基于自适应蝙蝠粒子滤波算法的WSN目标跟踪

WSN Target Tracking Based on Adaptive Bat Particle Filter Algorithm

摘要:

为了解决粒子滤波(PF)的无线传感器目标跟踪中样本贫化导致的精度较低的问题,提出了自适应蝙蝠粒子滤波的WSN目标跟踪方法。通过自适应的蝙蝠算法的滤波算法优化粒子重采样过程,结合最新的观测值定义粒子的适应度函数,引导粒子整体上向较高的随机区域移动。同时利用动态自适应惯性权重探索新的粒子位置更新为设计机制,引入动态适应惯性权重值, 有效调整全局探索和局部探索适应能力、改善粒子贫化和局部极值问题,增加粒子群多样化从而提高跟踪性能。实验结果表明,自适应蝙蝠粒子滤波算法重采样方法可以防止粒子的退化,增加粒子的多样性,减少跟踪误差,可以减少算法的运行时间,实时追踪性能大幅提高。与BA-PF算法和PF算法相比较,IBAPF 算法的计算时间是最短的,IBA-PF算法的位置和速度的平均平方根误差最小(位置0.0311、0.0202、速度0.0262、0.0101),PF算法的跟踪精度是最低的,而IBA-PF跟踪精度较高,IBA-PF算法被证明具有良好的跟踪性能。

Abstract:

In order to solve the problem of low precision caused by sample dilution in WSN target tracking based on particle filter (PF), a WSN target tracking method based on adaptive bat particle filter was proposed. The re-sampling process of particle filtering was optimized by the improved bat algorithm, and the fitness function of the particle was defined based on the latest observations to guide the particle to move to a higher random region on the whole. At the same time, the dynamic adaptive inertia weight is used to explore the new particle position update as the design mechanism, and the dynamic adaptive inertia weight value is introduced to effectively adjust the global exploration and local exploration adaptability, improve the particle dilution and local extreme value problems, and increase the particle swarm diversification to improve the tracking performance. The experimental results show that the resampling method of the adaptive bat particle filter algorithm can prevent the degradation of particles, increase the diversity of particles, reduce the tracking error, reduce the running time of the algorithm, and greatly improve the real-time tracking performance. IBAPF algorithm has the shortest computation time. Compared with BAPF algorithm and PF algorithm, IBAPF algorithm has the smallest mean square root error of position and velocity (position 0.0313, 0.0270, speed 0.02021, 0.0102), PF algorithm has the lowest tracking accuracy, and IBAPF algorithm has the highest tracking accuracy. The IBAPF algorithm is proved to have good tracking performance.