Zonotopic particle filtering based state estimation algorithm and its application on temperature recognition for Lithium battery

For solving the linear system state estimation problem, a zonotopic particle filtering based state estimation algorithm is proposed. By wrapping each particle with a recursive zonotope, the predicted particle is obtained according to the importance density function. Then, the particle prediction is updated by using the parameter feasible set constraint band, before analyzing the positions of each particle and the zonotopic. For the particles outside the region, a point randomly generated from the zonotope is used to update the particles close to the true value, increasing the particle diversity of the filter resampling process. The given case of temperature recognition for Lithium battery shows the feasibility and effectiveness of the proposed method.