Design, Control and Optimization of a Robotic Hand and Wrist for Autonomous Grape Harvesting

This thesis describes the development and improvement of a robotic hand and wrist designed for an autonomous grape-harvesting robot. The project builds on an existing robotic platform with the goal of increasing dexterity, reliability, and the ability to handle grapes gently and efficiently. The robotic hand uses linear actuators to move each finger independently, allowing precise control of motion and gripping force. A servo motor drives the palm to adapt the grasp to different grape clusters, while two additional servos control the wrist to provide the necessary orientation during harvesting. The control system is implemented on an NVIDIA Jetson, which manages motion coordination and communication with the motor control boards. Each board, based on an STM32WB55MMG microcontroller, controls local actuators and communicates with the Jetson through Bluetooth Low Energy (BLE). This setup allows real-time data exchange and feedback for smoother and more stable operation. Laboratory tests were carried out to evaluate the performance of the new hand and wrist in terms of motion accuracy, response time, and grasp stability. The results show improved precision and flexibility compared to the previous design, demonstrating the potential of this system for future applications in autonomous and selective grape harvesting.ist and hand of a grape harvesting robot