Study On Locating Algorithms Based On Time-related Measurements In Digital Cellular Systems

With the development of communication technology, the wireless locationtechnique in wireless cellular network has obtained more and more attention.Through the location system, people can get information about the position of aMS operating in a geographical are and process that information to form a locationestimate. There are several potential applications of wireless location system, suchas public safety, location-sensitive billing, cellular system design and resourcemanagement and intelligent transportation system.In July 1996 US Federal Communication Commission(FCC) issued the E-911mandate stating that all carriers of mobile communication must provide thegeolocation of an emergency 911 caller to the appropriate public safety answeringpoint. In Oct. 1999 the FCC issued the E-911 Third report and Order, which adoptthe revised standards for Phase 11 location accuracy –For network-based solutions:100 meters for 67% of callers, 300 meters for 95% of caller。But in cellularenvironments, because of the influence of several adverse factors, such asmultipath, non-line-of-sight propagation, noise, interference and channel frequencycharacteristics, the measurements attained from methods at present inevitably haveerrors, the performance of location algorithm may be significantly degraded andthe position estimation of MS is biased. So it becomes the focus on how to utilizeexisting network resource to achieve locate mobile station by low cost.By analyzing the previous research in wireless location extensively, aimed atthe core question of improve the accuracy of MS location, many methods arediscussed in detail. The TDOA technique is one of the most promising positionlocation techniques for cellular-type wireless communication system, which canprovide accurate position location information using the existing cellularinfrastructure without requiring additional hardware and software implementationwithin the mobile unit. So this thesis aim at the research of wireless locationalgorithm based on time-related measurements in Wire Cellular Network.First of all, the mathematical model of TDOA hyperbolic equations isestablished, Taylor-Series and Chan algorithm are importantly introduced, andsome other algorithm have been simply discussed. Meanwhile, locationperformance of several methods is evaluated based on the simulation in differentchannel environments. It is shown by experiments that Taylor-Series and Chanalgorithm are easy to realize and can achieve good location accuracy, but they fit invarious simulation environments. In order to meet the so-called E-911 demand,these algorithm must be modified.Secondly, because of the limitations of single location technique, people havebegan to study hybrid location algorithm. If the TDOA and AOA measurements areaccurate, we need only the home BS and another BS to location an MS. But inreality, due to propagation channel impairments, the TDOA and AOAmeasurements contain errors. As a result, more than two BSs are usually requiredfor a location estimate of reasonable accuracy. Therefore, this paper gives ageometry location method based on more than two BSs. Then, based oncharacteristics of AOA,Chan algorithm and Taylor-Series algorithm, a new hybridTDOA/AOA location algorithm is proposed, and the corresponding CRLBtheoretical bounds are derived. This method combines the advantages of both Chanalgorithm and Taylor-Series algorithm, promising that the accuracy is no worsethan that of Chan algorithm. Compared to Taylor-Series algorithm, if the AOAmeasurement is accurate, this proposed method can attain much higher locationsucceeding rate and much fewer iterative times. The simulation results show that,this hybrid location algorithm with angle measurements can make full use of theinformation provided by cellular network. Especially when the AOA measurementsobtained have relatively high accuracy, the location performance can be improvedsignificantly.Because of the influence of multipath propagation, it has obvious bias whenusing the General Crossing Correlation methods to evaluate the time delayparameters. While by the character that higher order cumulants of Gaussian noisesare zeros, we can use it to restrain the noises. And then a new TDOA parameterestimation algorithm is presented. The key of this method is to translate the TDOAparameter estimation in time domain to the frequency retrieval problem aboutcomplex harmonic in frequency domain. First the output of the channel is passedthrough the matched filter and then deconvoluted in the case of location signal.Using the higher order cumulants and some other correlated processing methods,the noises can be repressed in the process of the TDOA parameter estimation. Theresults of simulation show that this method is feasible. Compared with other timedelay methods, this method has many advantages:1. It provide some flexibility that translate the problem in time domain to infrequency domain.2. It has good noise-restrain ability. Because of the character of the higherorder cumulants, we can attain high accuracy although in low SNR about coloredGaussian noises.3. The computational costs are small. Some other methods such as MUSICmethod need one-dimension spectral peak research while the ESPRIT methoddoesn't need any spectral peak research.