Design of Intelligent Picking Robot Based on Embedded Technology

Abstract

Aiming at the practical problems of labor shortage and low manual picking efficiency in agricultural production, this paper designs
an intelligent picking robot system based on embedded technology. The system adopts a distributed control architecture of "upper computer
+ lower computer". It takes Jetson Nano or Raspberry Pi as the main control core to process visual information, and uses STM32 microcontroller as the underlying motion controller. Integrated with RGB-D camera, ultrasonic sensor, flexible gripper and other modules, it realizes
the complete operation process of target recognition, autonomous navigation and precise picking. In terms of visual recognition, the system
applies the YOLO series algorithm to detect fruits and judge maturity. In terms of motion control, the PID algorithm is used to optimize drive
control, and the RRT algorithm is combined for obstacle avoidance path planning of the manipulator. Experimental results show that the design
can effectively identify target fruits, achieve stable autonomous navigation and flexible grasping, significantly improve the intelligence level and
efficiency of picking operations, and provide a feasible scheme for the research and development of agricultural automation equipment.