CrunchBot: a mobile whiskered robot platform

Authors:
Charles W. Fox;Mathew H. Evans;Nathan F. Lepora;Martin Pearson;Andy Ham;Tony J. Prescott
Affiliations:
Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK;Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK;Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK;Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK;Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK;Active Touch Laboratory at Sheffield, University of Sheffield, Sheffield, UK
Venue:
TAROS'11 Proceedings of the 12th Annual conference on Towards autonomous robotic systems
Year:
2011

Citing 4
Cited 2

Biomimetic whiskers for shape recognition

Robotics and Autonomous Systems
Contact type dependency of texture classification in a whiskered mobile robot

Autonomous Robots
Mapping with sparse local sensors and strong hierarchical priors

TAROS'11 Proceedings of the 12th Annual conference on Towards autonomous robotic systems
Whisker-based texture discrimination on a mobile robot

ECAL'05 Proceedings of the 8th European conference on Advances in Artificial Life

Towards hierarchical blackboard mapping on a whiskered robot

Robotics and Autonomous Systems
Efficient coding in the whisker system: biomimetic pre-processing for robots?

Living Machines'13 Proceedings of the Second international conference on Biomimetic and Biohybrid Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

CrunchBot is a robot platform for developing models of tactile perception and navigation. We present the architecture of Crunch-Bot, and show why tactile navigation is difficult. We give novel real-time performance results from components of a tactile navigation system and a description of how they may be integrated at a systems level. Components include floor surface classification, radial distance estimation and navigation. We show how tactile-only navigation differs fundamentally from navigation tasks using vision or laser sensors, in that the assumptions about the data preclude standard algorithms (such as extended Kalman Filters) and require brute-force methods.