磁性微型机器人神经自适应轨迹跟踪控制
Neural Adaptive Trajectory Tracking Control for a Magnetic Microrobot
-
摘要: 针对复杂环境和不确定的动态模型问题, 提出了一种自适应神经网络控制算法, 并且基于李雅普诺夫理论保证了闭环系统的稳定性。相较而言, 基于模型的控制策略需要精准地知道微机器人的动力学以及周围环境参数, 而本文提出的基于径向基函数神经网络的状态反馈控制策略可以有效地根据状态与期望轨迹在线估计系统模型的不确定项。最后, 在所开发的磁场驱动的微型机器人系统上进行了2项实验并进行了对比, 以验证所提出的控制器的有效性。结果表明, 曲线与直线的轨迹跟踪均方根误差分别达到6.2204像素与6.4279像素, 明显优于传统的PID(比例-积分-微分)算法。Abstract: In response to challenges posed by complex environments and uncertain dynamic models, an adaptive neural network control algorithm is proposed, with stability of the closed-loop system ensured based on the Lyapunov theory.Compared to model-based control strategies that necessitate precise knowledge of the microrobot dynamics and surrounding environment, a state feedback control strategy based on a radial basis function neural network(RBFNN) is introduced, which can effectively estimate system model uncertainties online from both states and desired trajectories. Ultimately, 2 experiments are carried out and compared on the developed magnetic-field-driven microrobot system to validate the effectiveness of the proposed controller. The results demonstrate that the root mean square errors(RMSEs) for trajectory tracking in curved and straight paths achieve 6.2204 pixels and 6.4279 pixels, respectively, significantly surpassing the performance of the traditional PID(proportional-integral-derivative) algorithm.