Date of Award

1994

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

S. S. Iyengar

Abstract

In the present work, the author explores the issues surrounding the design and development of an intelligent wheelchair platform incorporating the "semi-autonomous" system paradigm, to meet the needs of individuals with severe motor disabilities. The author presents a discussion of the problems of navigation that must be solved before any system of this type can be instantiated, and enumerates the general design issues that must be addressed by the designers of systems of this type. This discussion includes reviews of various methodologies that have been proposed as solutions to the problems considered. Next, the author introduces a new navigation method, called Incremental Signature Recognition (ISR), for use by semi-autonomous systems in structured environments. This method is based on the recognition, recording, and tracking of environmental discontinuities: sensor reported anomalies in measured environmental parameters. The author then proposes a robust, redundant, dynamic, self-diagnosing sensing methodology for detecting and compensating for hidden failures of single sensors and sensor idiosyncrasies. This technique is optimized for the detection of spatial discontinuity anomalies. Finally, the author gives details of an effort to realize a prototype ISR based system, along with insights into the various implementation choices made.

Pages

150

Share

COinS