EP1350408A1

EP1350408A1 - Method and system relating to positioning of a mobile station

Info

Publication number: EP1350408A1
Application number: EP01270073A
Authority: EP
Inventors: Per Wirdemark; Magnus Sicking; Hans Grubeck
Original assignee: Cellpoint Systems AB
Current assignee: Cellpoint LLC
Priority date: 2000-12-04
Filing date: 2001-12-04
Publication date: 2003-10-08
Also published as: AU2002221221A1; WO2002047421A1

Abstract

The present invention relates to a method and system for determining a position of a Mobile station (190) in a cellular communications network (150) comprising a number of cells: a serving cell and a number of neighboring cells, wherein a set of date on each cell and neighboring cells is available. It is characterized by, for each cell, calculating a cell center and cell size, calculating a Point Distributuion Array (PDA) for a number of points in a coordinate system, and based on said PDA and statistical calculation determining the position of said mobile station.

Description

Title

METHOD AND SYSTEM RELATING TO POSITIONING OF A MOBILE STATION

Field of the invention

The present invention relates to a method and system for determining the position of a Mobile Station in a cellular communications network, which comprises a number of cells: a serving cell and a number of neighboring cells, wherein a set of data corresponding to each serving cell and neighboring cells is available.

Background of the invention

Nowadays, the importance of GSM (Global System for Mobile communications) based communication network is increasing on a daily basis. Alongside the increased usage of mobile phones, there are other areas where GSM can be used. One of the areas is mobile positioning. A mobile positioning centre is a flexible link between commercial as well as security-related applications. There is a number of activities in this area. The United States Federal Communications Commission (FCC), for example, requires that, by year 2001, all mobile communication networks should be able to locate a caller's mobile unit requesting emergency assistances.

There are several other areas where mobile positioning systems can be used. The advantage of mobile positioning systems is that the existing wireless telephony network can be used to obtain the position. Another reason that makes GSM-positioning so attractive is that GSM network delivers a number of data, which can be used in different positioning algorithms.

There are many efforts are made on these issues. Many make serious attempt to develop positioning methods that fulfill the market expectation.

The prior art present a number of examples: EP 0 290 725 is based on the object of keeping the technical expenditure for determining the approximate location of a mobile radio station low in a cellular radio telephone network. 2.2 To achieve the object, an individual base station number contained in a data message of the base station is evaluated by the mobile radio station and, if necessary, embedded into a voice signal to be sent out by the mobile radio station together with the subscriber number as location information. Voice signal and location information are transmitted via the base station and a line to a central evaluating station, which contains an evaluator for the location information, that is to say the base station number and, if necessary, the subscriber number. The base station number specifies that the mobile radio station is located in the radio cell of the relevant base station.

In European patent application No. 0 895 435, a method is disclosed of analyzing lists of neighboring cells in a cellular telecommunications system comprising a plurality of active mobile stations and a static network, the static network having a first cell and a plurality of cells neighboring the first cell, each of the cells having a base station. The method is applicable to GSM and GSM-like systems. In the case of a GSM system, the analysing method may comprise the steps of: extracting from the static network the GSM MEAS RES produced by the mobile stations in said first cell and producing a reporting list including, for each position in the GSM BA(SACCH [Slow Associated Control Channel]) list, the number of times that any of the base station identifiers has been reported for that position; extracting from the static network GSM HA DO CMD messages for handovers from said first cell, and producing a handover list of the GSM BCCHs and corresponding GSM BSICs in the extracted handover messages; correlating the reporting list and the handover list with respect to the BCCHs; and analysing the correlated lists to determine whether any of the control channels is affected by bad frequency planning.

WO 99/41854 describes a method and system for facilitating the timing (e.g., the known relative timing differences) of base stations (BSs) (BS1, BS2, and BS3) in asynchronous CDMA mobile communications systems. A plurality of mobile stations (MSs) (MSI, MS2, and MS3) measure the relative time differences between various pairs of BSs, and these measurements are stored by the BSs. A source BS sends to an MS, in a neighbor list message, estimates of the relative time difference between the source BS and each of the BSs on the neighboring cell list. Each BS on the list can maintain a relative time difference estimate table, which can be updated continuously from the reports received from MSs. Subsequently, the BSs can send entries from this table to the MS in the neighbor list message. Using this novel technique, the BSs have known relative timing differences. Consequently, when the MS initiates a cell-search for a candidate BS, the MS already has an estimate of the timing of that BS as compared to its source BS. As such, the resulting cell-search procedure has a lower level of complexity and thus can be accomplished much quicker than with prior procedures. In addition, the relative time difference estimates can be compared with corresponding time differences that are measured by a second mobile station. Based on this comparison, the propagation delays of signals between the second MS and various BSs can be calculated to determine the position of the second MS.

According to WO 99/46 950 standard mobile cellular telephones use shortened burst signal transmissions to synchronize the transmission of a call over a cellular telephone network. These shortened bursts as well as other traffic channel bursts can be used to locate the position of the cell phone, hi operation, a primary base station sends a traffic channel designation message to a selected mobile phone which then initiates selected traffic channel burst transmissions, such as shortened bursts, over an interval of five seconds or less. The primary station and at least two other neighboring base stations receive these burst transmissions and determine their respective distances from the cell phone based upon the time of arrival of the bursts or some other suitable distance related measurement. A command center determines the locations of the cell phone by triangulating the distance measurements of the base stations.

However, none of above-mentioned documents suggest a solution for enhanced position according to the present invention. Nor a combination of the documents guides a skilled person to reach the solution according to the present invention. Summery of the invention

The main object of the present invention is to provide a positioning method and arrangement, which allows very accurate positioning of a mobile station in a cellular communications network.

For this reason, the initially mentioned method is characterized by: calculating for each cell a cell center and cell size, calculating a Point Distribution Array (PDA) for a number of points in a coordinate system, and based on said PDA and statistical calculation determining the position of said mobile station. The PDA consists of a number of points distributed around said serving cell coordinates where each point rests at a specific distance and direction relative said serving cell. The PDA calculation is carried out by creating a PDA position and populating the PDA with a signal strength data. Both the direction and a specific distance are calculated from an index value with respect to said number of points.

The method further comprises the steps of calculating a cost weighting function for a number of points by: Loading an initial PDA, calculating a Cell Location (CL) Penalty Function for each point in the array, calculating an Absent Neighbor (AN) Penalty Function for each point in the array, looking up and loading relevant Timing Advance (TA) Penalty Function, calculating an Absolute Power (AP) Penalty Function for each point in the array, ccalculating a Relative Power (RP) Penalty Function for each point in array, summing up Merged Penalty Function with weighting factors, and returning coordinates of point with lowest penalty. The positioning uses one or several of PDAs: Time of Arrival (TA) PDA, Cell Center (CC) PDA and Statistical Probability (SP) PDA. The TA-PDA is populated with penalty values for TA relations. At least one PDA per TA value and PDA is arranged. Said CC PDA is populated with penalty values for cell centre relations. Moreover, SP-PDA is populated with a statistical probability summary of the index.

The invention also relates to a system for determining a position of a Mobile Station in a cellular communication network comprising a number of cells: a serving cell and a number of neighboring cells, wherein a set of data on each cell and neighboring cells is available. The system comprises a preprocessor, a Point Distribution Array (PDA) processing arrangement, a database and a position engine. The system is in communication with a user application arrangement. The preprocessor processes operator input data into an internal database and that in the preprocessor, the Cell Centers and Cell Sizes are calculated, based on which the initial PDAs are generated. The position engine receives position request and delivers position and accuracy estimations.

The invention can be implemented as a computer program for determining a position of a Mobile Station in a cellular communications network, which comprises a number of cells: a serving cell and a number of neighboring cells, wherein a set of data on each cell and neighboring cells is available, wherein the computer program comprises: code for calculating for each cell a cell center and cell size based on said available data, code for calculating a Point Distribution Array (PDA) for a number of points in a coordinate system, ands aid code further comprising a procedure, which based on said PDA and a statistical calculation determines the position of said mobile station.

Short description of the drawings

In the following the invention will be described in more detail in conjunction with an embodiment, illustrated in the attached drawings, in which:

Fig. 1 is a block diagram showing a cellular communications network employing the system according to the present invention,

Fig. 2 is a schematic illustration a cell structure in a communications network, used for center cell calculations,

Fig. 3 illustrates the distribution of points in a Point Distribution Array according to the invention, Fig. 4 is a flowchart illustrating the steps of positioning according to the invention,

Fig. 5 shows the cumulative distribution of the range error divided by the cell size for four different propagation environments, and

Fig. 6 shows the common curve, i.e. the normalized CER-function. Definitions

Positioning Input Data Classification

A: Cell Id Only

B: Cell Id + RxLevel for serving cell BCCH, BSIC + RxLevel for up to6 neighbor cells

C : Cell Id + RxLevel + T A for serving cell

BCCH, BSIC + RxLevel, e.g. for up to 6 neighbor cells D: Cell Id + TA (WAP standard)

Abbreviations

BCCH Broadcast Control Channel

BISC Base Transceiver Station Identity Code

Cellld Cell identity

RXLevel Transmission Level

TA Timing Advance

WAP Wireless Application Protocol

Cell Definitions

Picocell: Indoor cell

Microcell: Outdoor cell, with antenna below rooftop, possibly with lower radiated power Normal Cell: Outdoor cell, with antennas above rooftop

Umbrella Cell: Outdoor cell, with antennas normally mounted on high buildings or towers. Dualband cells in networks that does not enforce "coincidental handovers"

Detailed description of the embodiments

Fig.l is a schematic overview of a system 100, according to the invention arranged in communication with a communications network 150 The communications network 150 is a cellular network, such as GSM, AMPS, etc., function and structure of which is assumed to be known to a skilled person and not described here in greater detail. The communications network comprises a number of base station antennas 151 connected to controlling arrangement 152, function of which are assumed well known for a skilled person.

The system 100 according to the present invention mainly comprises a preprocessor 110, a Point Distribution Array (PDA) processing arrangement 120, a database 130 and a position engine 140. The system is in communication with a user application arrangement 160.

The preprocessor 110 processes operator input data and stores it in an internal database. In the preprocessor, the Cell Centers and Cell Sizes are calculated, based on which the initial PDAs are generated. The input data may comprise cell data, frequency plan, and neighbor cell definitions. A PDA consists of a number of points distributed around the serving cell coordinates where each point rests at a specific distance and direction relative the serving cell.

The preprocessor can be arranged as an offline system which can run manually each time new network configuration data is received from the operator. However, it is possible to run different parts of the pre-processor individually, for example importing a new frequency plan will not require all PDA's to be rebuilt. It may be set for automatic scheduled operation, which may require a direct connection into the operator's data environment. It may also be run iteratively in an optimization process where its output is visualized by a separate arrangement, which also allows modification of the input data.

When the preprocessor is operated "off-line", it can have lower requirements on availability and scalability than the position engine described below.

The position engine 140 receives position request and delivers position and accuracy estimates via a defined API (Application Protocol Interface). The position engine can be arranged as an "online" system, which has high requirements on availability, processing speed and scalability. These aspects are covered by the implementation.

The position engine provides the online position estimation as described below. Parts of the calculations will also cover the translation of BCCH/BSIC combinations to Cell Id's. In one embodiment, all position data external to the system are stored in, for example WGS 84. WGS 84 is an earth fixed global reference frame, including an earth model. It is defined by a set of primary and secondary parameters: the primary parameters define the shape of an earth ellipsoid, its angular velocity, and the earth mass which is included in the ellipsoid reference, the secondary parameters define a detailed gravity model of the earth.

In this case, the results from the WGS 84 need to be translated into, for example a Cartesian coordinate system for normal calculations inside the system. A preferred system is Universal Transverse Mercator (UTM). The accurate conversion between these two systems can be slightly time consuming, so an approximation can be used. It is however suggested to start with the accurate conversion and later simplify this as a part of code optimization.

The UTM grid is most appropriate for scales of 1:250,000 and larger. On large-scale maps, the simple numbers of the UTM grid make plotting precise locations easier than with the complex degrees, minutes, and seconds of latitude and longitude. For example, it is possible to readily use the UTM grid while hiking to report the location of an emergency by cellular. The UTM grid has tremendous value for emergency service organizations. UTM coordinates simply measure in meters east and north from two perpendicular reference baselines. A full UTM coordinate value defines a worldwide unique position ("map address").

In the system as much information as possible is pre-processed and stored. Preferably, but not exclusively, the pre-processor needs to parse through cell data twice, first iteration loads cell info from text files and calculated cell centres and sizes, and second iteration populates the PDAs. Fig.2 shows a diagram for the cell size calculation.

The cell radius is defined as:

If AntennaHBeamwidth = 360 l .^{j Q}. ^cellsize else

U ^cellsize where d is the distance to the closest cell coordinates within a sector +/- max(60,AntennaHBeamwidth/2) around the serving cells antenna direction, which has a celltype >= serving cell and a distance > 50 meters. If no cells are found or if the distance is larger than for example 30000, d is assumed 30000. C_ceιi_si_ze is a user defined constant. The value 0.33 is an assumed value.

The Cell Centre Coordinates is calculated as: if AntennaHBeamWidth = 360 or CellType = PicoCell BS coordinates else

BS coordinates + r * d * C_ceιicentre

Where r is a unit vector indicating the direction of the serving cells transmitting antenna, d is the distance to the closest cell as described above, Cceiicentr_e is a user defined constant, and a preferred value is 0.33.

The PDA calculations are carried out in two steps: 1. Creating the PDA position,

2. Populating the PDA with signal strength data.

A PDA consists of, for example 4000 points distributed around the serving cell coordinates where each point rests at a specific distance and direction relative the serving cell. Both the direction and distance are calculated from (in this case) the index value 0 — 4000. The distance depends on the PDA class 0 - 4, where the distance d as calculated above gives the PDA class as follows:

0 m < 2*d <= 4000 m => PDAclass 0 4000 m < 2*d <= 8000 m => PDAclass 1 8000 m < 2*d <= 16000 m => PDAclass 2 16000 m < 2*d <= 32000 m => PDAclass 3 32000 m < 2*d => PDAclass 4

The distance is then calculated as the index multiplied with 2 exponent PDA class:

Distance = index * (2 ^{PDA class}) The direction in radians is the index times square root of 2 multiplied with π:

Direction in radians = index * sqrt(2) * π

Fig. 3, illustrates the distribution of points in a PDA according to above calculations:

The preprocessor calculates the signal strength PDA for each cell. For this reason following mathematical method is used:

It is considered that the earth is a flat surface and the fix orthogonal coordinate system is e = \e_x e_y e_z), where e_x , e_y and e_z are the vectors representing the coordinate axes. The e_x and e_y are parallel to the earth surface, i.e. they lie in the plane represented by the earth surface, and e_z is perpendicular to this plane and represents the height above the surface. The lengths of the vectors are equal to one meter, i.e. |e_x| = e_y = |e_z| = 1 [m]. Any location in

space is represented by its position vector: r = x ■ e_x + y ^■ e_y + z • e_z = (x y z) . The corresponding location vector s = x ^■ e_x + y ^■ e_y + 0 • e_z = (x y 0) is defined as the vector in the e_z and e_y plane, which corresponds to the projection of r onto the e_x and e_y plane along e_z .

The MS 's position is determined by the position vector r₀ = (x₀ y₀ z₀ ) . However, the desired estimate of the MS location is determined only by its x and y coordinates. Hence, the desired MS location is determined by the MS location vector J₀ = (x₀ y_Q O)

The number of cells in the measurement report is N_M . The position vector of the serving cell antenna is ϊ = (x, y, z, )^τ and the corresponding location vector is s, = (x_x y_γ θ)^τ . The position vectors of the neighboring cell antennas in the measurement report are ⁼ _\ ^x _k y_k ^z _k ) ^and ^me corresponding location vectors are s_k = {x_k y_k 0) , where k = 2,...,N_M . A neighboring cell is related to the serving cells, either because its signal strength is high within the service area of the serving cell or because it is closely located to the serving cell.

There are neighboring cells, which are not in the measurement report. These cells are called absent neighboring cells, and the number of these cells is N^ . Hence, the total number of neighboring cells is N^ = N_M + N_Λ . The position vectors and location vectors for the absent cell antennas are r_k = {x_k y_k z_k )^τ and s_k = (x_k y_k θ)^τ , where k = N_M +l,...,N_M +N_N . For Power measurement predictions, let s₀ be an estimate of the MS location. Let P_k be the predicted received power at the estimated MS position from the signal transmitted by a cell k , k = l,...,N_M , of the measurement report. Then

10 • log{P_k)= 10 • log(R _O )+ G_k>AL + G_{k G} + G_k>R

where P_{k AO} is the amplifier output power of the cell.

G_kιAL is the gain in dB due to the cell specific antenna losses,

G_kΛG is the cell antenna gain in the direction of the estimated mobile location.

G_{k R} is the gain due to range and the propagation environment between the MS and the cell, then

^Gk,R (s₀) = ^Gk,w -³⁵ -^l°g _* -

where G_{k L0} is the, so-called, land use offset, which in principal is the path-loss gain at the range of one meter from the antenna and the gain from the MS antenna to the measurement report of the MS, and

G_k:LO = -42 [dB].

The path loss formula is intended for far-field measurement, i.e. at large distances from the cell antemia compared to the antenna size. The value of the G_{k L0} parameter is adjusted to fit the mobile station modules operating in, for example the 900 MHz band.

For Cell Location Penalty Function, it is likely that the measured cells of the measurement report are located closely to the positioned Mobile Station (MS). Hence, the locations of the measured cells yield an indication of the MS location. An assumption can be made that the center of gravity of the cell locations is the most likely location of the MS. Any deviation of the MS location from that center is penalized.

The location of the center of gravity s_g = [x_g y_g 0| of these locations is

The penalty function Ψ_CL for the deviation of the estimated MS location s_Q from s is

= ^WcL - - ⁱ(^χo - ^χ _g)² + ( o - y_s)² >

where

^WCI = 555

For Absent Neighbor Penalty Function, predictions can be made of the received power level at arbitrary locations on the earth surface. If the predicted received power level of a neighboring cell is high compared to the received power of the serving cell at a certain MS location, it is likely that the neighboring cell is one of the cells in the measurement report. If it is not part of the measurement report, it is not likely that the location corresponds to the MS location. Hence, such locations are penalized.

The penalty function Ψ_AN for estimated MS locations with high, predicted received power from the absent neighboring cells is

where

^WAN = ^l >

P is the predicted received power of the serving cell measured in watts, and

P_k , k = N_M + 1, ...,N_M +N_Λ, ι^'s the predicted received power of the absent neighboring cells, see below. In the Timing Advance (TA) Penalty Function, the TA is delivered from the serving cell at call setup. The TA value defines a circle around the location of the serving cell, where it is likely that the MS is located. Any deviation of the MS location from that circle is penalized.

The penalty function Ψ_TA of the deviation of an estimate from that circle is

where

0.1 w TA

555.{N_TA + Ϊ)

N_TA is the measured TA value, where N_TA e {θ,l,...,63}. s_TA is the radius of the circle defined by the TA value

s_TA = 110 + 555 - iV_M [m].

hi Absolute Power Penalty Function, power measurements are made at the serving cell and at the neighboring cells. A measured power level is dependent of the range between the cell and the MS. Due to the antenna gain profile and the propagation environment, the power measurement defines a closed curve around the location of the measuring cell, where it is likely that the MS is located. Any deviation of the MS location from that curve is penalized.

The penalty function Ψ_AP of the deviation of the predicted received power levels from the measured received power levels P_k , k = l,..., N_M , is

N_u

Ψ„(^o) = ^ - ^- ∑|lO -log(p; )-10 -log(p_;t(fo)) _J

where

w_AP = l . For Relative Power Penalty Function, the difference in the power measurements from two measuring cells defines an open curve between the cells, where it is likely that the MS is located. Any deviation of the MS location from that curve is penalized.

Consider TA penalty function case above. The penalty function Ψ^ of the deviation of the predicted received power level differences from the measured received power level differences is

* Rp ^Sθ) ⁼

where

w_RP = l .

For Merged Penalty Function, the penalty functions are merged into one single penalty function Ψ_M , where all measurements and all measurement conditions are considered:

Ψ (f₀) = Ψ_Ci( ₀) + Ψ^(f₀) + Ψ_r,(#₀) + Ψ„(f₀) + Ψ_Λ (f₀) .

The selected MS position s^* is the position, which minimizes Ψ_M {s₀) , i.e.

#₀ ^* = argmin(ψ ) .

The antenna gain can be modeled by using a number of reference antenna diagrams. These could have different lobe widths. The diagrams could be used to interpolate to a diagram with a desired lobe width.

The weights of the penalty functions may be tuned during testing. Therefore, these should be handled as constants when possible. The above-mentioned calculations relate to a simple 35 log d calculation plus an antenna masking. It is also possible to include either an Okumura-Hata style model or imported predictions from a real planning system. However, the problem with this is that there are many different suppliers of these systems.

Each PDA index can be populated with a value from -128 to +127 where the value is representing information for the geographic point depending on the PDA type. A PDA can be one of the following types:

RX-PDA

An RX-PDA (signal strength PDA) is populated with signal strength where the value is the signal strength in dBm minus an offset of 110 to comply with NMR RX level:

PDA value = signal strength in dBm - 110.

The RX-PDA can be populated offline.

Also, an LU-PDA (Land Use PDA) can be populated with an index representing the land-use type and an RM-PDA (Road Map PDA) can be populated with an index representing vehicle availability.

Preferably, the pre-processor will populate the PDAs.

It is desirable to have as much of the constants in the algorithms as possible configurable to avoid having to recompile the product when optimizing its performance. As a minimum the configuration file should read the following parameters:

^cellsize ~ κ⁾.55 cellcentre ⁼ 0.33

The position engine will calculate the estimated position and accuracy. The algorithm will depend on the input data over the API. Accuracy is in this implementation a direct relation to the cell size. It can be a best guess approximation. The following method can be used:

Accuracy is calculated as C_aCurracy[Desired percentage] * CellSize serving cell. Where C_accuracy is an array [1..99] of values which is loaded from a file:

1 x.xx

² y-yy

99 z.zz

When estimating the location of a mobile station, it is desirable to have an estimate of the accuracy of the position estimate. This will be delivered as a Circular Error Radius (CER), which defines a circle corresponding to a confidence area. The circle should be centered at the position estimate and the circle should surround the true position with a certain circular error probability (CEP). Therefore, an input to determine the size of the CER is the CEP. A larger CEP will increase the CER and vice versa.

The CEP is defined in an domain ranging from 1 to 99 in steps of integers, which corresponds to percent levels. The radius of the confidence area r_CER is determined by

^rCER ^{= r}S ^' J PCEP ) where r_s is the "cell size", which defines a radius of a circle, which surrounds most of the traffic in the corresponding cell, p_CEP is the CEP and / is the normalized CER-function defined for p = 1,2, ... ,99. The cumulative distribution of the range error divided by the cell size is plotted for four different propagation environments in Fig. 5.

The rural, suburban and urban environments have similar performance, while the curve for the dense environment is somewhat deviating. To form a common curve for all types of environments, a solution is to take the average level for each percent level of the curves corresponding to the urban and the dense environments. The common curve, i.e. the normalized CER-function, is shown in Fig. 6.

For positioning algorithm for input data Class A, preprocessed cell center coordinates are looked up and the result is presented as X and Y coordinates. Also, the cell size is looked up and returned.

The positioning algorithm for Input Data Class B, C and D, as shown in block diagram of Fig.4, comprises cost weighting function for a number of points by:

400: Loading the initial PDA (Point Distribution Array)

410: Calculate Cell Location (CL) Penalty Function for each point in array.

420: Calculate Absent Neighbor (AN) Penalty Function for each point in array. 430: Lookup and load relevant Timing Advance (TA) Penalty Function. TA penalty PDA can be pre-calculated for each PDAClass 440.

450: Calculate Absolute Power (AP) Penalty Function for each point in array.

460: Calculate Relative Power (RP) Penalty Function for each point in array.

470: Sum up Merged Penalty Function with weighting factors, 480: Return coordinates of point with lowest penalty.

Online PDAs used by the positioning algorithm are:

TA-PDA A TA-PDA is populated with penalty values for TA relations. There will be one PDA per TA value. The penalty equals to the distance from ideal radius, squared and divided with the TA value plus one. Then divided with the weight factor to be represented in correct penalty scale:

Ideal distance = 100 + TA * 555 meters

Fault distance = absolute (PDA index * (2 ^Λ PDA class) - Ideal distance) TA penalty = Fault distance * Fault distance / (TA value + 1) PDA value = TA penalty * (TA weight factor ~ 0.0015)

A CC-PDA (Cell Centre) is populated with penalty values for cell centre relations.

An SP-PDA (Statistical Probability PDA) is populated with a statistical probability summary of the index.

As for the configuration of the position engine, it is desirable to have as many constants in the algorithms as possible configurable to avoid recompiling the product when optimizing its performance. As a minimum the configuration file should read the weighting factors for each PDA.

Tables 1-3 are examples of information received from an operator of the system. Table 1 includes cell information, Table 2 the frequency information and Table 3 the neighbor information.

The first column of table 1 comprises the data entered by the network operator, either manually or automatically. Default values are values typical for a cellular network of GSM type.

Tables 4-6 are examples of information included in the database of the system. Table 4 includes cell information; Table 5 includes frequency information and Table 6 includes neighbor information.

The invention is not limited the shown embodiments; it can be varied in a number of ways without departing from the scope of the appended claims. The arrangement and the method can be implemented in various ways depending on the application, functional units, needs and requirements etc. Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Claims

1. Method of determining a position of a Mobile Station (190) in a cellular communications network (150) comprising a number of cells: a serving cell and a number of neighboring cells, wherein a set of data on each cell and neighboring cells is available, the method comprising the steps of: calculating for each cell a cell center and cell size based on said available data, calculating a Point Distribution Array (PDA) for a number of points in a coordinate system, and - based on said PDA and a statistical calculation determining the position of said mobile station.

2. The method according to claim 1, wherein said PDA consists of a number of points distributed around said serving cell coordinates where each point rests at a specific distance and direction relative said serving cell.

3. The method according to claim 1, wherein said PDA calculation is carried out by creating an PDA position and populating the PDA with a signal strength data.

4. The method according to claim 1, wherein that for said cell size a cell radius is defined as: If AntennaHBeamwidth = 360

1.J ^Q cellsize else

O. Ccellsize where d is the distance to a closest cell coordinates within a sector +/- max(60,AntennaHBeamwidth/2) around the serving cells antenna direction, which has a celltype >= serving cell and a distance > minimum distance, and Cceiisize is a user defined constant.

5. The method according to claim 4, wherein if no cells are found or if the distance is larger than an specified distance, the distance d is assumed to be said specified distance.

6. The method according to claim 1, wherein said Cell Centre Coordinate is calculated as: if AntennaHBeamWidth = 360 or CellType = PicoCell

BS coordinates else BS coordinates + r * d * C_ceιicentre

Where r is a unit vector indicating the direction of the serving cells transmitting antenna, d is the distance to the closest cell coordinates within a sector +/- max(60,A_ntennaHBeamwidth/2) around an antenna direction of the serving cells, which has a cell type >= serving cell and a distance > minimum distance, and C_ceιis_ze is a user defined constant.

7. The method according to any of claims 2-5, wherein both the direction and a specific distance (D_s) are calculated from an index value with respect to said number of points.

8. The method according to claim 6, wherein said specific distance (D_s) depends on a PDA classification, where the distance (d) gives the PDA classification:

First distance < 2*d <= second distance => first PDA class to max distance < 2*d => maximum PDA class.

9. The method according to any of claims 6 to 8, wherein that said specific distance (D_s) is calculated as the index multiplied with 2 exponent PDA class: D_s = index * (2 ^{PDΛ class}).

10. The method according to claim 8, wherein said classification comprises at least:

A: Cell Identity (Cell Id), ^'

B: Cell Id and reception level (RxLevel) for serving cell Broadcast Control CHannel (BCCH),

C: Cell Id and RxLevel and Time of Arrival (TA) for serving cell, and D: Cell Id and TA.

11. The method according to claim 10, wherein for the positioning comprising input data Class A, preprocessed cell center coordinates are looked up and returned as result in form of two-dimensional coordinates.

12. The method according to claim 10, wherein the steps of calculating a cost weighting function for a number of points by:

Loading an initial PDA,

Calculating a Cell Location (CL) Penalty Function for each point in the array, - Calculating an Absent Neighbor (AN) Penalty Function for each point in the array,

Looking up and loading relevant Timing Advance (TA) Penalty Function, Calculating an Absolute Power (AP) Penalty Function for each point in the array, Calculating a Relative Power (RP) Penalty Function for each point in array, Summing up Merged Penalty Function with weighting factors, and - Returning coordinates of point with lowest penalty.

13. The method according to claim 12, wherein the positioning uses one or several of PDAs: Time of Arrival (TA) PDA, Cell Center (CC) PDA and Statistical Probability (SP) PDA.

14. The method according to claim 13, wherein said TA-PDA is populated with penalty values for TA relations.

15. The method according to claim 14, wherein at least one PDA per TA value and PDA is arranged.

16. The method according to any of claims 13 to 15, wherein said penalty equals the distance from an ideal radius, squared and divided with the TA value plus one, divided with the weight factor to be represented in a correct penalty scale, whereby: Ideal distance = 100 + TA * a predetermined distance

Fault distance = absolute (PDA index * 2 ^{PDA class} _ ideal distance)

TA penalty = Fault distance * Fault distance / (TA value + 1)

PDA value = TA penalty * (TA weight factor ~ K), where K is constant value.

17. The method according to claim 13, wherein said CC PDA is populated with penalty values for cell centre relations.

18. The method according to claim 13, wherein SP-PDA is populated with a statistical probability summary of the index.

19. The method according to claim 12, wherein the CL penalty function Ψ_CL for the deviation of the estimated MS location s₀ from is

Ψct( o) = ^w. CL = W, CL ^■ J( ₀ -^χ _g)² +(yo -y_g

where

^WCL = 555

^s* " ° J

N_M is the number of cells in the measurement report,

20 The method according to claim 12, wherein the AN penalty function Ψ^ for estimated MS locations with high, predicted received power from an absent neighboring cells is

N_N+N_A

^X¥AN (SO) = WAN - — - ∑ ^ak >

N_A k=N_M+\

_a where w AN = 1,

P, is a predicted received power of the serving cell measured in watts, and P_k , k = N_M +1,..., N_M + N_A , is a predicted received power of the absent neighboring cells.

The method according to claim 12, wherein the TA penalty function Ψ_TA of the deviation of an estimate s₀ _S_j from that circle is

^xP_XΛ(s₀) = w_TΛ - -s_l -s_TAj where

0.1 w_w = 555.{N_TA +Ϊ)

N_TA is the measured TA value, where N_TA e {θ,l, ... ,63} . s_TA is the radius of the circle defined by the TA value s_TA = 110 + 555- N_M [m].

The method according to claim 12, wherein the AP penalty function Ψ_AP of the deviation of the predicted received power levels from the measured received power levels P_k , k = l,...,N_M , is

N,

Ψ^(*o) = w^ ^~- ∑|lO- log(p₄)-10.1og( (fo)]| ,

M k=l where

w_tf = l . 23 The method according to claim 12, wherein the RP penalty function Ψ^ of the deviation of the predicted received power level differences from the measured received power level differences is

RP {^SO) ^~

= w RP ^• ∑^ llO-log^ -lO-log^J-^O.log^^f_o^-lO-log^^^))) ,

N M k=\ l=k+l where

w_RP = l .

24 The method according to claim 12, said Merged Penalty Function Ψ is:

Ψ (f₀) = Ψ_Ci (_ ₀) + Ψ^( ₀) + Ψ_r, (f₀) + Ψ^(f₀) + Ψ_ΛP(f₀) .

25. A system (100) for determining a position of a Mobile Station (190) in a cellular communication network (150) comprising a number of cells: a serving cell and a number of neighboring cells, wherein a set of data on each cell and neighboring cells is available, said system comprising: a preprocessor (110), a Point Distribution Array (PDA) processing arrangement (120), a database (130) and - a position engine (140).

26. The system of claim 25, wherein the system is in communication with a user application arrangement (160).

27. The system of claim 25, wherein the preprocessor (110) processes operator input data into an internal database and that in the preprocessor, the Cell Centers and Cell Sizes are calculated, based on which the initial PDAs are generated.

28. The system according to any of claims 25 to 27, wherein the position engine (140) receives position request and delivers position and accuracy estimations.

29. A computer program for determining a position of a Mobile Station (190) in a cellular communications network (150), which comprises a number of cells: a serving cell and a number of neighbouring cells, wherein a set of data on each cell and neighbouring cells is available, wherein the computer program comprises: code for calculating for each cell a cell center and cell size based on said available data, - code for calculating a Point Distribution Array (PDA) for a number of points in a coordinate system, and said code further comprising a procedure, which based on said PDA and a statistical calculation determines the position of said mobile station.