Vibration-based terrain classification for electric powered wheelchairs

Authors:
Eric Coyle;Emmanuel G. Collins, Jr.;Edmond DuPont;Dan Ding;Hongwu Wang;Rory A. Cooper;Garrett Grindle
Affiliations:
Center for Intelligent Systems, Control and Robotics, Tallahassee, FL;Center for Intelligent Systems, Control and Robotics, Tallahassee, FL;Center for Intelligent Systems, Control and Robotics, Tallahassee, FL;Univ. of Pittsburgh, Pittsburgh, PA;Univ. of Pittsburgh, Pittsburgh, PA;Univ. of Pittsburgh, Pittsburgh, PA;Univ. of Pittsburgh, Pittsburgh, PA
Venue:
Telehealth/AT '08 Proceedings of the IASTED International Conference on Telehealth/Assistive Technologies
Year:
2008

Citing 3
Cited 2

Probabilistic neural networks

Neural Networks
Frequency response method for terrain classification in autonomous ground vehicles

Autonomous Robots
Vibration-based terrain classification for planetary exploration rovers

IEEE Transactions on Robotics

Adaptive Bayesian filtering for vibration-based terrain classification

ICRA'09 Proceedings of the 2009 IEEE international conference on Robotics and Automation
Life-logging of wheelchair driving on web maps for visualizing potential accidents and incidents

PRICAI'12 Proceedings of the 12th Pacific Rim international conference on Trends in Artificial Intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

Automated terrain classification for electric powered wheelchairs (EPWs) has two primary motivations. First, certain terrains (e.g., sand and gravel) make wheelchair mobility more difficult. To alleviate this problem the wheelchair control system can be manually tuned for maximum speeds and/or accelerations to help adapt to various terrains. Terrain classification can then be used to automate the switch from one control mode to another. Second, terrain classification can help yield a better understanding of the surfaces traversed by various groups of wheelchair users. This can provide the data needed to develop wheelchairs geared to specific groups of users. This paper presents an algorithm for vibration-based terrain classification on EPWs. This algorithm has been shown to be highly accurate in offline analysis of experimental data. Future work will stress online implementation and algorithm improvements.