A Comparison of Classifier Performance for Vibration-Based Terrain Classification

Abstract

The ability to recognize the encountered terrain is an essential part of any terrain-dependent control system designed for mobile robots. Terrains such as sand and gravel make vehicle mobility more difficult and thus reduce vehicle performance. To alleviate this problem the vehicle control system can be tuned for maximum speeds, turning angles, accelerations and other conditions to help adapt to various terrains. Terrain classification can be used to automate the switch from one control mode to another. This paper compares the performance of several classifiers on the problem of vibration-based terrain classification. The purpose of this comparison is to assess the strengths and weaknesses of these techniques in order to better understand the tools available in developing future vibration-based terrain classification algorithms.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 2008
Accession Number
ADA505696

Entities

People

  • Emmanuel G. Collins Jr.
  • Eric Coyle

Organizations

  • Florida State University

Tags

Communities of Interest

  • Autonomy
  • Energy and Power Technologies
  • Ground and Sea Platforms
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Artificial Intelligence
  • Classification
  • Control Systems
  • Dimensionality Reduction
  • Engineering
  • Frequency
  • Inertial Measurement Units
  • Kernel Functions
  • Machine Learning
  • Maximum Likelihood Estimation
  • Neural Networks
  • Pattern Recognition
  • Probability
  • Probability Distributions
  • Supervised Machine Learning
  • Vibration

Fields of Study

  • Engineering

Readers

  • Computer Vision.
  • Robotics and Automation.
  • Traumatic Brain Injury (TBI) and Cognitive Aging in the Guam and Border Populations Affected by Alzheimer's Disease and Tau-Associated Dementias.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy