20-000001002 Investigation of an Autonomous Cooperating Robot System for Maintenance of Naval Platforms
Abstract
We aim to develop a novel multi-robot system (MRS) to enhance the agility of performing maintenance tasks for Navy vessels such as ships and submarines. Typically, a small number of complex robots are deployed to carry out such tasks, which are expensive to build, maintain and operate in naturally hazardous marine environments, rendering task execution susceptible to robot breakdowns and run-time decision making quality. Our proposed MRS will, instead, consist of a fleet of simple uni-functional, cooperating robots (UCRs), which are expected to benot only more cost-effective, but also enable operational benefits through on-demand generation of desired autonomy capabilities.In particular, we plan to focus on the following research objectives: allocate a suitable set of UCRs using auction-based algorithms for any given maintenance operation to ensure robust task execution with (near)-minimal operating cost; develop a semantic-topological multi-robot simultaneous localization and mapping method for environments characterized by multi-scale features, time-varying states, and presence of unexpected objects; and, devise a multi-agent reinforcement learning framework to enable collision-free navigation of the UCRs in uncertain environments. While we intend to evaluate the autonomous UCR fleet for inspection of water tanks inside docked ships, our work is expected to be useful for a variety of undersea applications ranging from on-mission inspection and repair of vessel structural components to early detection of adversarial threats.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 06, 2021
- Source ID
- N000142112075
Entities
People
- Ashis Banerjee
Organizations
- Office of Naval Research
- United States Navy
- University of Washington