Sensor Failure Detection through Introspection

Abstract

The advancement of robot technology holds many opportunities for military applications. One area of research being done is simultaneous localization and mapping (SLAM). SLAM uses a robot's sensors to generate a map of the area while maintaining its current position within that map. SLAM research is built upon the assumption that all of the sensors are working correctly. Since field conditions are likely to cause erratic sensor function due to damage or inclement weather conditions, this assumption must be addressed. The goal of our research is to discover methods of effectively performing self-diagnostic checks on robots to detect failures and malfunctions in sensors. There has been little work in the area of error detection in sensors, and what little work has been done has limited applications. This thesis will perform a series of experiments using a variety of different error detection techniques. It is our hope that the methods developed will prove to be applicable to a variety of real world systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2007
Accession Number
ADA473472

Entities

People

  • Andrew Valerius
  • Jeremy Smeltz

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Autonomy
  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Autonomous Navigation
  • Cartography
  • Computer Science
  • Damage Detection
  • Detection
  • Detectors
  • Machine Learning
  • Maps
  • Probability
  • Robot Mapping
  • Robotics
  • Sequential Monte Carlo Methods
  • Simultaneous Localization And Mapping
  • Sonar Transducers
  • Test And Evaluation
  • Two Dimensional

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Computer Vision.
  • Systems Analysis and Design

Technology Areas

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