Acceleration-Level Pseudo-Dynamic Visual Servoing of Mobile Robots With Backstepping and Dynamic Surface Control
针对非完整移动机器人，提出加速度级伪动态视觉伺服结构，并设计了反步法和动态面控制两种自适应控制器，在未知深度信息下实现稳定跟踪，实验比较了两者性能。
In this paper, we propose an acceleration-level pseudo-dynamic visual servoing structure for the nonholonomic mobile robots, based on which we design two different adaptive controllers-backstepping and dynamic surface control (DSC) in the presence of unknown depth information. Different from existing kinematic controllers, which directly regard linear and angular velocities as control inputs, this paper designs acceleration control that is integrated to easily obtain smooth velocity signals to be accurately executed by the robot. Two controllers are designed and analyzed with Lyapunov techniques: 1) a backstepping controller yielding asymptotical stability and 2) a dynamic surface controller ensuring system errors to be ultimately uniformly bounded. The unknown depth is handled by designing an adaptive parameter estimation law in both methods. Finally, a comparison between backstepping and DSC is given based on the experimental results and the design procedures.