The Study Of Ultra Wideband Wireless Positioning Technology

There has been a great deal of interest in ultra-wideband(UWB) communications, especially after the UWB technology is approved of being employed in civil productions by the Federal Communications Commission(FCC). With the integration of mobile internet and the service of local life, people have an increasingly strong needs for the service based on the location. The UWB technique which being a novelty communications mechanism is becoming a potential one for wireless positioning with many advantages, such as anti-multipath, powerful penetration and high precision.This paper studied the techniques of UWB positioning systems which includes ranging and positioning. The study focued on time difference of arrival(TDOA) location model to locate the target node because of the characteristics of perfect immunity to time in UWB communication. There is a non-line-of-sight(NLOS) error to according to simulating the channel model of IEEE 802.15.4a during UWB ranging, and quantitative detection has little noise, so this paper applied quantization level detection to TOA estimation and combined with the smooth processing of linear Kalman filtering equation, the result showed that three level quantization detection had better TOA estimation performance. The research paid attention to the theory and error of Chan and Taylor localization algorithm, and aiming at the performance of Taylor’s method depends highly on the initial estimation, this paper proposed a new algorithm which combines the Taylor series expansion method with updated initial position. The problem of positioning was transformed into solving the unconstrainted nonlinear optimization problem, and according to slow convergence in steepest decent method and large computing in Newton algorithm, this paper presented a novel DFP algorithm based on Armijo step size to correct the initial position. Besides, the research analyzed the global convergence of novel DFP algorithm.Experimental results show that the proposed algorithm not only decreases the demand of localization optimize algorithm to initial position, but also achieves higher localization accuracy and more fast convergence speed than the steepest decent method. Finally, simulation results showe that hybrid algorithm which includes the novel DFP algorithm and Taylor algorithm has better performance on localization accuracy in cellular network.