Since its inception, the field of Robotics has striven to build versatile and reliable systems that demonstrate the capability to plan, adapt and survive in uncertain and unstructured environments as well as perform cooperative tasks. The intellectual merit of the proposed project lies in the definition and development of an integrated methodology for the study of the above problems in simulation, but under realistic physical constraints. In the above context, this proposal will develop algorithms to plan for robotic mechanisms under complex constraints by pursuing the combination of sampling-based planners with physical simulators. The proposal will also study motion generation and robot coordination in uncertain, unstructured and dynamic environments and develop frameworks to ensure the safety of the involved robots as well as compliance with physical constraints. The problems we propose to work on are interrelated; yet they can be treated separately. Considered together, they constitute a framework for augmenting the capabilities of modern robotics systems and are not constrained to address one problem in isolation or one particular system. The broader impact of the project will be implemented through the development and distribution of software for popular sampling-based planners, and a number of training and mentoring activities.
