doi: 10.3850/S201042862015001271

Backstepping Neural Control Design for BTT Missiles With Randomly Assigned
Parameters of Hidden Nodes

Hai-Jun Rong^1,a, Rong-Jing Bao^1,b and Xianghui Gao^{2 1}State Key Laboratory for Strength and Vibration of Mechanical Structures,
School of Aerospace, Xi'an Jiaotong University, No. 28 Xianning West Road,
Xi'an, 710049, China ^ahjrong@mail.xjtu.edu.cn ^bhjrong@mail.xjtu.edu.cn ²Faculty of Science and Technology, University of Macau, Taipa, Macau, China yb37433@umac.mo

ABSTRACT

According to the nonlinear, high coupling and uncertainty characteristics of BTT missiles, this paper presents a back stepping neural control approach for the attitude control of BTT missiles. In the proposed control method, the Single Layer Feed forward Networks (SLFNs) are used to approximate the nonlinear dynamics and uncertainties existing in the missile system. Then the resulting models are utilized to build the controllers. Different from the existing methods, the parameters between the input layer and the hidden layer are determined using the extreme learning machine (ELM) principle, where the parameters of hidden nodes are randomly assigned without human intervention. This can avoid the local minima and slow convergence suffered by back-propagation (BP) algorithm. Various simulation studies verify the feasibility of the proposed control approach.

Keywords: BTT missile, Backstepping, Neural Control.

---

Back to TOC