AU638160B2 - Digital radio-telephony process and apparatus, particularly for cellular radiotelephone for communication with mobiles - Google Patents

Description

Regulation 3.2(2)

AUSTRALIA

Patents Act 1990 86N~

ORIGINAL

COMPLETE SPECIFICATIOii STANDARD PATENT Application Number: LodgeC~: 0S S 0 000 0 00

S

606 a.

0 0 S*6 0 .0 6

S

S.

60 0 0 *6 Invention Title: DIGITAL RADIO-TELEPHONY PROCESS AND APPARATUS, PARTICULARLY FOR CELLULAR RADIOTELEPHONE FOR CCMMUNICATION WITH MOBILES 0 0 00.0 00 S. 0 0@ 0@ 0 O *0

SOS.

0 0005 00 S Oe 05 The following statement is a full description of this invention, including the best method of performing it known to us 1 I S1 DIGITAL RADIO-TELEPHONY PROCESS AND APPARATUS, PARTICU- LARLY FOR CELLULAR RADIOTELEPHONE FOR COMMUNICATION WITH

MOBILES

The invention generally relates to processes and apparatuses for digital radio-telephone communication whose implementation requires broadcast by a stationary infrastructure, periodically and at high rate, towards transmitter-receiver stations, serial numbers of fcequencies which constitute any sub-set, having a cardinal number or range which is at most equal to a predetermined value n, of a set of predetermined frequencies having a 15 cardinal number N.

The invention is particularly suitable for use in *s radio-telephone systems (the term "radio-telephone" being construed broadly and as relating to data transmission as well as to speech transmission) known as "cellular", whose standards are being prepared by the Conference Europ6enne des Administrations des Postes et T6lecommunications. A description of the general features of such systems may be found in the paper "Le systeme cellulaire numerique europeen de comnunication avec les mobiles", by B. GHILLEBAERT et al, in Echo des Recherches, n* 131, first quarter 1988. Those aspects of digital cell' -ar radio-telephone particularly concerned by the present invention are for instance described in recommendation e• GSM 04.08 "Mobile radio interface layer 3 specification", published in March 1990 by the special group "Public Mobile Systems of Radio-telecommunication" of CEPT.

In a cellular radio-telephone system as hereinbefore defined, the infrastructure should transmit, on broadcasting channels, two lists of frequencies which are repeated at regular intervals. A first list is that of frequencies on which, in a given cellule or zone, the mobile station is to operate with frequency shift keying.

A second list is that of the frequencies of the broadcasting channels that the station should monitor, in stand-by or waiting condition, for cellule selection, and in connected condition, for measurement and comparison needs.

As an indication, it is presently envisioned that the list of frequencies to be monitored will be communicated at a rate of about one per second. It is also envisioned to communicate, to a mobile transmitter-receiver station, in response to a request for allocation by the ~latter, the serial numbers of the frequencies which are dedicated to it for transmission, such frequencies cons- 1 5tituting a sub-set, having a cardinal number at most equal to n 16, of a set of N 210 1 024 frequencies.

The transmission mode should be such as to authorize adoption of any one of the possible sub-sets, Sw which excludes the process which would consist, in the case of a limited number of sub-sets, to dedicate an identification number which would be alone transmitted to each. Individual transmission of the serial number of each frequency of the selected sub-set would require, in the most unfavorable case, to encode 16 serial numbers, each on 10 bits.

It is a general object of the present invention co reduce the volume of binary data necessary for transmitting the serial numbers of the frequencies to be used, the serial numbers being possibly considered as elements belonging to a set It is a more particular object to provide a simple solution to the problem when the number of elements N is equal or slightly higher than a whole power of 2, i.e. when N is equal or slightly higher than 2 k (k being an integer).

For that purpose, the invention particularly provides a process of digital radio-telephone communication, including the steps of broadcasting, from a stationary infrastructure toward mobile transmitter-receiver

-YILI~-

3 stations, the serial numbers of working frequencies which constitute any sub-sets, having a cardinal number at most equal to a predetermined value n, of a set of predetermined frequencies of cardinal N; identifying, in each station in waiting condition, frequencies of the sub-sets from the serial numbers and causing the station to operate on the identified frequencies, characterized in that, for broadcasting the serial numbers of frequencies belonging to a set whose cardinal N is comprised between 2 k- and 2 k, the number V of the frequency for which half of the remaining frequencies has a lower serial number and half of the remaining frequencies have a serial number higher than modulo (total number of frequencies) is identified; the serial number of the selected frequency is identified over k bits; the serial numbers of the two divided groups of frequencies are renumbered; determination of the serial number of the frequency is repeated on each group and the two partition frequencies are defined on k-i bits; the operations of renumbering, determination and definition are repeated until the serial numbers to be transmitted are exhausted; all serial numbers of the partition frequencies are transmitted according to a predetermined o 25 sequence and the working frequencies are determined from the serial numbers, in the station.

The terms "lower" serial number and "higher" serial number should not be construed as indicating a relation in natural order, but rather as indicating that e0 30 the serial numbers belong to that half of the original set which follows the value or which precedes the value, respectively, in a circular arrangement.

The process is applicable for any value of N.

•However, it is of particular interest when N may be written as 2 kl since that removes the need for mixed radix arithmetics and consequently simplifies computation by substituting mere shifts to multiplication or division t 4 operations.

That particular case is frequently encountered.

Frequently the number of possible frequencies, i.e. the number N of elements in the set may be written as 2 k, due to reasons related to the use of binary mode transmission. Then the number of elements may easily be reduced from 2 k to 2 k- 1 by identifying one of the frequencies, arbitrarily selected in the set, with a single bit which is 0 or 1 depending whether the respective frequency is present or not.

When the number N of elements in the set is slightly higher than a whole power 2 k, another approach consists in using the above-defined process for 2ki elements (the first of the last elements of the list for instance) and identifying the serial numbers of the other elements by bit mapping.

When the number of elements in the set may be written as 2 k-i and when the cardinal number of the sub-set is optimum, then the process according to the invention makes it possible to carry out transmission with a total number of bits which is close to the theo- S. retical minimum. The gain is not as high in the particular case of cellular digital radio-telephone as provided 25 by the standard, but it is sufficient for authorizing transmission of all serial numbers of the 16 frequencies (maximum number) while are necessary, among 1 024, with 128 bits, i.e. 24 bytes, even when 2 bits should be S* reserved as a header.

The invention further provides a mobile radiotelephone station comprising emitter-receiver means capable to operate at a plurality of frequencies defined by serial numbers, and constituting a sub-set of frequencies, the sub-set having a cardinal number at most equal to n, belonging to a set of frequencies whose serial numbers constitute a set having a cardinal number N, suitable for implementing the above-defined process, characterized in that the movable station comprises means for matching respective predetermined frequency with each of the serial numbers in the set and a decoder capable to provide the actual serial number, in clear, to said means, from a message resulting from the above-defined process.

The means may consist of a processor, which is anyway necessary in the movable station for fulfilling 10 other functions, when the different possible frequencies i0 can be deducted from one another by an algorithm including the serial number. Such means may have other constructions and for instance may use a matching table stored in a programmable read only memory.

15 The invention will be better understood from the following description of a particular embodiment and of modified embodiments. The description refers to the accompanying drawings, wherein: Figures 1 and 2 are "loop" sketches indicating 20 how the first and second nodes are selected; 2O Figure 3 is a representation of a possible broadcasting format.

The description will first be given of the nrocess (and of the corresponding decoding process) making it possible to reduce the number of bits 25 Snecessary for transmitting the n serial numbers among N frequencies before the encoding and decoding mode for N 2 k-1 and a corresponding program are given.

3 •Encoding is equivalent to a process of identification of elements, whose maximum number is n, belonging 30 to a sub-set by a partition method including a tree selection approach, the code of an element depending on previously selected elements.

By selecting a "parent" element, a left group and a right group are formed and in each group a "child" is according to a selection criterion which is the selected according to a selection criterion which is the 6 same as for the parent arbitrarily, the second child may be considered as being the left-hand child. It is smaller than the parent and may be encoded with a number of bits which is lower by one unit. The right-hand child is greater than the parent but can be reduced to a lower value by a subtraction operation, thereby making it possible to encode it with the same number of bits as the right-hand one.

The first step of the partition process is schematically illustrated in Figure 1, on which seven elements constiting a sub-set to be identified and to be encoded are illustrated, the seven elements being among R elements R-1. A node will be defined as the value V which separates two groups the right-hand group, from V 1 to V L the left-hand group, from the preceding value, increased by one, to V-l.

Referring to Figure 1, the two groups may be defined on a loop, value R-1 being followed with the value 0.

It can be demonstrated that it is always possible to find an element whose value V is such that there are exactly L values in the left-hand group. That convention will generally be used for defining V. Another convention would however also be possible.

After the first partitioning has been carried out, a first and secnnd child, each in one of the groups, are determined. Arbitrarily, it will be assumed in the 30 following that the second child is the left-hand one, the third child is that at right.

Referring to Figure 2, first the elements of the left-hand group are renumbered beginning from V L Similarly, the elements of the right-hand group are renumbered from V.

The process is then carried out again and makes it 7 possible to determine the new child V1, using the same criterion as above.

At each partition, it is possible to encode the children with a number of bits which is lower by one unit than that necessary for the parent if initially k bits are necessary, k-i bits are sufficient for the two immediate children, k-2 bits for the four following children, etc.

Selection of the nodes may be carried out by a simple, however slow, process consisting in carrying out a test on each element in turn; however fast, more complex, algorithms can also be used.

It can be seen that the values assigned to all nodes have a number of bits which is independent from the values to be encoded, i.e. from the serial numbers, and depend on predetermined conventions only (stored on emission and reception) and of the cardinal number of the set.

As indicated above, particularly favorable results are obtained when the cardinal number of the initial set is equal to 2 k-l, k being any integer. Then k bits are just sufficient and necessary for encoding the first node S"while k-l, then k-2 bits, etc. are necessary for the g 25 nodes of the following generations. The structure is then wholly binary, and the complexity resulting from encoding and decoding with a mixed radix is avoided.

Decoding operations in the movable station, where S" the selected conventions are stored, are reverse from those which have just been described. A processor first S 5 identifies the first node, whose serial number is directly given in binary code. Then it carries out the numerations required for reconstructing the first two children.

Then the processor, using a predetermined algorithm which associates the frequency values with their serial numbers, delivers to the emitter-receiver portion of the station the values of the frequencies which correspond to 8 the channels to be monitored and/or co the channel on which transmission should be carried out for communication with the infrastructure.

The performances which are obtained immediately appear when the cardinal number of the set is equal to 2k-l. In that case, which is typical of the situation for cellular radio-telephone, where N 210-1 1 023 and where n 16, 10 bits are necessary for the first node, 9 bits are necessary for each of the following two nodes, 8 bits are necessary for each of the following four nodes, and similarly until six bits are necessary for the single node of the fourth generation. The total number is consequently 122 bits.

As indicated above, 1 supplemental bit is sufficient for increasing the total number of frequencies among which selection can be carried out to 1 024.

*For any value N of the cardinal number of the set, 2 the necessary number of bits is

N

N log n log i log 2 i 1 log 2 •where designates the whole part of x.

30 The gain on the number of bits, as compared to an encoding mode consisting in giving all values constituting the sub-set as a number having the maximum number of bits which may be necessary, depends on the number of the elements. As an example, some values are given thereunder: 1 2 3 gain (bits) 0 1 2 34 66 9 and the following table gives the number of bits which is necessary in a number of typical situations Range of Number of elements Number of bits the set in the sub-set required 1 023 14 116 210-1 16 122 511 20 106 22 114 29-1 125 126 As an example, there will now be given a sequence of partiting operations necessary for sending, to a mobile .station, a particular sub-set of elements, consisting of sixteen serial numbers belonging to a set of 210 serial 25 numbers of frequencies having the values f 0 f 1023 It will be assumed that the sub-set consists of frequencies whose serial numbers are 12,70,121,190,250,320,401,475, 520,574,634,700,764,830,905,980.

For decreasing the cardinal number or range of the Sset from 1 024 to 1 023, one frequency, arbitrarily selected, is identified with one bit. Presence or absence of the frequency in the sub-set is indicated by sending a 'single bit. That frequency has to be stored in the mobile 3:"I station. It is for instance ,f ou fl 023. Consequently, the maximum number of elements to be transmitted is sixteen and they consist of sixteen numbers between 1 and 1 023.

A first node then consists of that element V 520 whose "opposed value is [520 nodulo (210-1) i.e. 8: it is seen that there are eight elements between 9 Sand 519 (belonging to the left-hand group) 12,70,121,190- 250,320,401,475) and seven elements between 521 and 8, as indicated by the following table.

TABLE I 520 oposed 8) 12 574 (61) 634 121 (112) 700 190 (181) 764 250 (241) 830 320 (311) 905 401 (392) 980 475 (466) Arbitrarily, in the case (which is that of the above example) where the number of elements to be parted is even, the node is so selected that there is one element o. more in the left-hand group. The contrary possibility could 2 be used; It is only essential that it is stored in the station.

The first node is encoded with 10 bits, even when a lower number of bits would be sufficient.

The second node consists of the child of the first 30 node, arbitrarily in the left-hand group, renumbered starting from rine, i.e. from the opposite of 520. Renumbering is indicated between parentheses in the above table. It is seen that value 112 (corresponding to the element which was initially numbered 121) separates the group into tw, subgroups of four and three elements respectively.

The third node consists of the child of the first in the right-hand group. Then renumbering is carried out 11 from 521, again from 1 to 29-1 511 which gives seven elements numbered 53,113,179,243,309,384 and 459.

Value 53 separates that group into two sub-groups each of three elements, comprised between 54 and the value 309 opposed to 53.

In the new numbering, the second and third nodes can be encoded with nine bits, since their values are lower than 512.

It is proceeded with a method to obtain the following nodes. Arbitrarily, it will be considered here that each even node 2m is the left-hand child of node m while the odd node 2m+1 is the right-hand child of node m.

The successive nodes are then identified by the values of table II TABLE II Node Value Number of bits 1 520

N

9 63 68 8 6 7 59 0 8 9 57 2 11 59 7 12 22 13 74 14 2 15 Mixed radix .ncoding is very simple in that particular case since it is such that there is an exact correspondence between values and words having a decreasing number of bits.

The above values, plus one, are then encoded in the appropriate order, then transmitted as 10 bits, then 9 bits, then 8 bits, then 7 bits and, last, (for the single elements for a sub-set of 16 elements) 6 bits: Conversely, decoding is very simple. For instance, for the second node, the original value is: (520 511 113) mod 1 023 121 while for node 11, the original value is {520 511 [113 (69 60) smod 255] smod 511} mod 1 023 i.e. 250.

In the above formula (n smod p) designates the integer r for which there exists a number d such as 1 r p and n dp r.

That formula is used for differenciating from modulo, which is such that 0 4 r p-1.

S: 25 Encoding and decoding can easily be carried out by software. A decoding program corresponding to a specific situation and language will now be given, as an example only.

w i will be used to designate the ith word appearing 30 in the field. When w i is equal to zero, the description of the sub-set is ended and all following bits in the field are set to zero. When the iat of the broadcast message is of the type illustrated in Figure 3, the serial numbers of the sub-set are computed as follows.

Bit fo indicates whether frequency FO is present or not if present, it is immediately identified.

The serial numbers of the 16 frequencies F 1 to F16 I 1 13 or fo to f 15 (the number possibly being lower than 16) are given by: F, W, smod.1023 F, (W 1 512 W 2 smod 1023 73 W 3 smod 1023

F

4 512 (W 2 256 W 4 smod 511) smod 1023 FP (W 1

(W

3 256 smod 511) smod 1023 Fs 512 (W 2 smod 511) smod 1023

F

7 smod 511) smod 1023 F, (W 512 (W 2 256 (W 4 128 W, smod 255) smod 511) smod 1023 F, (W 3 256 (W 5 128 W smod 255) smod 511) smod 1023 FPo 512 (W 2 128 smod 255) smod 511) smod 1023 128 W) smod 255) smod 511) smod 1023 512 (W 256 (W 4

W,

2 smod 255) smod 511) smod 1023

F,

3 (Wi 256 smod 255) smod 511) smod 1023 F,4 512 4 smod 255) smod 511) smod 1023 F,s

(W

7 smod 255) smod 511) smod 1023 512 256 (W 4 128 64 smod 127) smod 255) smod 511) smod 1023 Each in bytes 2 to 17, constitutes the coded 00 value of a non-negative integer in binary format. If W(i) is zero for a value 1, then words W(1+1) to W(16) are also equal to zero. The adopted symbols are as follows 30 indicates addition of natural integers, 30 indicates multiplication of natural integers, n mod m indicates the remainder of the euclidian divison of n by m, i.e. 0 (n mod m) m-1, which implies that there exists a value p such as p k*m (n mod m).

I *1 14 The program may be as follows, using ADA language.

The declarative parts have been skipped since they are cbvious INDEX K7 J GREATEST POWEROF2LESSEROREQUALTO(INDEX); N W(INDEX); while INDEX>l loop PARENT INDEX J/2 (INDEX-J) mod if2 *INDEX 3*J then left child (N W(PARENT INDEX) 1024/J 1) mod (2048/J 1) 1; else right child N (N W(PARENTINDEX) 1) mod (2048/J- 1) 1; end if; INDEX PARENTINDEX; J J/2; end loop; F(K) Encoding is :arried out by a program which reflects the decoding process; consequently the encoding program may S~.be directly deducted from the decoding process which has just been given.

sNumerous modifications of the invention are possible and only some of them will now be briefly mentioned.

.:2 coding A first modificatibn consists in simplifying encoding by not fully using the mixed radix base arithmetic concu.pt, but instead by encoding a given value with an integral number of bits. For instance it is then possible to use 9 bits for encoding a value comprised between 0 and 30 350. That simplifies encoding and decoding since the divisions and multiplications are replaced with binary shifts.

The cost, as regards the number of bits, is about 1/2 bit S per encoded value for the above example, the number of bits which is required is the,. increased by an amount which is zero when N is equal to 2 k_ 1 whatever k if it is an integer.

4 l 4 Still another process making it possible simultaneously to preserve purely binary coding and performances when the range or cardinal number of the set is only slightly higher than 2 k-l consists in using a bit map for the values beyond 2 k-1.

Another modification makes it possible to adapt the above described encoding-decoding process to the situation where the possible sub-sets represent only a predetermined portion of all possible sub-sets of the set of N elements.

The modification implies to give, at the beginning of the format, an indication of the range or of the portion of the set within which all elements of the sub-sets to be encoded are found. Several approaches are possible.

A first solution consists in giving the lowest value and the highest value of the domain in which the elements are to be found. That consumes [2 log bits.

The remaining available bits are used for encoding the values. The relationship between the value k such that all sub-sets having at most k values can be coded and the domain (maximum value minimum value) is as follows, assuming a maximum number of 124 bits Domain max.number of elements 1023 S511 17 255 21 o 30 Obviously, it is more advantageous to use bit map 30 coding if the domain has 106 elements or more.

A second solution consists in firstly encoding the lowest value (with ten bits in the above example), then the range (highest value minus lowest value), for instance *with three bits. The relation then becomes Domain max. number of elements 1023 511 255 127 For a domain having less than 112 elements, it is preferable to enccde on a bit map basis.

It is seen that there is neither a gain, nor a loss for extended domains, but there is an advantage for domains having less than 150 elemenfts.

Last, still anotheic modification consists in not giving the exact first value, but rather indicating only a first bound among a plurality of bounds, indicating that the domain will not be lower than the bound. The bound may for instance be selected among 21 different values in the above-identified case.

S

S S 0e* b

S.

S S Sb 6

S

b Sb b.

S S 5- S

S

S.

5555*5

S

Sb..- S S S .5 S Sb bS i.

*5 S S

S.

bb@SSb b I epti-fy tha tW and the preceding,\ tpue and lo j of'pa_8 uO tho 8peoifiotIofl originaII4 1Q0E4Iod pages are I

Claims (7)

1. Process for digital radio-telephone communication, including the steps of broadcasting, from a stationary infrastructure toward mobile transmitter-receiver stations, the serial numbers of working frequencies which constitute any sub-set, having a cardinal number at most equal to a predetermined value n, of a set of predetermined frequencies having a cardinal number or range N comprising between 2 k- l and 2 k (k being an integrer., identifying, in each mobile station in waiting i 5. condition, the working frequencies of the sub-set by deriving them from the serial numbers and S- causing the station to operate on the identified frequencies, wherein step includes determining the serial number V of the frequency for which half of the remaining frequencies have a lower serial number and half of the remaining frequencies have a serial a sa number higher than -V modulo (total number of frequencies), identifying the serial number V of the selected frequency 421% with k bits, renumbering the frequencies in the two groups by changing their serial numbers, repeating determination of the serial number of a partition frequency on each group and defining the two partition frequencies with k-l bits repeating renumbering, determination and definition until the serial numbers to be transmitted are exhausted transmitting all serial numbers of the partition frequencies according to a predetermined sequence stored in the mobile stations and determining the working frequencies from the serial numbers, in the stations. eas *of *A a a 0*6 go 0 d a .a 04 r 0

2. Process according to claim 1, characterized in that it is used for broadcasting the serial numbers of a set of 2 k-1 frequencies while transmitting the possible additional serial number in a different way.

3. Process according to claim 2, characterized in -Pfreu encies that the number of eleme o is decreased from 2k to 2k-1 by identifying one of the frequencies, arbitrarily se- lected within the set, with a single bit which is set to zero or to one depending whether the relevent frequency is present or not.

4. Process according to claim 2, characterized in r e. e-pcle.s that, the number N of el being only slightly higher than the whole power of two, it is used for encoding 2 elements (for instance the first elements 15 or the last elements of the list) and the serial numbers of the other elements are identified as a bit map.

Mobile radio-telephone station for a cellular digital radio-telephone system using frequency shift modulation comprising emitter-receiver means over a plurality of frequencies defined by serial numbers and constituting a sub-set of frequencies having a cardinal number at most equal to n, belonging to a set of fre- quencies whose serial numbers constitute a set having a cardinal number N, suitable for implementing the process according to claim 1, characterized in that the mobile station comprises means for monitoring a broadcasting channel, frequency generating means for establishing a correspondence between a predetermined frequency to each of the serial numbers of the set and a decoder capable to provide the actual serial number, in clear, to said means, from a message containing the list of numbers encoded by the process according to claim 1.

6. Station according to claim 5, characterized in that the generating means comprise a process which generates the different possible frequency through an algorithm involving the serial number. *i 0 19

7. A process of digital telephone communication as illustrated in and as described with reference to the accompanying drawings. DATED this 19th day of August 1991. MATRA COM~MUNICATION WATERMARK PATENT TRADEMARK ATTORNEYS "THE ATRIUM" 290 BURWOOD ROAD HAWTHORN. VIC. 3122. 66 Ue 6 0 *0 S. a.. 06 S. 0 000 S 55 S *0 S. Be em C, So a a a. *OeS 0 *6g~ 51' 0 0a OS