Abstract:
Multi-robot teams show exceptional promise across applications like Search-and-Rescue, disaster-response, agriculture, forestry, and scientific exploration due to their ability to go where humans cannot, parallelize activity, operate robustly to failures, and expand capabilities beyond that of an individual robot. Collaborative Simultaneous Localization and Mapping (C-SLAM) is a fundamental capability for these multi-robot teams as it is required for them to plan, navigate, and, in turn, achieve their mission goals. A key component of the C-SLAM system is the back-end algorithm responsible for estimating the state of the robot team from their distributed, noisy measurements. However, existing C-SLAM back-end algorithms struggle to handle the practical conditions experienced by multi-robot teams deployed in the real-world.

During real-world deployments multi-robot teams require C-SLAM algorithms that are — 1) online to permit replanning as new information is gathered, 2) robust to outlier data that we expect due to perceptual aliasing, 3) resilient to network challenges like delayed messages and communication outages, and 4) provide accurate and consistent solutions to the robot team.

We propose the design of a C-SLAM algorithm that addresses all of these real-world challenges. We first explore robustness in the context of single robot scenarios and design an incremental algorithm riSAM that operates robustly to outlier measurements and provides online efficiency. We next explore networking conditions and design a distributed batch C-SLAM algorithm MESA that can tolerate fully asynchronous communication. We then extend this algorithm to operate incrementally and online to provide real-time solutions to a multi-robot team. Finally, we propose an algorithm that leverages these prior works to provide C-SLAM solutions that are accurate, robust, resilient, and online. Further, we propose an extensive evaluation of this algorithm in realistic settings to validate its performance and its ability to overcome the challenges of real-world multi-robot C-SLAM.

Thesis Committee Members:
Michael Kaess, Chair
Sebastian Scherer
George Kantor
Timothy D. Barfoot, University of Toronto