In this paper, we describe a real-time navigation assistance system dedicated to teleoperated vehicles for indoor applications. In order to obtain real-time reaction and motion continuity (no ''perception/move/step'' cycle), the system is decomposed in three levels working in parallel, with short cycle times. The lowest level, the pilot, continuously computes tangential and rotational speeds for the locomotion system, in order to follow a particular subgoal. This subgoal is dynamically generated by the second level, the subgoal generator, in order to avoid local obstacles, and tracking milestones positions (x,y,th). The highest level, the geometrical path planner, computes an optimal path composed of milestones to be roughly followed by the vehicle. In the first part of the paper, we present a global view of the navigation assistance system, its decomposition in three levels and its embedding in the overall system. In the second part, we give a more detailed description of each level, and we terminate with a presentation of results obtained in a real environments.

Date Published: 4 May 1993
PDF: 12 pages
Proc. SPIE 1831, Mobile Robots VII, (4 May 1993); doi: 10.1117/12.143796

Published in SPIE Proceedings Vol. 1831:
Mobile Robots VII
William J. Wolfe; Wendell H. Chun, Editor(s)