An intelligent particle filter for state estimation and fault detection

With the continuous development of computer science and control science, the complexity of the system also increased rapidly. Accordingly, people began to improve the security and stability of the systems, and fault diagnosis in time is an effectively method to reduce the loss of property. The reality systems are invariably more complex, nonlinear and non-Gaussian. The previous method cannot solve the problem very well. However, as an unique technology, the particle filter (PF) can apply to nonlinear and non-Gaussian systems effectively. The accuracy of state estimation is influenced by the particle impoverishment problem in the quintessential PF algorithm. In order to deal with the particle impoverishment, we propose an intelligent particle filter (IPF) algorithm which based on genetic algorithm optimization after analyzed the particle filter algorithm. The common PF is a special circumstances of IPF that relieves the particular parameters. Results of these two experimental applications of the IPF are given to illustrate it can increase the particles diversity and improve the state estimation results of fault diagnosis.