Kalman Filter-based Multi-station Positioning Technology Research

Passive locating by single station technology is not radiate electromagnetic wave when measuring the target's position; it has broad application in electronic reconnaissance. But passive locating by single station technology can't measure the target's position accurately when the target is far from the locating station; so it can't work effectively in long distance target location system. Multi-station tracking and positioning method makes use of more location measure information, so it can be more accurately determine the target's movement information. By using Kalman filter algorithm , the tracking and positioning accuracy can be more precisely. The dissertation is accomplished in this background.This paper explains the basic theories of radar measurement; introduces the system measure model of Multi-station tracking and positioning method; discusses the solutions for nonlinear equations and find out a better solution for nonlinear location equations; for more precise measurement of the moving target, the locating station distributing mode is optimized by computer simulation.As the basic applied filter theory of tracking and positioning system, this paper also presents the theory of adaptive filtering and Kalman filter.In this paper, the emphasis is on the Kalman Filter arithmetic in order to apply it to the tacking system. This paper chose the UKF(Unscented Kalman Filter) arithmetic as a main research direction after comparing the strongpoints and the demerits about EKF(Extended Kalman Filter) ,UKF(Unscented Kalman Filter)and MAUKF(Memory Attenuated Unscented Kalman Filter).For the UKF algorithm is developed for the nonlinear system application, and it has better accuracy than other filters such as EKF, so the UKF algorithm is the filter method for enhancing the positioning precision of this paper. In the compute of UKF,for simplifying the computational complexity of the tracing system, identity matrix is introduced into the Unscented Transformation, and the complicated matrix calculation is replaced by simple numerical calculation for matrix square root solving, Develop a new algorithm named FMSRUKF (Fixed Matrix Square Root Unscented Kalman Filter). Gives a detailed description on the improve process and algorithm characteristic of FMSRUKF; and analysis the arithmetic merit of FMSRUKF. The simulation results of variably accelerated motion target tracking and positioning of long distance under three dimensional coordinate show that FMSRUKF achieves better precision and robust in the nonlinear system with the white Gaussian noise.