Accession Number : ADA318335

Title :   Achieving Near-Optimal Sensor Allocation Policies Through Reinforcement Learning.

Descriptive Note : Final rept.,

Corporate Author : WRIGHT LAB WRIGHT-PATTERSON AFB OH

Personal Author(s) : Malhotra, P.

PDF Url : ADA318335

Report Date : OCT 1996

Pagination or Media Count : 11

Abstract : TACTICAL AIRCRAFT MUST FREQUENTLY PERFORM COMPLEX SEQUENTIAL TASKS IN WHICH THEY RELY HEAVILY ON THE INTEGRATION OF SENSORY DATA TO ASSESS STATE AND MAINTAIN SITUATIONAL AWARENESS. IN MODERN SYSTEMS, THE CONTROL OF THE SENSORS' INFORMATION-GATHERING ACTIVITIES IS CRITICAL-OPTIMAL PERFORMANCE IS DESIRED. BUT THIS IS MADE DIFFICULT BY THE REQUIREMENTS TO CONTEND WITH SOPHISTICATED FLEXIBLE SENSORY ASSETS, AND VOLATILE, UNCERTAIN ENVIRONMENTS. THIS PAPER INTRODUCES THE SENSOR MANAGEMENT PROBLEM AND THE PLAUSIBILITY OF LEVERAGING A MACHINE LEARNING ALGORITHM TOWARD THIS DIFFICULT CHALLENGE.

Descriptors :   *AVIONICS, *RADAR TRACKING, *IDENTIFICATION SYSTEMS, *DATA FUSION, ALGORITHMS, OPTICAL RADAR, AIR DEFENSE, OPTIMIZATION, DATA MANAGEMENT, AIR FORCE RESEARCH, FLIGHT CONTROL SYSTEMS, LEARNING MACHINES, SYNTHETIC APERTURE RADAR, AIR STRIKES, AERIAL TARGETS, VISUAL PERCEPTION, DYNAMIC PROGRAMMING, FLIGHT SIMULATION, SEARCH RADAR, MARKOV PROCESSES, TACTICAL AIRCRAFT, INFRARED SCANNING, RADAR RANGE COMPUTERS.

Subject Categories : Flight Control and Instrumentation
      Active & Passive Radar Detection & Equipment
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE