Particle filter based on simulated annealing for target tracking
Tao Wu · Journal of Optoelectronics·laser · 2011
This paper proposes a novel particle filter tracking method based on simulated annealing(SAPF).The importance probability density function of particle filter is improved by using the simulated annealing method(SA).In order to expand the searching scope of particles,the amount of disturbance is introduced into the equation of state transition.According to the rule of Metropolis,the indigence is solved.At the same time,for moving particles to the high likelihood region,the temperature adjustment factor is used as a controller.Therefore,the proposal distribution function can better reflect the posterior probability distribution of the system state.Experimental results show that the SAPF has higher tracking accuracy compared with the PF.