NASA SBIR 2004 Solicitation


PROPOSAL NUMBER: 04 A1.04-8006
SUBTOPIC TITLE: Automated On-Line Health Management and Data Analysis
PROPOSAL TITLE: Self-Repairing Flight Control System for Online Health Monitoring and Recovery

SMALL BUSINESS CONCERN (Name, E-mail, Mail Address, City/State/Zip, Phone)
SenAnTech, Inc.
5444 Silver Creek
Columbus, OH 43228-9062

PRINCIPAL INVESTIGATOR/PROJECT MANAGER (Name, E-mail, Mail Address, City/State/Zip, Phone)
Seung-Keon Kwak
5444 Silver Creek
Columbus, OH 43228-9062

In this SBIR project, a reliable self-repairing Flight Control System (FCS) will be developed. To achieve this goal, an artificial Neural Network based Sensor Validity Monitoring, Verification and Accommodation (SVMVA) scheme will be developed. Initially, the Extended Back Propagation Algorithm (EBPA) or Learning Vector Quantization (LVQ) will be evaluated and employed as the on-line real time learning, monitoring and estimating tools. For a feasibility study, sample flight vehicle dynamics will be assessed. Then, an optimal on-line estimator for the flight dynamics will be shown by applying our unique input vector discrimination procedure and network optimization technique. In Phase I, an in-house numerical simulation tool for SVMVA will be developed to demonstrate the feasibility of our approach. As a result, it is expected that the new self-repairing Flight Control System based on SVMVA scheme will meet or exceed the NASA's requirement for their new flight vehicles. The most significant advantage of the new self-repairing Flight Control System is that the system is able to mask failed physical sensors by providing flight-worthy flight dynamics information to most FCS of flight vehicles without needing redundant sensors or excessive power, weight and space.

The SVMVA product will be very attractive to NASA's space exploration missions. By providing reliable backup information of space vehicle dynamics without additional physically redundant sensors in case of failure of the primary or secondary sensor systems, it will considerably reduce the weight, space, and power requirement of vehicles and reduce the vulnerability of vehicle operations.

The new reconfigurable flight control system based on Sensor Validity Monitoring, Verification and Identification (SVMVA) scheme has great potential for both commercial and military markets. The applications of the proposed technology and related products are not only limited in the aerospace industries but also extended to various military and civil fault detection and health monitoring markets. Aircraft engine industry is also potential markets. Real-time health monitoring of expensive and mission critical rotating component can save money and life. Consequently, its application includes virtually anything that requires detection of sensor and actuator conditions in broadband frequencies.