基于新工科的移动机器人控制实验设计和教学实践

Design of the Mobile Robot Control Experiment and Teaching Practice Based on Emerging Engineering Education

针对机器人工程专业四年级学生开设的移动机器人控制实验,循序渐进地设计了基于深度学习的移动机械手抓取实验项目。从激光雷达同时定位与建图、自主导航与实时避障,到借助ROS的通信机制完成语音控制导航以及物体识别与检测,构建深度学习网络生成最优抓取位姿,并利用MoveIt功能包控制机械臂完成抓取动作。实践教学表明,基于项目式实验教学模式有利于把大学阶段多门课程知识有机交叉融合,本教学方式可提高学生解决复杂工程问题的能力。

Aiming at the mobile robot control experiment set up for the senior of robot engineering specialty, a mobile manipulator grasping experiment project based on deep learning is designed step by step. From simultaneous localization and mapping by lidar, autonomous navigation and real-time obstacle avoidance, to the completion of voice control navigation and object recognition and detection with the help of the communication mechanism of ROS, a deep learning network is constructed to generate the optimal grasping position and pose, and the manipulator is controlled by MoveIt to complete the grasping action. The practical teaching shows that the project-based experimental teaching mode really integrates the knowledge of multiple courses learned in the university and greatly improves the students’ ability to solve complex engineering problems.