Abstract
In the robot navigation problem, noisy sensor data must be filtered to obtain the best estimate of the robot position. The discrete Kalman filter, commonly used for prediction and detection of signals in communication and control problems, has become a popular method to reduce the effect of uncertainty from the sensor data. However, in the domain of robot navigation, sensor readings are not only uncertain, but can also be relatively infrequent compared to traditional signal processing applications. In addition, a good initial estimate of location, critical for Kalman convergence, is often not available. Hence, there is a need for a filter that is capable of converging with a poor initial estimate and many fewer readings than the Kalman filter. To this end. we propose the use of a Recursive Total Least Squares Filter. This filter is easily updated to incorporate new sensor data, and in our experiments converged faster and to greater accuracy than the Kalman filter.
Original language | English (US) |
---|---|
Title of host publication | Reasoning with Uncertainty in Robotics - International Workshop, RUR 1995, Proceedings |
Editors | Michiel van Lambalgen, Frans Voorbraak, Leo Dorst |
Publisher | Springer Verlag |
Pages | 221-234 |
Number of pages | 14 |
ISBN (Print) | 3540613765, 9783540613763 |
DOIs | |
State | Published - 1996 |
Event | International Workshop on Reasoning with Uncertainty in Robotics, RUR 1995 - Amsterdam, Netherlands Duration: Dec 4 1995 → Dec 6 1995 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 1093 |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Other
Other | International Workshop on Reasoning with Uncertainty in Robotics, RUR 1995 |
---|---|
Country/Territory | Netherlands |
City | Amsterdam |
Period | 12/4/95 → 12/6/95 |
Bibliographical note
Publisher Copyright:© Springer-Verlag Berlin Heidelberg 1996.