Abstract Robotic manipulation and transportation of payloads are finding widespread adoption, especially in industrial applications. This paper presents a novel approach towards multi-robot cooperative manipulation and transportation of payloads. The proposed hierarchical approach comprises of a sampling based global planner to generate feasible paths in the presence of static and dynamic obstacles. This path is then passed down as waypoints to a high-level model predictive controller (MPC) that accounts for the coupled kinematic constraints of the robot-payload system. The control commands generated by the MPC are passed to low-level controllers running on each robot to drive the payload along the desired path. The overall pipeline was successfully validated over different simulated maps, with static and dynamic obstacles. To further validate the applicability of the proposed approach in real-world conditions, a prototype system was built, consisting of two differential drive robots connected to a trolley using revolute joints. Experimental trials were conducted on three different layouts, where the environment was previously mapped, and the pose of the robot was measured in real-time using an indoor localization system. The path tracking performance of the proposed multi-stage controller was compared to that of an expert human teleoperator. The results indicate that in terms of path tracking performance in the presence of real-world disturbances like slip, the difference in length of the path followed under proposed approach and teleoperation is less than 14% across multiple trials over different maps.
Kunwar et al. (Wed,) studied this question.