Improved Particle Filter Algorithm Based on Averaging Likelihood Functions with Diverse Proportion
In the bearings-only tracking, the stochastic observation noise influenced the accuracy of particle weights.To solve the problem, an improved particle filter algorithm based on averaging likelihood functions with diverse proportion was proposed.At the step of particle weights updating, the new method used multi-observations instead of single observation to compute the likelihood functions of each particle, then averaged them with diverse proportion.The method reduced the influence of the stochastic observation noise to particle weights.Simulation results showed that, in the application of bearings-only tracking, the improved particle filter algorithm had better tracking performances than the original algorithm.
particle filterlikelihood functionaveraged with diverse proportionbearings-only tracking
NETL
NSTL
万方数据
维普
