Abstract:
This thesis develops sensor coverage algorithms for mobile robots that are scalable to large and complex environments. The core challenge is computing the shortest paths that can direct one or more robots to sweep onboard sensors over all accessible surfaces within an environment. This problem resembles the watchman route problem that is known to be NP-hard, and thus can be computationally prohibitive in large and complex environments. This thesis tackles two key tasks related to the sensor coverage problem – exploration of unknown spaces and inspection of known maps. For both, the efficiency of a solution is measured by coverage time and computational runtime. To find efficient solutions to sensor coverage problems, this thesis presents a hierarchical framework using dual-resolution reasoning. By first planning paths on a coarse resolution, then refining locally, this framework can produce shorter sensor coverage paths with lower computational cost than state-of-the-art methods. Extensive simulations and real-world experiments validate the framework, which also enabled robotic underground exploration in the DARPA Subterranean Challenge.

Thesis Committee Members:
Ji Zhang, Co-chair
Howie Choset, Co-chair
Stephen Smith
Henrik I. Christensen, University of California San Diego
John Leonard, Massachusetts Institute of Technology

More Information