带有未知死区微机械陀螺仪自适应神经网络量化容错控制  

Adaptive neural network quantization fault-tolerant control for micromechanical gyroscopes with unknown dead zones

作　　者：高丽文 祝贵兵 李彦朝 GAO Liwen;ZHU Guibing;LI Yanzhao(Marine Department,Zhejiang Institute of Communications,Hangzhou 311112,Zhejiang,China;School of Naval Architecture and Maritime,Zhejiang Ocean University,Zhoushan 316022,Zhejiang,China)

机构地区：[1]浙江交通职业技术学院海运学院,浙江杭州311112 [2]浙江海洋大学船舶与海运学院,浙江舟山316022

出　　处：《上海海事大学学报》2025年第1期1-7,共7页Journal of Shanghai Maritime University

基　　金：浙江省自然科学基金(2022C31039)。

摘　　要：针对微机械陀螺仪控制受执行器未知死区、执行器故障、不确定动态、未知时变扰动等实际因素的影响,提出一种自适应神经网络量化容错控制方案。采用量化器减少信号量化过程中的抖动,同时减少执行器更新控制指令的频率。使用单参数自适应技术抑制系统中不确定动态和未知时变扰动的集总扰动,使用有界估计和参数自适应等技术补偿未知死区、执行器故障和输入量化的影响,从而减少系统的计算负载。在反步法框架下,引入平滑函数,设计一种自适应神经网络容错控制律,以处理执行器未知死区非光滑非线性行为。理论分析证明,微机械陀螺仪轨迹跟踪闭环控制系统中的所有信号均有界,并且数值仿真证明了所提出的控制方案有效。Aiming at the influencing of practical factors,such as actuator unknown dead zones,actuator faults,uncertain dynamics,unknown time-varying disturbances in the control of micromechanical gyroscopes,an adaptive neural network quantization fault-tolerant control scheme is proposed.A quantizer is used to reduce jitter during the signal quantization,and to reduce the frequency of actuator updating control commands.The single-parameter adaptive technique is used to suppress the lumped disturbance of uncertain dynamics and unknown time-varying disturbances in the system,and the techniques such as bounded estimation and parameter adaptation are used to compensate for the effects of unknown dead zones,actuator faults and input quantization,thereby reducing the computational load of the system.Under the backstepping framework,a smoothing function is introduced,and an adaptive neural network fault-tolerant control law is designed to handle the non-smooth nonlinear behavior of the unknown dead zones of actuators.Theoretical analysis proves that all signals in the micromachined gyroscope trajectory tracking closed-loop control system are bounded,and numerical simulation proves that the proposed control scheme is effective.

关 键 词：微机械陀螺仪 容错控制 输入量化 神经网络控制 未知死区 

分 类 号：U666.12[交通运输工程—船舶及航道工程]