Loading...
Can you hear me now? Robotic networks for information acquisition
Schack, Matthew A.
Schack, Matthew A.
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2024
Date Submitted
Keywords
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
2026-04-04
Abstract
Robot teams are an effective tool for Information Acquisition—e.g., Search and Rescue or Remote Monitoring—since the team can collect information from multiple sources in parallel and form ad-hoc networks to transfer the information to human operators. However, communication is not guaranteed as occluding obstacles, distance, or environmental interference may prevent communication between team members; therefore, robots must reason about how their motion affects their communication. Requiring the team to maintain connectivity ensures that robots transmit obtained information to human operators at a base station. However, communication constraints limit the valid space so some information may not be
reachable. Conversely, allowing robots to disconnect and then rendezvous with the team lets robots visit previously unreachable areas, but robots must reason about when and where to rendezvous to ensure they deliver all data to the base station. We examine centralized algorithms for Communication Constrained Multi-Robot Information Acquisition, show the problem is PSPACE-hard, and construct an algorithm that offers linear scalability in many cases. Furthermore, we relax the communication constraints and construct decentralized algorithms that pre-plan or predict rendezvous locations with disconnected team members, allowing the team to reach areas that would be unreachable in the communication-constrained case. We show that our decentralized algorithms return partial maps back to the base station faster, improving a time-weighted information reward by up to 87% in exploration tasks compared to baseline works.
Associated Publications
Rights
Copyright of the original work is retained by the author.