Development of Collaborative Human Robot Systems for Inspection and Repair Tasks in Tightly Constrained Spaces
Abstract
NEEC Technical POC & Topic Number: Thai Tran; KPT-01 A variety of Navy shipyard inspection and repair operations require hazardous and/or labor-intensive tasks that are carried out in tightly constrained spaces. Examples include coating removal, welding and painting, enclosed tank and void repair, and internal piping inspection. It is hypothesized that mobile robots may be immensely useful, in terms of both safety and reliability considerations, to perform these tasks in collaboration with humans. However, certain key technical challenges need to be addressed to realize such mobile robotic operations. While the robots are typically equipped with a multitude of sensors and actuators that are quite accurate, they are not necessarily very precise. Hence, probabilistic models have to be used to recursively estimate the states of the robots and their workspaces based on the continuous sequences of measurements and actions. The robots also encounter difficulties in successfully completing new types of tasks, particularly in unknown spaces that have not been explored before. Consequently, collision-free trajectories need to be planned for the robots such that they are robust to action and perception uncertainties. Humans should also be involved in the control loops of the robots, e.g., using teleoperation to exploit human flexibility and knowledge to handle uncertainties in the constrained spaces. Nevertheless, teleoperation can be slow and the robots should be autonomously operated when possible. Therefore, we will also examine the use of mixed-initiative traded control between automation and teleoperation based on the estimated failure probability for the pure automation mode, to ensure safe stable actions while performing the tasks in an optimal manner. All these research efforts would directly lead to the primary goal of educating the next generation of workforce in Navy shipyards who would be skilled in implementing the latest robotics technologies.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 13, 2025
- Source ID
- N001742010003
Entities
People
- Ashis Banerjee
Organizations
- United States Navy
- University of Washington