基于PID神经元网络的四旋翼飞行器姿态控制系统  

Attitude control system of quadrotor based on PID neural network

作　　者：邹逸飞 刘武 严中稳 张卫平 崔峰(微米/纳米加工技术国家级重点实验室 上海市北斗导航与位置服务重点实验室 ZOU Yifei;LIU Wu;YAN Zhongwen;ZHANG Weiping;CUI Feng(National Key Laboratory of Science and Technology on Micro/Nano Fabrication,Shanghai Key Laboratory ofNavigation and Location-based Services,Department of Micro/Nano Electronics,School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电子信息与电气工程学院微纳电子学系,上海200240

出　　处：《传感器与微系统》2020年第1期68-70,74,共4页Transducer and Microsystem Technologies

基　　金：航空基金资助项目(2016ZD57006);青岛海洋科学与技术国家实验室开放基金资助项目(QNLM2016ORP0404);上海市自然科学基金资助项目(17ZR1414500);预研项目(LZY2016215);上海市平台项目(16DZ2290103)

摘　　要：针对PID控制对于四旋翼飞行器这类强耦合系统参数整定困难和控制稳定性不足的问题,提出了基于PID神经元网络(PIDNN)的新型控制算法,并设计了基于STM32微处理器的微飞行器。MATLAB系统仿真和飞行测试结果表明,相比于传统PID控制,PID神经元网络控制具有较快的稳定时间和较小的超调量,提高了系统的稳定性和自适应能力,同时实现了广义解耦。Aiming at the problem that the traditional PID controller has difficulty in parameter setting and lack of control stability for strongly coupled system such asquadrotor,a new type of control algorithm based on PID neural network(PIDNN)is proposed.A micro quadrotor is designed based on STM32 microcontroller.The simulation of the system by MATLAB shows that compared with the traditional PID control,PID neural network control has faster stable time and less overshoot,which improves the stability and adaptive ability of the system,and generalized decoupling is realized at the same time.

关 键 词：四旋翼 PID神经元网络 STM32微处理器 姿态角 控制器 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] TP212[自动化与计算机技术—控制科学与工程]