Distributed Planning of Collaborative Locomotion: A Physics-Based and Data-Driven Approach

• Published in: IEEE access Vol. 11; p. 1
• Main Authors: Fawcett, Randall T., Ames, Aaron D., Hamed, Kaveh Akbari
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Collaboration; Computational modeling; Legged locomotion; Legged Robots; Locomotion; Motion control; Multi-Contact Whole-Body Motion Planning and Control; Nonlinear control; Optimization and Optimal Control; Payloads; Planning; Predictive control; Quadrupedal robots; Reduced order systems; Rigid structures; Robot dynamics; Robots; Subsystems; Tracking control; Trajectory; Trajectory optimization; Trajectory planning
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2023.3332820

This work aims to provide a computationally effective and distributed trajectory planner at the intersection of physics-based and data-driven techniques for the collaborative locomotion of holonomically constrained quadrupedal robots that can account for and attenuate interaction forces between subsystems. More specifically, this work lays the foundation for using an interconnected single rigid body model in a predictive control framework such that interaction forces can be utilized at the planning layer, wherein these forces are parameterized via a behavioral systems approach. Furthermore, the proposed trajectory planner is distributed such that each agent can locally plan for its own trajectory subject to coupling dynamics, resulting in a much more computationally efficient method for real-time planning. The optimal trajectory obtained by the planner is then provided to a full-order nonlinear whole-body controller for tracking at the low level. The efficacy and robustness of the proposed approach are verified both in simulation and on hardware subject to various disturbances, payloads, and uneven terrains.