Share Email Print
cover

Proceedings Paper

Semiautonomous navigation of mobile robots
Author(s): Johan Forsberg; Thomas Hogstrom; Ake V. Wernersson
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The paper is on operations for semi-autonomous mobile robots. The robot is assumed to be remotely controlled by an operator. It is difficult for an operator to directly control the robot, especially over a communication link with low bandwidth and/or time delays. Therefore it is necessary to close the control loop in the robot with the operator giving high level commands. The operator is still needed as the fully autonomous robot does not exist today, except for limited scenarios. The scene around the robot is sensed using a scanning range measuring laser and a camera. The high level scene interpretation is done by the operator who also does the high level planning. Which operations the robot are to perform is indicated by pointing in the images or in a map created by the robot. Operations are functions that the robot can perform autonomously. They can be simple, like `Travel 2 m ahead', or more advanced, like `Follow the corridor and take the first door to the right'. Some typical operations are: (1) Lock the heading of the robot when driving on a `straight' line and preprogrammed 90 and 180 turns. (2) Automatically enter a camera defined line; the direction of the camera is used to drop a new coordinate frame at any time (using a knob on the keyboard). The robot will automatically enter this new line and also compensate for the overshoot. (3) Travelling along corridors. The operation is both robust and precise. The precision is about 1 cm at 1 m/s and the robot is not disturbed by people passing it in the corridor. (4) Command for passage through a door works within 1 cm and 0.5 degrees at a speed of 0.5 m/s. The range weighted Hough Transform on laser measurements extracts the walls in an indoor environment. This is used to create an internal map in the robot which is used for operations like corridor following or passing through doors. The map is also useful when presenting information to the operator.

Paper Details

Date Published: 9 January 1995
PDF: 11 pages
Proc. SPIE 2352, Mobile Robots IX, (9 January 1995); doi: 10.1117/12.198963
Show Author Affiliations
Johan Forsberg, Lulea Univ. of Technology (Sweden)
Thomas Hogstrom, Linkoping Univ. (Sweden)
Ake V. Wernersson, Lulea Univ. of Technology and Linkoping Univ. (Sweden)


Published in SPIE Proceedings Vol. 2352:
Mobile Robots IX
William J. Wolfe; Wendell H. Chun, Editor(s)

© SPIE. Terms of Use
Back to Top