Research on Mobile Grasping and Path Planning of Robotic Arms Based on Deep Learning
摘要
In view of the increasing aging of the Chinese population, this paper studies the mobile grasping robot for helping the elderly, because the environment in the bedroom is very complex, which requires the robot to have the ability to sense the indoor environment and pose estimation, so this paper studies the pose estimation algorithm, robotic arm grasping and mobile robot mapping and positioning in the environment, and the experiment proves that the size of our improved DOPE network model is reduced by 38.8% The inference speed is increased by 1.13 FPS, the planning time of the RRT algorithm is shortened by 9.5%, and the wheeled odometer fused with IMU information has a more robust positioning effect.