Towards Game-Theory for Tracking (ToGT4T)

Abstract

This is the final report for the Towards Game Theory for Tracking project, funded by AFOSR via grant number FA8655-22-1-7021. A particle filter was used to estimate states of a moving target. The assumed behaviour of the target included the capacity to attempt to evade being detected by the sensor. The experimental results show some ability to identify evasive behaviour. In situations involving pronounced obscuration, the are too few detections to make meaningful inferences. Recommended future work includes consideration of a more sophisticated particle filter, Monte-Carlo roll-out and reinforcement learning.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 12, 2023
Accession Number
AD1226695

Entities

People

  • Simon Maskell
  • Yifan Zhou

Organizations

  • University of Liverpool

Tags

Readers

  • Game Theory.
  • Robotics and Automation.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference