Affiliation:
1. AI and Autonomy Technology Center Agency for Defense Development Yuseong Daejeon Republic of Korea
Abstract
AbstractThis paper proposes a novel design strategy and task‐priority‐based control methodology for a robot to successfully complete a rescue operation in an extremely unstructured environment. The mechanical structure is designed to obtain both versatile manipulability and all‐terrain mobility. The regularized hierarchical quadratic program is used for whole‐body motion and force control. The optimization strategy is reasoning about regularization and thus it ensures convergence of the solution in the face of singularities while taking into account equality and inequality constraints. We demonstrate the effectiveness of the online optimization‐based control algorithms through extensive real‐world numerical and experimental results. Finally, we highlight that the rescue robot can successfully execute missions to extract a casualty and dispose of a dangerous object both indoor and outdoor environments.
Funder
Agency for Defense Development
Subject
Computer Science Applications,Control and Systems Engineering