Design of a Solar-Powered Car Based on Gesture Recognition

This paper presents the design of a smart solar-powered vehicle system based on gesture recognition, implementing real-time gesture recognition using the YOLOv5 algorithm on the K210 platform and controlling multiple motion modes through an STM32 main control module. The vehicle integrates omnidirectional movement with Mecanum wheels, ultrasonic obstacle avoidance, grayscale line following, and features a detachable solar module to enhance energy endurance. To improve recognition accuracy, the gesture dataset was balanced and annotated, and model optimization significantly increased recognition accuracy and real-time performance. Experiments demonstrate that the vehicle can quickly respond to user gestures and switch motion modes in various complex environments, achieving effective path and obstacle avoidance and line-following control, offering an innovative approach for low-power smart vehicles.