Knowledge-Based Cooperative Particle Filter

In view of the current possible degradation phenomenon of particles in the particle filter and the problem of particle diversity caused by the re-sampling, a novel particle filter algorithm based on knowledge plate and co-evolution is proposed in this paper. The main idea of this algorithm is to sample from the importance density function in order to generate particle samples which are further divided into several sub-sample groups. Particle swarm searches for each sample in different areas and finds the optimal state estimation of the dynamical system through the communication between sampling particle swarm. Theoretical analysis and experimental results show that the proposed algorithm has potential advantages in improving convergence rate and keeping population diversity.