Reachability Map-Based Motion Planning for Robotic Excavation

The trend towards automation of excavation has necessitated the development of an efficient method for excavator motion planning. Traditional approaches typically model the excavator as a 4-DoF (degrees of freedom) tandem manipulator, despite its three closed kinematic chains and linear actuation via three hydraulic cylinders. To address this limitation, this study proposes a novel search-based motion planning approach that is tailored to the excavator's unique characteristics. Specifically, we employ the excavator's reachability to construct a graph that encodes the feasible poses and moving direction of the excavator bucket in an efficient manner. Moreover, we integrate the displacement of hydraulic cylinders into the cost function to ensure the smoothness of the excavator's movement. Experimental analysis based on a typical digging task demonstrates the planner's ability to generate full-bucket motion while providing guarantees on the smoothness of driving.