A three-dimensional self-adaptive region fuzzy guidance law based on RBF neural networks

Abstract

In this paper, a new three-dimensional (3D) self-adaptive region fuzzy guidance law based on radial basis function (RBF) neural networks (SRFGLRBF) design method is proposed. Firstly, 3D motion equations for pursuit-evasion of missile and target are given. Secondly, the proposed method is applied to decreasing the miss distance, which is mostly from the fixed navigation rates of traditional proportional navigation guidance laws (TPNGLs). The line of sight (LOS) rate and the closing speed between the missile and the target are regarded as the inputs of the fuzzy controller. The output of the fuzzy controller is the commanded acceleration. Then, a non-linear, self-adaptive region function is introduced based on the RBF neural networks to change the region. This non-linear function can be changed with the input variables, thus it can realise dynamic change of the fuzzy variable region. Finally, two engagement scenarios are examined and a comparison between TPNGLs and the proposed PNGLRBF is made. The simulation results show that the proposed 3D SRFGLRBF can achieve ideal miss distance better than TPNGLs.