Autonomous golf ball picking robot design and development

This paper presents the methodology and the results on the design and development of an autonomous golf ball picking robot. The strategy followed to develop a commercial product is presented, based on prior identification requirements, which consists of picking up golf balls on a Driving Range in a safe and efficient way. A functional prototype was built and tested. It has autonomy of up to 8h per day. It is capable of collecting up to 1200 balls in one single journey. It weighs 130 kg and is capable to climb slopes of up to 22o. The maximum speed is 8 km/h and the robot takes 140 min to completely sweep a 25,000 m2 field at 7.2 km/h (2 m/s) average speed. A hybrid information system was implemented in order to provide a statistically relevant prediction of golf balls location. Autonomous navigation was developed and tested on a simulation environment. Preliminary results showed that the new path planning algorithm Twin-RRT* is able to form closed loop trajectories and improve the result over time. Kinematic constraints were already taken into account on the algorithm. This sampling based algorithm has potential usage in solving other TPP (Travelling Purchaser Problem) related problems. Two patents are pending concerning the technological novelties of this work.