System and method for analog cellular radio geolocation

[00002] The present application is related to co-pending U.S. provisional patent application Ser. No. 60/254,177 entitled “A METHOD FOR ANALOG CELLULAR RADIO GEOLOCATION” filed Dec. 11, 2000, which is hereby incorporated herein by reference.

[00003] A system and method for determining the geolocation of autonomous mobile appliances emitting analog waveforms is disclosed. More specifically, the inventive system and method is used to geolocate FM analog signals such as those used in the AMPS cellular radio air standard by using a time difference of arrival (“TDOA”) approach.

[00004] There is currently much focus on technology for performing geolocation of cellular phones and other mobile appliances. This has been largely motivated by an FCC mandate requiring the location of wireless users to be provided to the Public Service Answering Point (“PSAP”) when making an emergency 911 (“E911”) call or other 911-related transmission. There are currently many systems and patents that deal with the source location of wireless radio frequency emitters. To a great extent, these patents simply suggest the idea of location of a transmitter source by some means, which is often impractical to implement for an application such as an E911 event. Very often, factors such as cost, complexity, and required computation horsepower are not fully considered.

[00005] Prior art systems usually deal with the source location of either analog or digital signals. The analog signals referred to herein are those signals which are not encoded for the purpose of transmitting digital information, but rather analog waveforms through means such as frequency modulation (“FM”). Digital signals, on the other hand, are those that employ some form of information bit to pulse shaped symbol encoding for the purpose of transmitting digital information, theoretically allowing loss-less transmission of data. The pulse-shaping and modulation schemes used in digital waveforms can be exploited to reduce the cost and resources required for a location system, in a manner that is not available in analog location systems. The present inventive system and method focuses on a location method used specifically for FM analog signals such as those used in the AMPS cellular radio air standard. It is to be understood that the present invention is not limited to a particular type of emitter of FM analog signals, such as AMPS or the AMPS air standard. Rather, the present invention is applicable to any system emitting FM analog signals. The motivation for the present invention has been driven by the requirements to design a geolocation system that reduces system cost and latency while simultaneously maximizing geolocation accuracy.

[00006] Radio location systems and methods can be grouped into two major classes. The first class of radio location systems and methods uses the arrival angle of a radio frequency (“RF”) signal at an antenna array to determine a line of bearing from the array to the emitter of the RF signal. For example, a mobile radio transmits a signal which is received by multiple base stations in separate geographic locations. Each base station has an antenna array which measures the radio wave phase difference at different antenna elements in the array. An angle of arrival of the mobile radio's signal is calculated and a line of bearing to the mobile radio is determined. By obtaining multiple lines of bearing from multiple antenna arrays, the intersection of the lines of bearing provides the geolocation of the mobile radio.

[00007] The second class of radio location systems and methods uses the time differences of arrival from geographically separated sensors in order to estimate the emitter location through triangulation techniques. For example, a mobile radio transmits a signal which is received by multiple base stations in separate geographic locations. Each base station conducts time difference of arrival (“TDOA”) measurements of the received signal and the TDOA measurements are used to determine the location the mobile radio using conventional positioning algorithms. A global positioning system (“GPS”) or other time standard is typically used to provide a common time reference among the base stations and the mobile radio. Typically, a TDOA-based analog signal location system comprises multiple geographically separated sensors, referred to herein as Wireless Location Sensors (“WLS”) that are connected and controlled through some communication means such as telephone lines of high speed data communication lines such as trunked DS0 or ISDN. All of the WLSs are controlled by a central processing facility, referred to herein as a Geolocation Control Server (“GCS”) that tasks the sensors to simultaneously capture signals transmitted by a particular mobile radio, also referred to herein as a “mobile station”. The captured signals are then sent to the GCS via telephone lines or high speed data communication lines, along with the precise time measurement of when the signals were captured. The location of the mobile station is then calculated through cross-correlations of the signals received at the GCS from the WLSs—the peaks of the captured signals reveal an estimate of the time differences of arrival of the mobile station's signals at the various WLSs. The time differences of arrival are geometrically interpreted as branches of hyperbolic surfaces that intersect at the location of the mobile station.

[00008] Each of the WLSs includes a precise time source, such as those derived from a GPS disciplined oscillator or other time standard, a radio frequency receiver, digitizing circuitry, a digital processor, and other standard circuitry such as an analog to digital (“A/D”) converter, all of which operate to capture, store, and manipulate the received signals. However, in order to obtain an acceptable level of accuracy when performing a geolocation evolution, it is highly desirable to receive and process as much data as possible from the received signal. The more data that is used to determine the geolocation of the mobile station, the greater the accuracy of the geolocation estimate. One processing limitation that must be taken into account is the sampling rate. Due to the physical limitations of Nyquist sampled real signals, the sampling rate must be set to at least twice the rate of the highest frequency component of the received signal. Additionally, the level of quantization used at the analog to digital converter stage must be sufficient to capture a broad range of signal levels without significant distortion. All this data must be sent to the GCS, which requires that the data be sent over telephone lines of high speed data communication lines. Since the WLSs are located in geographically different locations, the data used for the geolocation calculation must be sent over telephone or high speed data communication lines from at least one WLS.

[00009] An example of the amount of data that must be sent from a WLS to the GCS, and the amount of time to send that data, consider a one second AMPS analog waveform received at a WLS and used to locate the mobile station to be geolocated. For the AMPS analog waveform, the baseband double-sided bandwidth is 30 KHz. Further consider that 16-bit A/D converters are used on each of two receive channels to provide approximately 96 dB of dynamic range and that the sampling rate is 40,000 complex samples/sec. Additionally, the data link between the WLS and the GCS is a DS0 high speed digital data communication line with a data transport rate of 64,000 bits per second (“bps”). Note that 40 KHz is a practical over-sampling rate which will allow sufficient excess bandwidth for filtering of adjacent signals and anti-aliasing, as is known in the art. In order to capture one full second of signal data for sufficient cross-correlating of the signals received at the GCS from the WLSs, the required amount of data in bits to be transmitted from a WLS to the GCS can be calculated as follows: $\left(40,000\frac{\mathrm{samples}}{\sec }\right)\times \left(1\sec \right)\times \left(2\mathrm{channels}\right)\times \left(16\frac{\mathrm{bits}}{\mathrm{sample}}\right)\times \left(2\mathrm{samples}\right)=2.56\mathrm{Mbits}$

[00010] The 2.56 Mbits of data is the total for two signals to be cross-correlated. Using a 64 kbps data rate for a DS0 high speed digital data communication line, the time to transfer the 2.56 Mbits of data can be calculated as follows: $\left(2.56\mathrm{Mbits}\right)÷\left(64\frac{\mathrm{Kbits}}{\sec }\right)=40\sec$

[00011] As shown above, the time it takes to transfer a sufficient amount of data from an AMPS signal from a WLS to the GCS in order to accurately geolocate the mobile station is 40 seconds. This is clearly an unacceptable amount of time for applications where a high throughput of location estimates is required. Certain prior art systems attempt to overcome the data transfer problem by either limiting the amount of data sampled or by using excessively lossy data compression schemes. Thus, there is a need for a geolocation system and method for accurately geolocating a mobile station in a practical and efficient manner.

[00012] The present inventive system and method increases the speed of location estimates without sacrificing accuracy of the geolocation estimate. The inventive system and method does not use smaller sample sizes or excessively lossy data compression schemes. The inventive system comprises multiple WLSs that are typically co-located with the base stations of the mobile station's communication network, and a centrally-located GCS. The WLSs operate in one of two modes, a primary mode and a secondary mode, the operation of each will be described in detail below. Generally, once the system receives a request to locate a wireless user, or mobile station, each WLS operates in primary mode to initiate the geolocation evolution and send information regarding the signal received from the mobile station to the GCS. Upon receipt of the signals from the multiple WLSs, the GCS selects one WLS to be the primary WLS, the significance of which will be discussed later. The primary WLS continues to operate as before as well as operating in the secondary mode. The remaining WLSs switch to and operate in the secondary mode. The details of the operation of the system is disclosed below.

[00013] Accordingly, it is an object of the present invention to provide a novel system and method of geolocating a mobile station transmitting FM analog signals such as those used in the AMPS cellular radio air standard from a plurality of WLSs located in geographically spaced-apart locations.

[00014] It is another object of the present invention to provide a novel system and method for geolocating a mobile station transmitting an AMPS analog signal by reducing the amount of data to be transmitted across data communication lines.

[00015] It is yet another object of the present invention to provide a novel system and method for geolocating a mobile station transmitting and AMPS analog signal by combining multiple signals received at a WLS to a single channel, demodulating the single channel, and compressing the single channel by use of a Fourier transform circuit.

[00016] It is still another object of the present invention to provide a novel system and method for efficiently and accurately geolocating a mobile station by parallel processing the received signal data at the WLSs rather than at the GCS.

[00017] These and many other objects and advantages of the present invention will be readily apparent to one skilled in the art to which the invention pertains from a perusal of the claims, the appended drawings, and the following detailed description of the preferred embodiments.

[00020] A preferred embodiment of a novel system and method of geolocating a mobile station transmitting FM analog signals such as those used in the AMPS cellular radio air standard from a plurality of WLSs located in geographically spaced-apart locations is described.

[00021] FIG. 1 is a notional schematic diagram of a geolocation system according to an embodiment of the present invention. The geolocation system 100 comprises a number of WLSs, 122A, 122B, and 122C, typically located at the base stations 120A, 120B, and 120C, respectively, which include the antenna arrays 121A, 121B, and 121C, respectively. The antenna arrays may be either a single antenna or an array of antenna elements. While the present invention is not limited to co-locating the WLSs with the base stations, it is convenient to do so since the base stations and their antenna arrays are typically in-place and the WLSs can then simply be added to the existing structures. It is to be understood that while only thee WLSs are shown in FIG. 1, the invention is not limited to any specific number of WLSs. Obviously, at least three WLSs are desired in order to accurately geolocate the mobile station 110.

[00022] Each of the WLSs is connected to the Geolocation Control Server (GCS) 130 via communication lines 125A, 125B, and 125C, respectively. The communication lines may be telephone lines or, preferably, high speed data communication lines such as DS0 lines. The present invention is not limited to any particular type of communication line. The GCS 130 is connected to a geolocation information user 140 which may be, for example, a Public Service Answering Point (“PSAP”) or any other system that would use the geolocation information produced by the GCS.

[00023] Operationally, the antennas 121A, 121B, and 121C receive FM analog signals, such as those used in the AMPS cellular radio air standard, from the mobile station 110. The received signals are sent to the respective WLSs 122A, 122B, and 122C, at each base station for processing. Once the signals are processed, the details of which will be described below, the signals are sent to the GCS 130 via the communication lines 125A, 125B, and 125C, respectively. The GCS processes the signals received from the WLSs and determines the geolocation of the mobile station 110. The GCS sends the geolocation information to one or more geolocation information users 140.

[00024] With reference now to FIG. 2, the detailed processing of signals at the WLSs and the GCS will be described. The block 200 represents the tasks that are performed at the GCS. The block 220 represents the tasks that are performed at all the WLSs for determining which WLS will be designated as the primary WLS. The block 240 represents those tasks performed by the primary WLS after it has been so designated. The block 260 represents those tasks performed by the secondary WLSs.

[00025] Upon receipt of an instruction to geolocate a mobile station, at block 201 the GCS tasks all the WLSs to begin the geolocation event in the primary mode. The instruction to geolocate a mobile station can be any typical instruction known to those of skill in the art such as receipt of a 911 call, receipt of another type of emergency call, a manual request, etc. The type and method of receipt of the instruction is not material to the present invention. All of the WLSs receive the initial primary mode tasking request at block 221. The primary mode processing encompasses the steps represented by blocks 222 through 225. At block 222, each of the WLSs that can receive the signal from the mobile station 110 substantially simultaneously capture the signals transmitted by the mobile station as received by the respective antenna array associated with the WLSs as shown in FIG. 1.

[00026] Each WLS that received the signal from the mobile station processes the captured signal at block 223 which includes digitizing the signals captured from each antenna on the respective antenna array, time stamping the captured signals, combining the captured signals to form one signal, and then location stamping the combined signal. If the mobile station's signal is only captured from one antenna or one antenna element, the step of combining captured signals is not performed. If the mobile station's signal is captured from more than one antenna element at a base station, the WLS associated with that base station combines the captured signals for two reasons. First, to reduce signal facing incurred from multipath and interference, the presence of either tends to decrease the accuracy of the geolocation estimate. Second, to reduce the amount of data to be transferred across the communication lines between the WLSs and the GCS. The algorithms used to achieve the combination of signals are algorithms that are well known in the art such as Equal Gain Combining (“EGC”), Maximal Ratio Combining (“MRC”), Projection Approximation Subspace Tracking (“PAST”), Constant Modulus Algorithm (“CMA”), etc. Any of these algorithms are applicable for reducing the effects of fading and/or interference typically encountered in wireless communication channels.

[00027] After the captured signals are combined and location stamped, the resultant signal, referred to herein as the “combined signal”, is demodulated at block 224. If the mobile station is transmitting FM analog signals, then the demodulation is an FM demodulation. The resultant demodulated signal is then transformed at block 225 to compress the data without loss of geolocation accuracy and the transformed signal is sent to the GCS for comparison with the transformed signals from all the WLSs which received the signal from the mobile station. The details associated with the transformation of the demodulated signal at block 225 are revealed next.

[00028] In the case of an AMPS analog signal transmitted from the mobile station, the demodulated signal bandwidth ranges from approximately 300 Hz to 3500 Hz. Additionally, there is a small amount of bandwidth occupied by the Supervisory Audio Tone (“SAT”) associated with the AMPS air standard. The SAT is located at one of three possible frequencies: 5970 Hz, 6000 Hz, or 6030 Hz. In order to exploit a reduction in bandwidth for the purpose of reducing the storage requirements and the data transfer requirements, a transformation circuit, preferably a Fast Fourier Transform (“FFT”) circuit, is used at block 225 to transform the demodulated signal from the time domain to the frequency domain. The transformed signal, in the frequency domain, now spans a number of frequency bins. Assuming that the number of frequency bins is 1024 the transformed signal resides in 1024 frequency bins. It is to be understood that the present invention can operate with any number of frequency bins and is not limited to 1024 frequency bins.

[00029] With 1024 frequency bins and a practical over-sampling rate of 40,000 complex samples/sec (from the example above), each of the 1024 frequency bins has a span of: $\left(40,000\frac{\mathrm{samples}}{\sec }\right)÷\left(1024\mathrm{bins}\right)=39.0625\mathrm{Hz}$

[00030] Of all the 1024 bins that contain information, only those bins that contain information in the 30-3500 Hz range for the demodulated signal as well as the bin for the SAT frequency are of interest for geolocation purposes. All the other bins represent noise and are of no use for geolocation. Consequently, 83 of the 1024 bins of information are kept and the rest are discarded. Of the 83 bins, 82 of the bins are for the demodulated signal (3200 Hz range divide by 39.0625 Hz/bin) and one bin is for the SAT.

[00031] An important point to note is that the output of the demodulator is a real signal which makes the spectrum exhibit conjugate symmetry. Therefore, we can preserve only half of the output of the FFT with no loss in signal representation.

[00032] We can now calculate the amount of transformed data to be transferred from a WLS to the GCS as follows: $\left(40\mathrm{Hz}\right)\times \left(1\sec \right)\times \left(16\frac{\mathrm{bits}}{\mathrm{bin}}\right)\times \left(2\mathrm{samples}\right)\times \left(83\mathrm{bins}\right)=106,240\mathrm{bits}$

[00033] Note that we used 40 Hz blocks of data rather than 39.0625 Hz so as not to lose any information. When compared to the amount of data to be transferred from the previous example (the raw samples taken from the A/D converter), the amount of data to be transmitted using the inventive method is only 4.15% of the data in the previous example:
106,240 bits÷2,560,000 bits×100=4.15%

[00035] The time to transfer the 106,240 bits of data, herein referred to as WLS reference data, over a DS0 high speed digital data communication line operating at 64 kbps (as in the previous example) can be calculated as follows: $106,240\mathrm{bits}÷64\frac{\mathrm{Kbits}}{\sec }=1.66\sec$

[00036] The 1.66 seconds required to transfer the WLS reference data is a vast improvement over the 40 seconds required to transfer the raw samples out of the A/D converter in the previous example.

[00037] After transforming the data as explained above, each WLS transfers the WLS reference data to the GCS. The GCS, at block 202, receives the WLS reference data from each of the WLSs that received a signal from the mobile station being geolocated and compares the signals to determine which WLS received the strongest signal from the mobile station. The GCS can determine which station received the strongest signal due to the location stamp added by each WLS. The WLS then sends a control signal at block 203 to the WLS with the strongest mobile station signal and designates that WLS the primary WLS, the effect of which is described below. The determination of which WLS received the strongest mobile station signal is performed by methods well known in the art and the specific method used is immaterial to the present invention.

[00038] It is important to note that some of the WLSs may receive the mobile station signal without sufficient quality to demodulate the signal. Typically, poor signal quality results from propagation loss, blockage or fading. As is known in the art, approximately a 12 dB signal to noise ratio is required for proper demodulation. To overcome this problem, the present inventive system and method incorporates the concept of primary and secondary WLSs as will be described next.

[00039] Once the GCS determines the WLS with the highest signal quality, that WLS is designated as the primary WLS. The GCS commands the primary WLS to continue processing blocks of captured mobile station signals as before and to buffer a copy of the captured signal prior to demodulation. Substantially simultaneously, the WLSs not designated as primary, referred to herein as secondary WLSs, continue to capture and process signals from the mobile station and buffer the processed signals prior to demodulation and transformation. When the primary WLS sends its WLS reference data, herein designated the primary reference data, to the GCS, the GCS routes the primary reference data to all the WLSs participating in the geolocation event including the primary WLS. Each WLS then inverse transforms and modulates the primary reference data to restore it to a replica of the mobile station signal as seen by the primary WLS. This replica signal is used by each WLS to cross-correlate with the processed signal received at that WLS to extract the time of arrival for the mobile station signal at that WLS. The cross-correlation helps improve the quality of the mobile station signal as received at the secondary WLS performing the cross-correlation. After processing all the blocks of data in the primary reference data, each WLS participating in the geolocation event, including the primary WLS, sends a signal representative of the time of arrival of each block of data back to the GCS. The GCS receives the times of arrival from the WLSs and determines the geolocation of the mobile station.

[00040] With renewed reference to FIG. 2, the above procedure will be explained with respect to FIG. 2. When the GCS determines the primary WLS in block 202, the GCS sends a control signal to the to-be-designated primary WLS at block 203. The to-be-designated WLS receives the control signal from the GCS at block 241 and commences to operate as the primary WLS. Initially, the primary WLS processes captured mobile station signals as before in blocks 222 through 225, which correspond to blocks 242 through 245. Additionally, the primary WLS buffers, at block 246, a copy of the processed captured mobile station signal prior to demodulation of that signal. The buffered signal will be used later as discussed further below. While the primary WLS is capturing and processing mobile station signals, the other WLSs participating in the geolocation event, now referred to herein as secondary WLSs, substantially simultaneously receive the mobile station signals at block 262 and capture and process at block 263 those signals. At block 266, the captured, processed signals are buffered in a similar manner as the primary WLS at block 246. Note that in FIG. 2 a dashed line connects the block 203 in the GCS to the block 262 in the secondary WLSs. The secondary WLSs do not receive a signal from the GCS designating the WLSs as secondary WLSs. The lack of receipt of a signal designating the WLS as the primary WLS initiates the WLSs to operate in secondary WLS mode.

[00041] Once the primary WLS transforms the mobile station signal at block 245, the primary WLS sends the primary reference data to the GCS. The GCS receives the primary reference data at block 204 and routes the primary reference data to all the WLSs, including the primary WLS, at block 205. The primary and secondary WLSs receive the primary reference data from the GCS at block 247 and block 267, respectively. The primary and secondary WLSs perform and inverse transformation, preferably an inverse FFT at block 248 and block 268, respectively. The primary and secondary WLSs then modulate the inverse transformed signal at block 249 and block 269, respectively. The output of the modulators is a replica of the mobile station signal as received by the primary WLS, which was initially designated the primary WLS based on a determination at the GCS as being the WLS with the highest quality of the received mobile signal. The replica signal is then used at each WLS to cross-correlate at block 250 and block 270, for the primary and secondary WLS respectively, with the signal received at that WLS to improve the quality of the mobile station signal received at the WLS. The cross-correlation method used is any standard cross-correlation algorithm known in the art. The output of the cross-correlation blocks 250 and 270 are signals representative of the times of arrival of each of the blocks of data in the primary reference signal. The times of arrival signals are then sent to the GCS, which receives the times of arrival signals at block 206. The GCS then determines, at block 207, the geolocation of the mobile station by standard time of arrival techniques well known in the art. The GCS may then send a signal to one or more geolocation information users as shown in FIG. 1.

[00042] While preferred embodiments of the present invention have been described, it is to be understood that the embodiments described are illustrative only and the scope of the invention is to be defined solely by the appended claims when accorded a full range of equivalence, many variations and modifications naturally occurring to those of skill in the art from a perusal hereof.