each sector is increased from AR to 3AR, it is apparent 9 that vehicles Vb, Vc and Vd will now exist in the new, innermost sector comprising original sectors 42-44. In this 11 manner, the width of each sector AR constitutes a priority 12 resolution with which a priority is assigned to the 13 participating vehicles. The coarser lower) the 14 resolution, the more vehicles will answer the 6lection criteria and be rated at a particular priority associated 16 therewith while the finer higher) the resolution, the 17 smaller the number of vehicles which ahswer the selection 18 criteria and are rated with the corresponding priority.

19 Thus, after a first step of targeting in which a small group of taxis is chosen, in a second step of targeting the 21 highest priority taxi, a coarse resolution (finer however 22' than that in the first step) is set as shown in Fig. 3 and a 23 call message is transmitted by control center 12 to all of 24 the participants. The call message also preferably defines a time interval AT which is divided into an. equal number of 26 time slots At generally of equal width such that the total 27 number of time slots is equal to the total number of 28 priorities: i.e. the number of sectors. In a further 29 preferred embodiment of the invention, frequency diversity can be used to define multiple slots at the same time, each 31 of the slots being at a different frequency distinguishable 32 by the control center.

33 Upon receiving the call message, each of the 34 participating taxis determines its priority in accordance with the selection criteria, which, in the simplest case, is 36 assumed to be solely the distance of the participant from 37 the customer and within the maximum radius Rmax. Thus, 38 vehicles Vb and Vc are both assigned the highest priority, SUBSTITUTE SHEET (RULE 26) I ld -I I I L- WO 95127963 PCT/E P95/01330 24 1 while vehicles Vd and Va (in that order) are assigned 2 successively lower priorities. It should be noted that 3 typically there may be hundreds of vehicles in the target 4 area 40; only a few are shown in the figure for the sake of clarity. Further, each vehicle may have a handset (see Fig.

6 9) having a disabling switch by means of which the driver 7 can prevent the transmission of a response message upon 8 receiving an call message from control center 12. By such 9 means he can go off duty, etc.

The active participants Va to Vd now transmit an 11 indication signal within the time slot At corresponding to 12 their priority. Thus, vehicles Vb and Vc transmit an 13 indication signal first; vehicle Vd transmits his indication 14 signal second; and vehicle Va transmits his indication signal third. In an actual situation, of course, there may 16 exist many time slots corresponding to a large number of 17 coarse resolution priorities and perhaps hundreds of 18 vehicles will transmit an indication signal in the same time 19 slot. This, in itself, is not important because all that matters during this first phase of the process is to 21 determine the first time slot (or more generally, the slot 22 representing the highest priority) in which a vehicle 23 transmits an indication signal.

24 This having been done, it is immediately clear which is the nearest sector to the customer in which at least one 26 vehicle is located and therefore all of the vehicles in all 27 of the other sectors may now be eliminated. In a practical 28 implementation of such a system, the broadcast and receive 29 time for transmitting the call message from the control center to the participants and receiving the first 31 indication signal therefrom takes a short time. Thus, in a 32 relatively snort time interval thousands of participants in 33 the field can be reduced to a small number of potentially 34 suitable participants for the task, without the use of excessive frequency spectrum.

36 Furthermore, if a full duplex communication system is 37 used, the control center need not wait for the entire time 38 AT, and can go on to the next iteration or the next phase SUBSTITUTE SHEET (RULE 26)

IL

-r-l-IIFI I WO 95/27903 PCT/EP95/01330 25 1 immediately when a first indication signal is received.

2 As explained above, this process is repeated 3 iteratively as often as required, each iteration having 4 successively finer priority resolutions sectors of successively decreasing width AR), until a predetermined 6 resolution is reached. At this point, the width of the 7 remaining sector is sufficiently small that only a small 8 number of participants are likely to be found therein. It 9 is, of course, not known how many participants there are in this remaining sector since regardless of whether only one 11 participant or many send an indication signal in a 12 particular time/frequency slot, the control center does not 13 receive a message which is capable of uniquely identifying 14 any one of those participants.

It should be noted that the receive time taken for the 16 control center to process a response from the highest 17 priority participants is a function of the number of time 18 slots At. Thus, as the resolution is increased, there will 19 be more time slots and, since each requires a minimum transmit time, it might take longer to identify the highest 21 priority time slot. There is therefore a tradeoff between, 22 on the one hand, increasing the resolution sp as to identify 23 the most suitable participant in fewer iterations and, on 24 the other hand, increasing the cycle time of a given iteration by doing so. The choice of initial resolution and 26 rate of increasing the resolution may be made based on the 27 number of participating vehicles and or a priori 28 expectations of the responses. Thus, the range of values of 29 a priority which are assigned a given sector may based on the number of expected units having that priority. If 31 distance is the sole criteria, the range of distance values 32 may be proportional to the distance so that the area of the 33 sectors assigned to each priority may be the same.

34 Fig. 4 shows a subsequent iteration of the targeting phase wherein the priority resolution is increased and a 36 further call signal is transmitted to the participants. The 37 currently targeted participants Vb or Vc in what is 38 currently the highest priority sector 42 are assigned new SUBSTITUTE SHEET (RULE 26) II I rl, I II I I a Il- L I; WO 95127963 PCT/EP95/01330 26 1 slots according to the finer priority resolution and again 2 transmit indication signals during time slots corresponding 3 to their priorities. As a result, it transpires that Vc has 4 a higher priority than V b and its indication signal is therefore transmitted first (or in a slot corresponding to a 6 high priority, even if such slot is not first). However, 7 from the perspective of the control center, there is no way 8 of knowing how many participants exist in what is now the 9 highest priority sector. All that can be known is that at least one participant has the priority.

11 Thus, while there may still be hundreds of 12 participants in the targeted sector, it is expected that 13 with the increased priority resolution only a small number 14 of participants will now be targeted. One of these is now to be identified to respond to the request for service.

16 During this second phase of identification, the control 17 center assigns a new time interval AT-ID and divides this 18 time interval into a number of equal width time slots At 19 (or time/frequency slots) related to the expected number of participants in the highest priority sector 42. The expected 21 number of participants in sector 42 is determined 22 statistically as a function of the resolution of the sector 23 AR and according to the application or according to the 24 metL.o described below. The only remaining targeted participant Vc in the highest priority sector 42 now selects 26 randomly one of the time slots and transmits, within the 27 randomly selected time slot, an identification message 28 whereby the sending participant can be uniquely identified.

29 In the more general case, where there are still a number of targeted participants, the control center receives 31 a plurality of identification messages some of which may, of 32 course, have been transmitted during the same randomly 33 selected time (or time/frequency) slot. It is understood 34 that where two signals are transmitted at the same time and frequency, no information on the identity of the transmit- 36 ting participant is obtained in the absence of a capture 37 effect. However, it is expected that at least one of the 38 identification messages can be uniquely identified and, in SUBSTITUTE SHEET (RULE 26) I -1 Barc~8a~s~8lr~Rlll~aan~-- 4--I WO 95/27963 PCT/E P95/01330 27 1 this case, thI task is allocated to a participant which can 2 be identified. Where possible, of course, in the interest 3 of speed, the task is allocated to the first uniquely iden- 4 tifiable participant.

If it is not possible to identify one of the particip- 6 ants uniquely, the communication protocol allows for appro- 7 priate action according to each particular situation. Thus, 8 it may be that during the final iteration in phase one, no 9 participants were targeted. This itself could be due to several different reasons: for example, the call message may 11 never have reached the participants or, more likely, the 12 response of the highest priority participants may not have 13 been received, possibly having been obstructed by an obsta- 14 cle in its path.

Alternatively, possibly too many participants were 16 targeted in the final iteration of phase one and an 17 insufficient number of identification time slots were 18 allocated during phase two. In this case, identification messages may collide during all of the identification time slots, rendering it impossible to identify any one 21 participant. In the more general case where more than one 22 participant is to be identified, it may also occur that too 23 few identification messages arrive in phase two owing to an 24 insufficient number of participants having been targeted in phase one.

26 The various strategies for dealing with each of these 27 possibilities from the point of view of the control center 28 will now be described with reference to Figs. 5 6B which 29 show state diagrams relating to the targeting and identification phases, respectively. In both of these diagrams the 31 following terminology is employed: 32 PHASE-1.x :xth iteration of phase 1; 33 PHASE-2.x :xth iteration of phase 2; 34 IB :Control center's Broadcast Message; RD :Responders' signal Detection and 36 signal processing; 37 IBPHl.x :Control Center's x th Broadcast 38 messages in Phase 1; SUBSTITUTE SHEET (RULE 26)

II'-

PCT/EP95/01330 WO 95/27963 28 1 IBPH2.x :control center's x th Broadcast 2 messages in Phase 2; 3 ATRTPHl.x :x th time interval for Responder's 4 Transmission activity in Phase 1; ATRTPH2.x :xth time interval for Responder's 6 Transmission activity in Phase 2; 7 x :Number of iterations in Phase 1 or 2 8 (application-dependent); 9 NIP :Total number of iterations performed in current Phase; 11 Limitl, Limit2:Application-dependent maximum number 12 of iterations for Phases 1 and 2, 13 respectively.

14 n predetermined number of successful iterations 16 PS Priority slot 17 Thus, referring to Figs. 5A and 5B, if during a 18 successive iteration, no indication signal is received (i.e.

19 a MISS is detected), the initiator requests at least once that all of the participants who have not yet been 21 identified transmit a respective indication signal and this 22 is repeated until an indication signal is received or for a 23 maximum number of iterations determined in accordance with 24 the protocol. Thereafter, whilst the resolution is higher than a minimum resolution determined in accordance with the 26 protocol (and the iteration process has not been terminated 27 for some other reason), further priorities having a coarser 28 resolution are assigned to all of the participants who have 29 not as yet been identified, or until the resolution reaches the minimum resolution.

31 Another way of checking if a MISS is true is to 32 provide an additional slot during which all of those 33 stations which should have broadcast during the designated 34 priority slots will broadcast again. If no signal is received during this slot, the MISS is verified.

36 Referring to Figs. 6A and 6B, if during any iteration 37 no identification message is received by the control center 38 and during preceding iterations fewer than the desired SUBSTITUTE SHEET (RULE 26)

M

WO 95127963 PCT/EP95/01330 29 1 number of identification messages were received so as to 2 permit identification of the respective participants or if 3 invalid data were received, there is further included the 4 step of the control center requesting at least once that any currently targeted participants who have not yet been 6 identified re-transmit their identification message.

7 If fewer than the desired number of valid 8 identification messages are received so as to permit 9 identification of the respective participants owing to the occurrence of more than one identification message arriving 11 in the same identification slot or to any other reason such 12 as receipt of erroneous data, thereby rendering it 13 impossible to determine the respective identifications, the 14 following courses of action may be taken.

One possibility is for the control center to allocate 16 to all of the targeted participants still remaining and who 17 have not yet been identified a greater number of discrete 18 identification time slots than the number previously 19 allocated, and to invite the targeted participants who have not yet been identified to transmit a respective 21 identification message during one of the new identification 22 time slots. In other words, the nuimber of targeted 23 participants is maintained but more identification time 24 -ots are allocated so as to increase the probability that the desired number of valid identification messages will be 26 received by reducing the probability of collisions.

27 Alternatively, the taxis can be required to choose a random 28 number which can then be compared to some reference number 29 to eliminate some of the taxis or which can be used in changing the priorities of the taxis to eliminate some of 31 them. Alternatively, an additional criteria may be added to 32 reduce the number of participants.

33 Alternatively, if the protocol allows a maximum 34 priority resolution, then so long as the current priority resolution is lower than the maximum priority resolution, 36 phase one can be repeated as required for a maximum number 37 of iterations determined in accordance with the protocol at 38 successively finer priority resolutions, until the maximum SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCT/EP95/01330 30 1 resolution is reached in respect of all of the participants 2 who have not as yet been identified. This causes fewer 3 participants to be targeted and again reduces the 4 probability of collisions in phase two when any newly targeted participants are identified.

6 If, on the other hand, during a successive iteration, 7 fewer than the desired number of valid identification 8 messages are received so as to permit identification of the 9 respective participants owing to an insufficient number of participants having been targeted during preceding 11 terations, then the opposite must be done. Thus, so long 12 as the resolution is higher than a minimum resolution 13 determined in accordance with the protocol, phase one is 14 repeated as required for a maximum number of iterations determined in accordance with the protocol at successively 16 coarser priority resolutions until an indication signal is 17 detected or the minimum resolution is reached. This process 18 is performed in respect of all of the participants who have 19 not as yet been identified and, by targeting participants in phase one who were not previously targeted, increases the 21 probability that the desired number of newly targeted 22 participants will subsequently be identified in phase two.

23 If, during phase one, no indication signal is received 24 in response to a call message, the control center requests at least once that the participants re-transmit an indica- 26 tion signal. On receipt of the call message, the particip- 27 ants assign themselves priorities and transmit respective 28 indication signals during a corresponding indication time 29 slot. This covers the possibility that the call message never reached the targeted participants or, alternatively, 31 that their responses never reached the control center.

32 In all of the above cases, data is stored in respect 33 of any participants who have already been identified and 34 subsequent iterations are performed only to identify additional participants.

36 The protocol includes at least one termination 37 condition whereby further iterations are not performed even 38 if no indication signal has been received and/or if fewer SUBSTITUTE SHEET (RULE 26) I C WO 95/27963 PCTEP95/01330 31 1 than the desired number of participants have been 2 identified. This is necessary to avoid an infinite loop 3 being executed in the event that, in a particular 4 application, there are not enough participants who can be identified.

6 Figs. 7A and 7B show a timing diagram relating to the 7 flow of information between the control center and the 8 participants Va, Vb, Vc and Vd during the example of Figs. 3 9 and 4.

It will be noted that in the initial phase of 11 targeting, each participart selects a time slot according to 3.2 his respective priority, such that participants with the 13 highest priority transmit first. Consequently, as soon as 14 the control center receives a response from the participants, the highest priority may be determined 16 immediately in accordance with which time slot data was 17 first received. Further iterations may now be effected, as 18 required, there being no need even to await the responses of 19 lower priority participants. This results in very rapid convergence of the targeting phase to the priority range 21 containing the most suitable participant. This requires a 22 full duplex system. Figs. 7A and 7B show the timing diagram 23 for a half duplex system.

24 In another embodiment of the invention priorities may be assigned according to a measured elapsed time since 26 participants have performed some activity. For example, 27 priorities are assigned to taxis according to the time they 28 have been idle.

29 In a first iteration of phase one, the mutually common priority scale relates to an elapsed time of say 3 hours and 31 the priority resolution is say one-half hour. Thus, each 32 interval in the priority scale corresponds to an elapsed 33 time of one-half hour.

34 In a second iteration of phase one, the mutually common priority scale relates to an elapsed time of one-half 36 hour and the priority resolution is 2.5 minutes. Thus, each 37 interval in the priority scale corresponds to an elapsed 38 time of 2.5 minutes. If during the first iteration a signal SUBSTITUTE SHEET (RULE 26) ~1 I WO 95/27963 PCT/EP95/01330 32 1 was received in the time slot of two to two and one-half 2 hours, then the time slots in the second phase may have a 3 resolution of 2.5 minutes and span the range between these 4 limits.

If this is considered to be sufficiently fine so that 6 not too many participants will have the same priority, the 7 process is terminated after only two iterations. It is now 8 appropriate to implement phase two wherein one of the 9 targeted participants is identified.

It will be understood that since, during the 11 identification phase, a participant may be, in effect, 12 selected randomly, it cannot be assured that the identified 13 participant is actually the one who has waited the longest.

14 However, it can be said with certainty that the identified participant has the highest priority to within the priority 16 resolution (in this case 2.5 minutes).

17 If, notwithstanding the above expectation, it becomes 18 impossible to identify a single participant in phase two 19 owing to too many participants having been targeted during phase one, then, as explained above, several options are 21 available. More identification time slots can be allocated 22 in phase two or, alternatively, a further iteration in phase 23 one can be performed at an even finer priority resolution, 24 for example 6 seconds, before repeating phase two in respect of a smaller number of targeted participants or one of the 26 other options described above may be employed.

27 In all of the embodiments described above, at least 28 two phases are required to identify a targeted participant.

29 Thus, during a first phase, participants are only targeted and are identified during a subsequent second phase.

31 However, according to another preferred embodiment of the 32 invention, provision may be made for identifying a 33 participant during the first phase by transmitting an 34 identification message as the indication signal. The identification message can be decoded in the particular 36 circumstance that only one participant has the highest 37 priority, so that only one indication signal is transmitted 38 in the highest priority time slot, and there is a SUBSTITUTE SHEET (RULE 26) ~I I- WO 95/27963 PCT/EP95/01330 33 1 sufficiently long time interval between receipt of 2 successive indication signals by the control center to allow 3 decoding of the identification signal before a lower 4 priority indication signal arrives in a subsequent indication time slot. Alternatively, the identification 6 time slots are made long enough so that different slots have 7 minimal or no overlap. In this particular case, the second 8 phase of identifying the targeted participants is 9 eliminated. It should be noted that, generally, this embodiment of the invention is less efficient than the 11 embodiment which uses non-information bearing signals in a 12 first, targeting, phase to reduce the number of vehicles in 13 a second, identification, phase.

14 Yet a further consideration relates to the possibility that the highest priority participant may not be targeted in 16 phase one owing to a malfunction. Thus, for example, his 17 indication signal may not be received, having been 18 obstructed by an obstacle in his path or his signal is 19 subject to fade. This may not matter if other participants having the same priority have nevertheless been able to 21 transmit indication signals, since if the indication time 22 slot having the highest priority is determined and all the 23 participants associated therewith are targeted, even a 24 participant whose indication signal was lost will still be targeted. However, if a sole participant's indication 26 signal is lost this could prevent correct determination of 27 the highest priority participant.

28 The protocol can take this possibility into account by 29 reserving, preferably, the first indication time slot in the next iteration for exclusive transmission therein by a non- 31 targeted participant having a higher priority than that of 32 the targeted participants. The control center then transmits 33 a call message inviting the targeted participants to trans- 34 mit a respective indication signal during any one of the indication slots except the reserved indication slot.

36 So far as newly targeted participants are concerned, 37 the process is essentially unchanged; each of the newly 38 targeted participants transmits an indication signal during SUBSTIPTE SHEET (RULE 26) I_~I _C WO 95/27963 PCT/EP95101330 34 1 one of the unreserved indication slots according to his 2 respective priority. However, any previously non-targeted 3 participant having a higher priority than that of the newly 4 targeted participants transmits a respective indication signal during the reserved slot. This slot is referred to 6 herein as an "Inter-Iteration Miss/Trap Control Slot." 7 In a further preferred embodiment of the present 8 invention, at least one iteration of the targeting phase 9 includes a control slot, which is reserved for simultaneous transmission by all the priority-bearing participants of the 11 iterative stage responding to the call sent from the control 12 center. According to this preferred embodiment, each 13 priority-bearing participant transmits twice, once during 14 its characteristic-indicating slot as described above and once during the control time slot. This control slot is 16 referred to herein as an "Intra-Iteration Miss/False Control 17 Slot." This control slot provides an independent indication 18 of transmission by at least one participant in response to 19 the control center call which initiated the iteration.

The additional indication obtained from the Intra- 21 Iteration Miss/False Control Slot will generally economize 22 on both the total processing time and the total 23 transmission time of the targeting phase. For example, if no 24 transmissions are received in this control slot, it may be assumed that any transmissions received in the 26 characteristic-indicatlng slots are due to a false alarm 27 error and, thus, further processing and transmissions 28 derived from the last iteration are avoided. Similarly, in a 29 "miss" error situation in which no participants are targeted, a transmission received in this control slot 31 indicates a possible "miss" and the last iteration is 32 preferably repeated. Thus, the use of a control time slot 33 improves the reliability of the targeting phase.

34 It should be noted that for improved reliability of the detection process at the control center receiver, the 36 remote station transmitter may perform a more sophisticated 37 transmission process that includes diversity techniques 38 based on randomly varying the amplitude and or the phase of SUBSTITUTE SHEET (RULE 26) L I I WO 95/27963 PCT/EP95/01330 35 1 the transmitted signal, so that any correlation between 2 transmitters will be reduced when detecting a sequence of 3 transmission slots.

4 To further improve the reliability of the targeting phase, a preferred embodiment of the invention employs a 6 majority voting techniques in which an odd number of slots 7 greater than one, for example three slots, are assigned to 8 each range of characteristic values. According to this 9 preferred embodiment, each participant of a given characteristic value transmits an indicating transmission 11 during each of the time slots assigned to the range. The 12 given characteristic value is taken into account only when 13 indicating transmissions are received in a majority of the 14 slots assigned to the value, for example by two slots out of three. When indicating transmissions are received only by a 16 minority of the slots assigned to the value, for example by 17 one slot out of three, the response is preferably ignored.

18 It should be appreciated that the odd number of slots 19 assigned to the given range of characteristic values may be distributed among slots that have minimum correlation.

21 It is appreciated, however, that such majority voting 22 schemes which improve the over-all reliability of phase one, 23 consume substantial transmission and processing time. There- 24 fore, such schemes are preferably applied selectively, in potentially problematic situations. For example, majority 26 voting may be employed only after a predetermined number of 27 previous false alarms and/or "misses" have been detected 28 using the control slot technique described above or other- 29 wise.

Within a single iteration of phase one, the assignment 31 of priorities to each participant may be performed in 32 respect of a different sub-set of selection criteria for at 33 least some of the participants. In effect, this permits 34 different search strategies to be executed each in respect of a respective indication slot. For example, the first 36 indication slot ha;ing the highest priority may relate to 37 all participants who are located within a radius of 10 m 38 from the customer without any further restriction; whilst SUBSTITUTE SHEET (RULE 26) I WO 95/27963 PCTIEP95/01330 36 1 the second indication slot may relate to all participants 2 who are located within a radius of 25 m and who have been 3 awaiting instructions for more than 20 minutes. By this 4 means Boolean OR search or other search strategies can be performed in a single iteration.

6 Furthermore, during each iteration the participants 7 may optionally assign themselves a priority having a 8 magnitude outside the priority scale so as not to be 9 targeted by the initiator. This can be done if, for example, a participant is otherwise occupied or for any 11 other reason does not wish to receive instructions.

12 During a particular iteration the priorities assigned 13 to each participant are generally absolute with respect to a 14 mutually common scale which itself is external to the participants and independent thereof. However, between 16 successive iterations the priority scale may well relate to 17 different combLnations of selection criteria. By this means 18 finely tuned search strategies can be performed whereby all 19 participants answering to a first combination of selection criteria are targeted during a first iteration, whilst all 21 of the targeted participants answering to a different 22 combination of selection criteria are targeted during a 23 successive iteration.

24 Alternatively, the combination of criteria which make up the priority may be information dependent and, for 26 example, different groups of slots may relate to different 27 combinations of criteria.

28 Once a sufficiently small number of participants are 29 targeted such that, in accordance with the protocol, identification of a desired number of participants is likely 31 to yield a successful outcome, the second phase described 32 above is commenced. The number of identification slots to 33 be allocated to the targeted participants is calculated by 34 first estimating the number of targeted participants remaining at the end of phase one. The number of 36 identification slots is then calculated according to the 37 estimated number of targeted participants who must transmit 38 respective identification messages, so as to reduce the SUBSTITUTE SHEET (RULE 26) 11 WO 95/27963 PCT/EP95/01330 37 1 total time required to identify the required number of 2 participants.

3 In this connection, it will be realized that there 4 exists a tradeoff between allocating too many and too few identification time slots. Specifically, allocating too many 6 identification time slots reduces the probability of a 7 targeted participant selecting an early identification slot, 8 thereby increasing the time required to identify the highesu 9 priority participants. On the other hand, allocating too few identification time slots increases the probability that 11 .iore than one participant will select the same 12 identification time slot. In this case, the resulting 13 collision of more than one identification message makes it 14 impossible to identify the respective participants, requiring further iterations and again increasing the 16 identification time. In practice the number of 17 identification time slots may be minimized by increasing the 18 maximum priority resolution in phase one in order to target 19 no more than the expected number of participants who are to be identified in phase two, or by using a random process to 21 eliminate some of the participants, such as that described 22 above.

23 In the specific embodiments described above the 24 process of assigning priorities to each of the participants is performed within the participating vehicles themselves 26 since only they know their locations relative to the 27 customer. Moreover, the onus of tracking the participants' 28 movements in terms of their location, availability, 29 occupancy, loading and all the other selection criteria which may be of significance is now passed to the 31 participating vehicles themselves as opposed to most 32 hitherto proposed systems wherein a central dispatcher had 33 to keep track of all these parameters.

34 As a result of the above, the communication channel between the control center and the participants may be of 36 relatively narrow spectrum width compared with that of 37 hitherto proposed systems. Additionally, the task of tar- 38 geting potentially suitable participants is distributed SUBSTITUTE SHEET (RULE 26)

I

WO 95/27963 PCTIEP95/01330 38 1 amongst the participants themselves rather than being deter- 2 mined solely by the control center. Such distribution 3 results in a reduction of computing power being required by 4 the control center.

While the selection criteria must obviously be known 6 to the participating vehicles, the manner in which this is 7 made known can be varied according to circumstances. Thus, 8 for example, the selection criteria may be fixed and known 9 in advance to the participants (in which case the selection criteria are not subject to change). Alternatively, the 11 selection criteria may be determined on-line by the control 12 center and then transmitted to all of the participants 13 together with the call message.

14 Thus, in the particular example described above, during the first phase of targeting it may be predetermined 16 that each sector has a width of 10 km and that in subsequent 17 phases, the width of each remaining sector is reduced, for 18 example, by a factor of 10 until the sector has a width of 19 only 10 m, whereupon all those vehicles within the 10 m width sector send an identification message; or, 21 alternatively, the width of each sector in each respective 22 phase of allocation may be transmitted to the participants 23 by the control center. In order to reduce the number of 24 participants in the identification phase, the resolution in the targeting phase may be increased artificially, i.e., 26 past the point at which it is meaningful.

27 It should also be noted that once a particular 28 participant has been uniquely identified to perform a task, 29 he is notified of this in the normal way by the control center, in any one of a number of ways which are well known 31 in the art as for example, by voice over a communication 32 channel or by text or other data transmission.

33 Furthermore, although in the preferred embodiment 34 described above, one participant is uniquely identified as the most suitable, in fact it may sometimes be appropriate 36 to omit the second phase of identification altogether. In 37 such cases, the participants having the highest priority are 38 not uniquely identified as individuals but all are SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCT/El'95/01330 39 1 identified as a group. One such situation relates to an 2 improvement of service in particular areas. In this 3 situation the number of taxis in a given area is monitored 4 and additional cars are sent in to the area if there are not enough cars in the area. The number of cars may be estimated 6 statistically, for example, from the number of slots having 7 responses.

8 In a further preferred embodiment of the present 9 invention, the number of participants complying with given criteria is estimated based on a down-sampling technique in 11 which only a portion of the complying participants actually 12 respond to a call from the control center. For example, the 13 participants may be assigned a given response probability 14 such that only a given percentage, for example ten percent, of the complying participants respond to the call. Response 16 time-slots are preferably randomly selected by the different 17 participants. If the number of response slots is sub- 18 stantially greater than the number of responses, i.e. the 19 number of complying vehicles times the response probability, the number of detected responses generally corresponds to 21 the number of responses which, in turn, is a down-sampled 22 indication of the number of complying participants.

23 Down-sampling is particularly useful for situations in 24 which the number of expected responses is high, such as for estimating the number of vehicles in a given area, when very 26 course selection criteria are applied. On the one hand, in 27 such a case the number of responses is large enough to be 28 statistically reliable, provided that the number of 29 responders is small enough compared to the number of slots dedicated for this purpose. On the other hand, the smaller 31 number of participants transmitting simultaneously reduces 32 the total transmission power and, thus, prevents occasional 33 bursts of powerful transmission which may be in violation of 34 FCC co-channel interference or other regulations.

Another application of the system of priority assign- 36 ment according to the invention is in the assignment of 37 available lines for car-phones or transceivers. Presently, 38 available lines are allocated on a when available basis.

SUBSTITUTE SHEET (RULE 26)

I

WO 95/27963 PCT/EP95/01330 40 1 Thus, one unlucky user may wait a long time while a lucky 2 user may get an immediate line. In a preferred embodiment of 3 the invention, when a user wants a line, he indicates this 4 either by pressing a call button or by lifting his receiver.

A computer chip associated with the car-phone notes the time 6 at which a line was requested.

7 Lines (communication channels) are allocated on a 8 waiting time basis. In operation, a control center 9 broadcasts a call for priorities in accordance with a targeting phase of the present invention. The priority is 11 assigned according to waiting time, and the individual 12 phones broadcast signals during time slots assigned 13 according to their waiting time or by some other special 14 priority. During a second identification phase one of the phones is identified, in the same manner as described above, 16 and is given the available line.

17 Referring now to Fig. 8 there is shown schematically 18 the principal features associated with the control center 19 shown in Fig. 2. Thus, there is provided a transceiver and modem 50 coupled to 4n antenna 51 for effecting bi-direc- 21 tional communication with the participants (taxis) and being 22 connected to a message processor 53 which is coupled to a 23 computer 54. Message processor 53 receives non-demodulated 24 signals from the transceiver and determines which slots contain signals for the targeting phase and identifies the 26 participant(s) in the identification phase. A preferred 27 embodiment of such a receiver is shown in Fig. 28 A service request is effected by customer 25 by 29 telephoninlg his nearest taxi rank and then dialing his telephone number, the request being routed to the computer 31 54 via a Public Switched Telephone Network (PSTN). The 32 computer 54 converts the customer's telephone number to a 33 corresponding location based on a data base stored in the 34 computer. Alternatively, such communication can be effected via an operator. A terminal 56 is coupled to the computer 36 54 for allowing an operator to enter commands and display 37 data. In addition the system also allows for voice signaling 38 to a dispatcher at the control station or for voice SUBSTITUTE SHEET (RULE 26) ~O~IIRUII~IAI~IIC- WO 95/27963 PCT/EP95101330 41 1 communication between the taxi driver to the customer.

2 Fig. 9 shows the principal components associated with 3 a participant allocation unit 60 located in each of the 4 vehicles. Allocation unit 60 preferably includes a transceiver 61 coupled to an antenna 62 for effecting bi- 6 directional cc,9munication with the transceiver 50 in the 7 control center 37. Transceiver 61 is connected to a vehicle 8 computer 64 coupled to a microphone/handset 65 providing a 9 human interface between the vehicle computer 64 and the corresponding taxi driver.

11 A Global Positioning System 66 (GPS) or other position 12 determining system as known in the art receives positioning 13 data via an antenna 68. The Global Positioning System 66 is 14 coupled to the vehicle computer 64 and functions as a positioning means for providing positioning information 16 relative to a predetermined origin in respect of the 17 corresponding participant. Thus, once the location of the 18 customer is provided to the vehicle computer 64, The 19 latter, being coupled to the Global Positioning System 66 is able to determine the relative location of the participant 21 to the customer and thus determine the participant's 22 priority.

23 Associated with the vehicle computer 64 is a storage 24 means 70 for storing the protocol according to which priorities are assigned. Also stored in the storage means 26 70 are any singular areas which can affect the actual route 27 e.g. obstructions such as rivers, road blocks and so on 28 which result in the actual route distance being longer than 29 it would otherwise be. As explained above, the handset allows the driver to assign himself a priority outside the 31 range of the priority scale and, by such means, to exclude 32 himself from the process of targeting. It also includes a 33 microphone for establishing voice contact with the control 34 center, as well as paging means for obtaining a text message th.refrom.

36 The system described above may include a full duplex 37 broadcast network such that the control center does not need 38 to await responses from all of the participants before SUBSTITUTE SHEET (RULE 26)

I

WO 95/27963 PCT/EP95/01330 42 1 targeting the highest priority participants. Thus, 2 specifically, as soon as a valid response is received by the 3 control center, the participants corresponding to the 4 response can immediately be targeted or identified whilst informing other participants to stop transmitting indication 6 signals or identification messages. This permits the steps 7 of targeting and/or identifying participants to be effected 8 more quickly. However, the invention may also be employed 9 in a simplex half-duplex) broadcast network, albeit at the expense of longer targeting and identification times 11 since the control center cannot transmit to the participants 12 until all their responses have first been received and 13 validated.

14 It will be appreciated that, instead of employing a Global Positioning System, other systems for determining a 16 participant's location can equally well be employed. For 17 example, a route scheduler based on dead reckoning 18 responsive to each participant's location can be used for 19 determining a route having minimum distance. Such a route scheduler might possibly comprise sensors located at 21 intervals along the road for sensing a passing vehicle's 22 presence and for transmitting to the vehicle data 23 representative of its location relative to a specified 24 location for error correction. Typically, such a route scheduler has a memory for storing therein a scaled contour 26 map so that an optimal route can be determined taking into 27 consideration the nature of the terrain. Likewise, 28 prevailing traffic conditions can be red into the route 29 scheduler at regular intervals of time, so that traffic jams, roadwork and so on can be considered when determining 31 the optimal route.

32 In the foregoing description it has also been assumed 33 that a single channel broadcast network is employed.

34 However, this is by no means essential and a centralized controlled trunking system having at least two channels may 36 equally well be employed. This permits more than one task 37 to be handled simultaneously each on a different broadcast 38 channel. Thus, in the case of a two channel broadcast SUBSTITUTE SHEET (RULE 26) I, I M WO 95/27963 PCT/EP95/01330 43 1 network, for example, having first and second channels, each 2 call message is transmitted via a broadcast control channel 3 so as to be received by all the participants associated with 4 the first channel. Upon determining that he has not been targeted by the control center, a participant starts to 6 measure elapsed time and waits a predetermined elapsed time 7 locked on to the first channel and thereafter returns to the 8 broadcast control channel for receiving further call 9 messages.

The period of time during which a non-targeted 11 participant remains locked on to the first channel is of 12 sufficient duration to allow an updated priority to be 13 assigned to the participant. Owing to the dynamic variation 14 in a participant's status, it may occur that, with an updated priority, a previously non-targeted participant 16 becomes targeted in the next iteration. Thus, the period of 17 time during which a non-targeted participant remains locked 18 on to the first channel must further be of sufficient 19 duration to allow a corresponding indication signal and/or identification message to be transmitted by the participant 21 to the control center, whereby the control center may target 22 and/or identify the participant.

23 Alternatively, the call message may be transmitted via 24 a broadcast control channel so as to be received by all the participants associated with the first channel and, upon 26 determining that he has not been targeted by the control 27 center, a participant receives from the control center an 28 instruction to return immediately to the broadcast control 29 channel. This immediately frees a non-targeted participant to participate in a subsequent search strategy on the second 31 channel relating to a different task.

32 According to yet another variation, an initial call 33 message is transmitted together with the selection criteria 34 via a broadcast control cha nel so as to be received by all the participants associated with the first channel. Each of 36 the participants receiving the call message assigns to 37 himself from the priority scale a respective priority repre- 38 sentative of his relative suitability in accordance with the SUBSTITUTE SHEET (RULE 26)

II

WO 95/27963 PCT/E P95101330 44 1 selection criteria and transmits an indication signal during 2 a respective indication slot. Only the targeted particip- 3 ants remain switched to the first channel and subsequent 4 call messages are transmitted only to those participants who have been previously targeted by the control center. This 6 again frees a non-targeted participant to participate in a 7 subsequent search strategy on the second channel relating to 8 a different task.

9 It will be appreciated that whilst the invention has been described with particular application to a taxi 11 dispatching service, the invention has more general 12 application wherever one or a group amongst a plurality of 13 participants is to be targeted in accordance with their 14 respective suitabilities based on at least one selection criterion. It will further be understood that, whilst the 16 preferred embodiment has been described for the sake of 17 simplicity with regard to only two selection criterion (i.e.

18 distance and waiting time), in practice a large number of 19 selection criteria may be employed, all having different relative weights, whereby an integrated search strategy may 21 be implemented.

22 It will also be understood that whilst the invention 23 has been described with particular reference to 24 2 dimensional terrain, it can equally well be applied in 3 dimensional space and is thus suitable for air or space 26 travel, as well as land and sea.

27 Mention should also be made of the variable parameters 28 in association with which the protocol functions. These are 29 generally application dependent and typically are provided with default values built into the protocol. Thus, if 31 distance is one of the selection criteria, this fact may be 32 represented by a default value of an associated parameter.

33 Likewise, the lower and upper bounds of the priority scale 34 and the priority resolution associated with each iteration in phase one can been assigned to respective parameters each 36 having corresponding default values.

37 Any unassigned parameters must, of course, have values 38 assigned thereto prior to initiation of phase one. This can SUBSTITUTE SHEET (RULE 26) at I WO 95/27963 PCT/EP95/01330 45 1 be done during the initiation of the process prior to 2 transmitting the first call message to the participants.

3 However, in certain applications, all the parameters may 4 have pre-assigned default values which are acceptable for the application. In this case, the call message merely 6 starts the process enabling the participants to determine 7 the appropriate priority scale and assign themselves 8 respective priorities at the appropriate priority 9 resolution; there being no need to inform any of the participants of the boundary values of the priority scale or 11 of the priority resolution or indeed of the selection 12 criteria.

13 While the invention has been described with particular 14 reference to a wireless broadcast network, it will be appreciated that the invention is capable of much more 16 general application. For example, hard-wired communication 17 systems may also employ the principles of the invention in 18 which case the indication signals need no longer be CW. In 19 such cases the dynamic variables would generally not be position; however the system is generally applicable to 21 systems with any set of dynamic variables.

22 The principles of the invention may also be used in a 23 routing system, for example to a system which identifies 24 buses or other vehicles which are delayed and adjusts the speed and/or location of other buses to compensate therefor.

26 In the first (targeting) stage of this utilization of the 27 invention, the priority would for example be based on the 28 amount of time that a vehicle is behind schedule. Vehicles 29 which are behind schedule more than a predetermined amount would then be targeted and identified in a second 31 (identification) stage. Preferably, the identified bus would 32 then be asked for its exact position. Due to the fixed 33 lineal nature of bus routes, the position of the bus on the 34 route is a one dimensional function, the distance along the path.

36 A query would then be sent to other busses on the same 37 bus line asking them for their positions and, optionally, 38 where they stand in relation to their schedule. Based on SUBSTITUTE SHEET (RULE 26) WO 95127963 PCTIEP95/01330 46 1 this information, a control center would determine 2 corrective action to provide improved service, which may 3 include steps such as speeding up some buses, as for example 4 by operating them in a skip-stop fashion, slowing some buses down, keeping some buses from leaving the terminal or adding 6 new buses to the route, perhaps at some intermediate point 7 on the route. Some indication of occupancy of the buses 8 would help to avoid sending full or almost full busses to 9 load additional passengers when less full buses are available. Such indication, which could, for example, be 11 keyed in by the bus driver, would help in optimizing 12 routing decisions. Suitable instructions would, of course, 13 be transmitted to these busses after a corrective action 14 plan is formulated.

Alternatively, information on deviations from schedule 16 are ignored, and a revised schedule is based only on the 17 position of the buses and optionally on their occupancy.

18 The principles of the invention are also applicable to 19 a routing system for determining slow areas of traffic and rerouting traffic around such areas. In such a system a 21 large number of participating vehicles are queried as to the 22 delays they are experiencing, and the delay time is one 23 example of a "characteristic value" for the first phase of 24 this embodiment. When a vehicle experiences a delay above a threshold, the position of the vehicle is determined in the 26 second phase. It should be noted that no identification 27 signal per se is transmitted in the second phase, instead a 28 position signal is broadcast. Preferably, the delay is also 29 verified by the driver of the vehicle to avoid false alarms.

Once the position of the targeted delayed vehicle is 31 determined, a new first (targeting) stage determines those 32 vehicles that are close to the specific delayed vehicle, and 33 determines, by successive second stages, the extent of the 34 delay as a function of the time of the delay. Furthermore, by making multiple queries, the traffic conditions can be 36 estimated. Based on this information, the seriousness of the 37 delay may be determined and corrective action, such as re- 38 routing of other vehicles, may be started. In particular, SUBSTITUTE SHEET (RULE 26) e II-- WO 95/27963 PCT/EP95/01330 47 1 information on the traffic conditions and the geographical 2 extent of the delay may be transmitted to vehicles which 3 have routing apparatus of types which are known in the art, 4 to be used by these apparatus for determining the optimum route for the receiving vehicle.

6 In an alternative preferred embodiment of the 7 invention, the first query requests responses only from 8 vehicles which are experiencing delays greater than a given 9 time (or which are moving at an average velocity of less than a given velocity). Those vehicles which meet the 11 criteria then broadcast a signal in a time or time/frequency 12 or frequency slot which is indicative of the absolute 13 position of the vehicle. As in some of the previous 14 embodiments of the invention, it is expected that more than one vehicle will broadcast in a particular slot and the 16 system is interested, at least at this stage, only in 17 determining if there are vehicles which are experiencing 18 delays of a given magnitude.

19 Fig. 10 shows an initial map generated by such a method, wherein the area represented by a pixel (slot) may, 21 for example, be of the order of 250 to 1000 meters square.

22 In a preferred embodiment of the invention, the system 23 then determines, based, inter alia, on the extent of the 24 various contiguous areas which shows positive responses, a smaller area or areas for further study. Preferably, the 26 system then broadcasts a further query requesting those 27 vehicles within the more restricted area which have at least 28 a given delay (which may be the same as or different from 29 that used in the first query) to broadcast in a position slot using a finer rcsolution, for example, 100 to 250 31 meters. Based on the responses to this query a second map 32 such as that shown in Fig. 11 is generated. As can be seen 33 from Fig. 11, various branches of a road network radiating 34 from an intersection, designated as A-F in Fig. 11, can be identified. To improve the usefulness of the display, a 36 background map, such as a road map may be displayed 37 underlying the displays of any of Figs. 10, 11 or 13.

38 In the event that additional information relating to SUBSTITUTE SHEET (RULE 26) 1 'C'-C WO 95/27963 PCiE P95/01330 48 1 the delay is desired, further queries can be made. For 2 example, vehicles which are traveling toward the 3 intersection can be requested to broadcast in a slot which 4 corresponds to the slot they are in and to their velocity toward the intersection. This allows for generati.on of the 6 graph shown in the lower portion of Fig. 12. Additional 7 slots may be used for the generation of other information 8 regarding the responding stations. Such information may also 9 be graphed as shown in the upper portion of Fig. 12.

Alternatively or additionally, a map which shows the 11 average velocity of the vehicles toward the intersection as 12 a function of the position can be generated. Such a map is 13 shown in Fig. 13. To acquire the information needed for 14 generating such a map, a number of queries may be made, each requesting an indication from all vehicles within the area 16 of interest having a given average velocity toward the 17 intersection. The responding vehicles would broadcast their 18 indication signals in slots corresponding to their position.

19 In the map of Fig. 13 the velocity for a given pixel is determined, for example, as the average velocity of the 21 reporting slots for that position. In a display of the map 22 of Fig. 13, the velocity toward the intersection can, for 23 example, be displayed as a gray scale value or as a color, 24 with for example red being the highest delay and blue being a minimum displayed delay.

26 Fig. 14, which is a generalized block diagram for a 27 system useful for performing the IVHS function described 28 above, shows a base station or control center 91 having a 29 control center transmitter 79 which broadcasts queries and optionally other signals to vehicles on command from a 31 control computer 80. A remote vehicle 85 (only one vehicle 32 is shown for simplicity) receives the query at a vehicle 33 receiver 84 and transmits commands to a microprocessor 86, 34 based on the queries it receives from the control center.

Microprocessor 86 also receives information regarding 36 the status of the vehicle from one or more information 37 generators and sensors indicated by reference numeral 88.

38 This information may be sent by the sensors on a regular SUBSTITUTE SHEET (RULE 261 ill WO 95127963 PCT/IEP95/01330 49 1 basis or may be sent on command from the microprocessor.

2 Microprocessor 86 is then operative to command vehicle 3 transmitter 90 to transmit indication signals (or if 4 required information bearing signals) in a suitable slot in accordance with the information received by microprocessor 6 86.

7 The indication (or other) signals are received by a 8 control center receiver 92 and processed by receiver 92 and 9 computer 80. While the operation and construction of the apparatus designated by reference numerals 82, 84, 86 and 11 is straightforward and needs no further explanation, the 12 operation of receiver 92 is usefully expanded upon with 13 reference to Fig. 14 The system described above is based on a central decision maker which receives information from vehicles, 16 plans the routing for each vehicle and then broadcasts a 17 route or route changes to the individual vehicles. This type 18 of system has the advantage that the routing for each 19 vehicle takes account of the routing for the other vehicles and the control center in computing the routings can balance 21 the routings to cause minimum delays. The disadvantage of 22 such a system is the large bandwidth required to notify the 23 individual vehicles of their individual corrected routes.

24 A second approach for routing systems which has been suggested is to have each of the vehicles compute its own 26 route, based on some information about the present status of 27 traffic which it receives from a central transmitter. While 28 such systems require only a limited bandwidth, the routes 29 computed by the individual stations cannot take into account the future effects of the routes of other vehicles.

31 In a preferred embodiment of the invention, vehicles 32 compute their own routing and then report, in response to a 33 query, their expected time of arrival at locations that are 34 known to have a high incidence of traffic jams and slowdowns (and preferably also additional locations which do in fact 36 have such slowdowns). Preferably, such reporting is 37 performed using the slot method of transmission which does 38 not identify the individual vehicles. Since a large number SUBSTITUTE SHEET (RULE 26) -I I- WO 95/27963 PCT/EP95/01330 50 1 of vehicles is involved, down-sampling as described below 2 may be effectively used to estimate the numbers of vehicles 3 which pass the locations.

4 Additionally or alternatively, the future development of existing slowdowns can be estimated from the prior 6 development of the slowdowns, the rate of change of the 7 length of the slowdown and the average speed of the vehicles 8 which are within the slowdown. Such information can be made 9 available to the vehicles based on comparison of the development of slowdowns which are detected by the methods 11 which have been previously described above.

12 Based on the estimates of the numbers of and times of 13 arrival of the vehicles at the trouble spots, information on 14 future expected traffic jams is generated by the central station and broadcast to the vehicles which update and 16 recalculate their individual routings. This recalculation of 17 routs, broadcast of times of arrival at trouble spots and 18 estimations of future traffic jams and slowdowns gives each 19 vehicle the information required to make a distributed system effective in avoiding future problems, without the 21 huge bandwidth requirements of central calculation of the 22 routes for the vehicles.

23 Such a distributed method may be applied to fleet 24 management and other systems.

For example, in a preferred embodiment of the 26 invention, the redistribution of a a fleet of taxicabs from 27 a, present, actual distribution to a desired redistribution 28 is accomplished by having the taxicabs choose, based on a 29 predetermined algorithm and information transmitted to them by a central station, which taxicabs will move to a new 31 location. In this redistribution procedure, a present 32 distribution (containing numbers of taxicabs in a region, 33 without necessarily any indication of the position of 34 individual identifiable taxicabs) R~nd a new desired distribution is transmitted to all of the taxicabs. Based on 36 a predetermined protocol, each of the taxis will decide 37 locally if they are to move to a new location. Preferably, a 38 decision by each taxi is based on a statistical model SUBSTITUTE SHEET (RULE 26) 1 C II- -I WO 95/27963 PCT/EP95/01330 51 1 whereby generally approximately the correct number of taxis 2 decide, on their own, to move to new sites. This protocol 3 may consider parameters such as the location, idle time and 4 distance of individual taxi from the area which need additional taxis, as well as the present distribution of 6 taxis and what, statistically, they will choose to do, based 7 on the protocol.

8 Another similar application of the invention is bus 9 fleet management, where bus distribution information, occupancy information, connection times, location 11 distributions, etc., is broadcast to the bus fleet to enable 12 the buses to make distributed decisions. In particular, 13 based on the information received, a bus may, in effect, 14 instruct itself to skip a bus stop, wait (or not wait) for a connection with another bus, to leave a starting point early 16 (or late), etc.

17 Generally speaking, the RF signals transmitted by the 18 vehicle may be at any frequency slot. It is to be expected 19 (both for the IVHS application and for the dispatching application described above) that there will a certain 21 amount of frequency diversity caused by the imperfect 22 accuracy and stability of the vehicle transmitters 90. The 23 slots are wide enough to accommodate this diversity.

24 Furthermore, often the system utilizes very large numbers of vehicles. If too many of these vehicles (in some 26 particular situation) transmit in the same slot, then the 27 total power transmitted may exceed authorized ERP or dynamic 28 range restrictions. To overcome this problem longer, lower 29 power, pulses may be used for indication signals.

Furthermore, if a single receiver is used for receiving 31 signals for all of the slots, intermodulation effects may 32 cause spur.ous signals to appear in slots for which no 33 actual signals have been received.

34 These problems as well as near-end to far-end transmission problems are substantially solved by the system 36 shown in Fig. 15 and by certain constraints placed on the 37 system which are not shown in Fig. 38 With respect to excess power problems, if it is SUBSTITUTE SHEET (RULE 26) L I 111 1 r- I WO 95/27963 PCT/EP95/01330 52 1 expected that many vehicles may transmit in a particular 2 slot, the queries can be designed so that fewer than the 3 total number of vehicles will respond, whenever this is 4 possible. This can be accomplished, for example by having the vehicles choose, statistically, which vehicles will 6 respond within a given percentage of the total number of 7 vehicles. The power transmitted by the vehicles can be 8 adjusted to a minimum based on either the known distance 9 between the vehicle and the control receiver, with each vehicle transmitting just enough power so that detection of 11 the signal by the control station is assured. A further or 12 alternative power adjustment may be made by the vehicle 13 transmitter based on the power received from the control 14 station, for example, during the query. Finally, a closed loop system in which the query includes instructions as to 16 the power levels to be used may be used. It is not desired 17 that such closed loop system result in exactly the same 18 power level being received from each remote station be 19 perfect since this would increase the probability of amplitude correlation between the signals and resultant 21 destructive interference, usually in a situation where a 22 strong line of sight transmission exists between a few 23 vehicles and the base station. A balance should be struck 24 between a reduced variation in the power level received by the control center from the various remote vehicles and 26 keeping the chances of destructive interference low.

27 Increased pulse duration can also reduce the 28 transmitted power for a given ratio of detection probability 29 to false alarm probability especially in the receiver shown in Fig. 15 and described below.

31 Preferably, the amplitude of the signals broadcast 32 during the time slot is shaped over the broadcast period to 33 reduce the side-lobes of the signals and avoid false signals 34 in adjacent frequency slots, which may be a problem when large numbers of vehicles broadcast at the same time.

36 Alternatively or additionally, frequency windowing at the 37 receiver may be used to reduce cross-talk between channels.

38 Dynamic range limitations can b2 reduced by providing SUBSTITUTE SHEET (RULE 26)

I

h WO 95/27963 PCT,'EP95/ 1330 53 1 multiple receivers, each covering only a portion of the 2 frequency band. Finally the novel receiver of Fig. 15 may be 3 used to determine the presence or absence of signals in 4 particular slots.

Fig. 15 shows a receiver system corresponding 6 generally to reference number 92 and to a portion of 7 computer 80 of Fig. 14. In general such a receiver is also 8 useful for the first phase of the dispatching system 9 described above as well as for the IVHS system.

An antenna 94 (or an array of antennas) receives 11 signals from a plurality of vehicles simultaneously and 12 passes them to receiver and (optionally) AGC 96. Receiver 13 and AGC 96, which may be of conventional design, 14 downccn-verts the received signals from RF to IF frequencies.

The threshold levels of the detection process may be 16 dependent on the AGC process. The IF signal is digitized by 17 an A/D jystem 98 and further down converted by a 18 downconverter 100 to base band. It should be understood that 19 this receiver/downconverter system does not demodulate the incoming signals, but only downconverts the RF so that the 21 same relative frequency differences of the signals is 22 present at the output of convertor 100 as in the incoming 23 signals, except that the absolute frequency has been reduced 24 to a low frequency from the RF frequency of the transmi:tted signal. At these lower frequencies digital systems can be 26 used to analyze and detect the signals.

27 The low frequency band signals are fed to a series of 28 correlation filters 102 (correlation-type receiver), each 29 of which has a very narrow bandwidth which is related to the correlation time of the correlation filter. Preferably, the 31 frequency bandwidths of adjacent receivers 102 overlap so 32 that the entire bandwidth of each of the slots is covered by 33 one set of receivers 102. The output of each of the 34 receivers is compared to a threshold 104 to determine if a signal is present at the frequency of the respective 36 receiver 102 and the outputs of all of threshold detectors 37 for a given slot are OR gated to determine if any signal is 38 present in the slot. Alternatively, the outputs of the SUBSTITUTE SHEET (RULE 26) m I WO 95/27963 PCT/EP95/01330 54 1 correlation receivers can be summed and this sum signal used 2 to determine if any signal is present in the slot. However, 3 this will generally result in increased noise.

4 In an alternative preferred embodiment of the ir -ention, the strongest output of the set of correlation 6 receivers is chosen for comparison with a threshold, with or 7 without post-detection integration.

8 Use of a plurality of overlapping narrow band 9 receivers in this manner also reduces the extent of side lobes of the detection process outside the band of the slot.

11 This allows for closer frequency spacing of the slots since 12 interference between slots having adjacent frequencies is 13 reduced.

14 One set of receivers 102, threshold detectors 104 and an OR gate is provided for each slot and is referred to 16 herein as a slot detector unit. Slot detector units for all 17 of the slots feed a data processor 108 which, together with 18 computer 80 processes the data as described above. When 19 large numbers of vehicles are used in the system and ntermodulatton becomes a problem (or if AGC is used, and 21 low level signals are lost), it may be neceesary to provide 22 a plurality of front end pori'.ons of receiver 92 (the front 23 end being defined as receiver convertor 98 and converter 24 100), where each front end receives signals from only a portion of the entire frequency band including one or many 26 of the slots. The function of correlation receivers 102 may 27 also be implemented, for example, using set of DFTs or an 28 FFT (for CW signals), matched filters or other correlation 29 receiver methods or other optimum receiver methods, depending on the transmitted signals. Other methods such as 31 energy detectors radiometers) with or without 32 tracking may also be used, however, they will give less 33 optimal results, because of practical limitations on input 34 band-pass filter designs.

It should be understood that using a plurality of 36 correlation receivers for the same slot may increase the 37 false alarm probability and hence the threshold for positive 38 detection may be adjusted to provide a desired low false SUBSTITUTE HEET (RULE 26) ~li~t*aranna~-~- P L- WO 95127963 PCTIEP95/01330 55 1 alarm probability.

2 For all the above applications of the invention, 3 destructive signal interference between units which 4 broadcast in the same slot can be further reduced by performing transmission diversity operations such as random 6 phase (for example, 180* and 00) and/or amplitude changes, 7 at the transmitters of the remote stations.

8 In a preferred embodiment of the invention, an 9 improved probability of detection may be desired for some of the slots, such as, for example, the control slots. For 11 these slots repeated transmission of signals using 12 transmission diversity may be performed and detection 13 enhancement methods such as post detection integration may 14 be used to improve the detection probability.

The system may also be provided with a display 110 for 16 displaying the data, such as the maps and graphs of Figs.

17 10-13 and with a user interface 112 which is used by an 18 operator to control both the operation of the system. The 19 user interface also preferably co rols the display and the memory to allow for the opeyztor to review the maps 21 previously generated or to generated new displays based on 22 information previously received.

23 Information may be sent by the control center to the 24 vehicles to enable them to minimize average travel delays.

This information may consist of the above mentioned maps or 26 of travel delay information at various intersections. The 27 vehicles can then use this information to optimize their 28 route. Alternatively, the control center may send routing 29 information to some of the vehicles in order to equalize traffic delays. In either event, the fast response of the 31 system in a matter of seconds allows for real time 32 supervision, adjustment and continuous stabilization of 33 traffic patterns with additional iterations. As described 34 above, in a distributed system only prospective traffic patterns is broadcast by the control center and each vehicle 36 calculates its own route.

37 The IVHS system described above is also useful in 38 tracking situations such as for fleet management.

SUBSTITUTE SHEET (RULE 26) LL IL

L-

WO 95/27963 PCT/EP95/01330 56 1 In a further preferred embodiment of the invention the 2 position and other characteristics of a large number of 3 vehicles can be mapped and tracked in near real time using a 4 relatively narrow bandwidth. In this embodiment each vehicle is assigned a number of slots according to a predetermined 6 protocol which are used only by that vehicle.

7 The vehicles are preferably first mapped with a 8 preferred mapping phase of a mapping and tracking procedure.

9 A way to perform this phase, is to devote a small matrix to each vehicle to be tracked. This smaller matrix is part of 11 the entire matrix of slots assigned by the control center.

12 The smaller matrix represents, for example, (in mapping of 13 spatial coordinates) a square area which is divided into 14 nine sub-areas, each of the sub-areas represented by one of nine slots assigned to a particular vehicle. In a first 16 iteration a large area is divided into nine sub-areas and a 17 vehicle broadcasts a signal in the slot which corresponds to 18 its present position. In a second step of the mapping phase, 19 the area in which the vehicle previously broadcast is expanded to fill the nine slots, with a finer resolution.

21 Alternatively, the area which is zoomed into the nine slots 22 is slightly larger than the area of the previous broadcast 23 to avoid a situation in which the vehicle was at the border 24 of the area and left the area between steps.

This identification of one sub-area and consequent 26 convergence to a higher resolution is repeated several times 27 until the required resolutioi is achieved. The highest 28 practical resolution, as will become clear below, is the 29 distance that a vehicle could travel in the time it takes to perform a tracking cycle as described below. Within five 31 iterations the individual resolution can be improved from a 32 10 km square (divided into nine 3.3 km squares) to 33 resolution of about a 40 meter square area.

34 In a second, tracking, phase of the mapping and tracking procedure, performed periodically after the 36 required resolution is reached, for example, nine slots, 37 representing a 3 x 3 area of minimum resolution areas, are 38 used to track additional movements of the vehicle. The SUBSTITUTE SHEET (RULE 26) II WO 95/27963 PCTIEP95/01330 57 1 central one of the nine areas corresponds to the area 2 occupied by the vehicle at the end of the mapping phase (or 3 during a previous periodic updating iteration of the 4 tracking phase). During each periodic update, each vehicle broadcasts in a slot which corresponds to either its 6 previous position (the slot corresponding to the center area 7 of the 3 x 3 group of areas) or one of the adjoining areas.

8 In the next iteration, the newly chosen area is the center 9 of the 3 x 3 matrix, according to a predetermined protocol.

In a further preferred variation of this embodiment of 11 the invention, only 5 slots are utilized to map into the 3 x 12 3 area. One of these slots represents, as a reference, one 13 of the corner (or the center) areas of the 3 x 3 area and 14 the other 4 slots represent north-south or east-west variations. In this the vehicle may broadcast during one or 16 two of the five slots, depending on the deviation, if any, 17 from the corner (or center) area chosen as reference. If the 18 vehicle is in the reference area, broadcasting takes place 19 only in the slot which represents the reference area. If the vehicle is in the areas north-south-east or west of the 21 reference, then the vehicle broadcasts in only one slot 22 representing such deviation from the reference. If the 23 vehicle moves into an area diagonally shifted from the 24 reference, the vehicle will broadcast during two slots representing, for example, the east-west and north-south 26 deviations of the area in which the vehicle is situated.

27 If a vehicle does not respond or its response was not 28 detected during a given iteration if the vehicle is 29 lost: or an erroneous code is received) a number of remedial steps are possible. The particular vehicle (or some or all 31 the vehicles) may be requested to retransmit the particular 32 iteration, or may be asked to return to perform the previous 33 iteration or a sequence of previous iterations or to operate 34 at a lower mapping resolution. In some situations it may be desirable to start the process over, at the lowest 36 resolution for the particular vehicle or for all of the 37 vehicles, for example to repeat the entire process or some 38 of the steps of the process to increase reliability even if SUBSTITUTEn SHEET (RULE 26) Illi WO 95127963 PCT/EP95/01330 58 1 no errors are detected. The process may be repeated using 2 the previous position information or with present 3 information with a lower resolution especially to find 4 "lost" vehicles.

Fig. 16B shows the slots in which a signal would be 6 broadcast in the tracking phase (or possibly in the mapping 7 phase) to indicate each of three positions of the vehicle 8 shown in Fig. 16A, while Fig. 16C shows the slots which 9 would be used if a corner area were used as the reference.

It should be noted that for either case the vehicle was in 11 the center area during the previous tracking iteration.

12 While it appears from Figs. 16A-16C that it is 13 desirable to have north south deviations represented by a 14 change in slot frequency and east west variations represented by a change in slot time, it is actually more 16 practical to use the same frequency for all small matrix 17 slots used by a particular vehicle, since this requires only 18 one transmitter per vehicle.

19 While it is desirable to dedicate particular transmission slots for each vehicle, it is possible to have 21 overlapping assigned transmission slots. For example, if one 22 slot for one vehicle is the same as a slot for another 23 vehicle, then if a signal is received in the shared slot, 24 the systems checks if a signal was received in one of the unshared slots for one of the vehicles. If it was, then the 26 signal in the shared slot is considered as coming from the 27 other vehicle. If no signal is received in unshared slots 28 for either vehicle, then the signal is considered as coming 29 from both vehicles.

In a preferred embodiment of the invention, nine areas 31 are represented by a four bit word which is more than 32 sufficient to define the 3 x 3 matrix of elements. In this 33 or similar cases the "physical slots" described above may be 34 represented by any convenient code.

Furthermore, the logical slots may have different 36 meanings or resolutions in the same logical matrices. For 37 example, the position resolution of the logical elements may 38 depend on the maximum expected velocity of the vehicle and SUBSTITUTE SHEET (RUtlE 26) WO 95/27963 PCTIEP95/01330 59 1 the resolution in the two mapped directions need not be the 2 same. Additionally, for mapping and tracking along a road, 3 one dimension may be the (one dimensional) position along 4 the length of the road and other logical slots, if any, may, for example, represent the lane in which the vehicle is 6 traveling.

7 While the preceding mapping and tracking system 8 implies that each vehicle in the system is a priori included 9 in the map, this is not necessarily a requirement of the system. Each of these vehicles would then be assigned slots 11 for use in the mapping and/or tracking phases. The vehicles 12 from whom identification is requested may be chosen in 13 accordance with a criteria of the vehicle determined 14 according to any one of the procedures outlined above in which vehicle having certain characteristics are determined.

16 As indicated above, bus tracking, for buses which 17 follow a particular route, requires only one dimensional 18 tracking information. This means that only three logical 19 slots (or two logical slots, if only forward movement is assumed) are required to track these vehicles. Other slots 21 may be used to provide other information about the bus such 22 as occupancy level. In such a situation after an 23 initialization phase similar to that for position, one slot 24 may represent no substantial change in the occupancy, with two other slots representing an increased occupancy of the 26 bus and another slot representing a decreased occupancy.

27 Such allocation allows for occupancy to be tracked in a 28 simple way, similar to the position tracking described 29 above, simultaneously with the position tracking or at a lower frequency, interleaved with the position tracking 31 responses.

32 In general, one or more base stations may be used for 33 broadcasting calls and/or receiving responses from remote 34 stations. If more than one base station is used, each station preferably performs a reduction of the data which it 36 receives by either choosing its best candidate for 37 performing the task or by performing a mapping function of 38 its nearby region or of its associated vehicles. The base SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCT/EP95/01330 60 1 stations then preferably send this reduced information to a 2 central base station which makes the final decision, 3 constructs the desired map or performs any other final 4 analysis. Furthermore, the central base station would, in a preferred embodiment of the invention, instruct each of the 6 base stations as to which additional queries they should 7 make. In this situation the subsequent queries need not be 8 the same for all the base stations.

9 In a further preferred embodiment of the invention, a pager system having an appointment making capability 11 operates on principles similar to those described above. For 12 example, the pager system may broadcast a request to one or 13 more pagers which also incorporate an appointment calendar.

14 The individual pagers broadcast a signal in one or more of a matrix of slots which correspond to busy times. The 16 appointment may then be made for other times. The 17 individuals are then notified, by pager, that an appointment 18 has been made for them.

19 In many of the above embodiments of the invention, the system is triggered and/or synchronized according to a synch 21 signal broadcast by the control station. Other sources of 22 synchronization, which synchronize both the remote and 23 control station, such as GPS received signals or other 24 timing signals, can be used to trigger and/or synchronize the system.

26 The invention has been described herein using examples 27 in which the indication signals are transmitted in time, 28 frequency or time and frequency slots. Other types of 29 transmission slots are also useful in the invention such as frequency hopping and other spread-spectrum transmission 31 slots. The term "transmission slots" or "slots" as used 32 herein includes all these types of slots.

33 The following is an analysis of the communication 34 resources required by some of the above described inventive apparatus and method compared to polling, for cases where 36 more than one remote station may transmit non-modulated in 37 the same slot.

38 SUBSTITUTE SHEET (RULE 26)

M

WO 95/27963 PCTIEP95/01330 Abstract A novel method for information collection from a. multiple sources (stations) system which drastically reduces the channel capacity requirements is presen=ed and analyzed. The mair types of applications are: dispatching, traffic surveillance, trafic control.

The communication resources required using the described method are lower by orders of magnitude relative to conventional methods. Method's power is due to effective use of the computation power of the remote and central stations, effective transfer of information in the links and collisions robustness of the reverse link.

1 Introduction This paper presents and analyzes a novel method for information collection from a multiple sources (stations) system which drastically reduces the channel capacity requirements.

1.1 System The system under consideration consist of a central base station and many remote stations (information sources). A forward communication link is used for transmissions from the central base station to the remote stations and a reverse communication link is used for transmissions from the remote stations to the central base station. The reverse link is divided in channels according to a multiple access scheme which will be described.

In traditional multiple sources (stations) system the information from the sensors is gathered, processed and disseminated. The information gathering process require a lot of communication resources such as time and frequency. The novel method permits a drastic reduction of the communication resources required to obtain the information of interest.

1.2 Distributed Calculation The novel method is base on distributed capability of processing. It is assumed that each remote stations may collect data from its own sensors such as GPS, speedometer, timer and others and calculate a characteristic value based this data and other attributes of the respective station. This is a reasonable assumption because in future the vehicles will have several sensors to determine parameters such as speed and position and its own processing capability.

The characteristic of interest and its selected attributes are determined by the base station and broadcasted to the remote stations in the forward link. The station can calculate its own characteristic value based on the selected attributes according to a predetermined method.

The characteristic value is in fact a quantitative criterion to the remote station status. For example the characteristic value may be the distance of the remote station from a specific target, its speed, its availability for service and so on or a function of these parameters.

SUBSTITUE SHEET (RULE 261

I

WO 9127963 PCT/EP95/01330 62 1.3 Novel Method Description The novel method is based on two modes of operation which can be combined according to the application: Discrimination. In this mode the remote stations are discriminated according to their characteristic value. In this mode of operation the remote stations are interrogated and they respond by an un modulated signal in one of reverse link channels. The channel is chosen by the remote station according to its characteristic value. The resolution of the discrimination can be improved by several iterations.

Identification. The stations which have been detected with the proper characteristic values are identified. In this mode of operation the remote stations respond by a data signal on the reverse link.

1.4 Applications In the following discussion we refer to two types of applications: a. Search of preferred or optimal remote stations according to a characteristic value based on selected attributes of the respective stations. For example the search of a free ;a:d closest to a specific destination.

b. Surveillance. Observance of remote stations conditions according to their relative characteristic values. Detection and analysis of incidents and local status. For example detection and analysis of traffic congestion areas.

Paper outline The novel protocol will be called in the rest of the paper characteristic values discrimination (CVD) protocol. In, Section 2 the CVD protocol is described in details and in Section 3 it is compared with a conventional polling algorithm. In Section 4 the performance of the discrimination mode of operation of the protocol are analyzed and numerical results are given for different channel conditions. The analysis is performed for a simple but representative model of the characteristic value distribution, by using an lower bound for the probability of protocol correct detection. The analysis conclusions are given in Section 2 Protocol Description 2.1 Forward and Reverse Links The protocol is based on the two way communication link. In each step of the protocol a message is transmitted by the base station to the remote sensors in the forward link. and a reply is received in the reverse link.

In the reverse link two types of operation are possible: SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCT/EP95/01330 63 Pulse transmission. A multiple access communication protocol provides M 2" channels such that in the interrogation replay, an unmodulated pulse can be transmitted on one channel by a remote station. and ;he base station has to detect their presence.

When two ore more remote stations respond in the same channel of the reverse link collision occurs, however the effect of the collision depends on the signal distribution.

For the Rayleigh distribution of each signal the total distribution is also Rayleigh with an average power which is the sum on the signal average powers. Therefore for the Rayleigh distribution the collision is not a problem.

Data transmission. Digitally modulated signals are transmitted by the remote sensors and the base station has to detect their messages.

2.2 Discriminatio.i Mode The discrimination mode of operation has the purpose to discriminate among remote stations with different characteristics values. The discrimination mode of operation consists of interrogation steps; in each step the remote stations are addressed in the forward link and they may respond in the reverse link. The interrogation in the forward link defines: the characteristic value of interest which can be measured by each one of the remote stations, the range of the characteristics value, the number M 2" of divisions of characteristics value range.

The interrogation range of the characteristics value is di and the resolution of characteristics value division is r; d;/M.

The communication reverse link has M 2 m channels such that in the interrogation replay an unmodulated pulse can be transmitted on one channel by a remote station. The M channels corresponds to the M divisions of the characteristic value. A remote stations chooses to reply in the channel which corresponds to its own characteristic value.

After each interrogation step the base station analyzes the reply channel to determine the parameters of the next step. If we search for the best remote stations, let's say with the smallest characteristic value and assume no failures in the reply analysis, then the parameters of the next step are: the same characteristic value of interest, the range of the characteristics value, di+, ri, the same number M of divisions of characteristics value range.

Therefore in each step of the discrimination mode the resolution of the characteristic value of the optimal stations may be improved by M times.

SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCTIEP95/01330 2.3 Identification Mode After several steps of the discrimination mode, the resolution is fine enough such that only one or few remote stations respond. The stations which have been detected with the proper characteristic values are required to identify themselves by a message in the forward link and the remote stations respond by a message on the reverse link which contain an identification number. To avoid collision more then one channel may be used for reply.

3 Comparative analysis of the CVD protocol We would like to compare the time required to find the optimal station using CVD method and a conventional polling method. The comparison is done first with simple assumptions.

3.1 System Parameters We assume the following parameters of the system: The number of remote stations is N 2".

The optimal L 2' are required to be identified. These are the remote stations with she best characteristic value.

Characteristic value y in the range Y y 0.

Characteristic value resolution with R 2' levels.

3.2 Communication Reverse Link The communication reverse link has an allocated bandwidth and a multiple access communication protocol with M channels such that during a time duration T An unmodulated pulse can be transmitted on each channel in the discrimination mode of CVD method.

One bit can be transmitted on each channel in the identification mode of CVD method.

One bit can be transmitted on each channel in the polling method.

The same duration T was assumed for all three cases to simplify the comparison, however, this is a reasonable assumption as is explained in the Appendix.

SUBSTITUTE SHEET (RULE 26) /n 3 PCrT/f'po'/1f ln VV WJ .3 1YJU65 3.3 Search of optimal stations 3.3.1 Calculation of Novel Search Duration The CVD search method has two phases: Discrimination. In this phase the resolution of the characteristic value of the optimal stations is improved in I iterations of duration T each, such that in each iteration the resolution is decreased m times. The number of iterations is I and the total m duration is "n- 2 Identification. The L stations which have been detected with the best characteristic values are identified. Assuming identification messages which include n bits identification name, r bits for resolution and a random access protocol slotted Aloha with 10L slots, the identification phase duration is T 1 T. For L 1 there is only one remote station to be identified and in this case the identification phase duration is T.

The total duration of CVD search is TTQ n T for L 1 and for L 1 it is TrQ T.

3.3.2 Calculation of Polling Search Duration Each of the N remote stations transmits a message of r bits with a duration rT. The total duration of polling process is Tpo, 3.3.3 Comparison of Search Methods For a simple comparison of search methods we farther assume: -L=1 r na n. The resolution is enough to discriminate among the remote station characteristic values.

The duration of the search method is determined by the reverse link.

The ration between CVD search duration and the duration of polling process is given by: TTrQ r T 1 M m m As a numerical example with M 16 for N 5 the CVD method is superior to the polling method. For N 128 the CVD method is superior to the polling method by a factor of SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCTEP95/01330 66 3.4 Surveillance 3.4.1 Description of surveillance method The CVD method can be used for surveillance of remote stations in a specific area. For this applications only the discimination mode of operalion is used because the is no need to identify the specific remote stations but only the local situations like traffic congestion areas. It is also of interest to analyze the analyze the local status of the remote stations and determine existence of accident heavy, traffic, etc.

3.4.2 Calculation and Comparison of Surveillance Process Duration The same assumptions as in Section 3.1 are used for comparison of the CVD method and polling method. The discrimination mode of operation is used in several iteration: ko iterations will use an initial resolution of discrimination on ko areas of interest. If the whole area is investigated then ko 1. The analysis of these iterations determine k, smaller areas for farther investigation.

k, iterations will use a better resolution of discrimination on kL areas of interest. The analysis of these iterations determine k 2 smaller areas for farther investigation.

kI iterations will use a better resolution of discrimination on k2 areas of aiterest, The analysis of these iterations determine k 3 smaller areas for farther investigation, The process continues until the desired resolution is obtained.

The total time is TTQ (ko k k 2 The total duration of polling process is as explained in Section 3.3.2: Tpol The ration between CVD method search duration and the duration of polling process is given by: Tpol Nr T M 2 TTQ (k 0 o -I +k 2 M(ko k As a numerical example with M 100, N 10 5 R 106, ko 1, kt 9, k 2 40, kJ the CVD method is superior to the polling method by a factor of 4 Discrimination Protocol Model Analysis The discrimination protocol described in Section 2 discriminates among remote stations with different characteristics values. At step i the interrogation range and resolution of the characteristics value are di and r; d; /i respectively where M is the number of channels of the reply link and of divisions of the characteristic value. In each step of the discrimination mode the resolution o- the characteristic value of the optimal stations may be improved by M times. and at next step the range of the characteristics value may be reduced such that d+i

L

ri.

SUBSTITUTE SHEET (RULE 26) WO 95/27963 PCTIEP95/01330 4.1 Failure mechanisms and modes Failure mechanisms in the forward link are: a. No detection.

b. Detection of message with errors detected by the error detection code.

c. Detection of message with errors not detected by the error detection code.

Failure mechanisms in the reverse link are: a. Channel false alarm b. Channel miss detection If we search for the best remote stations with a characteristic value co, then the failure modes in the reverse link are: FM#l. The channel with the best characteristic value detected has a characteristic value better than co and the detection in this channel is a false alarm. Caused by false ala.in a channel corresponding to a better characteristic value than that of the best remote cannel.

FMr2. The channel with the best characteristic value detected has a characteristic value worse than co and the detection in this channel is a false alarm. Caused by mris detection of the best remote channel and false alarm in a channel corresponding to a characteristic value worst than that of the best remote cannel and no detection in a channel with a characteristic value better than that of the false alarm channel.

FMNI3. The channel with the best characteristic value detected has a. characteristic value worse than co and the detection in this channel is a true detection. Caused by miss detection of the best remote channel and true detection in a channel corresponding :c a characteristic value worst than that of the best remote cannel and no detection in a channel with a characteristic value better than that of this true detection channe.

FM#4. Miss detection of the best remote channel and no other detection.

False alarm when there are no signals transmitted.

4.2 Auxiliary channels in the reverse link Two optional auxiliary channels may be added: a. Miss detection trap channel (MD TC). This channel should be used by a remote channel whose characteristic is better then the range specified in the interrogation. This channel is useful for recovery from FM#2 and FM,3.

b. Common range channel (CRC). This channel should be used by all the remote stations in the interrogation range.

SUBSTITUTE SHEET (RULE 26) '1 TII ~C9 p WO 95/27963 PCTIEP95/01330 4.3 Interrogation replay decisions The base station decides based on the interrogation reply how to proceed to the next intersgation steps. The decision is based on the individual decisions in the M resolution channels (RC) and in the auxiliary channels MD TC and CRC. The decisions rules are given in Table 1 for the case we search for the best remote station.

RC MD TC CRC Next interrogation Note range r normal case r d MD TC trusted, second best choice S(1) r RC trusted d di- 1 Table 1: Base station decision rules based on individual decision on RC, MD TC and CRC.

and indicates respectively pulse detection, no detec:ion and don't care conditions one a channel. indicates that in only one RC a pulse was detected.

4.4 Analysis of Protocol Model The perfrrmance measure of practical interest is the probability that the protocol will make correct detection of the preferred remote station. A correct detection is made even if during the steps of the protocol errors occurred but were corrected by the described protocol.

Another performance measure is the probability of no failure during the protocol. This probability is easier to calculate. The probability of no failure PVF is a lower bound for the probability of protocol correct detection PPCD: PVF PPCD (3) The quantitative analysis which follows is of the P.vF-, the lower bound of PPCD and is performed for few examples assuming a simple model of ucriform distribution of the characteristic value. Example A assumes a typical uniform distribution of the characteristic value while examples B and C assume a worst case one. Examples A and B assume a auxiliary channels while example C assumes auxiliary channels.

SUBSTITUTE SHEET (RULE 26) 3 Uo~a- WO 95/27963 PCT/EP95/01330 69 Assumptions for example A.

The range of the characteristic value has 64 values.

16 remote stations have 16 consecutive values from 19 (010011) to 34 (100010).

The remote station with the smallest characteristic value i_ searched.

The number of divisions of characteristics value range per iteration or the number of resolution channehl is M 4. The channels are 00, 01, 10 and 11.

The forward channel messages are received with no error.

No auxiliary channels: MD TC and CRC.

When no failure events occur there are three interrogation steps. The potential failure events are given in Table 2. The probability PE that the process ends with no failures is: Iteration I FA in channels: I MD in channel: I MDs 1 00 01 13 2 00 1 3 00, 01, 10 11 1 Table 2: The potential failure events in the three interrogation steps of example A with no failure.

PvF 3 2 (1-PFA) 4 1-2PW-MD F.r.

The probability of no failure PIVF is a lower bound for the probability of protocol correc: detection PPCD: PvF PPCD Assumptions for example B.

The same as example A but the consecutive values of the 16 remote stations are from 47 (101111) to 62 (111110). When no failure events occur there are three interrogation steps.

The potential failure events are given in Table 3. The probability PVF that the process ends Iteration FA in channels: I MD in channel: MDs 1 00, 01 2 00, 01, 10 11 1 3 00, 01, 10 11 1 Table 3: The potential failure events in the three interrogation steps of example B with no failure.

SUBSTITUTE SHEET (PL'LE 26) PCTIEP95/l01330 ii7/- nrr/'t"Tf VYJ lUJ with no failures is: PVF (1 PMDo)(1 PFA4)I 2 1 3 PMD 12PFA The probability of no failure is worse in example B. It can be shown that for 16 remc:e stations with consecutive characteristic values example B is the worst case situation.

Assumptions for example C.

The same as example B but with auxiliary channels: MD TC and CRC. When no failu:e events occur there are three interrogation steps. The potential failure events are given i: Table 4. The probability Ppr that the process ends with no filires is: Iteration I FA in channels: MD in channel: I MDs 1 00, 01 10 1 2 00, 01, 10, MD TC 11, CRC 1 3 00, 01, 10, MD TC 11, CRC 1 Table 4: The potential failure events in the three interrogation steps of example C with no failure.

PF (1 PD)j(1 PA) 1 5PoI 14PA The probability of no failure is worse in example C then in B. However PvF is only a bound for PPCD and the improvement due to the auxiliary channels is in PPCD Model Detection Parameters The detection parameters in the forward link are: a. No detection probability Pa.

b. Detection of message with errors detected uy the CRC with probability PDEc. Detection of message with errors not detected by the CRC with probability PDE.

d. Correct detection with probability Pc.

These probabilities sum to 1: Po PDE PVDE Pc 1.

The channel detection parameters are the failure probabilities in the reverse link a. Channel false alarm probability PF.4.

b. Channel miss detection probability PMD.

For the detection parameters in the forward link as a irst approximation we assume that Po PDE PNDE 0 and Pc 1. As a second approximation we assume that Po PDE 0, PNVDE 0 and Pc 1 Po PDE.

The false alarm probability P.

4 and the miss detection probability PMD can be calculated for SUBSTITUTE SHEET (RULE 261 I WO 95/27963 PCT/EP95/01330 71 E/No Signal dist-ibution Noise characteristics The values of PA. and Po i- PMD for Rayleigh distributed signal in white Gaussian noise are given in Table 5 are from [1J. The results are for systems with and without diversity.

E/No Diversity PFA Po IA PD System Sd.B 1 order performance 1 10- .98 10- 4 .99 Fair 2 10- 8 .98 10- 6 .99 Fair 2 10-10 .998 10- 6 .998 Good 2 10- 1 0 .9996 10- 6 .9998 Excellent Table 5: System performance versus E/No and diversity order.

4.6 Numerical results The CVD protocol performance where estimated for simple distribution of the characteristic values of the remote stations. For the example C in 4.4 an worst case of uniform distribution of the characteristic values of the remote stations were assumed. The probability of no failure P,VF is a lower bo-ad for the probability of protocol correct detection PPCD and is given for this example in Taule 6.

E/VNo Diversity PA PD PNF System [dB] order peormance 1 10-" .98 .95 Fair 2 10-8 .98 .95 Fair 2 10-10 .998 .99 Good 2 10-10 .9996 .999 Excellent Table 6: System performance versus E/No and diversity order.

SUBSTITUTE SHEET (RULE 26) I Y- L-II WO 95/27963 PCTIEP95/01330 72 Conclusions In multiple sources (remote stations) system the information rom the sensors is gathered.

processed and disseminated. The information gathering process requires a lot of communication resources such as time and frequency.

The CVD method may be used for two types of applications: a. Search of preferred or optimal remote stations.

b. Surveillance.

The communication resources required using CVD method are lower by orders of magnitude relative to conventional method. The improvement obtained is proportional to the number of remote sources.

Appendix: Communication Reverse Link Description The communication reverse link has an allocated bandwidth and a multiple access communication protocol with M 2 m Two types of operation are described: Pulse transmission. Unmodulated pulses are transmitted by the remote sensors and the base station has to detect their presence.

Data transmission. Digitally modulated signals are traiu.m-tted by the remote sensors and the base station has to detect their message.

5.1 Pulse transmission Pulse transmission is used in the discrimination mode of CVD method. The multiple access communication protocol is TDMA or FDMA or TDMA/FDMA such that in a time frame T and ba-dwidth B there MT time slots and MF frequency slots. The total number of slots or in fact virtual channels, is MT MF M 2m.

The time slot duration is Tp and one pulse may be transmitted by a remote station in a time slot. If the time accuracy is T, and the frequency accuracy is Fa then: T Mr(Tp T.) B M( F) (8) BT M( +TF T.Fa) The last equation can be simplified BT Mg (9) where g is the loss factor due to uncertainties.

SWe may conclude that in a bandwidth B there are M channels and in a time frame T one pulse may be transmitted per channel.

SUBSTITUTE SHEET (RULE 26) _I WO 95/27963 PCT/EP95/0 133() 5.2 Data transmission 7 3 73 Data transmission is used in the identification mode of CVD method. and in the poliza method. The multiple access commuaiction protocol is TDMA or FDMA or TDMA/'FDM.

such that in a time frame Ta-trme and bandw'dth B:here M time slots and M.F frequerzy slots. The total number of slots or in fact virtual chanaels, is M- VIF Mf 2.

A message of N6 bits, each of duration Tb, may be transmitted by a remote station in a time slot with a duration VbT Tr,,a,b where is the preamble duration. If :He time accuracy is T and the frequency accuracy is F, then: Td-frw,,, IIT(NbT Tpeimble T) B M BTdata-rame MN1(l TCFL+ Tamb+T F) The last eauation can be simplified BT g() where g is the loss factor due to uncertainties and T Td(..f~n Nb We may conclude that in a bandwidth B zhere are 2: channels and in average durig seconds one bit may be transmitted per channel.

In sections 5.1 and 5.2 the following entities have the same symbolic notation T: The duration of the reply in the reverse channel in Phase 1 of CVD search method. The reply consists of an unmodulated pulse which is detected by the base station with a probability of detection Po and a probability of false alarm PFA.

Bit duration in Phase 2 of CVD search method.

Bit duration in polling method.

The same duration T was assumed for all three entities to simplify the comparison, however, this is a reasonable assumption. As an illustration the performance are shown in Table for a Rayleigh fading channel. The same transmission power was assumed such th-t the energy of a bit Eb is the same as the rcply pulse energy E, E, E. The performance of the system, P6 for the data detectizn in Phase 2 of CVD search method or in polliz method and PFA and Po for pulse presence detection in Phase 1 of CVD search method.

are similar for the same conditions of E/No and diversity order. The values of PFA and FP in Table 5 are from References [1j J. V. DiFranco, W.L. Rubin Radr Detection Artech House, 1980, pp. 319 and 395.

SUBSTITUTE SHEET (RULE 26)

Claims (27)

1. A method of tracking changes at a plurality of remote stations each having a varying attribute affecting a characteristic value computed according to a predetermined procedure, involving: assigning at least one transmission slot to each of the remote stations, wherein the presence or absence of energy in a particular slot defines a single bit of information; determining, by the respective stations, of their characteristic values relative to a previously determined charactieristic value; and transmitting, by the respective stations, of the difference between their determined characteristic •values and a previously determined characteristic value that was transmitted in at least one previous transmission slot, in said at least one assigned transmission slot. 20 2. A method according to claim 1, involving ;repeating and wherein said previously determined characteristic value is the characteristic value determined in the previous performance of step eS 25 3. A method according to claim 1 or claim 2 involving periodically repeating and wherein said previously determined characteristic value is the characteristic value determined in the previous performance of step

4. A method according to any one of preceding claims wherein all the stations complete steps and before repeating step

5. A method according to any one of the preceding claims wherein a characteristic value region associated with the determined characteristic value is divided into a H:\Cowty\Keep\Nck\22162.95.doc 22101/99 ~0 75 plurality of contiguous regions and wherein transmitting the difference between the characteristic values involves transmitting a signal in one or more of the transmission slots which slot or slots indicate which of the regions contains the determined characteristic value.

6. A method according to claim 5 wherein the extent of the associated regions is established based on an expected maximum rate of change in the characteristic value of the remote station.

7. A method according to any one of the preceding claims wherein the respective stations synchronously carry o out the step of transmitting.

8. A method according to any one of the preceding claims wherein the at least one transmission slots have a 'capacity of less than 5 bits per transmission. 20 9. A method according to claim 8 wherein the capacity is no more than 2 bits. 4 A method according to any one of the preceding claims wherein determining said previous value involves: 25 assigning at least one transmission slot to each of the remote stations, wherein the presence or absence of energy in a particular slot defines a single bit of information; determining, by the respective stations, of their characteristic values; initially transmitting, by the respective stations, of a signal responsive to their determined characteristic values in said at least one assigned transmission slots, said transmitted characteristic value having a first characteristic value resolution; and subsequently transmitting, by the stations, of a signal related to their respective characteristic I{,\Cgowty\Keep\tliCk\22162.95.doc 22101199 la 76 values in said at least one assigned transmission slot, said subsequent transmission having a finer characteristic value resolution relative to said previously transmitted characteristic value.

11. A method of mapping a plurality of remote stations each having a varying attribute affecting a characteristic value computed in accordance with a predetermined procedure, involving: assigning at least one transmission slot to each of the remote stations, wherein the presence or absence of energy in a particular slot defines a single bit of information; determining at the respective stations, of 15 their characteristic values; initially transmitting by the respective stations, of a signal responsive to their determined characteristic values in said at least one assigned slot, said transmitted characteristic value having a first characteristic value resolution; and subsequently transmitting, by the stations, of a signal related to their respective characteristic values in said at least one assigned transmission slot, e* 25 said subsequent transmission having a finer characteristic 25 value resolution relative to said previously transmitted characteristic value.

12. A method according to claim 10 or claim 11 wherein said at least one assigned transmission slot is not capable of transmitting a desired characteristic value with a desired characteristi value resolution.

13. A method according to any one of claims 10-12 involving repeating with successively finer characteristic value resolutions until the characteristic value is transmitted with a desired characteristic value resolution. H:%CgowtY\KseePNNcck\2162.95.doc 22/01/99 I I I 77

14. A method according to any one of claims 10-13 wherein the finer characteristic value resolution less than twice as fine as that of the previous characteristic value resolution. A method according to any one of claims 10-13 wherein the finer characteristic value resolution is less than twice as fine as that of the first characteristic value resolution.

16. A method according to any one of claims 10-16 wherein the respective stations synchronously carry out the step of initially transmitting.

17. A method according to any one of claims 10-15 wherein the respective stations synchronously carry out the step of subsequently transmitting. 20 18. A method according to any one of claims 10-17 wherein all the stations complete the step of initially "transmitting prior to any of the stations performing the step of subsequently transmitting. 25 19. A method according to any one of claims 10-18 wherein a characteristic value mapping space is divided into a fixed number of portions and wherein said initial transmission indicates which of said portions contains the characteristic value. A method according to claim 19 wherein a portion of characteristic values, somewhat larger than said initially transmitted portion, is divided into a fixed number of portions of a smaller size and wherein said subsequent transmission indicates which of said portions of smaller size contains the characteristic value. 1:\Cgowty\Keep\Nlck\22162.95.doc 22/01/99 s I I L II I 78

21. A method according to claim 19 wherein a portion of characteristic values, substantially equal to said initially transmitted portion, is divided into a fixed number of portions of a smaller size and wherein said subsequent transmission indicates which of said portions of smaller size contains the characteristic value.

22. A method according to any one of the preceding claims wherein the remote stations are identified by: first assigning a transmission slot to each of a relatively large number of remote stations; and transmitting, by those stations which meet a given criteria of a signal in their assigned transmission slot. .23. A method according to claim 22 wherein further assignment of transmission slots additional to said first assigning is made only to those stations which meet the criteria.

24. A method according to any one of the preceding claims including repeating at least one step of transmitting, at a coarser characteristic value resolution. 25 25. A method according to claim 24 wherein said step of repeating is performed, when a valid signal is not received from a remote station.

26. A method according to any one of the preceding claims wherein each said transmission slot carries information only by said presence of absence or energy.

27. A method according to any one of the preceding claims and including determining, by a central station, of the presence of the transmitted signal in a transmission slot, involving: determining the strength of a signal in each of a H:\Cgowty\Keep\Nlk\22162.95.doc 22/01199 ~II 79 plurality of subchannels into which the transmission slot is divided; and providing an indication of the presence of the signal in the transmission slot if at least one of the determined signal strengths is greater than a threshold.

28. A method according to any one of the preceding claims and including determining by a central station, of the presence of the transmitted signal in a transmission slot, involving: determining the strength of a signal in each of a plurality of subchannels into which the transmission slot is divided; and providing an indication of the presence in the 15 signal in the transmission slot if the sum of the determined signal strengths is greater than a threshold.

29. A method according to any one of claims 1-27 and including determining by a central station, of the presence of the transmitted signal in a transmission slot, involving: determining the strength of a signal in each of a plurality of subchannels into which the transmission slot is divided; and 25 providing an indication of the presence of the signal in the transmission slot if at least one of the determined signal strengths is greater than a threshold. A method of according to any one of the preceding claims wherein the transmitted signal is a CW signal.

31. A method according to any one of the preceding claims wherein the characteristic value is the location of a mobile remote station.

32. A method according to any one of the preceding claims wherein transmission of signals is by on-off keying. H;\Cgowcy\Keep\NlC k\2216 .95,doc 22/01199 I IY 80

33. A method according to any one of claims 27-29 wherein determining the strength of the signal involves: receiving of the transmitted signals by the central station; and feeding a signal derived from the received signals to a series of energy detectors each having a bandwidth related to the bandwidth of the subchannels.

34. A method according to claim 33 wherein determining the strength of a signal involves: feeding a signal derived from the received signals to a series of energy detectors each having a bandwidth related to the bandwidth of the subchannels

35. A method according to claim 33 or claim 34 wherein the energy detectors are correlation type receivers. *e0*90

36. A method according to claim 33 or claim 34 20 wherein the energy detectors are matched filters. *9 0 o0

37. A method according to any one of claims 34 to 36 including deriving said derived signal by downconverting the signal to a low frequency signal. .9 0

38. A method according to claim 33 of claim 34 wherein the step of feeding a signal involves performing a digital Fourier transformation on the signal. 3i 39. A method of tracking changes as claimed in claim 1 and substantially as herein described with reference to the accompanying drawings. A method of mapping a plurality of remote stations as claimed in claim 11 and substantially as herein described with reference to the accompanying drawings. I, I r 81 Dated this 8th day of February 1999 HELEN LEW and YOSEF MINTZ By their Patent Attorneys GRIFFITH HACK Fellows Institute of Patent and Trade Mark Attorneys of Australia 09 .06:0, c- 1 'I eI61