The proposed project will tackle the following inter-related research topics regarding localization, monitoring, and motion coordination of autonomous indoor service robots: (1) adaptive coverage-planning of arbitrary and uncertain non-convex indoor regions, (2) accurate and robust indoor localization of service robots using vision-based technologies, (3) deep learning based depth estimation and high-precision path-tracking control of service robots, (4) intelligent detection of certain robot system and environmental states. The tools to be used in devising solutions for these problems include geometric and graph theoretical optimal path planning algorithms, various sensor data fusion techniques, model-predictive optimal control, artificial intelligence and machine learning approaches including reinforcement learning. The goal is to produce solutions to the four general problem topics above and to apply these solutions to provide reliable autonomous motion of Avidbots robots cleaning semi-structured non-convex indoor areas, with precise on-line monitoring, maximal area and quality of the cleaning, guaranteed collision avoidance, without requiring frequent calibration and fine-tuning.
Baris Fidan;William Melek;Soo Jeon;Stephen Smith;Ehsan Hashemi
Niraj Reginald;Megnath Ramesh;Omar Al-Buraiki;Xiule Fan
Professional, scientific and technical services
University of Waterloo
Find the perfect opportunity to put your academic skills and knowledge into practice!Find Projects