Research in Neural Network Based Adaptive Control

Abstract

The most significant theoretical accomplishment has been the development of a new approach for dealing with control limits and nonlinearities in adaptive systems. This approach both prevents the Maptire system from doing harm to an otherwise stable system, and also allows adaptation to continue while the control is saturated. We regard this as a major step towards flight certification of adaptive controllers. The approach is more general in that it permits a broad class of input nonlinearities, including such effects as discrete and bang/bang control. In the area of output feedback, we continue to refine our curlier work, and have begun to take steps in the direction of decentralized adaptive systems in a state feedback setting. Our most significant interactions have been with NASA Marshall and NASA Ames. In particular, we arc fully exploiting our research in limited authority adaptive control in the areas of autopilot design for launch vehicles, and propulsion control for commercial aircraft subject to partial or total loss of conventional flight control.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2000
Accession Number
ADA390892

Entities

People

  • Anthony J. Calise

Organizations

  • Georgia Tech

Tags

Communities of Interest

  • Energy and Power Technologies
  • Space
  • Weapons Technologies

DTIC Thesaurus Topics

  • Adaptive Systems
  • Aerospace Industry
  • Aircraft Industry
  • Aircrafts
  • Airframes
  • Automatic Pilots
  • Commercial Aircraft
  • Control Systems
  • Dynamic Pressure
  • Feedback
  • Flight Control Systems
  • Launch Vehicles
  • Neural Networks
  • Tilt Rotor Aircraft
  • Turboshaft Engines
  • Vehicles
  • Wind Tunnels

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Aerospace Engineering.
  • Robotics and Automation.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks