Research On The Techniques Of Target Localization Based On Base Stations

With the development of the modern technology, more and more measurement systems apply network technology. In the networked localization system, it is very important to capture the target position information. This paper mainly develops the research on the techniques of target localization based on the base stations, which break stations layout,time difference measurement,location algorithm and track prediction technique, realizes precise localization and track prediction of the target, and solves the questions that include distributing station difficultly,time difference measurement accuracy strongly affected by noise,low localization precision and hard track prediction, in order to improve test efficiency, decrease the experimental times and save the research funding.According to the time difference localization theory, this paper designs the target localization system including radio beacon, base stations, long-distance data forwarding and center processing platform. It analyses the system error sources from signal propagation, station measurement, localization parameter estimation, the localization algorithm and other aspects. Synchronously, it indicates that base station layout,time difference measurement and localization algorithm are the important factors which affect localization precision, and explains the contents which the paper mainly researches on.Firstly, aiming at the problem that embattling difficultly in the localization of target, a inclined parallelogram base station layout scheme is put forward based on star, diamond, inverted triangle and other embattling mode methods. In addition, it conducts the preliminary research on base stations layout using adaptive genetic algorithm. Experimental results show that the new scheme has superior performance in the two localization precision technology index of geometry precision factor and Cramer-Rao Bound. Secondly, aiming at the problem that the time measurement strongly affected by noise in the localization of target, and on the basis of the time difference measurement model, it focuses on six species of generalized cross-correlation time difference measurement methods including cross-correlation method,roth pulse shock response method, smooth related transform method, phase transformation method, HB weighted method and the maximum likelihood method and so on, and variable step length adaptive filter time difference measurement method based on the sigmoid function, probability curve, tongue-like curve, hyperbolic tangent function. The simulation results show that: in the conditions without noise, gaussian noise and colored noise, it can finish time difference measurement accurately by adopting the six adaptive filter algorithms. However, it can finish time difference measurement accurately only using the directly related law in the conditions with gaussian noise, and only using other generalized correlative methods in the condition with colored noise.Again, in allusion to the problem that it has low localization accuracy by the algorithm in the target localizaiton, a new target localization method is proposed on the basis of particle swarm optimization and its improved algorithm according to the analysis of traditional localization algorithms such as Chan, Taylor and Newton. In allusion to the shortcoming of Taylor and Newton localization algorithm sensitivity for the initial value, it puts forward mixed localization algorithm successively such as Chan - Taylor, Chan-Newton and so on. The experimental results show that: it has faster positioning speed by Chan and its hybrid algorithm with Taylor and Newton, and it has higher positioning accuracy by the grouping particle swarm target localization algorithm on the basis of diversity feedback.Finally, in allusion to the problem that it is difficult to predict the track in the target localization, the extended Kalman filtering algorithm is realized via the Taylor series development and linearization. Meanwhile, on the basis of building target flight trajectory model, it completed the target flight trajectory repair and drop point prediction based on the extended Kalman filtering. Experimental results show that the precision of track repair and drop point prediction is less than 100m.