Design and Development of a Mechatronic Vineyard Harvesting Robot with a Piezoresistive Sensor Network for a Human-Inspired Actuated Hand.

Considering the increasing difficulty in finding seasonal workers to harvest vineyards, adopting new autonomous technologies is essential to sustain the wine supply industry for both small and large producers. The goal of this master’s thesis is to design, develop, and test a mechatronic vineyard-harvesting robot. Specifically, the robot must recognize grapes in a vineyard environment using an AI vision model, grasp them with a human-inspired actuated hand equipped with piezoresistive sensors, and then cut the stem above each grape to complete the harvesting task before moving on to the next target. Since this project is carried out with five colleagues, this particular work focuses on prototyping the control of multiple servomotors using PWM signals generated by a PCA9685 16-Channel Servo Driver, connected to an STM32NUCLEO-F401RE microcontroller via the I2C protocol. Additionally, a scalable piezoresistive pressure sensor network is developed using a specialized analog sensing circuit based on a flip flop based shift register. To achieve this, a compact PCB is designed for each sensor, with force values applied to the sensors measured qualitatively through an analog-to-digital converter. These initial boards (the PCA and the NUCLEO) are then replaced with custom PCBs based on the STM32WB55V microcontroller, which includes a Bluetooth module for communication with an NVIDIA Jetson Nano board. This substantially reduces the space occupied by the driving and sensing elements, facilitating the efficient integration of the system into the robotic hand while managing the positioning of the various boards, connections to the motors and sensors, and power management. Subsequently, several tests are conducted on the complete system in a controlled laboratory environment that simulates a vineyard. During these tests, data are collected to evaluate the effectiveness of the sensors used and to determine if additional or alternative devices are necessary. Finally, potential future work is proposed to further enhance the system.