Selection of Fuzzy Controller Membership Functions Based on Adaptive Gaussian Cloud Transform

基于自适应高斯云变换的模糊控制器隶属度函数选取

¹ Aeronautics and Astronautics Engineering College, Air Force Engineering University, Xi’an 710038, China
² Northwestern Polytechnic University, Xi’an 710038, China

Received: 8 May 2017

Abstract

Aiming at the boundedness of existing methods of selecting membership functions, an adaptive Gaussian cloud transform algorithm which is guided by the threshold values of hybridization degree is proposed to construct concept hierarchy from original sample data, and then the number, shape and coverage area of membership functions can be derived from the distribution of Gaussian cloud. To test and verify the effectiveness of membership function that is extracted based on adaptive Gaussian cloud transform algorithm, a six-degree-of freedom model of unmanned aerial vehicles(UAV) is constructed, and a fuzzy controller of pitching angle is established with the platform of Simulink. The simulation results show that the fuzzy controller which includes membership functions derived from the distribution of Gaussian cloud transform can achieve perfect control performance of pitching angle and meanwhile obtain good dynamic response characteristics.

摘要

针对现阶段模糊控制器中隶属度函数选取的局限性，利用自适应高斯云变换算法在概念含混度阈值的引导下对工程样本数据进行概念层次划分，根据最终所得的高斯云分布情况确定出隶属度函数的数目、形状及覆盖范围。为了验证自适应高斯云变换算法提取的隶属度函数的合理性，构建了无人机六自由度模型，搭建了基于simulink仿真平台的无人机俯仰角模糊控制器用于仿真验证。数值仿真结果表明嵌套有自适应高斯云变换算法所确定的隶属度函数的无人机俯仰角模糊控制器可有效实现俯仰角的控制，同时具有良好的动态响应特性。