KR100737516B1 - Method and system for estimating the position of a mobile device - Google Patents

Abstract

 CLAIMS 1. A method of estimating a location of a mobile device, comprising: collecting location information; Selecting at least one of a number of different methods to provide a location estimate; And providing a position estimate based on the at least one selected position determination method.

Description

METHOD AND SYSTEM FOR ESTIMATING THE POSITION OF A MOBILE DEVICE}

The present invention relates to a location determination method, and more particularly to an overall method of determining the location of a base station in a wireless communication network.

 Wireless cellular communication networks are known. In such networks, the area covered by the network is divided into a number of cells. Each of these cells is associated with a transceiver base station. The transceiver station is arranged to communicate with mobile devices located in the cells. The mobile devices can take any suitable form and are generally mobile telephones.

The need for efficient and accurate positioning of mobile phones is increasing. In the United States, network operators must be able to provide the location of mobile subscribers for emergency calls. Similar proposals are currently being considered in Europe. Additionally, such as tracking services (ie, emergency services, people location, valuable assets location, etc. mentioned above), search / guidance services (yellow pages, direction indicators, points of interest locator, etc.) Commercial services such as proximity services) and announcement services (designated advertisements, traffic alerts, weather alerts, bus / train location, guided tours, etc.) are currently being proposed.

In the GSM (Global System for Mobile Communications) standard, cell identity and timing advances, arrival times, enhanced observed time differenc (E-OTD) and GPS (global positioning system technology as standalone GSP or assisted GPS) Four positioning methods based on this have been included.

The time of arrival method can locate the handset with standard software, but requires the installation of new network elements, such as position measuring units, at all base stations. The parallax observation enhancement method requires installation of location management units and software modification in the handset for every two to five transceiver stations. The auxiliary GSP method requires the installation of a GSP receiver and a possible position measuring unit in addition to the integration of the GPS receiver into the handset. All these methods require the introduction of a new network element or a corresponding function capable of responding to a location calculation called a serving mobile location center SMLC.

The temporal deployment of location services for all users, including those with handsets that do not include the necessary software or hardware, requires that measurements already available in cellular networks can be used. These techniques are important when they allow operators and service providers to start providing location based services at minimal additional cost while waiting for more accurate and sophisticated location techniques available. From a technical point of view, when technology and auxiliary GPS, such as parallax observation enhancement, are all available, network-based software solutions still require a backup method when new standard solutions fail or the required accuracy can be combined with such a method. Can be. Network-based software techniques can also be used as an initial inference about the algorithms used to implement one of the standard solutions for improving the accuracy or speed of convergence of such algorithms.

According to a first aspect of the invention, a method provided for position measurement of a mobile device comprises: collecting position determination information; Selecting at least one of a plurality of different location determination methods for providing a location estimate; And providing a position estimate based on the at least one selected position determination method.

According to a second aspect of the present invention, a system for position estimation of a mobile device includes: means for collecting position information; Means for selecting at least one of a number of different location determination methods for providing a location estimate; And means for providing a position estimate based on the at least one selected positioning method.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT For a precise understanding of the present invention and how the present invention can be effectively performed, only the accompanying drawings will be described in detail.

1 illustrates a wireless cellular network to which an embodiment of the present invention may be applied.

FIG. 2 illustrates a mobile device provided by three transceiver stations in the wireless cellular network shown in FIG. 1.

3 shows a topographical representation of a transceiver station and mobile device in the wireless cellular network shown in FIG. 1.

4 illustrates one embodiment of antenna gain in the transceiver shown in FIG.

5 shows another example of antenna gain in the transceiver shown in FIG. 4.

6 shows an approximate radiation pattern as found in the transceiver shown in FIG. 1.

FIG. 7 shows an Okumura-Hatta Path Loss Graph for the modeled wireless cellular networks shown in FIG. 1.

FIG. 8 illustrates topographical specification for a trusted region of location estimation for a serving cell in the wireless network shown in FIG. 1.

9 illustrates a confidence region of position estimation provided by cell identity of consecutive cells.

FIG. 10 shows a topographic indication of cell coverage in the wireless cellular network shown in FIG. 1.

FIG. 11 shows a continuous plot of the probability density function of the position estimate for various values of path loss / decay in the wireless cellular network shown in FIG. 1.

FIG. 12 shows a plot of I _i0 for path loss / attenuation in the wireless cellular network shown in FIG. 1.

FIG. 13 shows an example of a serving area of a cell in the wireless cellular network shown in FIG. 1.

14 shows a circular crown confidence area for position estimates generated from constant TA values.

FIG. 15 shows a more topographical view of the confidence zone for the position measurement shown in FIG. 14.

16 illustrates the terrain used to calculate the confidence area in CI location measurements.

17 shows a flowchart detailing the steps for providing position estimation in an embodiment of the present invention.

A description will be given with reference to FIG. 1 which schematically shows a wireless telecommunication network 2 to which embodiments of the invention may be applied. The area covered by the network is divided into cells 4. Associated with each cell is the base station 6. In another embodiment of the present invention, base station 6 may generally be referred to as a transceiver base station (BTS). The base station 6 is arranged to be able to communicate with the user equipment 8 via a wireless connection. The user equipment is generally a mobile phone, a computer, a personal digital assistant (PDA) or a mobile device such as this. In another embodiment of the invention, the user equipment may also be referred to generally as a mobile terminal (MS).

Embodiments of the present invention are configured to estimate the location of the mobile terminal MS and the associated confidence area. The trusted area is an area in which the mobile terminal MS can be predicted to be located at a given probability. That is, the estimated position of the mobile terminal MS can be calculated but the estimate will not be 100% accurate. The trusted zone defines an area in which the mobile terminal MS can be quite certain that it is located.

Embodiments of the present invention are configured to combine multiple positioning algorithms, which will be described in more detail below.

Positioning algorithms used in embodiments of the present invention use cell identity (CI), timing advance (TA) and received signal strength (RX) measurements.

Preferred embodiments use three algorithm types:

1) Algorithms based on cell identity. These are referred to as CI algorithms;

2) Algorithms based on cell identity and timing advance information. These are referred to as CI + TA algorithms; And

3) Algorithms based on cell identity and received signal strength measurements. These are referred to as CI + RX algorithms.

These different algorithms will be described in more detail below. For convenience, these algorithms fall into two categories. The first includes CI + RX algorithms, the second includes two other types of algorithms: CI and CI + TA algorithms.

In some networks, it should be appreciated, for example, that the GSM network, cell identity, timing advance and received signal strength data are already available and used for other purposes. This means that at least some algorithms can be implemented without requiring any modification to the existing handset.

The positioning algorithms used in the embodiments of the present invention require a specific set of simultaneous equations to be found. The simultaneous equations, which depend on the statistical assumptions of the measurement, can be nonlinear or linear in unknowns. If the simultaneous equations are linear, there is a closed solution, whereas if the equations are nonlinear then an iterative method must be applied to find the solution. Closed algorithms are easier to compute than iterative algorithms.

CI + RX Algorithms

The layer of algorithms based on cell identity and received signal strength will now be described.

The following describes the principle of position estimation based on the received signal strength with reference to FIG. Receive signal levels are measurements of received signal strength levels from the serving transceiver station (BTS) 6 (transceiver station (BTS) with which the mobile terminal MS is associated) and the six most powerful neighboring transceiver stations. Of course, other numbers of transceiver base stations may be used. These measurements are performed by the mobile terminal (MS) 8 and recorded in a fixed part of the network when the mobile terminal (MS) is in dedicated mode. In the idle mode, mobile device 10 measures the received signal level from the best server, namely the pending connected transceiver base station (BTS) and the six most powerful neighbor transceiver stations. However, since there is no connection with any transceiver station, these measurements may not be reported to a fixed part of the network.

The more precise attenuation experienced by the level or signal received by the mobile device depends on the mutual location of the transceiver station associated with the mobile device. The signal levels received at the multiple transceiver bases may then be combined to estimate the location of the mobile device.

Embodiments of the invention belong to the layer of a network-based software solution (NBSS). In particular, the embodiments of the invention described in this section are based on the use of signal level measurements and focus on applications where coverage prediction maps are not available. However, embodiments of the present invention may be used where coverage prediction maps are available. Coverage prediction maps are used in some methods where the mobile device location is measured as the location of the coverage map where the numerical value of the predicted signal strengths best matches the actually measured levels.

The measurement / information necessary to implement the algorithms described herein may be separated into the following network formation parameters: base station coordinates; Sector-oriented and basic transceiver antenna radiation patterns for sectorized cells; And measurements such as the maximum transceiver station downlink transmit power and the level of signals received by the mobile device. In addition to that data, knowledge of the appropriate models is needed to link the level measurements for the transceiver station inter-location used in the positioning procedure with the mobile devices to be located. These will be discussed in more detail below.

Here, the location determination methods outlined above estimate mobile device coordinates at a location where a particular function has its minimum value. The function is obtained by combining level observations, which are received by the mobile device and estimates of the attenuation experienced by the signal transmitted by the associated transceiver station. Level observations are the distribution of levels received by the mobile device (depending on the mutual angular position between the mobile device and each transceiver station) and the distribution of the transceiver station antenna radiation patterns, and dependent on the distance between the mobile device and each transmit / receive base station. I) by estimating other constant factors such as path loss and transceiver station transmit power, cable loss, antenna loss, and so on.

Clearly, the accuracy of the results provided by the algorithm depends on the accuracy of the information available. The accuracy of the results can be improved by including precise definitions of the antenna radiation patterns and finely tuned delivery models. These models, in principle, need not be represented in an analytical fashion, but can be included in the positioning algorithm as a lookup table format where the antenna gain can be restored when a particular mobile device is given to the destination transceiver station angle. However, such additional matters may be omitted in embodiments of the present invention.

It is desirable that network information such as transceiver base station coordinates, sector information, antenna radiation patterns, and other adjustable parameters be constantly updated. In preferred embodiments of the present invention, the models used to represent signal level measurements can be continuously adapted to changes in the wireless environment, preferably in an automatic way. One way to do this is to adjust the models based on a statistical comparison between the particular set of measurements collected and the corresponding actual quantities. However, this method requires the exact location of the known mobile device to calculate the actual quantity and to compare it with the measurement. When the measurement models are derived offline, drive tests may be performed by simultaneously collecting level measurement and mobile device positioning information with a Global Positioning System (GSP) receiver co-located with the mobile device. For example, other techniques may be implemented to provide more accurate location techniques using estimated observation time differences, arrival techniques, or assisted GPS location estimation that allows measurement model determinations online.

The algorithms described below are applications of the principle of maximum likelihood for processing observation sets from level measurements to estimate coordinates of a mobile device. Different maximum likelihood approaches are undertaken.

The location determination algorithms use signal level measurements to estimate mobile device coordinates, x and y, with a slow fading change σ _u ² that is assumed to be the same for all associated base stations.

As described in the embodiment of the present invention, the input data required by the algorithms is collected by a mobile device (MS), network parameters and an appropriate model for the proper path-loss law, ie wireless transmission characteristics. Signal strength measurements.

Average power of signals received by the MS from N transmit / receive base stations (BTS's) measured in decibels, P ⁱ _r :

(One)

Coordinates of N BTS's measured in meters, (x ⁱ , y ⁱ ):

Azimuth of N transceiver base stations (BTS's) sectors (in the case of sectorized cells) measured in radians counterclockwise in the x direction, φ ⁱ _B :

Maximum radiated power from N BTS's measured in decibels:

The maximum radiated power, P ⁱ _{t, max} , represents the maximum power at the output of the ith BTS antenna in the direction of maximum gain. This includes the transmit power P ⁱ _t , the maximum gain of the BTS transmit antenna, G ⁱ _{t , max} , antenna loss, cable loss, etc. all measured in decibels. To simplify the mathematical model, P ⁱ _{t, max} also includes the maximum gains of the MS receive antennas, G ⁱ _{r , max} .

The combined transmit-receive antenna pattern for the i th BTS (i = 1, ..., N) is decibel:

The combined radiation pattern, AP ⁱ _tr (ψ ⁱ (x, y)), represents the gain presented by the antenna installed in the BTS region and the antenna installed in the handset. The gain depends on the mutual orientation of such antennas. In the embodiment represented by the above equation, φ _M is the orientation of the antenna installed in the MS, φ ⁱ _B is the orientation of the antenna installed in the i-th BTS, and ψ ⁱ (x, y) is transmitted by the i-th BTS And the angle of arrival of the signal received by the mobile device MS, measured in radians counterclockwise in the x direction (shown in FIG. 3).

The combined radiation pattern AP ⁱ _tr is a function of the MS coordinates (x, y). This includes the radiation pattern AP ⁱ _t (θ) of the transmission antenna installed in the i-th BTS and the radiation pattern AP _r (θ) of the antenna installed in the mobile device MS. In an embodiment of the present invention, the antenna of the MS is omnidirectional and the orientation of such antenna is unknown; Thus AP _r (θ) = 0dB and φ _M = 0rad.

Path-loss law for transfer between MS and i-th BTS measured in decibels:

The path-loss PL ⁱ (d ⁱ (x, y)) represents an attenuation by the signal transmitted by the i th BTS when it transmits further away from the transmit antenna. This is represented by the distance function d ⁱ between the MS and the i-th BTS, which in turn depend on MS coordinates.

The output of the positioning algorithm includes:

Estimate of the MS location in meters:

Estimate of log-normal slow fading variance in decibels:

As mentioned above, the algorithms apply the principle of maximum likelihood estimation. They estimate the variance of slow fading by minimizing mobile device (MS) location and certain scalar functions. The function includes level observations determined by the transceiver base station antenna radiation patterns (depending on each position between the mobile terminal and each transceiver station) and the path loss (depending on the distance between the mobile terminal and the base transceiver station), where The level observations are obtained by combining the measurements of attenuation experienced by the signal received by the mobile device MS with corresponding prediction quantities. By minimizing a particular cost function (the cost function is a term generally applied in the mathematics for the explanation of optimization problems). The algorithms are unknown parameters that overall minimize the difference between the observed and predicted attenuation. Find the figures. The cost function is a measure of the predicted errors in the position estimation of the mobile device MS. Thus, by minimizing the cost function, the predicted position error is also minimized. Embodiments of the present invention use the following algorithms that depend on the assumed characteristics of the wireless environment.

Embodiments of the present invention perform a first algorithm A where it is assumed that the wireless environment includes correlated slow fading characteristics and where the slow fading is assumed to have the same distribution statistics. That is, the mobile device (MS) signal transmitted by each transceiver base station (BTS) has similar but not identical characteristics. This type of assumption is accurate where the transmission paths between the transceiver stations (BTS) and the mobile device are similar.

Algorithm A: Maximum likelihood estimation assuming the same variance and slow fading correlated with level observations

Calculate the ⁱ level observation, L ⁱ by extracting the maximum power P ⁱ _{t, max} radiated by the i th BTS from the i measured received power, P ⁱ _r :

L ⁱ is the total attenuation experienced by the signal transmitted by the i th BTS during delivery to the MS. The total attenuation depends on the path-loss, the gain introduced by the BTS antenna and the MS antenna, the variation of the wireless channel, and the like.

2. Loading the level observations from N BTS's into vector L:

3. Solve the minimization problem:

Here, the cost function F (x, y; σ _u ² ) is defined as:

And

ρ _u ^{i, j} (x, y) is BTS ⁱ And cross-correlation of slow fading affecting the signal from BTS ^j to MS.

Embodiments of the present invention assume that the wireless environment includes uncorrelated fading characteristics, and the slow fading performs the second algorithm B where it is assumed to have the same distribution statistics. That is, the signal transmitted by each transceiver base station (BTS) to the mobile device (MS) has properties that are not related. This hypothesis type is accurate where the transmission paths between the transceiver base stations (BTS) and the mobile device (MS) do not contain similar components.

Algorithm B: Maximum likelihood estimation assuming uncorrelated slow fading and covariance with level observations.

Calculate the i-th level observation by extracting the maximum power P ⁱ _{t, max} radiated by the i-th BTS from the i-th measured received power, P ⁱ _r :

2. Loading the level observations from N BTS's into vector L:

3. Solve the minimization problem:

Here, the cost function F (x, y) is defined as:

And D _xy is the domain of the existence of x and y. Some possible definitions of D _xy are given later.

가 계산된다:4. As follows

Is calculated:

Algorithm B is different from algorithm A in the definition of the cost function. Different models for slow fading that affect the signals transmitted by two BTS's are based on different definitions. It is assumed in algorithm A that the fading is correlated, while in algorithm B it is assumed that the fading is uncorrelated; As a result, the definition of the cost function of algorithm B becomes simpler.

Embodiments of the present invention perform a third algorithm C where the wireless environment includes uncorrelated slow fading characteristics, where the slow fading is assumed to have the same distribution statistics. That is, the signal transmitted by each transceiver base station (BTS) to the mobile device (MS) has unrelated characteristics. This hypothesis type is accurate where the transceiver base station (BTS) and mobile device (MS) do not contain similar components.

Algorithm C: Maximum likelihood estimation assuming uncorrelated slow fading and equal variance with level difference observations.

Calculate the i-th level observation by extracting the maximum power P ⁱ _{t, max} radiated by the i-th BTS from the i-th measured received power, P ⁱ _r :

2. By extracting the j-th level observation from the level observation L ¹ obtained as a reference, and calculating the j-th level difference observation:

3. Loading N-1 differences of level observations in vector D:

4. Solve the minimization problem:

here

And

D _xy is the domain of x and y presence. Some possible definitions for selecting D _xy are given below.

이 계산된다.5. As follows

This is calculated.

Algorithm C differs from algorithms A and B in the definition of observations and consequently in the definition of the cost function. The observations considered in Algorithm C differ in the attenuation experienced by the signal transmitted by the two different BTS's and received by the MS. Moreover, to simplify Algorithm C, slow fading processes affecting the signals transmitted by both BTS's are assumed to be uncorrelated and analogous as in Algorithm B.

Algorithms A, B and C described above are nonlinear algorithms.

The positioning algorithms are described in more detail below.

3 shows the basic terrain of the problem. 3 includes a mobile device (MS) 8 and an i-th transceiver base station (BTS) 6. The MS 8 and BTS 6 are in the area defined by the Cartesian coordinate system 301. All angles are defined as defined from the x-axis in the counterclockwise direction. The i-th BTS 6 is located at a point defined as (x ⁱ , y ⁱ ) and arranged to broadcast and receive in the maximum gain direction 303 at an angle defined by φ ⁱ _B. The mobile device MS 8 is located at point (x, y) and is also arranged to transmit and receive in the maximum gain direction 305 at an angle φ _M. The MS 8 is arranged in association with the i-th BTS by a line 307 comprising a length d ⁱ (x, y) and an angle ψ ⁱ (x, y). In a more specific embodiment of the present invention, an alternative coordinate system to the Cartesian system is used. In another embodiment of the present invention, the polar reference system comprises MS and BTS ⁱ defined in radical coordinates. Distance, d ⁱ And a BTS ⁱ having an arrival angle, ψ ⁱ , of the signal received by the mobile device defined in angular coordinates. Focus on

The distance between the mobile device MS and the ith transceiver base station BTS is defined according to equation (30).

The arrival angle of the signal transmitted by the i-th BTS and received by the mobile device MS is defined according to equation (31).

The conversion between x, y (Cartesian) and d ⁱ , ψ ⁱ (radical) coordinates is completed by the formulas described below.

As described in the above embodiment, the location determination algorithms process specific level observations to estimate the MS location. Such observations are obtained from signal strength measurements performed by the MS. This zone derives a model for level observations from the model of signal strength measurements performed by the MS. Such measurements are estimates of the average power of the received signals by definition. The general model for calculating the average power received by the mobile device from the i < th > base station at a specific coordinate (x, y) is as follows (all quantities are in dB):

here,

P ⁱ _r (x, y) is the power of the signal received by the MS at the location (x, y);

P ⁱ _t is the BTS transmit power;

G ⁱ _r (x, y) is the MS antenna gain in the direction of the i th BTS;

G ⁱ _t (x, y) is the antenna gain of the i th BTS in the direction of the MS.

PL ⁱ (x, y) is the path-loss determined by the transfer path between the MS and the BTS;

The term "losses ⁱ " takes into account losses by antenna feeder, cable, duplex, splitter, and the like.

u ⁱ (x, y) is the shadow fading that affects the signal transmitted by the i th BTS. It is generally modeled as a random variable with a log-normal distribution (ie u ⁱ (x, y) measured in decibels is a Gaussian random variable) with a standard deviation σ ⁱ _u . Typical values for σ ⁱ _u range from 5 to 10 decibels:

Equation 33 is the starting point for studying a technique for positioning mobiles by observations derived from observed level (or received level RXLEV) measurements. This is generally a general enough model to allow the introduction of some parameters that affect the transmission, such as antenna radiation patterns, path loss and random variations.

Antenna gains are generally represented analytically as a function of angle θ describing the angular distribution of power radiated by the antenna when the antenna is connected with the transmitter. In the case of directional antennas, the direction of maximum gain is identified by direction θ = 0:

The antenna gain G (θ) may be separated by a sum of a constant term defining a radiation pattern AP (θ) that describes an angular distribution of radiated power with the maximum antenna gain G _max .

4 shows a graph of antenna gain for an antenna used in an embodiment of the invention. The graph of FIG. 4 includes a y-axis representing the antenna gain G (θ) and an x-axis representing the orientation from the angle of maximum gain. The antenna gain is symmetrical about the θ = 0 line. The antenna gain graph includes a main lobe (or beam) 401 and four side lobes (or beams) 403, 405, 407, 409 that are each side of the main lobe. The main lobe 401 includes a maximum gain G _m centered at θ = 0. The main lobe antenna gain rapidly decreases to the zero side of θ = 0. The side lobes 403, 405, 407, 409 include a smaller maximum gain, and the first side lobe 403 (side lobe immediately adjacent to the main lobe 401) is the second side lobe 405 ( And a maximum gain lock greater than the first side lobe 403 and the side lobe close to the third side lobe 407. The second side lobe 405 in turn has a greater maximum power than the third side lobe 407 (side lobes close to the second side lobe 405 and the fourth side lobe 409), and the third side lobe 407 has a greater maximum gain than the fourth side lobe (the side lobe close to the third side lobe 407).

When analytical representations of antenna radiation patterns are not available, an approximation is required.

The orientation φ ⁱ _B of the BTS antenna (as shown in FIG. 3) details the direction in which the maximum power is radiated. This downlinks the transmission gain to the MS, where a given (x, y) MS coordinate, ψ _i (x, y) (eg, the angle at which the BTS “sees” the MS) can be calculated and expressed as It means you can do it.

In principle, if the MS has a directional antenna and the orientation of the antenna is known, an analog formula can be used to account for the gain of the receiving antenna on the MS side.

Although mobile devices have a directional antenna, it is not practical to assume that the orientation of the MS antenna is known; Therefore, in another embodiment of the present invention, the MS antenna radiation pattern may be modeled as an average omnidirectional pattern by imposing limitations to be described later.

Substitutions of (33) to (36) and (37) give the following compact representation of the received power.

Incidental definitions P ⁱ _{t, max} and AP ⁱ _tr have been previously defined separately in equations (4) and (5).

Some analytical representations of path-loss PL ⁱ in equation (33) have been proposed in the scientific literature. In fact, this term is based on transmission loss prediction models based on cellular operators designing their networks. In order to accurately calculate the attenuation experienced by the signal movement from the base station to the mobile, a precisely tuned prediction model, including topographic and morphological information of the environment, must also be considered. However, when terrain maps are not available, simplified models should be used. Such available transfer models are known in the art and the transfer model in an embodiment of the present invention is a path loss as a function of distance from the arrival angle (AOA) to the MS to BTS.

With the locations considered to remember, given coordinates (x, y), the average power received from the i-th base station ⁱ P _r by a mobile device (MS) in a can be expressed as follows.

Where the previous definitions are; Maximum radiated power from the ith transceiver station P ⁱ _{t, max} [equation (4)], combined transmit-receive antenna pattern AP ⁱ _tr [equation (5)], mobile device (MS) and ith transceiver station (BTS) Path loss PL ⁱ [equation (7)] and log-normal shadow fading u ⁱ [equation (34)] affecting the signal transmitted by the i th BTS (all quantities are in dB):

When moving from the model described by equation 41, the observations that can be derived from level measurements in a mobile wireless network and can be used for MS location purposes as used in embodiments of the present invention are statistically determined. Characterize and induce.

As used in the embodiment of the present invention for location purposes in algorithms A, B and C, one measurement or observation is the average power P ⁱ _r received by the mobile and the maximum power P radiated by the ith BTS ⁱ is the difference between _{t and max} .

This definition is justified by the fact that the MS position-independent information is based on the difference P ⁱ _r -P ⁱ _{t, max} and that the maximum radiated power P ⁱ _{t, max} is a reasonable parameter that is easy to calculate given the network shape.

According to the transfer model 41, L ⁱ (d ⁱ (x, y), ψ ⁱ (x, y)) is a random variable because of the probabilistic nature of u ⁱ (x, y). Since u ⁱ (x, y) is a Gaussian random variable, L ⁱ (d ⁱ (x, y), ψ ⁱ (x, y)) is also a Gaussian random variable.

The mean value μ ⁱ _L (x, y) and the standard deviation σ ⁱ _L (x, y) can be derived as follows.

The average value of L ⁱ (x, y) is

The variance of L ⁱ (x, y) is

The mean value and standard deviation of the i th observation depend on the MS coordinates (x, y). The probability density function (pdf) of L ⁱ adjusted by the x and y coordinates can be expressed as follows.

In an embodiment of the present invention for MS location estimation purposes, level observations from different BTS's are used. For this reason, the covariance of two level observations is important. The two levels of covariance, in turn, depend on the cross-correlation of slow fading processes that affect signal transmission from different BTS's. Often the slow fading processes are considered uncorrelated. However, as is known, certain correlations exist between signals transmitted at different BTS's. Therefore, in the embodiment of the present invention, a model for cross correlation of slow fading is defined as follows.

Thus, the covariance of L ⁱ (x, y) and L ^j (x, y) can be seen as follows (x and y can be ignored for ease of display).

Level measurements from multiple BTS's, BTS's ¹ , ..., BTS ^N may be collected and used in a further embodiment of the present invention for measuring the MS position. As shown in equation 42, single-level measurements can be loaded into the N × 1 vector of observations, L.

It has a multivariate Gaussian distribution.

In an embodiment of the present invention the covariance matrix of the mean value and L can be easily calculated using the results from the previous equations.

The average value of L (x, y) can be written as

The covariance matrix of L (x, y) can be written as

The general elements of R _L (x, y) are as follows (see equation (28)):

는 공분산 매트릭스 R_L(x,y)의 행렬식을 나타낸다.)Thus, the probability density function of L adjusted by x and y can be written as in equation (54), where

Denotes the determinant of covariance matrix R _L (x, y).)

In a further embodiment of the present invention, one simplified assumption can be made by considering slow fading from different BTS's when they have the same variable:

This assumption is correct where the transfer occurs in a single communication environment. That is, where the communication environment is consistent and similar. Thus, slow fading affecting different BTS's has statistically identical properties. The same covariance assumption for all slow fading links results in a modified structure of the covariance matrix R _L (x, y). The structure of the covariance matrix is the product of the slow fading variance σ _u ² (common to all BTS's) and the matrix r _L (x, y) dependent on the position dependent cross-correlation ρ _u ^{i, j} (x, y).

Using the assumptions defined above, the probability density function of L adjusted by x and y in equation 54 can be written as

In a further embodiment, the slow fading processes are assumed to be uncorrelated as well as to have the same variables. In such an embodiment, r _L (x, y) is an N × N identification matrix that is independent of MS coordinates.

And the probability density function 58 is changed as follows.

Where (see equations (51) and (54)),

Model for Level Difference Observation

The i th level observation L ⁱ defined by equation 42 is obtained from the level measured by the MS for the signal transmitted by the i th BTS. In a further embodiment of the invention, the observations may be derived from the difference between the level measured by the mobile devices from one BTS obtained as a reference and the level measured by the mobile device from another BTS.

Using the difference in levels helps to eliminate unknown common propensities in complete level measurements. To approach similarly those discussed above, the statistical characterization of the level difference measurement is shown below.

The difference in level observation is the difference between the level observation from the i-th BTS as defined in (42) and the level observation from the reference BTS (hereinafter defined as index i = 1), in a further embodiment of the present invention. Is defined.

D ⁱ is a Gaussian variable due to the stochastic nature of log-normal fading.

The mean value and covariance of D ⁱ (x, y) can be derived as follows.

The average value of D ⁱ (x, y) is

Covariance of D ⁱ (x, y) and D ^j (x, y)

Covariance between a pair of level difference observations is required in an embodiment of the present invention because the location algorithm uses combined multi level difference measurements. The general covariance definition between D ⁱ (x, y) and D ^j (x, y) is ignored (x and y are ignored to simplify the notation):

When i = j is a variable of the i-th difference observation, the resulting term (σ ⁱ _D ) ² is:

Where ρ ^{1, i} _u is the cross-correlation between slow fading affecting transfer from BTS ¹ and BTS ⁱ defined in equation (47).

The term for i ≠ j can be calculated as follows.

Therefore, in an embodiment of the present invention, the covariance between two level difference observations can be summarized as follows.

The probability density function of D ⁱ adjusted by x and y can be expressed as follows.

Measurements from multiple BTS's, BTS ¹ , ..., BTS ^N can be collected and used to estimate the MS location. In a further embodiment of the invention, the level difference observations may be loaded with (N-1) × 1 vector D of observations.

It has a multivariate Gaussian distribution.

The mean value and covariance of D can be easily calculated from the results defined above.

Average of D (x, y)

Covariance matrix of D (x, y)

General elements of R _D (x, y) (see equation (68)).

The result of the following probability density function of D adjusted by x and y is as follows.

In a further embodiment of the invention, slow fading from different BTS's is assumed to have the same variance.

The elements of the covariance matrix R _D (x, y) are

Furthermore, in a further embodiment of the present invention, if slow fading processes are uncorrelated, ρ _u ^{1 , i} in equation (77) because i, j ≠ 1 and ρ _u ^{i, j} = 0 when i ≠ j. = ρ _u ^{1, j} ; Thus, the general element of R _D becomes

The covariance matrix can be written as follows.

Where I and 1 are two (N-1) × (N-1) matrices.

Using the assumption of slow fading that is uncorrelated and has the same variable, the probability density function (pdf) 75 can be written as follows.

The determinant of matrix r _D and its inverse can be computed explicitly.

By substituting the representation of r ^{D -1} obtained from the argument of the exponent in definitions 70 and 72 and 81, the following results are obtained.

Positioning algorithms used in the embodiment of the present invention are applications of the maximum likelihood (ML) principle for level and level difference observations defined above. As discussed above, the maximum likelihood principle is a widely used method in estimation theory. A brief review of the maximum likelihood principle follows.

If x ∈ D is an unknown random parameter defined in a particular domain D, y is the observed random parameter. The principle of maximum likelihood (ML) maximizes the combined probability density function (pdf) of x and y to provide an estimate of x.

The ML estimates can be calculated by maximizing the characteristic algorithm of f (x, y), Λ (x) = lnf (x, y); In general, it can be referred to as a log-likelihood function:

f (x, y) is the product of the posterior probability density function y of the observation given the unknown x, f (y│x) and the previous probability density function f (x) for the unknown x, that is, the probability density of the first unknown x The function f (x) is multiplied by the first unknown x by the probability density function f (y | x) of the conditional observation y. A similar extension of the log-likelihood function is Λ (x) = ln f (y | x) + ln f (x). In both cases, the ML estimates can be calculated in an embodiment of the present invention to solve the problem described below.

If the previous probability density function for unknown x is not available, f (x) can be ignored (or equally, x can be assumed to be uniformly distributed in domain D), and equation (87) This results in a simplified form of equation (86) shown in.

In the context of an MS location with signal level measurements, the observation y is a vector of level observations defined in (49) or

Or a vector of level difference observation D defined at 70.

The parameter to be an estimated x includes estimated MS coordinates (x, y) with an estimated variance σ _u ² of slow fading in an embodiment of the invention.

It is shown above that the level observations L and the level difference observations D mode can be modeled as multivariate Gaussian random variables. In such circumstances, the ML criterion 84 may be simplified further. If the observation y∈R ^{M × 1} adjusted by the unknown parameter x is the M-variable Gaussian random variable and the correlation matrix R _y ∈R ^{M × M,} then the probability density function of the observation y adjusted by x is

Here the dependence of R _y and m _y from x is clearly shown. f _{y | x} (y | x) is

And the ML estimate of x can be calculated according to ML criteria 89 as follows.

Thus, when level observations and level difference observations follow the Gaussian statistical model derived above, equation 93 provides an ML criterion for MS location estimation. For simplicity, the examples described below are limited when the slow fading variances from different BTS's are assumed to be the same (statistic models for the observations are described above).

When N level observations are considered and the variance of slow fading is assumed to be the same for all observations, the ML criterion 93 for the Gaussian unknown is y, m _y (x), R _{y in} the embodiment of the present invention. Follow the definitions below for (x) and M.

및 R_L(x,y)은 레벨 관측 L의 공분산 매트릭스이다. 따라서 x, y 및 σ_u ²를 추정하기 위한 ML 기준은:here,

And R _L (x, y) is the covariance matrix of the level observations L. Thus, the ML criteria for estimating x, y and σ _u ² are:

When N level observations are considered and the slow fading is assumed to be uncorrelated with the same variance for all observations, the ML criterion 93 for the Gaussian unknown is y, m _y (x ), R _y (x) and M are defined below.

The ML criteria for estimating x, y and σ _u ² are:

Or analogously by inserting the values of L-mL given in equation 61.

의 (엄밀히 양수의)최소값을 찾는 것과 동일하다. σ_u ²의 추정값은 다음과 같다. In an embodiment of the present invention, sigma _u ² can be estimated separately. For fixed x and y, this estimate of σ _u ² is a function

Equivalent to finding the (strictly positive) minimum of. The estimated value of σ _u ² is

In the embodiment of the present invention, ML estimation of x and y can be found by solving the minimization problem described later.

Where D _xy is the domain of x and y present.

The domain D _xy may be determined in any further embodiment of the invention using, for example, any additional location information available for the implementation of Network Based Software Solutions (NBSS's). For example, if the cell identity of the serving BTS is known, in an embodiment of the present invention, D _xy may be defined as the geographic area provided by such BTS. If additionally timing advances are also available, in further embodiments D _xy may be defined as a topographic area determined by the intersection of the serving areas of the BTS identified by a given CI. In a further embodiment, D _xy , having an origin in the serving BTS coordinates and having an inner / outer radius determined according to the TA value, for example using techniques known in the art, can be defined as a circular crown or ring. In an embodiment of the invention, for example, restricting the domains of x and y along CI and TA has two advantages: first, TA information is considered an implicit function; Second, convergence of the minimization algorithm is faster.

When N-1 level difference observations are considered and slow fading is assumed to be the same variance and uncorrelation for all observations, the ML criterion 93 for the Gaussian unknown is y, m _y (x), R _y ( applies to the definitions of x) and M:

Here, r _D is a (N-1) × (N-1) matrix (independent of MS coordinates) to be described later.

ML criteria for estimating x, y and σ _u ² can be derived using the result 83.

here

And

D _xy is the domain of x and y presence. Some possible definitions for D _xy are given later.

For example, limiting the domain of x and y according to the values CI and TA has two advantages: first, the TA information is considered an implicit function; The second is that convergence of the minimization algorithm is faster.

In a further embodiment of the invention, there is an alternative interpretation of the ML criteria. The combined probability density function f (x, y) of the unknown x and the observation z can be written as the product of the conditional probability density function f (x | z) and the marginal probability density function of the observation. The log-likelihood function can be written as Λ (x) = lnf (x | z) + lnf (z), and the ML estimate can be calculated to solve the problem to be described later.

The second identity is maintained because f (z) is not dependent on x. If the maximum likelihood estimate of x is inside domain D, it can be calculated as the root of the equation to be described later.

The method of positioning used in a further embodiment of the invention, wherein the observation z includes CI and TA information from the serving BTS, CI's and RXLEV's (received levels) values from all associated BTS's, and x is the unknown coordinate of the MS. Based on maximum likelihood criteria as defined at 106 where

In a location service application, the CI knowledge of a particular cell includes not only the geographical coordinates of the BTS antenna, but also other known parameters such as antenna orientation, cell width, transmit power, and the like.

In a further embodiment of the present invention, ML estimation of MS coordinates is determined by solving the minimization problem described below.

Or, alternatively, it is determined by calculating the roots in domain D of the simultaneous equations described below.

In order to apply the ML principle, in the form expressed above, the domain and conditional probability density function f (x, y | z) of solution D needs to be determined. These are described below.

D is the domain where the location determination methods are described in the document look for solution x = [x, y] ^T. At 109, D may be defined using some existing information in the area where the handset may be located. There are several probabilities: four such methods used in embodiments of the present invention are described below.

1.D determined from the cell identity (CI) of the serving cell

In this case, D represents the topographic area where the handset is connected to the serving cell in which it is most suitably located. Thus, D may be defined as a trusted region associated with a location estimate based on serving cell CI information. The method for determining such trust zones is described in detail below. 8 illustrates such a trust zone. The confidence zone includes the difference in the area between the circle compartments 801, 803 of two radii that are not equal. The two compartments 801, 803 having a common origin 805 and a common arc angle 807 are defined by a set of parameters to be described later: origin 805 is coordinate (x ₀ , y ₀ ), inner radius R ₁ , at an unspecified radius R ₂ , azimuth angle 809 α and inclusion angle 807 β.

2. D. determined from cell identity (CI) and timing advance (TA) of the serving cell.

 As shown in FIG. 8, a method for defining parameters of a trusted area is used by a TA from a serving cell when CI and TA information is available from the serving BTS (in particular, the TA affects the radius R1 and R2). When possible, D can eventually be determined in the manner provided by ignoring cell partitioning later (ie, assuming α = 0 and β = 2π).

3. D. determined from the cell identity (CI) of all cells involved in the position calculation.

The coordinates of the BTS's associated with the location estimate provide themselves with an indication of the geographical area in which the MS is located. Thus, D may be defined, for example, by a convex polygon having a vertex at the coordinates of the outermost BTS's associated with the position calculation. The concept is shown in FIG. FIG. 9 shows six transceiver base stations (BTSs) at the boundaries of D defined by four BTS's 901, 903, 905, 907, and are thus located within quadrilateral 909 and D region boundary 909. FIG. To define two BTS's 911, 913.

4. D determined from convergence prediction maps

According to the same criteria proposed in application PCT / EP01 / 01147 to determine the confidence region of the position estimate, convergence prediction maps for serving and / or neighboring cells can be used to determine D. If TA information is available from the serving BTS, the circular crown centered on the serving BTS coordinates as inner radius R ₁ and outer radius R ₁ + R ₂ (as shown in FIG. 8) converges to determine D. Can be used in addition to the map. 8 shows this area as the area difference.

Representations for the probability density function f (x, y | z)

Some representations for the probability density function f (x, y | z) are described in detail below. In an embodiment of the present invention, observation z includes information resulting from reception level observation (RXLEV).

Given below definitions:

N is the number of reception level observations (RXLEV's) used to estimate the MS coordinates.

P ¹ _r , ..., P ^N _r are reception level observations (RXLEV's) measured by the MS from the associated N BTS's, expressed in decibels.

P ¹ _{t , max} , ..., P ^N _{t , max} are the maximum radiated power values from N BTS's measured in decibels. The i th maximum radiated power P ⁱ _{t, max} represents the maximum power at the output of the i th antenna in the maximum gain direction. As defined in equation (4), the numerical value P ⁱ _{t, max} includes the transmit power P ⁱ _t , the maximum gain G ⁱ _{t, max} of the BTS transmit antenna, antenna loss, cable loss, and the like; All are measured in dB. The total (positive) attenuation experienced by the signals transmitted by each of the associated N BTS's can be expressed in decibels during delivery to the MS as follows.

The vector z = [z ¹ , ..., z ^N ] ^T represents an observation based on that the MS location is estimated according to criteria 109 and 110. Therefore, the probability density function of the interest is

Using Bayes' theorem, f (x, y | z) can be expressed as:

F (z | x, y), which is required at (113), gives the result from the region where so-called multilevel observation L = [L ¹ , ..., L ^N ] ^T corresponds to -z as defined in this document. Can be calculated using The most common expression for f (z | x, y) = f _{L | x, y} (-L | x, y) is

는 동일 신호의 도착 각도(AOA)이고,

는 MS와 i번째 BTS 사이의 거리이며, (x¹,y¹),...,(x^N,y^N)은 연관된 N BTS's의 x,y 좌표들이다. 임의의 다른 이전 정보 없이, 상기 MS 좌표는 D에 균일하게 분포된다고 가정될 수 있다. 이 가정은 후술한 바와 같이 정의된 x 및 y의 결합 확률 밀도 함수를 야기한다:Where R _z (x, y) = E {zz ^T } -m _z m _z ^T = R _L (x, y) is a covariance matrix of z, m _Z (x, y) = E {z} = [ PL ¹ (d ¹ (x, y) + AP ¹ _tr (ψ ¹ (x, y), ..., PL ^N (d ^N (x, y)) + AP ^N _tr (ψ ^N (x, y) )] ^T = -m _L (x, y) is the mean of z (x, y), and PL ⁱ (d ⁱ (x, y)) and AP ⁱ _tr (ψ ⁱ (x, y)) are combined Path-loss associated with the transmit-receive antenna pattern and the signal transmitted by the ith BTS, each

Is the arrival angle of the same signal (AOA),

Is the distance between the MS and the i th BTS, and (x ¹ , y ¹ ), ..., (x ^N , y ^N ) are the x, y coordinates of the associated N BTS's. Without any other previous information, the MS coordinates can be assumed to be uniformly distributed in D. This assumption leads to a combined probability density function of x and y defined as follows:

Where M (D) is the size of D, and D may be determined using one of the methods described above.

Substitution of (115) in (113) causes the following expression for f (x, y | z).

This can be calculated using 112.

In practical implementations, it is very difficult to use 116 in the minimization problem 109. For this reason, alternative approximation definitions for f (x, y | z) need to be determined. f (x, y | z) can be written as the product of the probability density functions adjusted by each single measurement attenuation. f (x, y | z ¹ ), ..., f (x, y | z ^N ):

In 117, the i th probability density function represents the likelihood of (x, y) given the attenuation z ⁱ measured from the i th cell. When the signal received by the MS from the i th BTS undergoes attenuation z ⁱ , physically f (x, y | z ⁱ ) represents the spatial distribution of the MS. One embodiment of the present invention described below is defined by f (x, y) in such a way that the x, y coordinates are uniformly distributed in an area and defined by the radio convergence characteristics of the observed z ⁱ and / or i th cell. y | z ⁱ ) Further embodiments of the present invention described below do not assume that the handset is uniformly distributed because it is assumed that the handset is angularly uniformly distributed from the BTS position but not radially uniformly. It is assumed that the i th cell is omni-directional.

f (x, y | z ⁱ ): uniform distribution over area D _i

Assuming that the MS is uniformly distributed in the specific region D _i associated with the i th cell, f (x, y | z ⁱ ) is expressed as

Where M (D _i ) is the size of region D _i . Some probabilities exist to determine D _i .

1. D _i from coverage maps

If is generated by the coverage prediction tools coverage if available, D _i of 118 may be defined as the audible range H _i of the i-th cell. The audible region identifies the topographic region where signals emitted by the i < th > BTS reach the handset at the signal strength, which is the MS sensitivity level.

The definition 118 only considers the identity of the i th BTS and does not use the actual observed attenuation. A further embodiment of the invention readjusts the definition of f (x, y | z ⁱ ), including the measured attenuation, and the topography in which signals emitted by the i th BTS arrive at the MS with the attenuation observed from the i th BTS. It replaces the audible area with the coverage area N _i of the i-th cell identifying the in area. In practical terms, it may be beneficial to consider a range of values for attenuation instead of a single value; e. z ⁱ Δz ⁱ ± example, it is preferable to take into account the random variations in the attenuation Δz ^i. In this case, f (x, y | z ^z ) has the same definition as in 118 with H _i replaced by N _i .

2. Analytical Expressions for D _i

Without coverage maps, at _i can be represented by a number of analytic functions. One probability is to define D _i as the generated compartment as shown in FIG. 10. FIG. 10 shows the origin given in coordinates 1003 of i th BTS (x _i , y _i ), sector cell orientation 1005 φ ⁱ , sector width 1007 of second angular width and (forward) radius 1011 R. An area defined by D _i is shown as an area defined by a circle section 1001 having ⁱ _F. The defined area is also contained within the area 1009 of a smaller circle having the same origin 1003 and (rear) radius 1013 R ⁱ _B. Thus, the area D _i can be defined as:

는 i번째 BTS에 의해 전송되는 신호의 도착 각도이다. here,

Is the arrival angle of the signal transmitted by the i th BTS.

Has a definition 119 for D _i , and M (D _i ) in 118 is (R ⁱ _F ) ² Δφ ⁱ + (π−Δφ ⁱ ) (R ⁱ _B ) ² . In the most general definition, by setting Δφ ⁱ = π and R ⁱ _B = 0, D _i represents a sector cell, but not an omni-directional cell.

In an alternative embodiment of the invention, D _i may be defined as an ellipse having a center at the reference point of coordinates (x ⁱ _R , y ⁱ _R ) and having semi-axes σ _{x, i} and σ _{y, i} :

및 타원적인 기반을 가지는 3차원 원통이다. 일반적인 타원적인 정의는 만들어진 섹터 셀들의 커버리지에 대한 근사를 허용하나, 전방향 셀들에 적용될 수는 없다.(i번째 BTS 좌표에 원점 (x_R ⁱ,y_R ⁱ)을, (120)에서 σ_x,i=σ_y,i=Rⁱ _F를 설 정함으로써, D_i는 Δφⁱ=π이고 Rⁱ _B=0에 대하여 (119)로 얻어진 것과 동일하다.)Has a definition 120 for D _i , and f (x, y | z ⁱ ) in (118) is a constant height

And a three-dimensional cylinder with an elliptic base. The general elliptic definition allows an approximation of the coverage of the created sector cells, but cannot be applied to omni-directional cells (origin (x _R ⁱ , y _R ⁱ ) at the i th BTS coordinate, and σ _x at (120). By setting _i = σ _{y, i} = R ⁱ _F , D _i is Δφ ⁱ = π and is the same as that obtained in (119) for R ⁱ _B = 0.)

f (x, y | z): Non-uniform distribution for omnidirectional cells

In a further embodiment of the invention, the probability density function f (x, y | z ⁱ ) is the combined probability of the distance d ⁱ between the MS and the i th BTS and the arrival angle ψ ⁱ (x, y) from the same BTS. From the density function f (d ⁱ , ψ ⁱ | z ⁱ ) it can be calculated as follows.

here

In the example described below, the case of the omnidirectional cell is considered. In this example, the MS angular coordinates ψ ⁱ can be assumed to be independent of the radical coordinates d ⁱ and evenly distributed over the area [−π, π]. Using three assumptions, a simplified representation can be defined below for f (d ⁱ , ψ ⁱ | z ⁱ ).

Strictly speaking, the i-th observed attenuation z ⁱ generally depends not only on the distance between the MS and the i-th BTS but also on the gain of the transmitting antenna in the direction of the MS. If the cells are omnidirectional as assumed here, the antenna gain distribution can be ignored, which means that the BTS antenna radiates uniformly in all directions and the attenuation can only be considered as a function of the MS-BTS distance. Because there is. In this respect, the i th attenuation z ⁱ is equal to the i th so-called path-loss PL ⁱ (d ⁱ ), which depends only on the distance between the MS and the i th BTS:

Thus, in summary, the probability density function f (x, y | z ⁱ ) with omni-directional cells can be obtained by inserting (122), (123) and (124) into (121) as follows.

This has been shown in experimental and theoretical studies in which the average received signal power in the internal and external environment decreases logarithmically with the distance between the transmitter and receiver; Thus, the path-loss at distance d≥d _o can be expressed in decibels as

Where n is the environment-dependent transfer index, d _o is “close to the reference distance” and PL (d _o ) is the average path-loss experienced at the distance d _o from the transmitter. In free space, the value of n is 2, but the value n increases as the density of obstacles increases. Table 1 lists typical path-loss indices in different environments.

Table 1

Environment Transfer index, n Free space urban area cellular wireless shadow urban area cellular wireless 2 2.7 ÷ 3.5 3 ÷ 5 1.6 ÷ 1.8 4 ÷ 6 2 ÷ 3

d _o should also be chosen based on the environment. When the cells are large, d _o is generally set to 1 km; In the case of micro-cells, the "close to reference distance" is generally smaller (which can vary from 1m to 100m). Thus PL (d _o ) is calculated in the embodiment of the present invention using experimental data. When this is not possible, PL (do) can be estimated in further embodiments using the free space path loss law; If d _o is close enough to the transmitter, the ideal condition of propagation in free space can be assumed (λ = c / f is the signal wavelength, c is the speed of light, f is the frequency):

In the model 126, u represents shadow fading that affects the signal transmitted by the i < th > cell. It is generally modeled as a random variable with a log-normal distribution with a standard deviation σ _u (ie, measured in dB is a Gaussian random variable: u ~ N (0, σ _u )). Typical values for σ _u range from 5 to 10 decibels.

If you define it,

The model 126 can be rewritten as follows.

A well known model for path-loss according to (128) is the Okumura-Hatta model, where d is the distance between MS and BTS measured in kilometers, A and B are signal frequency f, base station effective antenna height h _B , Mobile terminal antenna height h _m and city type functions (either "big city" or "small / middle city"). The Okumura-Hatta model is close to the formula 126 given by d _o = 1 km, A = PL (d _o ) and B = 10n.

For a given path-loss PL, the distance d≥d ₀ has the following expression.

For using a given path-loss figure and the assumption of log-normal slow fading, the probability density function 129 of d can be calculated with a known random variable transformation technique:

을 위해 도입된 정규화된 인자이다. Where C (d ₀ ) is

Normalized argument introduced for

By inserting (130) into (125), the probability density functions of x and y for a given observed attenuation from the i th omnidirectional cell (see 124) are:

에 대하여 유효하다.

Valid against.

FIG. 11 shows a continuous plot of probability density functions f (x, y | z) using equation 132 for different values of attenuation z ⁱ = PL ⁱ . FIG. 11 shows that the probability density function has circular symmetry around the vertical axis through the i th BTS coordinates.

It can be seen that the probability density function is dispersed as the path-loss increases. At low values of attenuation, the probability density function curves approach the i-th BTS and curve the peak falling sharply to zero as the distance from the i-th BTS increases. For higher values of attenuation, the peaks of the probability density function curves move further away from the i th BTS coordinates. As the value of attenuation increases, the distribution becomes flat—ie, the peak probability density function value is smaller, but the rate of increase and decrease of the probability density function is lower. This variance produces a proposal that at higher levels of attenuation the probability of handsets is farther than the BTS location increase.

f (x, y | z ⁱ ): empirical Gaussian distribution for omnidirectional cells

In an embodiment of the present invention, the probability density function f (x, y | z ⁱ ) 132 is approximated by a two-variable Gaussian probability density function defined as:

here

To derive a Gaussian probability density function at (133), m and R (more precisely, its determinant | R | and its inverse R- ¹ ) must be determined. It can be seen that the mean values of x and y with the probability density function f (x, y | z ⁱ ) defined at 132 are as follows.

therefore,

In order to obtain a representation for the correlation matrix R of 136, the predicted values E {x ² | z ⁱ }, E {y ² | z ⁱ } and E {xy | z ⁱ } need to be calculated. .

;

;The analytical representation of E {x ² | z ⁱ } can be obtained as follows (the following second equation is obtained to solve the integral at the polar coordinates:

;

;

와 I_i는 각각 상기 제1 및 제3 적분들 나타낸다. here,

And I _i represent the first and third integrations, respectively.

The following definitions are used for the complementary error functions.

In a very similar development, it can be proved that E {y ² | z ⁱ } has the expression described below:

Eventually, an analytic representation of the correlation term E {xy | z ⁱ } can be obtained:

When using results 139, 142, and 143 in the definition for a given R in equation 136, R, which will be described below, obtains the following result.

이고 그것의 역행렬은 다음과 같다.The determinant of R is

And its inverse is

With the above conclusions, it is possible to define a Gaussian probability density function as shown in equation 139 as follows:

는 1이 되려 한다(실제로, (0.5)erf{c(-∞)}=1), Cⁱ(d₀)→1, 그리고In a further embodiment of the present invention, the use of the path-loss model 128 is extended to a distance less than or equal to a distance (e.g., d <d ₀ ), or alternatively the functions are maintained to work well and d is When we go to zero, we create a tendency for the proximity to be d ₀ → 0 so that the probability density function tends to be zero. When limit d ₀ → 0, term

Is going to be 1 (actually, (0.5) erf {c (-∞)} = 1), C ⁱ (d ₀ ) → 1, and

here

When inserting the above result for d ₀ → 0 into 133, the approximate Gaussian probability density function, which will be described later, obtains the following result.

12, the route - the role of I _i0 is shown as a loss / attenuation function z ^i. Figure 12 shows a graph in the range from I _i0 path loss / attenuation with respect to the z ⁱ 0 of 110 to 145dB 8. The graph plot is, z ⁱ = _i0 I have a slightly higher than zero in 110dB is similar to that of the exponential type is a slow increase in the initial ⁱ z increases rapidly as it passes through the 140dB plot.

f (x, y | z ⁱ ): nonuniform distribution for omnidirectional cells

In a further embodiment, the positioning algorithms may be applied to a non-uniform probability density function in the omni cell. By taking the algorithms of the terms from equation 117, the minimization problem can be rewritten as follows.

Or, alternatively

As mentioned above, equation 138 defines a probability density function f (x, y | z ⁱ ) in an environment that can be generated by omnidirectional cells based on a logarithmic path-loss model 128. Using the definition of the probability density function f (x, y | z ⁱ ), the representations for the partial derivatives to be described later can be known as follows.

here

Thus, the MS position can be estimated by solving a set of nonlinear simultaneous equations to be repeated later in further embodiments of the present invention.

Equation 154 defines Algorithm D, which is a nonlinear algorithm.

f (x, y | z ⁱ ): empirical Gaussian distribution for omnidirectional cells

The first two Gaussian approximations for the probability density function f (x, y | z ⁱ ) are defined. The first approximation is given by f _G (x, y | z ⁱ ) in equation (146). By using such a representation, the partial derivatives described below yield the following result.

로 주어진다. 따라서 상기 MS 위치는 본 발명의 실시예에서 반복적으로 후술할 비선형 연립방정식들의 세트를 풀어 추정될 수 있다. Where R's determinant

Is given by Therefore, the MS position can be estimated by solving a set of nonlinear simultaneous equations to be repeatedly described later in the embodiment of the present invention.

Equation 156 defines Algorithm E, a nonlinear iterative algorithm. The second approximation for the probability density function f (x, y | z ⁱ ) given in this section is f _G0 (x, y | z ⁱ ) in equation (149). This probability density function is obtained as a limit of the first Gaussian approximation f _G (x, y | z ⁱ ) that limits to the condition d ₀ → 0. Using such a representation, the problem is significantly simplified; In fact, the partial derivatives described below yield the following results:

Where μ _{x, i} = x ⁱ , μ _{y, i} = y ⁱ and I _i0 defined at (148) depends on the i th attenuation z ⁱ . As a result of this, the MS position estimate can be calculated as closed in the embodiment of the present invention as follows.

Equation 158 defines algorithm F, which is a linear algorithm with a closed solution.

Closed extensions

In the algorithm defined by equation 100 and used in the embodiments of the present invention, the estimated x (and angular y) MS coordinates are weighted of the x (and angular y) BTS coordinates whose signal is received by the MS. Obtained as an average. This algorithm can be extended by the formula described below.

Where w ¹ , ..., w ^N are appropriate weights assigned to each of the associated N BTS's.

The weights in algorithm 158 are calculated as the inverse of the terms I ₁₀ , ..., I _NO . The weights, calculated in such a way that the weight decreases as the attenuation increases, can be seen in FIG. 11. It is possible to suggest that signals transmitted by BTS's proximate to the MS experience less attenuation than signals transmitted by BTS's further away from the mobile device. This proposal may be considered by imposing a rule where the weights in the generalized closed algorithm 159 are defined in such a way that w ⁱ is high if the attenuation of the signal received from the i th BTS is small and vice versa.

The following three empirical definitions for the weights w ¹ , ..., w ^N in accordance with the proposed criterion used in further embodiments of the invention are described below.

1. Definition of Weights

The first definition of the weights reflecting the rules includes defining the i th weight as the inverse of the attenuation experienced by the signal transmitted by the i th BTS.

The definition of these weights is further enhanced in a further embodiment of the present invention by the introduction of auxiliary variable parameters that are determined using experimental measurements.

2. Definition of weights

의 역수를 사용하는 것이다. An additional definition of the i th weight used in the embodiment of the present invention is the estimated distance between the MS and the i th BTS obtained from the i th observation attenuation z ⁱ .

Is to use the inverse of.

에 의존적이다:The level of signal (and attenuation) received by the MS depends on the distance between the MS and the BTS as well as the gain of the transmit antenna in that direction. However, if the N BTS's involved in the position calculation are omni-directional, then the contributions of the BTS antennas can be ignored and the attenuation can only be considered to be a function of the MS-BTS distance. For this approximation, the i th attenuation z ⁱ is equal to the i th path-loss PL ⁱ (d ⁱ ), only the distance between the MS and the i th BTS

Depends on:

When ignoring slow-fading in model 128, the following expression for weights, as given at 162, produces the following.

An additional alternative may be found when the weights defined at 164 are adjusted to match the experimental measurements. The two observations described below can be made to produce more accurate weights to be found.

은 위치 추정에 기여한다. 이는, 만약 동일 상대 오류가 위치 계산에서 사용되는 각 거리를 추정에서 만들어진다면, 결과적인 위치 추정은 영향받지 않는다는 것을 의미한다. 상기 자동 오류 취소 인자는 외부에서 내부로 또는 도시의 개방 영역에서 좁은 도시 협곡으로 가는 것은 위치 정확성에 지대한 영향을 미쳐서는 안된다는 것을 의미한다. 본 발명의 실시예에서, 상기 에러 취소 인자는 위치 추정의 정확성을 향상시키고, 고도로 최적화되고 파라미터화된 경로-손실 모델을 사용할 필요를 최소화시킨다. Since the MS coordinates are estimated using 159 with the weights given by 162, the absolute value of the distances is not necessarily necessary. In the algorithm defined in equation (159), only the ratio

Contributes to the position estimate. This means that if the same relative error is made in the estimation of each distance used in the position calculation, the resulting position estimate is not affected. The automatic error cancellation factor means that going from outside to inside or from an open area of the city to a narrow urban canyon should not have a significant impact on location accuracy. In an embodiment of the invention, the error cancellation factor improves the accuracy of the position estimate and minimizes the need to use a highly optimized and parameterized path-loss model.

For position estimation calculations using algorithm 159 and weights 164, only a functional behavior of that distance is required as a function of received signal strength. As a simplified example, if the base station antenna height, center frequency, and maximum radiated power are equal to each BTS used for position calculation, most of the Okumura-Hatta path loss model terms will be canceled and the result used for position calculation Typical weights are simplified such that the weights can be found by equation (165).

3. Definition of weights

A further embodiment of the present invention experimentally determines the functional relationship between weights w ⁱ and attenuation z ⁱ . It is possible to collect a sufficient amount of reception level (RXLEV) measurements corresponding to the GPS coordinates used to refer to the correct MS location, and to solve the inverse coordinate problem from this information: if the MS coordinates are known in equation (159) If so, the weights w ¹ , ..., w ^N can be determined as a function of attenuation to minimize the error between the estimated MS coordinates and the correct MS coordinates. This approach fails where experimental measurements are not available.

The positioning methods described above apply to a network made from omni-directional cells, where the BTS's transmit antennas radiate isotropically in all directions. For sector cells, the assumptions are made only in all omnidirectional cell environments that maintain a reasonable level of approximate error when the MS is located in the main lobe region (MLR). On the other hand, when the mobile device is in that area, it is radiated by the main beam of the BTS (directional) antenna. A sector cell with a transmit antenna having a half power beam width (HPBW) of 60 ° has an MLR extending approximately ± 60 degrees at an angle with respect to the direction in which the BTS antenna is directed.

In a further embodiment of the invention, equations (154), (156), if the N BTS's associated with the position calculation beyond the total number of values received by the MS radiate the MS with the main beam of transmit antennas, The algorithms defined in 158 and 159 apply.

Unfortunately, it is not possible to determine whether the MS is in the MLR of a particular BTS from a single BTS received value, to obtain this information where the MS's location is required. However, in an embodiment of the present invention, it is possible to determine to have a low error rate if the MS is in the MLR of a particular sector cell when the MS measures and compares the levels (RXLEVs) received from different interlocated sector cells. Do. In this case, when the BTS antennas of the co-located sector cells have different orientations, the MS is likely to be located in the MLR where the signal suffers the lowest attenuation between the co-located sector cells of the cell.

With the knowledge collected by the method, if the N BTS's involved in the position calculation are selected other than those measured by the MS according to the procedure described below, the approximations found from the use of omni-directional cells are accurate (eg, the present invention). Embodiments may apply to the algorithms defined in equations (154), (156), (158), and (159).

1. Select all omni cells measured by MS.

2. Select all sector cells, where no inter-located cells are measured by the MS at the same time.

3. Select the sector cell with the lowest attenuation outside of the inter-located sector cells, where the inter-located cells are detected by the MS.

If one or more discarded inter-located cells produce a received level numerical attenuation that differs from the cells selected by the amount by a certain range (so-called 5 dB), select such inter-located cells and select an embodiment or present invention. In a further embodiment of the invention the attenuation average uses the attenuation values of all selected interlocated cells.

CI and CI + TA based algorithms

These algorithms will be described in more detail below. These algorithms are linear and only rely on serving cell information. Passes these location estimates and confidence regions for location estimates.

These algorithms are contemplated for operating mobile devices to operate when provided by a general measurement or generalized sector cell having coverage or analytically defined audible areas. The generalized cell is a simplification of the effect provided by the BTS antenna angular gain for transmission and reception as shown by the substantial antenna gain shown in FIG.

As previously described and shown in FIG. 10, the generalized cell is characterized by successive parameters including sector orientation φ, sector angular width Δφ, sector front radius R _F and sector rear radius R _B at an angle from the x-axis. Is defined.

Parameters defining the generalized sector can be found from coverage maps or coverage prediction tools. The input information required by the method includes a cell identity (CI) of the serving BTS, serving BTS coordinates and a map of the serving cell.

The steps performed to calculate generalized cell parameters from the coverage maps of cells include:

1. Use CI to select the coordinates of the serving BTS: x _s , y _s .

Given a CI of a serving cell, the xy coordinates of the cell are identified by CI, (x _s , y _s ) and recovered from a database of BTS coordinates.

2. Select the serving cell map: S

Analogously in the previous step, the CI is used to select the serving map S corresponding to the current CI. S is typically determined by a network planning tool using a large set of information such as BTS formation parameters, three-dimensional regional maps, coordinated delivery models, etc., taking into account the presence of other BTS's in the network to determine the serving area. do.

The serving map represents the geographic area provided by the cell identified by the CI. In an optional x-y Cartesian system, S can be obtained by dividing the region into elements of region (Δx) × (Δy) and representing each element by its central coordinate.

The coordinates (x _n , y _n ) represent the centers of the nth pixel of the region Δx × Δy provided by the cell.

An example of the serving area is shown in FIG. 13. 13 illustrates a base transceiver station (BTS) 1301, a main coverage area 1303, and a minor coverage area 1305. The transceiver base station 1301 is in the main coverage area 1303. The smaller coverage areas 1305 are positioned to be close to but not in contact with the edges of the main coverage area 1303. All coverage areas are divided into a number of elements 1309 by grid system 1307. The sum of all coverage areas dependent on the BTS is known as the serving area for the BTS.

3. Determine the center coordinates of the multiple serving maps: x _mc , y _mc

The (x, y) coordinates of the multiple centers of the serving cell are defined formally as follows:

Where M (S) is the area of the serving map.

By using separate elements defined in equation (166), multiple center coordinates can be calculated as follows:

4. Preferred estimates are much easier to calculate in the pole reference system ρ _i , θ _i starting at the serving BTS of the coordinates (x _s , y _s ). In this polar coordinate system, ρ _i is the distance of the point (x _i , y _i ) from the BTS, and θ _i is the angle measured counterclockwise from the x axis.

In the pole reference system, the serving map is represented by a set of points.

5. Determination of the Main Direction for the Entire Serving Map: θ _mc

The main direction θ _mc provides an indication of the azimuth angle φ _s of the serving sector. This angle is used as the reference direction to normalize the angular coordinates of the coverage area elements of the cell. [theta] _mc can be approximated as the orientation of the multiple centers of the cell from the BTS coordinates.

In the pole reference system ρ, θ, the majority center of the serving area has the following coordinates ρ _mc , θ _mc .

6. To further simplify the actual calculation of the estimates, the new pole reference system is rotated in such a way that the 0 and 2π angles are as far as possible from the most important serving area. After the rotation, everything needed for estimation can be calculated using known simple alignment algorithms.

The rotation of the mth pixel with angular coordinates θ _m is defined as:

here

와 연관되고, θ_mc에서 가장 먼 방향은

및

와 연관된다. The result δ _m 's is in the range 0 <δ _m <2π. The direction δ _m closest to the main direction θ _mc is

Associated with, the direction farthest from θ _mc is

And

Associated with

The next step is to organize the points in order of increasing distance from the serving BTS:

The serving map at the location is represented by an ordered set of points.

7. Determination of forward radius R _F

Where [●] rounds ● to the nearest integer to positive infinity (using a known sealing operation that pulls all numbers towards the next positive integer), where γ 'is from the serving BTS smaller than R _F. Determine the fraction of the desired point in the set S _polar to have a distance. In some embodiments of the invention, a good value for said parameter γ 'is between 0.95 and 0.98.

8. Determine the angular width as the rear radius R _B , the orientation as a function of the distance from the serving BTS φ _s (d) and the function of the distance from the serving BTS (Δφ _s (d)).

It is helpful first to determine the parameters of the circular crown shown in FIG. 14 shows a constant timing advance (TA) area defined by the difference in the area between two concentric circles with a common origin 1401. There is a first circle 1403 having a radius R ₁ 1407 and a second circle 1405 having a radius R ₁ + R ₂ .

The mathematical definitions of the crown in the pole reference system are:

Where R _inf is the inner radius and R _sup is the outer radius. In the example shown in FIG. 14, the first circle radius 1407 is R ₁ = R _inf and the second circle radius is R ₁ + R ₂ = R _sup .

Thus, this crown is represented by a set of points.

9. The steps described below to determine the rear radius R _B are as follows:

(a) set i = 0

(b) The number of points in the crown defined in equation (179) is calculated using R _inf = i.Δ and R _sup = (i + 2) · Δ, where Δ is a digitized element in the x and y directions Street

(c) If the serving region is omnidirectional at a distance R _sup , then there are approximately N _ideal = π (R ² _sup- R ² _inf ) / Δ ² points in the crown. If the number of actual points in the S _crown , N _crown is smaller than γ'N _ideal , the serving area is no longer considered omnidirectional. The value of γ '<1.0 is based on a good numerical experience of 0.75. If the serving area is not considered omnidirectional, the rear radius is chosen to be R _B = R _sup .

(d) If the serving area can be considered omnidirectional at distance R _sup , then set i = i + 1 and go to (b), otherwise an estimate of the rear radius is prepared.

10. In order to determine the orientation φ _s (d), the angles in the circular crown are arranged in order of increasing or decreasing angle.

From the definition of the crown R _int = d-2Δ and R _sup = d + 2Δ. Thus the estimate for φ _s (d) is the median of the angles and it is obtained as follows.

11. The angular width Δφ _s (d) as a function of distance from the serving BTS is calculated using the same alignment set of points in the circular crown as used for φ _s (d). The angular width Δφ _s (d) is calculated as follows:

12. As a final step, it is necessary to rotate Δφ _s (d) and φ _s (d) back to the origin coordinate reference system. This is done by simply adding θ _R to the obtained numerical values.

Where coverage maps are not available, it is possible to analytically estimate the cell front radius R _F and the cell rear radius R _B in the manner shown below.

As described above, the average received power P _R , such as by equation 33, can be expressed from the serving BTS by the distance d by the mobile device as follows:

The BTS transmits a known transmit power P _T and the gain of the antenna installed in the serving BTS in the direction of the MS is G. PL (d) is the path-loss that affects the signal power when transferring from the serving BTS to the MS. Accepted models, such as those used in equations 126-129 above, assume path-loss that increases logarithmically with distance d:

The sector forward radius is defined as the distance at which the average power P _R received by the MS is greater than a particular threshold P ^th _R defining the cell edges. Possible shadow fading effects are allowed, including the shadow fade margin FM _σ (all quantities are algebraic units).

가 반지름 추정의 신뢰성인 것과 같다. Under the assumption of log-normal slow fading, the fade margin can be defined as FM _σ = zσ, where σ is the standard deviation of slow fading and z is

Equals the reliability of the radius estimate.

According to equations 183 and 185 the front cell radius can be expressed in kilometers as follows.

Where G _m is the maximum antenna gain in dB.

The sector back radius can be calculated from the front radius including information of the radiation pattern of the antenna installed at the serving BTS location. FIG. 5 shows an annotated version of FIG. 4, where the plot of antenna gain in decibels relative to angle is measured in degrees. The antenna gain symmetrical with respect to angle θ = 0 includes the main lobe 501 and four side lobes 503, 505, 507, 509. The main lobe 501 is θ = 0 and a maximum value of G _m as well as the time of none does not have a larger gain than ρG _m of the side lobes (503, 505, 507, 509). ρ is the maximum rear-to-front ratio, the maximum gain between the antenna gain G _m in the maximum radial direction and the antenna gain in the outward direction of the main lobe. The cell front radius R _F is associated with a maximum gain G _m of the main lobe direction, the cell back radius RB is associated with a maximum gain in the outward direction of the main lobe ρG _m.

An approximation of the simple propagation model 183 and the approximate antenna gain G (θ) described below is given in decibels (G _m = 10 log (g _m )):

The upper limit for R _B , described below, can be written as:

미터를 부여한다. Typical values using the GSM900 are P _τ = 50 dBm and P ^th _R = -95 dBm. For a 30 meter BTS height, typical values for the variables are A = 124.5 and B = 35.7. In a mobile wireless environment, σ is often assumed to be 8 dB. For this example, the BTS antenna has an HPBW 65 degrees, a rear-to-front ratio of -18 dB (ρ = 0.0158) and a maximum gain of 12 dB. With these values, for z = 0.675, the formulas (186) and (188) yield R _F = 5.7425 kilometers and

Give the meter.

From the examples shown, it is clear that the representation of the sector as shown in FIG. 10 can be approximated poorly to actual cell coverage. In particular, Δφ _S is generally too large to represent the sector width in an area far from the BTS coordinates. In order to improve the inaccuracy in the estimation, it is possible to define the serving sector width as a function Δφ _s (d) of the distance from the serving BTS. In view of this, it is possible in the positioning algorithm to ignore an area that is clearly outside of the serving area. This applies equally to the sector direction φ _s which can be assumed to be the distance function φ _s (d) from the BTS position.

R _F , R _B , φ _s and φ _s (d), which in turn depend on the distance d, generate an analytical estimate of the boundaries of the serving cell as follows:

In an embodiment of the invention, the estimates of the generalized cell parameters are used to estimate the position of the mobile device MS. The algorithm used in the embodiment of the present invention is referred to as the classic CI-TA positioning algorithm as opposed to the map-assisted CI-TA positioning algorithm.

As previously described, the input parameters for the algorithm used in the embodiment of the present invention are the confidence coefficient ξ of CI, TA and position estimate (if available). Also as previously described, the output of the calculation is a position estimate and a confidence region (eg, an internal topographic region where the actual MS position is located within the confidence angle ξ). The result of the position calculation is different if the cell is sectored or omni-directional.

Embodiments of the invention described below may provide these results when CI and TA are available, and further embodiments described below also provide results only when CI is available.

의 추정 및 서빙 BTS 위치로부터의 MS의 각도 좌표

의 추정과 함께 서빙 BTS 좌표들 x_s 및 y_s를 조합하여 계산된다:If the serving cell is sectorized, the position estimate is MS to serving BTS distance

MS coordinates from the estimated and serving BTS locations of

Is calculated by combining the serving BTS coordinates x _s and y _s with an estimate of:

및

는 각각 이하에서 설명되는 추정된 거리 및 각도이다.

And

Are estimated distances and angles, respectively, described below.

Given the TA information, the method described below is used to estimate the distance between the MS and the serving BTS and to determine the radius of the circular crown C centered on the BTS coordinates at which the base station can be located with confidence γ. Can be used for

가 실제 거리 d의 50번째 백분위(또는 중간값)로서 계산된다. 즉,

는 다음과 같다.The distance estimation

Is calculated as the 50th percentile (or median) of the actual distance d. In other words,

Is as follows.

에 대한 다른 아날로그적인 정의들이 가능하다. 여기서 기준에 통합되는 유럽 특허 출원 번호 102251에 설명된 결과들을 사용할 때 중간 추정 거리는 다음과 같다. same as d

Other analog definitions of are possible. Using the results described in European Patent Application No. 102251 incorporated herein by reference, the median estimated distance is as follows.

Where X _1/2 is the median of TA measurement error X = d _TA -d, and

T _b = 3.69 ms is the bit period and c = 3 x 10 ⁸ m / s is the luminous flux.

는 φ_s와 동일하게 설정될 수 있다. 만약 섹터 방위가 서빙 BTS로부터의 거리의 함수라면,

는 TA에 기초하여 추정된 것과 동일한 거리에서의 섹터 방위이다. 즉,

이다. In the absence of any other information, the angular coordinates of the MS can be estimated in the orientation of the sector. In other words,

May be set equal to φ _s . If sector orientation is a function of distance from the serving BTS,

Is the sector orientation at the same distance as estimated based on the TA. In other words,

to be.

는 MS에 의해 수행되는 신호 레벨 측정들(RXLEVs)을 처리함으로써 결정되고, 미래에, 도착 각도를 사용하여 정보가 서빙 BTS 위치에 설치된 (스마트) 안테나 배열들에 의해 이용가능하도록 만들어진다. Additional meanings for estimating the MS angular coordinates can be incorporated into the algorithm found in the embodiments of the present invention; For example, in a further embodiment of the present invention

Is determined by processing signal level measurements (RXLEVs) performed by the MS, and in the future, using the arrival angle, information is made available by (smart) antenna arrays installed at the serving BTS location.

As shown in FIG. 8, the confidence zone R is associated with the estimated MS location and includes the parameters to be described below: an origin located at a point having coordinates (x ₀ , y ₀ ); Inner radius R1; Indeterminate radius R2; Azimuth angle α measured in the counterclockwise direction from the x-axis and inclusion angle β defining the width of the sector.

Analogously for distance estimation, the confidence area can be determined using the method described in European Patent Application No. 102251, which is incorporated by reference.

The origin of the trust zone is at the serving BTS location.

)은 알려진 맵 보조 CI-TA 추정 기술에 의해 제공되는 TA 측정의 통계적 인 특성들로부터 결정될 수 있다. 상기 거리 추정에 대한 신뢰 간격은 다음과 같이 정의되는 r_i, 및 r_s에 의해 결정된다. Given the confidence coefficient γ for the measured TA and the distance estimate, the confidence interval of the distance estimate (

) May be determined from the statistical characteristics of the TA measurement provided by known map assisted CI-TA estimation techniques. The confidence interval for the distance estimate is determined by r _i and r _s defined as follows.

백분위이다. r_i 및 r_s는 실제 MS-서빙 BTS 거리 d가 확률 γ를 가지고 신뢰 간격

내에 해당하는 것과 같다(도 15에 도시):Where X _{(1 ± γ) / 2} is the TA measurement error

Percentile. r _i and r _s are confidence intervals where the actual MS-serving BTS distance d has a probability γ

Same as within (shown in Figure 15):

Figure 15 shows a location estimate within the area defined by a transceiver base station (BTS) 1501, two calls 1503, 1505. 1507 is shown. Larger arcs have a radius d _sup and smaller arcs have a radius d _inf .

보다 더 작거나 또는 더 큰지 여부에 따라 다소 상이하게 계산된다. In an embodiment of the invention, the confidence zone parameters R1, R2, α and β are the lower bounds of the confidence interval for which the serving cell rear radius R _B is for distance estimation.

It is calculated somewhat differently depending on whether it is smaller or larger.

The cell rear radius R _B is strictly required, but provides sufficient information to improve the reliability of the method. Thus, in a further embodiment of the present invention, the value of the rear radius R _B is set to zero. In this case, only a portion of the cell in the direction of the main radiation lobe of the serving BTS antenna is included in the trusted area.

로 설정된다. 그러나

가 음수가 될 수 있고(즉, 만약

이고 TA 오류 통계가 긴 테일을 가진다면,

는 d_TA보다 클 수 있다.) 만약 R₁이 음수로 계산된다면, R₁은 0으로 설정되어야 한다. 더욱이, TA=0일 때 서빙 BTS 좌표들에 중심을 갖는 원으로서 신뢰 지역을 정의하는 것이 현명하다; 여기서, R₁의 경우 0으로 설정된다.The inner radius of the confidence zone is

Is set to. But

Can be negative (i.e. if

If the TA error statistic has a long tail,

Can be greater than d _TA .) If R ₁ is calculated as negative, then R ₁ must be set to zero. Moreover, it is wise to define a confidence zone as a circle centered on serving BTS coordinates when TA = 0; Here, in the case of R ₁ is set to 0.

In summary, R ₁ is defined as follows.

The unspecified radius R2 of the confidence zone is set equal to the confidence interval width r _i + r _s of the distance estimate. However, when the adjustment result is made at 197 to determine R ₁ , the exact definition of R ₂ is as follows.

The azimuth angle α of the confidence zone is determined by the azimuth and sector angle width of the serving sector. The width β of the confidence region is twice the width of the serving sector.

는 서빙 BTS로부터의 거리 d에 대한 셀 폭의 종국적인 의존성을 보존적으로 고려하기 위하여 정의된다. here

Is defined to conservatively consider the final dependence of the cell width on the distance d from the serving BTS.

로 정의된 호에 위치된다는 가정이다. 이는 또한 원형 크라운 γ를 사용하는 신뢰 계수가 거리 R₁ 및 R₂를 정의하기 위해 사용되는 신뢰 계수 ξ와 동일하다는 결과를 낳는다. The confidence zone has an angular width that is twice the angular width of the serving cell. Therefore, sectorization of the cell is considered. The fundamental reason for this is that in communications with BTS of interest, all mobile devices are angular (with the exception covered by equation (189) when R _B = 0).

It is assumed to be located in the arc defined by. This also results in that the confidence coefficient using the circular crown γ is equal to the confidence coefficient ξ used to define the distances R ₁ and R ₂ .

If d _INF is less than the serving cell rear radius R _B , there is a high probability that the actual position of the MS is in the rear radius region. In this case, the position estimate is previously calculated, but the confidence region is defined in such a way that it encompasses the entire rear radius region. This means that when the width β of the confidence zone is 2π and the confidence arc is defined as a full circle, the azimuth angle α provides extra information (α is set to zero). Moreover, internal and unspecified radii are given as:

In practical terms, the definitions are simply mentioned that the effects of the rear radius area are considered by treating the serving cell as omnidirectional when the MS is very close to the serving BTS position.

If the serving cell is omnidirectional, the concepts of sector orientation and angular width are meaningless. The best position estimate is provided using the coordinates given by the serving BTS position.

The confidence region is defined by a circle having the same radius as the upper limit of the confidence interval of the distance estimate.

및 r_s는 방정식 (192) 및 (195)를 사용하여 섹터화된 셀의 경우로서 계산된다.

And r _s are calculated as in the case of sectorized cells using equations (192) and (195).

, r_i 및 r_s는 오직 TA 양자화 규칙에 기초하여 정의될 수 있다.The calculations described above use statistical information from the TA measurement error. If TA measurement error statistical information is not available, the definitions for location estimation and the given confidence zone still apply; But the confidence coefficient is meaningless

, r _i and r _s can be defined only based on the TA quantization rule.

The algorithm shown above estimates MS coordinates and confidence region when CI and TA are available. If no TA information is available, the MS location estimate and its trusted area can still be defined using only the information carried by the CI. In particular, the confidence region can be defined if an estimate at cell radius R _F is available, and a position estimate can be provided even though R _F is unknown.

Where no TA information is available, embodiments of the present invention test the serving BTS to determine whether the serving cell is capable of sectorization.

If the cell is omni-directional and TA is not available, the position estimate is in accordance with embodiments of the present invention obtained in BTS coordinates.

에 의한 스케일링 R_F에 의해 얻어진다. 상기 인자는 반지름 R_F의 원형 셀의 모든 영역의 부분 ξ이 신뢰 지역에 포함되도록 선택된다. The confidence region of the estimate is a circle centered on the BTS location. The radius of such an area is

It is obtained by scaling R _F by. The factor is chosen such that the part ξ of all areas of the circular cell of radius R _F is included in the confidence zone.

If the cell is sectorized and the TA is not available, there are two alternatives to the position estimate.

The first alternative used in the embodiment of the present invention selects the MS location estimate to be at the serving BTS location. In this estimation, no cell radius R _F information is required.

에 의해 스케일 된다. If R _F is given, embodiments of the present invention compute the estimated MS position as the multiple centers of the simplified cell in the same form as shown in Figure 10, but only to ensure that a fraction ξ of the total area of the original cell is taken into account. Factor with the rear radius and the front radius

Scaled by

The coordinates of multiple centers of the serving cell are defined as follows.

를 가정함으로써, 다음과 같이 정의된다. Where S is the boundary of the home cell as described in equation 189 and M (S) is its region. Angular width of constant sectors, ignoring sector orientation and dependence of distance width

Is assumed to be defined as follows.

및 극 좌표들에서 푼 (208)에서의 적분들은 위치 추정

및

에 대한 후술할 표현을 부여하는 것이 보여질 수 있다.

And the integrals at solve (208) in polar coordinates are position estimates

And

It can be seen to give a representation to be described later.

The confidence zone has the same shape as shown in FIG. 8 with an inner radius of 0 (R ₁ = 0). The origin (x ₀ , y ₀ ) of the confidence zone is not located at BTS coordinates (x _s , y _s ), but along an axis defined by the sector orientation φ _s by the distance R _B ′ from the BTS coordinates ( In the direction opposite to the front side of the cell). Using this definition of the confidence zone, the rear radius zone is included (at least in part). The angular width β of the trusted region, calculated in light of the different positions of the origin and serving BTS coordinates of the trusted region, is defined as twice the angle Δφ _s ' ≠ Δφ _s . The azimuth α of the confidence zone is defined according to the cell orientation φ _s and the new variable Δφ _s '.

R _B ′ and Δφ _s ′ are determined from the terrain as shown in FIG. 16.

를 갖도록 정렬된 제2측면(1627)을 포함한다. 제1삼각형(1601) 및 제2삼각형(1605)은 또한 공통 우측 각도 꼭지점(1631)과 공통 측면(1625)을 포함한다. 상기 공통 측면은 두 삼각형의 제1측면들에 수직인 라인으로서 정렬되고 제1 및 제2 삼각형의 측면들의 교차하는(1635) 포인트에 연결된다. 상기 세그먼트(1607)는 원(1603)의 원점 x_s,y_s에 원점을 가지는 길의 R_F의 두 반지름과 상기 반지름의 종단들 사이에 정의되는 호를 포함한다. 제1반지름은 제2삼각형의 제2측면(1605)에 공통되고, 제2반지름은 제1반지름으로부터의 각도

이다. 그에 따라 단순한 삼각법을 이용할 때, 제1 및 제2삼각형 측면의 높이는

와 동일하게 보여질 수 있다. 도 16은 더욱이 원의 원점(1611)에서 제1삼각형의 제1측면(1623)에 연결되며 수직인 라인(1651)과, 상기 라인의 교차점 및 제1삼각형의 제2측면에서의 제2종단을 가진다. 상기 라인(1651)은 길이 h를 가진다. 16 illustrates the terrain used in an embodiment of the present invention to calculate a confidence zone. The shape includes a circle 1603, a first triangle 1601, a second triangle 1605, and a circular segment 1607. The circle has an origin 1611 at the BTS locations x _s , y _s and has a radius R _B. The first is the first vertex 1609 at position x ₀ , y ₀ , the first side 1623 at which one end is connected to the first vertex 1609 passing through the origin x _s , y _s of the circle, and A second side 1621 connected to one end of the first vertex and aligned so that the angle at the first vertex 1609 defined by the two sides 1623 and 1621 is Δφ _s '. The vertex 1609 of the first triangle is located at a distance R _B ′ from the origin of the circle. The second triangle 1605 lies in an area of the first triangle 1601 and is connected to a first vertex located at the origin of the circle, and connected to a first vertex of the second triangle at one end thereof, A first side 1629 that forms part of the first side and an angle between the first side 1629 and the second side 1627 connected to the first vertex at one end;

And a second side 1627 aligned with. The first triangle 1601 and the second triangle 1605 also include a common right angle vertex 1631 and a common side 1625. The common side is aligned as a line perpendicular to the first sides of the two triangles and connected to an intersection point 1635 of the sides of the first and second triangles. The segment 1607 includes an arc defined between the two radii of R _F of the path having an origin at the origin x _s , y _s of the circle 1603 and the ends of the radius. The first radius is common to the second side 1605 of the second triangle, and the second radius is an angle from the first radius.

to be. Thus, when using simple trigonometry, the heights of the first and second triangle sides

Can be seen the same as FIG. 16 further shows a line 1651 perpendicular to the first side 1623 of the first triangle at the origin 1611 of the circle and a second end at the intersection of the line and the second side of the first triangle. Have The line 1651 has a length h.

가 π/2보다 더 큰지 또는 더 작은지 여부에 따라, R_B' 및 Δφ_s'를 결정한다. As described below, embodiments of the present invention

Depending on whether is greater than or less than π / 2, R _B ′ and Δφ _s ′ are determined.

＞π/2이면, R_B'가 R_B와 동일한 것으로 정의되는 경우 신뢰 지역의 외부에 남겨지는 R_B에 의해 정의되는 단지 원의 매우 작은 영역으로 보여질 수 있다. If R _{F '} ' R _B

> Is π / 2, when R _B 'is defined to be equal to R _B may be viewed in a very small area of the circle, which is defined only by the R _B are left on the outside of the confidence region.

에 대한 분석적인 표현이 도 16에 도시된 지형으로부터 계산될 수 있다. 그 결과는 다음과 같다. Using the definition for R _B '

An analytical representation of may be calculated from the terrain shown in FIG. 16. the results are as follow.

Using the terrain shown in FIG. 16, Δφ _s ' is obtained as follows.

And using this result R _B 'is easy to calculate as follows.

로서 근사될 수 있다. 이 근사값은 단지, 상기 신뢰 지역은 너무 작고 반지름이 R_B인 원의 충분히 큰 일부분을 포함하지 않거나(즉, R_B'의 수치는 너무 작다), 역으로 R_B'는 너무 큰 신뢰 지역을 이끌어내기에 너무 크다는 것을 검사한 후 수행된다. 이러한 경우들에 대하여 검사하기 위하여 두 선택된 파라미터들이 도입된다: δ_min(너무 작은 RB' 예시에 대비하는 것)여기서 1.5는 δ_min에 대한 전형적인 선택이다) 및 δ_max(너무 큰 RB' 예시에 대비하는 것-그 4-5는 δ_max에 대한 전형적인 선택이다).The length h is

Can be approximated as This approximation only indicates that the confidence region is too small and does not contain a sufficiently large portion of the circle with a radius of R _B (ie, the value of R _B 'is too small), or vice versa, R _B ' leads to a too large confidence region This is done after checking that the bet is too big. Two selected parameters are introduced to check for these cases: δ _min (vs. too small RB 'example), where 1.5 is a typical choice for δ _min ) and δ _max (vs. too large RB' example). 4-5 are typical choices for δ _max ).

If R _B ′ is found to be too small, ie, δ _min R _B > R _B ′, R _B ′ is redefined as follows in an embodiment of the present invention.

And Δφ _s ' is obtained using the terrain as shown in FIG.

If R _B 'turns out to be too large, i.e. δ _max R _B > R _B ′, R _B ′ is redefined as follows.

And Δφ _s ' is calculated from equation (217).

A proposed location procedure incorporating at least some location determination methods using CI, TA or Rx information is shown in FIG. The procedure can be integrated into the represented location determination methods (CI, CI + TA and CI + RX algorithms). The above process is described below: First, in step S1, timing advances and received signal levels measured by the base station from the serving and neighboring cells are collected.

Next, in step S2 wireless network parameters such as associated algorithm input parameters and transceiver station coordinates for the measured cells are collected.

In step S3, the measurements and network data are analyzed to select which measurements should be used in the position calculation. In order to improve position accuracy, cells that do not contribute significantly are removed from the measurement set. As a result of the cell selection procedure, serving cell information may potentially be removed, and / or neighbor set information may be removed wholly or partially as possible.

In step S4, a cell identity / cell identity plus timing advance position estimate is made. If serving cell information is available, timing advances based on cell identity or cell identity and a positioning algorithm as described above may be used to determine location estimates and / or confidence regions for the serving cell. The resulting position estimate is called in this document "CI / CI + TA position estimate".

In step S5, an RX position estimate is made. If neighbor cell information is available, CI + RX based on the location determination algorithm is applied. This may be any of the algorithms A to F described above, or may actually be any other suitable algorithm. This is used to estimate the base station location. In selecting CI + RX based on algorithms, criteria to be described below may be considered.

Cell Sectorization. For omni cells, CI + RX based on algorithms can be used. For example, algorithms D (Equation 154), E (Equation 156) or F (Equation 158) can be used.

For sector cells, CI + RX based on algorithms such as Algorithms A, B and C described above can be used directly. Alternatively, algorithms D, E, and F can be extended to the case of sector cells as described above.

Absolute / Relative Level Observations

Algorithm C, described below, uses the difference in relative level observations, i.e., received signal level measurements from a transceiver station pair. It should be recognized that the algorithms A, B, D, E and F use absolute level observations.

 Closed / Repeated Algorithms

Algorithms A, B, C, D and E are iterative algorithms. Algorithm F is a closed algorithm. This determines the base station coordinates as the weighted averages of the coordinates of the transceiver stations from which received signal level measurements are collected. Three empirical approximations for one analytic expression (Equations 163 and 164) and weights are also provided as discussed above.

 RX-based domain determination

The RX-based algorithms may require a limited "domain" determination in order for the RX-based position estimate to be placed. Such a domain can be defined, for example, using any of the definitions provided above.

The obtained position estimate is referred to as "RX position estimate" for convenience.

In step S6, the RX direction position estimate is obtained. If serving cell information and "RX location estimation" are available, the direction from the serving transceiver base station to "RX location estimation" is calculated. This direction is used instead of the actual antenna direction as input to the algorithm based CI + TA, as described above. The resulting position estimate is referred to as "RX direction position estimate" for convenience.

In step S7, the virtual BTS position estimate is calculated.

If serving cell information and "RX-direction position estimation" are available, the coordinates of the "RX-direction position estimation" are used as additional (omni-directional) neighbor cells for the RX based algorithm. The RX level measurement associated with the additional neighbor cell and the coordinates may be referred to as "virtual transceiver station" and "virtual RX level", respectively. The value for the virtual RX level measurement may be selected, for example, as the average RX level from the measured neighbors or as the maximum RX level from the measured neighbors. The resulting position estimate is called "virtual BTS position estimate."

It should be appreciated that the position estimate obtained by any of the position determination methods discussed may be used as additional or virtual base station measurement. The virtual base station measurement is associated with a virtual measurement such as, for example, a virtual RX level, actual measurements and a set of virtual measurements, and the virtual measurement is reprocessed using any of the described location determination methods.

In step S8, any one of the obtained position estimates-CI / CI + TA position estimation, RX position estimation, RX-direction position estimation, RX position estimation, RX-direction position estimation or virtual BTS position estimation- It can be selected as the position estimate passed by.

Embodiments of the present invention are particularly flexible. For example, if TA measurements and / or serving cell information are not available, the process is still possible to provide location information such as RX location estimate. As such, if neighbor information and / or RX level measurements are not available, the process may still send location estimates using CI / CI + TA location estimate.

Embodiments of the present invention enable RX levels and TA measurements to be used effectively. This information is currently provided for standard GSM networks. The combined use of timing advance and RX level measurements "virtual BTS location estimation" gives the most accurate location estimates in the most dense areas of mobile users. In an embodiment of the invention, one of the above estimates is selected. Where one or more estimates are obtained, averaging or alternatively weighting the results may be provided.

Where a number of different position estimates are determined, only one of the above methods can be used. Any suitable criteria can be used to determine that the criteria of the methods are used.

While embodiments of the invention have been described in particular in the context of alignment in which many different methods are used, in some embodiments one of the positioning algorithms described above may be used alone.

Claims (26)

A method of estimating a location of a mobile device in an area covered by a network and divided into a plurality of cells, the method comprising: Collecting location information; Selecting at least one of a plurality of different location determination methods to provide a location estimate, wherein the selected one method includes determining a virtual base station location estimate; And Using the collected position information, providing a position estimate based on the at least one selected position determination method. The method of claim 1, wherein the at least one of the plurality of different positioning methods is as follows: Using cell identity information; Using cell identity information and received signal strength; Using cell identity information and timing advance information; And And at least one of a method of using cell identity information, received signal strength information, and timing advance information. delete 3. The method of claim 2, wherein the virtual base station position estimation comprises at least one of the following methods. 2. The method of claim 1, wherein the virtual base station position estimate is associated with at least one virtual measurement and at least one actual measurement, wherein the at least one virtual measurement is processed using a location determination method. 6. The method of claim 5, wherein the at least one actual measurement and the at least one virtual measurement are processed using the positioning method defined in claim 2. 6. The method of claim 4 or 5, wherein the value for the virtual measurement is one of measured levels, a combination of measured levels, and an average of measured levels. 2. The method of claim 1, wherein the at least one location determination method is selected depending on available location information. 2. The method of claim 1, wherein a plurality of position estimates are determined from the plurality of different positioning methods, and at least one estimate is used to provide the position estimate. The method of claim 1, wherein said location determination information is collected by said mobile device. 10. The method of claim 9, wherein the mobile device is adapted to measure the level of at least one type of information. 2. The method of claim 1, wherein the positioning information comprises at least one of timing advance information and a received signal level. 13. The method of claim 12, wherein the received signal level is an absolute received signal level or a relative received signal level. delete 2. The method of claim 1, wherein said information is collected for a serving cell of a mobile device. 14. The method of claim 13, wherein said information is collected for at least one neighbor cell. The method of claim 13, further comprising the step of selecting the cell or each cell to which the location information is to be collected. The method of claim 1, wherein the position estimation is performed in the following steps: The signal transmitted by the i th BTS during the transfer of the i th level observation to the base station L ⁱ by extracting the maximum power P ⁱ _{t, max} radiated by the i th BTS from the i th received received power P ⁱ _r To calculate the total attenuation experienced:

Loading level observations from N BTS's into vector L:

Steps to solve the minimization problem:

Is provided using an algorithm that includes, Wherein the cost function F (x, y; σ ² _u ) is

And

Position estimation method, characterized in that defined by. The method of claim 1, wherein the position estimation is performed in the following steps: The signal transmitted by the i th BTS during the transfer of the i th level observation to the base station L ⁱ by extracting the maximum power P ⁱ _{t, max} radiated by the i th BTS from the i th received received power P ⁱ _r To calculate the total attenuation experienced:

 Loading level observations from N BTS's into vector L:

Steps to solve the minimization problem are as follows:

Where the cost function F (x, y) is

And D _xy is the domain of x and y presence; And

of

Position estimation method, characterized in that provided using an algorithm comprising the step of calculating as. The method of claim 1, wherein the position estimation is performed in the following steps: The signal transmitted by the i th BTS during the transfer of the i th level observation to the base station L ⁱ by extracting the maximum power P ⁱ _{t, max} radiated by the i th BTS from the i th received received power P ⁱ _r To calculate the total attenuation experienced:

Calculating the j-th level difference observation by extracting the j-th level observation from the level observation L ¹ taken as a reference:

Loading the difference N-1 of level observations into the vector D:

Steps to solve the minimization problem:

Is provided using an algorithm comprising:

And

And D _xy is the domain of the presence of x and y. The method of claim 1, wherein the position estimate is provided using an algorithm for obtaining the equations of x and y:

here

 Position estimation method, characterized in that. 2. The method of claim 1, wherein the position estimate is provided using an algorithm for obtaining the following equations of x and y.

2. The method of claim 1, wherein the position estimate is provided using an algorithm based on the following equation.

2. The method of claim 1, wherein the position estimate is provided by either an iterative method or a closed method.  2. The method of claim 1, wherein the position estimate is provided by a linear method and a non-linear method. A system for estimating a location of a mobile device in an area covered by a network and divided into a plurality of cells, the system comprising: Means for gathering location information; Means for selecting at least one of a plurality of different location determination methods to provide a location estimate, wherein the plurality of different methods use cell identification information, and the selected one of the methods selects a virtual base station location estimate. Including determining; And Means for providing a position estimate based on the at least one selected position determination method, using the collected position information.

Images