Control and optimization of a specialized cutting tool for a grape harvesting robot

This thesis work concerns the partial mechanical redesign and subsequent modification of the control algorithm of a lightweight manipulator to enable its use as a specialized cutting tool for a grape harvesting robot. A preliminary study of the components and factory configuration have been conducted, along with the analysis of the programming code from the manufacturer. Several kinematic chains have been investigated in order to allow the manipulator to properly handle a pair of harvesting scissors as an end-effector. Then, specific brackets have been designed and manufactured to install the scissors and their actuator onto the main structure of the manipulator, providing it with an additional degree of freedom. Eventually, an appropriate control algorithm has been developed. The design process of brackets has been supported by finite element analyses carried out with the software Inspire by Altair. It is a software oriented towards additive manufacturing applications. A kinematic control algorithm based on the inverse Jacobian matrix has been firstly developed with Matlab, then it has been programmed into the manipulator control board through the Arduino IDE. A simulator has been also built in Matlab to predict the posture of the manipulator, to study the trajectory of the tool center point, to optimize the algorithm parameters and to implement an obstacle avoidance strategy. Tests have been performed to tune the parameters of the internal PID controllers of servomotors and achieve a good level of accuracy, as well as to validate the control algorithm. Future developments to extend the use and dexterity of the manipulator have been also discussed.