摘要
在基于卡尔曼滤波及其一些改进算法中,由于测量方差预先设定,从而导致信息资源的浪费和状态估计精度的下降,为此提出一种动态加权下测量方差自适应的同质多传感器融合算法.该算法依据各传感器当前时刻的滤波精度合理地分配权值,同时测量方差的时变特性,使得每次测量的信息得到充分的利用.仿真结果表明,该算法显著地提高了对机动目标的跟踪效果并具有实时性的优点.
In the algorithm based on kalman filter and its extension, the presupposition of the measurement variance leads to the waste of information and descent of state estimation accuracy. Hence, this paper presents a new algorithm of multisensor fusion based on self-adaptive measurement variance under dynamic weight. The algorithm reasonably distributes weight value according to the filter accuracy renewed of each sensor, meanwhile, the self-adaptive property of measurement variance makes innovation obtained each time sufficiently utilized. The simulation shows that this algorithm can obviously improve the efficiency of maneuvering target tracking on the basis of possessing real-time merit.
出处
《河南大学学报（自然科学版）》
CAS
北大核心
2005年第3期72-76,共5页
Journal of Henan University:Natural Science
基金
国家自然科学基金项目(60272024)
河南省高校杰出科研人才创新工程项目(2003KYCX003)
河南省高校创新人才培养工程项目.
