Novel particle filter algorithm based on invasive weed optimization

A kind of novel particle filter algorithm based on InvasiveWeed Optimization (IWO) is presented to solve particle degradation of the typical particle filter. The algorithm uses UKF to generate the importance proposal distribution, then the InvasiveWeed Optimization algorithm is used to optimize the importance resampling particles. The principle of weight jitter is introduced into the process of optimization, a new way to multiply that particles which have high weight use small diffusion variance and those with low weight use big diffusion variance to guide particles reproduce dynamically, increasing the diversity of particles. By comparison with PF, PF-IWO and UPF, the simulation results show that the algorithm proposed in this paper reduces the degree of degradation of the particles, improving the tracking accuracy. Copyright © 2015 Binary Information Press.