A method and system for locating a mobile radio receiver in a radio system with multiple transmitters

Method and system for positioning a mobile radio receiver in a radio system with multiple transmitters

The present invention relates to a multi-transmitter radio system, in particular, it relates to locating a mobile radio receiver in a system with a plurality of transmitting stations, each of which is located in a different geographic location, so as to provide the relevant area with Adequate radio coverage.

Traditional systems that use multiple transmitters to locate a radio receiver use different techniques to measure the characteristics of the received signal in order to find the location of the receiver. The characteristics of these technologies are time reference, phase reference or spatial reference, or a combination of these three types of technologies.

In time reference systems, such as the Loran C system established by the United States Coast Guard, radio transmission takes the form of radio pulses, which can accurately measure the time when these radio pulses reach the mobile receiver. The receiver is a special type of receiver that can determine the time difference between receiving pulses from each transmitter of a pair of transmitters in the system. This information defines a curve connecting equal time difference points, and the receiver must be located on the curve. Through similar measurement of multiple receiver pairs, further making the curve intersection can determine the receiver position as the curve intersection point that best satisfies all the measured time differences.

An extended form of this system is the Global Positioning System (GPS), which uses signals received from a satellite transmitter that is synchronized in time and also transmits its position to the receiver. The receiver measures the time of arrival at a defined point in the data stream received from each satellite with reference to its internal time source. Each measurement can define a sphere on which the receiver must be located, and by radio The wave from the satellite's flight time fixes the radius of the sphere. Using different satellites at different locations to make similar measurements can create multiple intersecting spheres. By solving the simultaneous equations of the distance to the satellite and the time offset of the receiver, the position of the receiver can be determined to best satisfy the spherical intersection point of all the measured time delays and the receiver clock offset.

Systems that use phase measurement can include Decca navigation systems and Omega systems. The mobile receiver measures the relative phase of signals received from a special transmitter with phase-synchronized transmission, or relative to each other with respect to an internal phase reference. By knowing the location of the transmitter and the predicted phase received at all points in the coverage area, the location of the receiver can be determined that is consistent with a specific set of phase measurements.

All of the aforementioned systems require a dedicated and expensive transmitter system with specialized receivers.

Another prior art method, such as US-5859612-A, uses the directional nature of the transmitted signal. This point has been applied to a method that uses an identifiable modulated beam that scans a defined area synchronously from a known starting position and starting time. By knowing the time when the beam starts its scan and the receiving time of the defined modulation, the position of the receiver can be determined as a radial line from the source transmitter in a specific direction. By determining points that best satisfy all radial orientations at the same time, similar reception from multiple sources can fix the two-dimensional position of the receiver.

A spatially fixed beam of radio transmission with a narrow emission pattern from a fixed antenna can also be used to determine the position of a receiver. By knowing which directional pattern of the overall emission pattern is received, the receiver can be determined The position within the sector centered on the radio station. By searching for points that best satisfy all the beam directions to the launching station, receiving narrow beams from multiple overlapping emitter patterns can similarly fix the receiver in space.

The above systems also require dedicated transmitters, but these receivers are usually limited to the specific area concerned. When the beam width and therefore the precise orientation of the receiver increase, the accuracy of the system decreases rapidly. This receiver is usually less specialized than a time or phase measurement receiver, and can share its positioning function with other functions such as weather forecast reception.

The above-mentioned final method of using directional patterns transmitted from fixed base transmitter stations can be applied to digital radio cellular systems. In this type of system, a plurality of fixed transmitting and receiving stations (usually co-located) are used to provide services for the area to be covered by the communication service. The area covered by a single fixed transmitter point on one of its directional antenna patterns is called a sector or a unit. The combination of all unit positions has been designed so that a large number of mobile transmitters and receivers in a large geographic area are in a continuous communication, and at least one of these fixed transmitter-receivers (servo base station) serves as the mobile station edge Move in any direction. The fixed transmitter-receiver in this system is usually referred to as a base station, and the mobile receiver can receive the transmission of the adjacent base station because its antenna pattern overlaps with the antenna pattern of the servo base station. In any particular situation, when the receiver moves relative to the fixed network, the base station used by the mobile receiver to connect to the basic service network will change accordingly. This procedure is called handover, which ensures that the communication traffic of the mobile receiver is smoothly transferred from one base station to another, so that there will be no delay in the service.

In order to realize the handover procedure, the cellular mobile receiver receives from a plurality of base stations Receive and launch. The related art shows how these transmissions for telecommunications purposes can also be used by a mobile receiver, or in a processing device in communication with the receiver, in order to calculate the location of the receiver. These prior art methods and their shortcomings will now be explained.

As described in US-6108553-A, a mobile receiver can measure the relative time delay between the two signals received from a time-synchronized base station and those transmitted by it. When taken from several pairs of different base station launches, these measurements form a set of intersecting hyperbolas. The point where the intersection of all these curves best meets can predict the location of the receiver.

The mobile receiver can also measure the signal received from each base station with reference to its internal clock, for example, see WO 97/11384 or WO 99/21028. This reception time can then be combined with measurements made by other external measurement receivers in the network for the same emission. The combined measurement information can then be combined in a mobile phone or other generalized computing component to provide a set of intersecting circles centered on the transmitter. The position of the mobile phone can be calculated as being at the intersection of the circle.

The aforementioned technology requires the installation and maintenance of additional equipment at the receiver in the network to implement timing measurement, and the normal communication function of the mobile phone must also be modified to perform additional timing measurement.

The prior art described in US-5293642-A also describes a position estimation technique using the spatial characteristics of radio waves applied in a cellular system. In these systems, the base station receiver measures the radio propagation parameters of the signal transmitted by the mobile station transmitter. These parameters can include signal strength and travel time. Then use these parameters and the knowledge of the base station antenna pattern to derive the probability density function of a mobile receiver (centered on each base station), from And provide the predicted signal strength of any location. By combining this probability density function with a similar probability density function (centered on other base stations), the maximum value of the combined probability density function can be derived, which can be interpreted as the most likely position of the mobile station.

In other prior art, such as US-5613205-A, a mobile receiver measures the relative signal strength of control signals received from a plurality of base stations. The system can then calculate the distance between the receiver and each base station as a function of the base station's transmit power, received power, and the signal attenuation factor in the radio path from each base station to the mobile receiver.

It also illustrates a simpler system for locating mobile receivers. One of these systems uses the center of mass of the servo unit as the location of the mobile phone. Another system measures the signal strength of adjacent units and uses the centroid of the unit with the largest signal level as the best indication of the receiver position, as described in WO 98/35524.

These specific prior art systems and methods are not reliable for estimating the position, because the center of the servo unit may not be the most approximate receiver position, and the shape of the unit is uncertain, and abnormal propagation conditions may be encountered. The signal level measurement of a single base station signal may also be wrong for similar reasons. These technologies suffer from the highly unpredictable nature of the radio propagation path and the imprecision of mathematical models. In particular, the estimated power level usually does not conform to a simple Gaussian distribution, especially when subjected to multipath attenuation. Of course, by using a mobile test set for calibration of location and signal strength, the actual power received at several locations in the network can be determined in advance. However, making such measurements is very expensive and time-consuming, and does not take into account the nature of the actual network that can change over time between measurements.

The present invention is designed to overcome many of the shortcomings of the aforementioned prior art systems. In particular, the location of a mobile receiver in a communication network can be determined with relatively high accuracy using only the measurements made by the receiver, without the need to provide additional equipment in the network, and without the need to implement signal reception cyclesSexcalibration.

According to the present invention, there is provided a method for determining one or multi-dimensional position of a mobile receiver in a radio system including a network of transmitters whose positions are known. The method includes the following steps: a) Compiling the receiver A list of detected signals; b) sending the list to a computing device; c) constructing a position function related to the probability that the receiver is at a given position, the position function corresponding to each component of the list Each component includes a predetermined probability function depending on the position of the mobile receiver; and d) evaluate the position function to determine a position corresponding to the highest position probability, and define the position as the position of the mobile receiver.

The detection of the signal by the receiver may include receiving the signal transmitted by a specific transmitter of the network and decoding information from the signal. The information also includes the identification of a specific signature in a received signal above a given threshold power, such as a synchronization pattern or a preamble, or only a signal above a given threshold received at a specific frequency power.

The list of signals detected by the receiver may, for example, include decoding information, such as transmitter identification, or may be determined by the frequency of receiving signals above a given critical power. The list of rates is composed of, or can be composed of a list of references to another list that is known to the computing device and contains information about the transmitter.

The computing device can be attached to the mobile receiver, or it can be located anywhere that can receive the list sent by the receiver.

The predetermined probability function may be based on a radio wave propagation model between each transmitter and the receiver assumed to be at that location, which considers variable transmission characteristics such as transmitter power, transmitter antenna pattern, travel distance, and signal propagation ( Such as attenuation) and other characteristics. Preferably, the predetermined probability function is proportional to the probability of detecting the signal from the transmitter, and therefore proportional to the eligibility of the list of possible components. Alternatively, the predetermined probability function may be based on a heuristic function derived from the detection obtained from the calibration point. For example, in practice, it can be found that the signal from the transmitter (located at a specific point in the area covered by the transmitter) can be detected by a two-dimensional Gaussian function centered near the middle of the area. Approximate value of probability. Using such heuristic functions can provide very good results because they are based on actual measurements rather than theoretical models.

If the signal propagation characteristics are better, the list can be augmented, or a second list can be constructed to include information from signals that have not been reported by the receiver but have been detected. In such cases, the position function can be modified to take into account the probability of not appearing in the list.

The position function may also be affected by the reliability of the receiver's detection of each signal. The reliability estimate can be, for example, proportional to the difference between the received power level and the detection threshold, or it can be a measure of the number of times a given signal is detected in a series of repeated measurements. Can be transmitted from the mobile receiver The reliability metric is transmitted in the message as part of the list or in a separate message.

In some applications, it may be advantageous for the receiver to sort the list before transmitting it. For example, the received signal power or the aforementioned reliability metric can be used for ranking.

Digital communication systems often measure the round-trip flight time of the signal from the servo transmitter to the mobile receiver and from the mobile receiver to the servo transmitter. The measurement can be used to advance the timing of the signal transmitted by the mobile phone transmitter so that the signal can reach the base station at a known time. When this so-called timing advance (TA) is available, it is preferably included in the position function, for example by assigning an additional probability that the mobile receiver is at a given position (given the measured value).

The position function can be used in several ways to calculate the position of the receiver in order to find the position corresponding to the highest probability. Preferably, the position function is a "cost function", which describes the logarithm of the reciprocal of the probability that is a function of the receiver position and the eligibility of the list component. The lowest value of this cost function corresponds to the most likely to provide a list of measured detected signals, including (if available) the list of undetected signals, the measured TA, and the sorted position of any reports in the list. The minimum value can be found by any of several well-known general mathematical minimization techniques. These techniques include (a) an analysis method that can manipulate position functions to directly provide solutions, and (b) usually seek to "tilt" in one direction. The iterative method of moving the current estimated value of the solution to a position closer to the minimum point, and (c) the "grid method" in which the position function is evaluated at each point of a position grid and will be compared with The position corresponding to the minimum value is taken as the result. These methods are explained in many textbooks, for example, C++ Numerical Methods Second Edition by William H. Press et al., Cambridge University Press, Chapter 10, pages 401 to 429.

Therefore, the present invention provides a method for obtaining one or multi-dimensional position of a mobile receiver by estimating its position in, for example, a geographical area with a number of receivable and identifiable radio transmissions established within it. This method does not rely on detailed knowledge of the type or content of the emission modulation or time delay measurement. Therefore, the method can be applied to any radio system of any modulation type without requiring knowledge of system timing.

A mobile receiver at a given location receives signals from one or more transmitters in the network. Launch stations that can be detected are called "visible" launch stations and can be added to the list of visible stations maintained. The location of the receiver affects which transmitter stations it can see. A system according to the present invention can also maintain a list of transmitters that cannot be received during position measurement and are called "invisible", but are known to exist and are active. The location of the receiver affects which transmitting base stations are not visible.

At any given base station, it is theoretically possible to predict which base stations are visible and which base stations are invisible, depending on factors such as the distance from the transmitting station, the transmitting power of the transmitter, and the detection of the transmitter's radiation pattern. According to an embodiment of the present invention, a "cost" function is assigned to each transmitter in the visible and invisible list. This cost function describes the combination of transmitter power, transmitter radiation pattern, distance and azimuth from the transmitter to the receiver, radio propagation characteristics, and (if available) other characteristics, such as timing advance or local layout data, It best simulates the probability of a mobile receiver detecting a signal. Based on mathematically convenient considerations, it will make the combination The cost function of all characteristics that affect the visibility of the transmitter is proportional to the negative of the logarithm of the probability of detecting its signal (ie, the logarithm of the reciprocal of the probability). Therefore, after a transmitter is isolated, when the receiver is close to the transmitter and directly coincides with the forward lobe of the transmitter antenna, for example, as shown in FIG. 1, the cost function of its visibility has a low value. As the receiver moves directly away from the base station, or moves tangentially at an angle to the direction of the forward lobe at a fixed distance from the transmitter, the cost function gradually increases. The corresponding cost function of each transmitter is combined into a sum to give an overall cost function that illustrates the combined visibility of the transmitter shown in FIG. 2, for example. The minimum value of this overall cost function is the location of the receiver according to the present invention.

According to the present invention, there is provided a mobile receiver device capable of detecting signals from a plurality of transmitters and configured to calculate a list of signals received by the mobile receiver and send the list (in use) to a Computing device.

In addition, according to the present invention, a computing device is provided for determining the position of a mobile receiver. The computing device is configured to: a) receive a list of signals detected by a receiver, and b) construct a A position function of the probability that the receiver is at a given position, the position function having components corresponding to each component of the list, each component including a predetermined probability function that depends on the position of the mobile receiver; and c) evaluating the position Function to determine a position corresponding to the highest position probability and define this position as the position of the mobile receiver.

According to the present invention, a system is provided, which includes a mobile receiver device and a computing device.

The computing device can be attached to the mobile receiver, or it can be remote from the mobile receiver but communicate with the mobile receiver.

An existing GSM mobile phone network has been used in accordance with the present invention to implement a mobile receiver positioning system (see Figure 3). The system 300 includes a plurality of transceiver base stations 301, 302, 303, 304 serving as the above-mentioned transmitters; a mobile phone 310 serving as a mobile receiver; a location calculation unit (LCU) 320; and a management database (MDB) )330. The base station provides signal transmission, which is detected by the mobile phone 310 and processed in the LCU 320 using the following mathematical algorithm.

The mobile phone 310 detects signals from several nearby base stations 301 to 304, compiles a list of detected signals, and transmits the list to the server base station 304 via a message through the link 340. 401 in Figure 4 shows the format and content of the message. It is defined by the GSM standard and includes the following fields: broadcast control channel (BCCH), base station identification code (BSIC), mobile network code (MNC), mobile country code (MCC), unit identification item (CI), local The area code (LAC), the number of measurements taken N, and the measurement quality indicator Q. For each base station signal it detects, the message 401 includes some or all of these fields. Sort the "neighbor cell" information according to the power level of the received signal, that is, the data corresponding to the base station instead of the servo base station through which the mobile phone communicates in both directions. The communication links 340 and 341 can be used as part of the telephone system.

Additional data can be supplied from a management database 330 to LCU 320, which Additional information includes the identification (CI, BSIC, etc.) and location of the fixed base station, the corresponding antenna pattern of the fixed base station and its transmitter power. These data can be used as part of the standard network configuration contained in the management system database 330 and are communicated to the LCU via the link 350. The LCU 320 uses the message 401 to create a list 402 of visible base stations. The table shows the identification items of the base station, the quality value associated with the detection, the type of the transmitting antenna pattern, the indication direction of the main lobe of the transmitter antenna, and whether the detected signal comes from the servo base station (S) or The flag of the adjacent base station (N) and the location of the base station.

The LCU 320 also uses the management information to generate a list 403 of base stations known to be active in the system but not present in the list 402 of visible base stations.

The mobile phone 310 performs time synchronization communication with its server base station 304, and also transmits the TA value from the network to the LCU 320 via the link 341 in another message (not shown in FIGS. 3 and 4).

In order to determine the location of the mobile phone 310, the LCU 320 needs to use the visible base station list 402, the invisible base station list 403, and the TA value associated with the servo base station to construct an indication that the mobile phone is at a given horizontal position in the GSM network. A function of probability (given TA value and component qualification for visible and invisible lists). Return the position corresponding to the highest probability as the position of the mobile phone.

The calculations involved in this particular embodiment will now be explained.

To determine the position of the mobile phone, LCU 320 finds the position C that minimizes the following function_TOT=C_TA+C_V+C_N, (1) where C_TOTIs a function whose minimum value indicates the most likely position of the phone, C_TAIs the function associated with the timing advance measurement, C_VIs the function associated with the base station in the visible list, and C_NIt is a function associated with the base station in the invisible list.

Function C_VWith C_NAs the sum of all N base stations, they are similarly evaluated as:<maths><img file="TWI345641B_D0001.tif" /></maths>Where P_nIs the probability that the n-th base station is visible (in C_VUnder the circumstances). There are two better ways to calculate this value, namely the Gaussian version and the power version, as described below.

<b>Gaussian version</b>

In this version, the Gaussian probability distribution is used to simulate the probability of the mobile phone detecting the signal from the transmitter. A two-dimensional Cartesian coordinate system is established with the base station as the center, in which the x-axis is oriented along the forward lobe of the transmitting antenna. The Gaussian distribution is based on the position on the x-axis (x₀,0) is the center. The probability P that the mobile phone at position x of the vector detects the signal from the transmitter at position (0,0) is given by:<maths><img file="TWI345641B_D0002.tif" /></maths>in<img file="TWI345641B_D0003.tif" />, X and y are the coordinates of the receiver, and ()^TRepresents the transpose of the matrix in brackets. That is, the position x of the distribution on the x-axis₀Has its peak at and has a standard deviation σ along the x direction_<i>x</i>, And has a standard deviation σ along the y direction_<i>y</i>, And there is no covariance item.

Match a cost function f(x,y) proportional to the negative of the logarithm of P(x,y) More convenient, that is: f(x,y)=-2log_eP(x,y). (4)

The above definition of f(x,y) uses coordinates related to the position and orientation of a specific base station. When combining the distribution of all the units in the visible list, it is more convenient to use a reference coordinate system that has an arbitrary origin somewhere near the expected position of the mobile phone and where the x-axis is aligned to the east and the y-axis is aligned to the north. , So it is necessary to apply a standard conversion to each term in the sum formula. If the base station k has a reference coordinate and an azimuth measured in radians counterclockwise from the east<img file="TWI345641B_D0004.tif" />Related position (x_k,y_k), then the contribution of the base station to the cost function will be equal to:<b><i>f</i>(<i>x</i></b>_<b>k</b><b>,<i>y</i></b>_<b>k</b><b>)=(<i>x</i>-z</b>_<b>0</b><b>)</b>^{<b><i>T</i></b>}<b>. B. (<i>x</i>-z</b>_<b>0</b><b>)</b>, (5) where,<maths><img file="TWI345641B_D0005.tif" /></maths>Parameter x₀, Σ_<i>x</i>And σ_<i>y</i>It varies from unit to unit and can be derived using heuristics. In one embodiment of the present invention, the results of measurements made at many known locations across the network are used to find the value that provides the lowest overall error when recalculating the location using this method. Once calibrated in this way, this method can be used to locate a mobile phone at an unknown location.

The form is f(x_k,y_kThe addition formula of the function of) makes:<b>(<i>x</i>-z</b>_<b>1</b><b>)</b>^{<b><i>T</i></b>}<b>. B</b>_<b>1</b><b>．(<i>x</i>-z</b>_<b>1</b><b>)+(<i>x</i>-z</b>_<b>2</b><b>)</b>^{<b><i>T</i></b>}<b>. B</b>_<b>2</b><b>．(<i>x</i>-z</b>_<b>2</b><b>)=(<i>x</i>-z</b>_<b>R</b><b>)</b>^{<b><i>T</i></b>}<b>. B</b>_<b>R</b><b>．(<i>x</i>-z</b>_<b>R</b><b>)+<i>e</i></b>, (6) where B_R=B₁+B₂as well as<b>z</b>_<b>R</b><b>=</b><img file="TWI345641B_D0006.tif" /><b>. (B</b>_<b>1</b><b>. z</b>_<b>1</b><b>+B</b>_<b>2</b><b>. z</b>_<b>2</b><b>)</b>. The term e is a constant, so it is an irrelevant term, because we only check the position of the minimum value of the function instead of the absolute value of the function. Interested in value. At the minimum point, the gradient of the function is zero. Because the gradient of the constant is always zero, the term e does not contribute.

Based on this, it is obvious that the total cost function C of all base stations can be simplified into a simple expression of the following form:<b><i>C</i>=(<i>x</i>-z</b>_<b>$</b><b>)</b>^{<b><i>T</i></b>}<b>. B</b>_<b>$</b><b>．(<i>x</i>-z</b>_<b>$</b><b>)</b>(7) where z_$With B_$The value of is derived from the positions and orientations of all base stations in the visible list by reusing Equation 6. Function C is at x=z_$There is a unique minimum at the point, so there is no need to perform a minimization by using an iterative numerical technique in this case.

<b>Power version</b>

In order to calculate P using this method, the average power received by the mobile phone from a given base station is first required. The average value is used here, so there is no need to consider the antenna pattern of the mobile phone receiver, which will point to any direction in actual use. If the mobile phone is separated from the base station by a distance r and is at an azimuth angle θ with the main lobe axis of the transmitter, the average signal power received by the mobile phone is given by the following equation:<b><i>R(r,</i>θ<i>)</i>=<i>G(</i>θ<i>)W</i></b>_<b>0</b><b><i>γr</i></b>^<b>-β</b>, (8) where R(r,θ) is the average received signal power, G(θ) is the transmitter antenna gain, W₀Is the transmitted power, γ is a logarithmic normal random variable (see below), and β is an index of power loss. Assuming that the mobile phone has an omnidirectional antenna pattern, this is not true in practice anyway, but it is a good model for averaging. The logarithmic normal variable represents the variability of the signal power characteristics of the radio path (see below).

Taking the logarithm on both sides of equation (8) can produce: l<b>og</b>_<b>e</b><b>(<i>R</i>(<i>r</i>,θ))=log</b>_<b>e</b><b>(<i>G</i>(θ))+log</b>_<b>e</b><b><i>W</i></b>_<b>0</b><b>+log</b>_<b>e</b><b>γ-βlog</b>_<b>e</b><b><i>r</i></b>。(9)

The logarithmic normal variable can be simulated as:<b><i>γ</i>=<i>γ</i></b>_<b>0</b><i>e</i>^{<b><i>m</i></b>}, (10) where γ₀Is the median value, and m is a variance of σ²The normal distribution of random variables. Substituting equation 10 into equation 9, we get:<b>log</b>_<b>e</b><b>(<i>R</i>(<i>r</i>,θ))=log</b>_<b>e</b><b>(<i>G</i>(θ))+log</b>_<b>e</b><b><i>W</i></b>_<b>0</b><b>+log</b>_<b>e</b><b><i>γ</i></b>_<b>0</b><b>+<i>m</i>-βlog</b>_<b>e</b><b><i>r</i></b>。(11)

As explained in the following paragraphs, this formula indicates the logarithmic-Gaussian distribution variable of the received power level, and its probability density function (PDF) is as follows:<b>PDF(log</b>_<b>e</b><b><i>R</i>)=(2πσ</b>^<b>2</b><b>)</b>^<b>-1/2</b><b>e×p[-(log</b>_<b>e</b><b><i>R</i>-</b><img file="TWI345641B_D0007.tif" /><b>)</b>^<b>2</b><b>/(2σ</b>^<b>2</b><b>)]</b>, (12) where<img file="TWI345641B_D0008.tif" /><b>=log</b>_<b>e</b><b>(<i>G</i>(θ))+log</b>_<b>e</b><b><i>W</i></b>_<b>0</b><b>+log</b>_<b>e</b><b><i>γ</i></b>_<b>0</b><b>-βlog</b>_<b>e</b>^{<b><i>r</i></b>}。

The mobile phone detects a base station, and if the received power level is higher than the minimum detectable signal power R_min, The base station is listed as "visible". Figure 5 outlines this point. The graph shows that the log R of the received power is plotted against the log r of the distance from the base station transmitter. The two are on axes 500 and 501, respectively, and the result is a straight line 502. An example point indicated by mark A is separated from the transmitter indicated by line 503 by a distance, and this point has a predicted received power given by line 504. When the received power is measured multiple times at the same distance from the transmitter, it can be found in practice that the result will change, generally following the probability distribution shown by the curve 505. This is a normal distribution on a logarithmic graph, so the logarithmic normal variable γ can best illustrate this variability, as described above.

Line 506 is the minimum received power R that can be detected by the receiver_min. Therefore, the probability P of detecting a signal at point A is given by the area below the curve 505 and above the critical line 506. Mathematically, this system is given by the following integer: P=log_eR&gt;log_eR_minThe probability of (13) is<maths><img file="TWI345641B_D0009.tif" /></maths>Such integers are well known in relation to the "error function" Erf (see, for example, C++ Numerical Methods by William H. Press et al., Cambridge University Press, Chapter 6, p. 225). Therefore, P is determined by<b><i>P</i>=(1+<i>Erf</i>((</b><img file="TWI345641B_D0010.tif" /><b>-log</b>_<b>e</b><b><i>R</i></b>_<b>min</b><b>)/</b><img file="TWI345641B_D0011.tif" /><b>σ))/2</b>given. (15) Just like the Gaussian version mentioned above, it is mathematically convenient to use a "cost" function C (which is the negative logarithm of P), namely:<b><i>C</i>=-log</b>_<b>e</b><b>(<i>P</i>)</b>。(16)

This function has a minimum value at the position with the greatest probability. In Figure 1, assuming that the antenna pattern G(θ) is a cos²Draw this cost function under the condition of the θ function. The x and y horizontal axes are the position axes related to the position of the base station. The vertical axis is the value of the cost function shown on any scale from 0 (low cost) to 10 (high cost).

The surface drawn in Figure 1 has several different characteristics. The first feature is that there is a low center area 101 around the location of the transmitter base station, in which the cost detected by the receiver is a minimum. This is because the power level here is so high that even if the channel attenuation changes to an unusual degree, the received power level is still above the minimum. The second feature is that along the antenna zero The direction has a high cost 103.

When this analysis is extended to include all signals from all BTSs and combined into a single cost function, the shape is similar to the shape shown in Figure 2. The axis in this diagram is centered on the position 202 of the mobile phone.

In this preferred embodiment, the corresponding cost function C associated with the timing advance measurement is_TAThe system is shown schematically in Figure 6. This figure depicts a view of the base station 601 and a graphical representation of the timing advance distance 602. Ideally, the distance can be calculated by multiplying the measured TA value by the speed of the radio wave. In practice, in a GSM system, the reported TA value measured at a specific radial distance 602 is quantified into units of approximately 550 meters or half a GSM "bit" by a measurement method. This point represents the radial distance 602 (D_TA) Has an equivalent distance error of one plus/minus one quarter bit on average, as shown by the center circle 603. As shown in the figure, the cost functions 604 to 606 applied to the specific embodiment are superimposed on this schematic diagram. In short, any position closer to a quarter bit of the base station than the measured TA distance (D_TA-225m) is assigned a cost value 604 proportional to the difference relative to this position. A zero cost value 605 is assigned to the position within the ambiguity band of the measured distance plus/minus a quarter of a bit. A quarter of a bit beyond the TA distance (D_TA+225m) is assigned a value 606 proportional to the square of the difference relative to this position.

In this preferred embodiment, the combined cost function C can be obtained as follows_TOTThe minimum value. Project the positions of the base stations on a horizontal plane, and identify points that form a grid of equal distances along two horizontal directions on the common horizontal plane of all base stations. This point is schematically shown in Figure 7, where The grid 700 shown in the figure is superimposed on the common horizontal plane of all base stations. An example point 701 is identified on the plane. Evaluate the total cost function C at each of these grid points_TOT, And find the corresponding to C_TOTThe lowest value point. In FIG. 7, this point is shown as an example point 701. Then take this point as the center of another square array of grid points that is one-tenth of the interval 702. Evaluate C again at each point of this sub-grid_TOT, And find the minimum value. Then use this point as evaluation C_TOTThe center of another group of grid points, and so on, until the sub-grid interval is equal to 10 m. Will correspond to C_TOTThe point with the lowest value is used as an estimate of the position of the mobile receiver.

According to the specific specific embodiments described in the "Gauss version" and "Power version" above, the tests have been implemented in several GSM networks. For example, the Gaussian version has been used in one of the experiments. A total of 319 test points randomly scattered on the test area are selected, and at each point, the mobile phone position measured according to the present invention is compared with the real position measured by a more accurate method ("ground truth") . The results show that two-thirds of the locations are within 220 m of ground truth, and nineteen-twentieth of the locations are within 380 m.

<p>101Low Center Area</p><p>103High cost</p><p>202Mobile phone location</p><p>300Mobile receiver positioning system</p><p>301Transceiver Base Station</p><p>302Transceiver Base Station</p><p>303Transceiver Base Station</p><p>304Transceiver base station</p><p>310Mobile</p><p>320Position calculation unit</p><p>330Management database</p><p>340Link</p><p>341Link</p><p>350Link</p><p>401Message</p><p>402List</p><p>403List</p><p>500Axis</p><p>501Axis</p><p>502Straight</p><p>Line 503</p><p>Line 504</p><p>505Curve</p><p>506Critical Line</p><p>601Base station</p><p>602 Timing advance distance</p><p>603Concentric circles</p><p>604Cost function</p><p>605Cost function</p><p>606Cost function</p><p>700Grid</p><p>701Sample point</p><p>702Interval</p>

Figure 1 illustrates an exemplary graph of the position function ("cost") associated with a single transmitter.

Figure 2 illustrates an exemplary graph of a position function ("cost") associated with multiple transmitters.

Figure 3 illustrates an exemplary multiple transmitter system.

Figure 4 illustrates an exemplary clearing of the detected signal transmitted by the mobile phone List, and an example of a list of visible and invisible signals maintained at the position calculation unit.

Figure 5 illustrates how the received power varies with the distance from the transmitter.

Figure 6 illustrates a graph of an exemplary fixed transmitter and position function components associated with timing advance.

Figure 7 illustrates a calculation grid.