System-Level Integration and Performance Evaluation of Distributed SLAM for Application-Focused Robotic Deployment
This research presents a system-level integration and performance evaluation of Kimera-Multi, an open-source library for distributed Simultaneous Localization and Mapping (SLAM), in a multi-robot robotics context. The study investigates the applicability of distributed SLAM in real-world deployments by integrating Kimera-VIO, Kimera-Semantics, and Kimera-Distributed within a ROS 2-based system. A single-robot multi-agent framework was used to simulate distributed operation, with custom datasets collected on a university campus. Experimental evaluation focuses on identifying performance bottlenecks and system-level requirements to make distributed SLAM pipelines more robust and application-focused.
Advisor: Siavash Farzan
Zoom: Zoom meeting has ended.