Chopthin重采样粒子滤波的目标跟踪算法

Particle Filter with Chopthin Resampling for Target Tracking

针对粒子滤波中传统重采样存在的滤波性能不稳定、有效粒子数波动剧烈、样贫的缺点,提出了一种基于Chopthin重采样粒子滤波的目标跟踪算法。与传统重采样相比,Chopthin重采样产生的粒子权重不相等,粒子相对集中时,对边缘粒子的舍弃力度更小,因此能够改善传统重采样存在的样贫。Chopthin重采样可以在每一个迭代周期进行,不必在有效粒子数低于阈值时才进行,有效粒子数和滤波性能更加稳定。仿真实验表明,在不增加计算量的前提下,所提算法克服了传统重采样的缺点。

Traditional resampling methods in particle filter suffer from the disadvantages of unstable filtering performance,severe fluctuation of Effective Sample Size(ESS)and sample impoverishment.A target tracking algorithm based on particle filter with Chopthin resampling is proposed.Compared with traditional resampling,Chopthin resampling produces a set of unequal weighted particles.It discards less edge particles when the particles are relatively concentrated,which is helpful for solving the sample impoverishment problem of the traditional resampling.Different from the traditional resampling that only occurs when ESS is below a threshold,Chopthin resampling can be implemented in every iteration cycle.Therefore,the ESS is and filtering performance are more stable.Simulation shows that the new method can overcome the disadvantage of traditional resampling without increasing the calculation cost.