Adaptive Bernoulli Filter For Single Target Tracking In Uncertain Detection Environment

Bo Li,Shuo Wang,Xu Jia

2017-01-01

Abstract:As a well-known approach, the Bernoulli filter can complete the task of single target tracking using the conventional measurements. However, the detection probability and the clutter rate are usually uncertain during the whole detection process. Once the detection model parameters mismatch, the tracking performance would become unstable and imprecise. In view of these problems, an adaptive Bernoulli filter and its particle implementation are proposed in this paper. First, we introduce the arbitrary value to represent the detection probability, and then compute the actual target component and the clutter component respectively in the extended state space. Employed the adaptive filtering scheme, the uncertain detection parameters are adaptively learnt. Numerical simulations demonstrate that the proposed filter can track the maneuvering single target in the uncertain detection environment with promising performance.