NASA SBIR 2006 Solicitation

FORM B - PROPOSAL SUMMARY


PROPOSAL NUMBER:06 X2.03-9178
SUBTOPIC TITLE:Software Engineering Technologies for Human-Rated Spacecraft
PROPOSAL TITLE:Analytical Methods for Verification and Validation of Adaptive Systems in Safety-Critical Aerospace Applications.

SMALL BUSINESS CONCERN (Firm Name, Mail Address, City/State/Zip, Phone)
Contek Research, Inc.
615 Nash Street, Suite 220
El Segundo, CA 90245-2827
(310) 414-6720

PRINCIPAL INVESTIGATOR/PROJECT MANAGER (Name, E-mail, Mail Address, City/State/Zip, Phone)
Fola    Soares
fola@contekresearch.com
P. O. Box 88758
Los Angeles, CA  90245-2827
(310) 414-6720

TECHNICAL ABSTRACT ( Limit 2000 characters, approximately 200 words)
A major challenge of the use of adaptive systems in safety-critical applications is the software life-cycle: requirement engineering through verification and validation. Adaptive systems incorporate learning to adapt the control system to the current operating conditions of the system, certifying their performance is a complex and tedious process. Ongoing effort in the development of tools for verification and validation of adaptive control systems, there is little research directed at the development of analytical methods. Learning rules for adaptive systems derivation using Lyapunov's second method, is based on the derivation of an energy-type function whose derivative must be negative to guarantee convergence therefore the asymptotic stability of the system. The first problem is that Lyapunov's second method provides a sufficient condition for stability thus the synthesis of an appropriate Lyapunov function for a particular application is a major challenge. The second problem in many applications, including the design of adaptive neural flight control systems, it is only possible to prove that the derivative of the Lyapunov function is non-positive, rather than being negative. For practical purpose, it is only possible to conclude that the control system errors are ultimately bounded, and it not possible to estimate the magnitude of these errors or the time it takes for these errors to converge to their steady-state limits. The objective of this research project is to develop analytical methods for the analysis of adaptive neural networks (ANN) based flight control systems including analytical estimates of the settling time and the steady-state magnitudes of the error dynamics. The magnitudes of the error bounds will be related to the performance handling qualities of the system and provide very important information about the performance of the closed-loop system.

POTENTIAL NASA COMMERCIAL APPLICATIONS ( Limit 1500 characters, approximately 150 words)
The proposed technology is designed to be compatible with current safety requirements for conducting aircraft systems ground and flight tests in order to take advantage of the state of the technology in expert system. A goal of NASA is to establish new levels of autonomy and robustness within aerospace vehicles. NASA believes these higher levels can be realized through the incorporation of intelligent systems. If these envisioned levels of autonomy could be achieved, significant aerospace business and research opportunities will be enabled. As example, components of the Intelligent Vehicle can include:
 Intelligent Mission Manager
 Intelligent Vehicle Manager
 Intelligent Health Manager
 Intelligent Flight Control System
 Intelligent Propulsion System
At NASA Dryden and other centers, there is need for quantitative performance measure of adaptive neural networks based control system for aeronautics and aerospace applications.

POTENTIAL NON-NASA COMMERCIAL APPLICATIONS ( Limit 1500 characters, approximately 150 words)
This research has significant importance for the use of adaptive NN in safety critical application in aerospace industry, medical fields, power industry and other industries where full knowledge is unable to predict system failures, failure scenarios and the magnitude or criticality of the failure. This research provides the foundation, that when fully understood, the commercial application of adaptive NN will be a more robust application technology for the future. We envisage further application in the aviation control towers of major airports, process industry (chemicals, food, etc.).

NASA's technology taxonomy has been developed by the SBIR-STTR program to disseminate awareness of proposed and awarded R/R&D in the agency. It is a listing of over 100 technologies, sorted into broad categories, of interest to NASA.

TECHNOLOGY TAXONOMY MAPPING
Airport Infrastructure and Safety
Guidance, Navigation, and Control
Intelligence
Operations Concepts and Requirements
Pilot Support Systems
Simulation Modeling Environment
Testing Facilities
Testing Requirements and Architectures


Form Printed on 09-08-06 18:19