基于PIDNN的六旋翼无人机飞行控制算法研究  

Research on flight control algorithm for six-rotor unmanned helicopter based on PIDNN

作　　者：徐会丽 石明全 张霞 王晓东 井光辉 

机构地区：[1]中国科学院大学中国科学院重庆绿色智能技术研究院,重庆400714

出　　处：《传感器与微系统》2017年第12期25-27,共3页Transducer and Microsystem Technologies

基　　金：基于虚拟现实技术的机械系统动力学仿真系统开发及应用项目(CSTC2012GGB60001);重庆市北碚区应用开发计划项目(Y62Z120G10)

摘　　要：针对六旋翼无人机比例—积分—微分(PID)控制器参数优化困难的问题,采用了PID神经网络(PIDNN)控制方法,利用其非线性映射和自学习的特性,实现了姿态控制参数的动态调整,增加了系统的自适应性。为验证方法的有效性,通过Matlab的Simulink模块构建了六旋翼无人机数学模型;利用S函数实现了基于反向传播(BP)算法的PIDNN控制器;将仿真结果与传统PID控制效果进行对比,结果表明:在缩短姿态调整时间与减少超调量方面,PIDNN方法控制效果优于PID方法。Aiming at problem that parameters of six-rotor unmanned helicopter proportion integration differentiation（ PID） controller are difficult to be optimized,a PID neural network（ PIDNN） method which has characteristics of nonlinear mapping and self-learning,is developed and employed. It can adjust attitude control parameters dynamically and promote the self-adaptability of controller system. To prove the effectiveness of the method,the mathematical modeling for six-rotor is established by Simulink module in Matlab,PIDNN controller based on back propagation（ BP） algorithm is obtained by S-function. Simulation results are compared with control effects of PID,verifying that control effect of PIDNN is more valid in reducing attitude adjustment time and decreasing overshoots.

关 键 词：六旋翼无人机 比例—积分—微分神经网络算法 姿态控制 

分 类 号：V249.1[航空宇航科学与技术—飞行器设计]