作者: Xin Li , Chenglong He , Wei Wu , Zhaohua Xiong
关键词: Association (object-oriented programming) 、 Track (disk drive) 、 Artificial intelligence 、 Radar 、 Tracking system 、 Computer science 、 Computer vision 、 Rotation (mathematics)
摘要: In real distributed multi-radar multi-target tracking systems, uneven sensor bias might bring great challenge to track association because it would make an rotation of all the radar tracks, which always causes mistakes. Unfortunately, correcting tracks in traditional method using a single average estimation lead false some regions. Unlike most published previous works, this paper for first time proposes adaptive technique track-to-track against bias. The algorithm consists three stages: analysis, online and adjustment with removal. We also present anti-bias flow. simulation results show effectiveness technique, could accurately remove significantly improve probability, suggesting value engineering.