Neuro-Augmented Observer Based External Angular Rate Sensing Approach for Gyrowheel System
Neuro-Augmented Observer Based External Angular Rate Sensing Approach for Gyrowheel System
复制标题
DOI：
10.1109/access.2023.3262676
复制
发表时间：
2023
期刊：
影响因子：
3.9
通讯作者：
Yuyu Zhao;Chao Suo;Yuxiao Wang
中科院分区：
文献类型：
--
作者：
Yuyu Zhao;Chao Suo;Yuxiao Wang
Gyrowheel是一种创新的仪器，可以作为动量执行器，同时测量小型航天器的角速率。与传统的陀螺仪不同，陀螺轮动力学是非常复杂和非线性的，受到各种不确定性和干扰。这导致了利用陀螺轮实现精确角速率感测的挑战。为了克服现有方法的缺点，本文研究了外部角速度传感问题。采用径向基函数神经网络，并进一步结合增广观测器来处理陀螺飞轮系统的非线性、不确定性和干扰。采用神经增强观测器，本文提出了一种实用的角速度传感方法，具有令人满意的精度，而不需要一个确切的知识的陀螺轮动力学。仿真实验验证了该方法的有效性和优越性。
Gyrowheel is an innovative instrument that can function as a momentum actuator and simultaneously measure angular rates for small spacecrafts. Unlike traditional gyroscopes, the gyrowheel dynamics is significantly complicated and nonlinear, subject to various uncertainties and disturbances. This results in challenges to realize accurate angular rate sensing with the gyrowheel. To overcome the drawbacks of the existing methods, the external angular rate sensing problem is investigated in this paper. A radial basis function neural network is adopted and further combined with an augmented observer to deal with the nonlinearities, uncertainties and disturbances of the gyrowheel system. Employing the neuro-augmented observer, this paper proposes a practical angular rate sensing approach with satisfactory accuracy and without the requirement of an exact knowledge of the gyrowheel dynamics. Simulation tests illustrate the effectiveness and superiority of the proposed method.