基于移动机器人的视觉伺服综合实验系统设计与实现OA

The experimental teaching in robotics engineering has long faced the challenge of visual perception and motion control being treated as isolated modules,making it difficult to gain a coherent system-level understanding.An integrated visual servoing laboratory system is established using the RoboMaster EP mobile robot as a platform.Target detection and feature extraction are carried out in the HSV color space.A monocular ranging model is derived from pinhole camera geometry,combined with error propagation analysis and EWMA filtering to reduce measurement uncertainty.A discrete PID controller is used to convert pixel-domain errors into gimbal angular-velocity commands,while a finite state machine orchestrates the two-degree-of-freedom gimbal and the Mecanum-wheel omnidirectional chassis through a four-phase sequence of searching,locking,tracking,and firing.Experimental data show that within the 0.5～2.0 m working range,the coefficient of determination for the ranging model is 0.998,with measured errors matching the theoretical prediction trend.The gimbal step response exhibits a rise time of approximately 0.4 s and a settling time of approximately 0.7 s,with a steady-state RMSE of 2.4 pixels.During coordinated gimbal-chassis tracking of a moving target,the pixel error remains within±15 pixels.The platform integrates visual modeling,closed-loop control,and multi-actuator coordination into a single experimental workflow,providing an effective tool for comprehensive engineering laboratory instruction.