Digital Commons Network^™

Real-Time Edge Computing For Autonomous Systems, Junlin Ou

Theses and Dissertations

This research presents systematic studies and algorithmic and hardware development to enable real-time edge computing for autonomous systems, in particular, for robot path planning and localization applications. First, a low-cost, indoor, multi-camera positioning system is proposed to localize ground robots using the ArUco marker. In this system, the normalized coordinates of the marker processed by OpenCV are utilized to calculate the camera coordinates. Data-driven models are developed to establish a mapping relationship between the camera coordinates and the world coordinates of the marker. Fifty data pairs of the camera and world coordinate systems within one camera view (150 pairs in …