Automatic Target Recognition Using Nonlinear Autoregressive Neural Networks

Abstract

Accurate combat identification is critical to military interactions. Laser radar for vehicle identification is a rapidly developing field that could possibly assist in combat identification by providing information about operating characteristics of a particular vehicle based on measured vibrations. This research focuses on simulated laser radar data collected from mounted vibrometers on idling vehicles. An approach to identify vehicles using nonlinear autoregressive neural networks for classification is developed and employed. The resulting algorithm combines the trained neural networks across three dimensions of vibration readings. This method offers improved performance over literature in successfully identifying a vehicle through vibration measurements alone.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 27, 2014
Accession Number
ADA610293

Entities

People

  • Marc R. Ward

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Air Force
  • Algorithms
  • Artificial Intelligence Software
  • Classification
  • Computer Programs
  • Computers
  • Detection
  • Detectors
  • Dimensionality Reduction
  • Information Processing
  • Information Science
  • Laser Radar
  • Machine Learning
  • Neural Networks
  • Spreadsheet Software
  • Target Recognition
  • Test And Evaluation

Fields of Study

  • Engineering

Readers

  • Computer Vision.
  • Instructional Design and Training Evaluation.
  • Unmanned Aerial System (UAS) Autonomous Capabilities and Mission Reconnaissance.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - Neural Networks
  • Directed Energy