Multi-target track system performance evaluation simulation method

Abstract

The invention discloses a multi-target track system performance evaluation simulation method. According to the method, target tracks are loaded by adopting a plurality of different paths, so that the diversity of the target track paths can be realized; through user-defined target track, the demands on the diversified track type configuration by a user can be met to the greatest degree; a radar sensor data source and an infrared sensor data source are adopted, so that the performance evaluation of different multi-target track systems can be realized, a server / client mode is adopted, an estimated algorithm can run on the client, the confidentiality of the algorithm can be realized, the mode of single-time and multiple-time Monte-Carlo method combination is adopted, so that not only can the statistical property of the performance of the estimated track algorithm be obtained, but also the estimated algorithm can be debugged, and discrete points of the algorithm can be found.

Description

technical field [0001] The invention relates to a performance evaluation method, more specifically, a method for evaluating algorithm performance indicators in a multi-target tracking system. Background technique [0002] In the multi-target tracking system, with the development of key technologies such as data association and multi-sensor fusion, it has gradually developed from single-sensor multi-target tracking to multi-sensor, and the types of sensors have also developed from single radar in the past to radar, infrared multiple type. [0003] After searching the existing technical literature, it is found that most of them analyze and evaluate a certain link of the multi-target tracking algorithm, and rarely evaluate the performance of the entire algorithm, such as improving and optimizing the performance of the data association part. method, limited evaluation indicators, lack of a third-party evaluation system, and is not objective. The only article that analyzes the ...

AI Technical Summary

Problems solved by technology

The only article that analyzes the entire system is "Multi-Target Tracking (MTT) Performance Evaluation System" published by Zhu Ziqian and Zhang Shifang in "Fire Control Radar Technology" Volume 32 in 2003. The disadvantage of the performance evaluation method of radar multi-target tracking system is that there is only one sensor of radar, the types of target source trajectories are limited, and the loading method is single, which is only limited to some typical tracks. It is not conducive to the confidentiality of the algorithm, and there is no single-step debugging function for the evaluated algorithm

Images