基于异步多传感器采样量测的最优状态融合估计

Optimal State Fusion Estimate Based on Asynchronous Multisensor Sampled Measurements

针对一类异步多传感器采样系统,在对量测方程形式统一描述的基础上,给出以采样量测顺序为基准的序贯滤波算法的基本思想;为了避免现有右同步提升技术所可能引发的系统非因果这一重要问题,选择将融合周期内所有异步采样量测进行左同步提升,并发展了一种集中式最优状态融合估计算法,同时从理论上严格证明了基于左、右同步提升技术的融合估计算法在精度上的等价性;将三种融合估计算法对两个多速率采样系统进行仿真比较,在验证三种算法估计精度相同的同时,还对三种算法计算代价进行了分析和比较.相关结论对异步多传感器各最优状态估计融合算法的实际应用具有一定指导意义.

Optimal state fusion estimate for a class of asynchronous multisensor sampled systems is considered.A unified description of measurement equations is formally completed and the idea of sequential filtering algorithm based on the order of sampled measurements is given.All asynchronous sampled measurements in one fusion cycle are left synchronously lifted,based on this,an optimal centralized state fusion estimate algorithm which can avoid the drawback of non-causality of the right synchronously lifted system is proposed.The equivalence of estimation accuracy for the two algorithms based on right and left synchronous lifting techniques is proved.Simulation results of two multirate sampled systems are presented to verify the equivalence of estimation accuracy and to analyze the computational cost of the three algorithms.The conclusions have certain guiding significance on the practical engineering application of asynchronous multisensor optimal fusion estimation algorithms.