Grasping Lightweight Objects With Chat-PM: A Rotorwash-Aware Motion Planning Method

Abstract

Grasping lightweight objects with aerial manipulators remains challenging due to the rotorwash interference caused by spinning rotor blades. This letter details a rotorwash-aware motion planning method for Chat-PM, a hybrid aerial-terrestrial vehicle equipped with a robotic arm. With proposed method's adjusting the vehicle's thrust and pose on flat surfaces timely, the rotorwash can be reduced and directed away from targets. To achieve this, we first model the rotorwash velocity distribution of Chat-PM in aerial and surface modes with computational fluid dynamics (CFD) simulations. The model further functions as an optimization objective of nonlinear model predictive control (NMPC) method, alongside other objectives including aerial-surface reference tracking and target grasping. An aerial-surface path searching method for the vehicle is then introduced, which takes into consideration both rotorwash velocity and kinodynamic feasibility to generate a high-quality reference for NMPC to track. Experimental results showcase the effectiveness of our proposed planning approach for aerial-surface grasping with rotorwash optimization.