Dissertation > Excellent graduate degree dissertation topics show

Algorithm Research and Application of Target Tracking

Author: XuZhiGang
Tutor: WuXiaoJun
School: Jiangnan University
Course: Computer Software and Theory
Keywords: Target tracking Multi-platform tracking Interacting Multiple Model Algorithm Distributed Fusion
CLC: TN953
Type: Master's thesis
Year: 2011
Downloads: 336
Quote: 2
Read: Download Dissertation

Abstract

With the development of computer technology, sensor technology, target tracking technology has become a highly important research topic. A core part of the target tracking technology is filtering algorithm, a good algorithm to target high-precision real-time tracking. With the great progress of the development of science and technology in recent years, hardware, algorithms and computer performance, target tracking technology has been rapid development. Enormous challenges continue to target tracking. Therefore, in the building of national defense, aerospace, and other various fields, new technologies continue to be applied to the target tracking to adapt to a more complex environment and application. This research work is carried out based on a distributed target tracking. First, the article describes the development of target tracking technology profiles and target tracking algorithms. Second, in a multi-platform multi-target tracking, based on the basic principles of the interacting multiple model algorithm (IMM) algorithm to improve the ideological and, combined with joint probabilistic data association (JPDA) algorithm to achieve a single platform for multi-target tracking algorithm, and then of further development into the distributed multi target tracking algorithm (DIMM_JPDA). Tracking Maneuvering Multiple Targets the current statistical model algorithm is fused to the IMM algorithm, the algorithm of maneuvering targets good results. The major modified work as follows: 1 in maneuvering target. Distributed interacting multiple model algorithm based on the Lang Hong (DIMM_CA / CV), the algorithm used in the single-target environment, do not apply in the multi-target environment. Constant acceleration model (CA) with uniform velocity model (CV) and the algorithm phase interaction, a larger error in the motor case. Of current statistical model (CS) algorithm proposed new algorithm (DIMM_CS, / CV). The new algorithm in tracking maneuvering target, especially strong maneuvering targets good results, simulation results are compared with DIMM_CA / CV algorithm, analyze its performance advantages. 2, in the multi-objective. Proposed a new algorithm of multi-target tracking model, called DIMM_JPDA algorithm. JPDA algorithm based on multi-target tracking fusion applied to the the IMM algorithm IMM_JPDA algorithm. Then use the data fusion method, will be expanded into a distributed multi-target tracking algorithm (DIMM_JPDA). Since the introduction of the JPDA algorithm model, makes amount of DIMM_JPDA computation not suitable to be applied in practice. 3, data fusion, covariance matrix P must be reversible. But the actual program is running, there will be irreversible. The author introduces eigenvalue perturbation method to solve the problem. The experiment showed that after a perturbation method, the matrix becomes nonsingular matrix irreversible. 4, DIMM_CS / CV algorithm is a good solution to the target mobile, but do not apply for multi-objective. The DIMM_JPDA algorithm to solve the multi-target tracking, but are too computationally intensive. Therefore, the authors combined the DIMM_CS / CV the algorithm with DIMM_JPDA algorithm to solve the target application in the case of motorized and multi-target.

Related Dissertations

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Radar > Radar tracking system
© 2012 www.DissertationTopic.Net Mobile