文献类型： 会议论文

第一作者： Hao Qiu

作者： Hao Qiu ¹ ; Xi Zha ² ; Xiaohua Jiang ³ ; Xiucai Huang ¹ ;

作者机构： 1.Institute of Artificial Intelligence, School of Automation, Chongqing University, Chongqing, China

2.Chongqing Academy of Agricultural Sciences, Chongqing, China

3.CISDI Information Technology (Chongqing) Co.,Ltd., Chongqing, China

关键词： Backstepping;Autonomous systems;Stochastic systems;Simulation;Heuristic algorithms;Neural networks;Probabilistic logic

会议名称： International Symposium on Autonomous Systems

主办单位：

页码： 1-6

摘要： In this paper, our focus is on prescribed performance control (PPC), a method that specifically targets a class of uncertain stochastic nonlinear systems exhibiting a pure feedback form. Firstly, we employ the mean value theorem to decouple the pure feedback form inherent in the system. In addition, the neural network and dynamic surfance techniques are exploited to identify the unknown dynamics. And utilizing the backstepping technique, a novel neuroadaptive control algorithm is developed to achieve the desired performance while tracking the reference signal. The proposed adaptive PPC scheme guarantees that all closed-loop signals maintain uniform ultimate boundedness in probability. Furthermore, the proposed scheme ensures tight tracking of the reference signal, thus attaining prescribed performance control. Lastly, simulation results are presented to demonstrate the practical feasibility of our approach.

分类号： tp2-53