| 引用本文: | 贾桐,李秀智,张祥银.车载惯性稳定平台的神经网络滑模控制[J].控制理论与应用,2021,38(1):13~22.[点击复制] |
| JIA Tong,LI Xiu-zhi,ZHANG Xiang-yin.Neural network sliding mode control for vehicle inertially stabilized platform[J].Control Theory and Technology,2021,38(1):13~22.[点击复制] |
|
| 车载惯性稳定平台的神经网络滑模控制 |
| Neural network sliding mode control for vehicle inertially stabilized platform |
| 摘要点击 2697 全文点击 897 投稿时间：2020-06-05 修订日期：2020-09-01 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2020.00329 |
| 2021,38(1):13-22 |
| 中文关键词 神经网络控制 反演控制 滑模控制 斯特里贝克模型 惯性稳定平台 |
| 英文关键词 neural networks control backstepping sliding mode control stribeck friction model inertially stabilized platform |
| 基金项目 国家自然科学基金项目(61703012), 北京市自然科学基金项目(4182010). |
|
| 中文摘要 |
| 三轴车载惯性稳定平台为复杂的MIMO非线性系统, 针对其在不确定扰动下的伺服控制问题, 本文设计了
一种神经网络反演滑模控制器(NNBSMC). 首先, 选用反演法对其解耦, 同时引入滑模控制律增加系统的抗干扰性;
其次针对框架间的非线性摩擦力与系统耦合选用RBF神经网络作为扰动估计器, 以便实时估计与补偿; 然后采用前
向增稳通道应对建模参数不精确以保证系统的稳定性. 最后, 利用Lyapunov定理分析了闭环系统的稳定性, 在伺服
控制与姿态锁定的仿真实验中分别与双环PID、滑模控制和反演滑模控制作对比, 结果验证了提出的控制算法的有
效性和鲁棒性. |
| 英文摘要 |
| Three-axis vehicle inertial stabilization platform (VISP) is a complex MIMO nonlinear system. Aiming
at the problem of uncertain disturbance in servo control of three-axis VISP, a neural network backstepping sliding mode
controller (NNBSMC) is designed. Firstly, backstepping control is used to decouple the three-axis VISP, and sliding mode
control law is introduced to increase the anti-interference of the system. Secondly, the RBF neural network is selected as
the disturbance estimator for the nonlinear friction between the frames and the system coupling for real-time estimation
and compensation. Thirdly, forward stabilization channel is used to deal with the inaccuracy of modelling parameters to
ensure the stability of the system. Finally, the stability of closed loop system is analyzed by using Lyapunov theorem.
In the simulation of servo control and attitude locking, NNBSMC was compared with the dual-loop PID controller, SMC
controller and BSMC controller. The results verify the effectiveness and robustness of the proposed control algorithm. |
|
|
|
|
|