| || || Chand, Aneesh Neeschal.|
| || || Development of a book retrieval robot|
Author:Chand, Aneesh Neeschal.
Institution: University of the South Pacific.
Call No.: Pac Z 699 .C43 2006
Copyright:Under 10% of this thesis may be copied without the authors written permission
Abstract: This thesis presents the design paradigm and development of a mobile robot intended to eliminate the human component in the book retrieval process by autonomously navigating to and retrieving a specific book from a bookshelf. Notably a salient feature of this robot is the employment of an economical and effective book detection and identification technique, bar coding technology, in the formulation of a time-optimal and computationally elementary method in contrast to vision-based techniques which has been used by researchers in similar work. The formulation of the mechanical structure of the robot with the constructional details is presented first. This includes the robot framework, steering mechanism, the gripper and book retrieval mechanisms. The robot moves along a predefined path leading it to the books. Various electronic technologies are implemented for, including but not limited to, the steering, locomotion, navigation, speed control and motor actuation. A typical scanning device scans the barcodes of books to match the scanned book call number with the user entered book call number. An affirmative match results in a positive identification and the robot then executes a systematic chain of book retrieval manoeuvres. The software development for the robot intelligence and cognition is presented following that. An onboard laptop computer running Linux operating system is the principal controller of the robot with an additional PIC16F877 microcontroller used as a secondary controller. The results of the experimentation to validate and verify the performance are then given. In addition the merits, constraints and biasing considerations of the robot are also discussed. The research concludes with the attainment of a robot prototype with a book identification accuracy of 98% and performance efficacy (percentage success) of 53% in physical retrieval of books. Recommendations for areas of further research are also made.