Gesture Recognition-Based Robot Arm Teaching System

Robot arms are widely applied in various fields of human-computer interaction. Most mature robot arm products rely on sensors for control, leading to high costs and strong dependency on hardware. This paper uses gesture recognition to control the movement of a robot arm. It employs the D-H method for robot arm modeling, utilizes a camera to capture gesture information, and transmits image data via a USB interface. The Baidu Qianfan large model API is invoked through a wireless network card to achieve gesture recognition results. During the control process, the trajectory of the robot arm is recorded in real-time, enabling teaching functionality for the robot arm. This system can complete the control of the robot arm by gestures.