Optical Object Localization and RRT Motion Planning for Human-Robot Collaboration

In this paper, an approach for detecting objects and individuals to enable an industrial robot to conduct safe and reliable trajectories within a Human-Robot Collaboration (HRC) context is proposed. This implementation combines a depth camera image analysis and classification combined with a rapidly exploring random tree (RRT) motion planning algorithm, and a Particle Swarm Optimization (PSO) method to solve the inverse kinematics. The aim is to provide a solid framework for the HRC research community to conduct empirical studies with a robot arm. Apart from the concept itself and the technical implementation, the strengths and disadvantages of this method such as the calculations intensive trajectory planning process are discussed.