An improved algorithm of Ground Moving Target Tracking based on Obstacle Information

Abstract

Standard Interactive Multimodal Model Algorithms (IMM) have problems such as poor model probability matching, slow model switching and poor tracking accuracy due to the use of fixed transition probability matrices. This paper presents an IMM algorithm to update the transition probability matrix adaptively. First, the prior information of the ground obstacles is added to the model probability update phase to improve the matching degree of the model probability. Secondly, the Markov transfer probability is corrected in real time by using the model likelihood function value to enhance the matching model and weaken the influence of the mismatch model. The simulation results show that the tracking accuracy of this method is significantly better than that of traditional IMM algorithm.