A Physics-Based Software Framework for Detection, Location, Isolation and Masking of Faults in Composite Sensor-Computer Networks

Period of Performance: 08/05/2002 - 08/04/2004

$600K

Phase 2 SBIR

Recipient Firm

MPI Software Technology, Inc.
101 S. Lafayette #33
Starkville, MS 39759
Principal Investigator

Abstract

This SBIR Phase II Proposal seeks to develop the physics-based fault tolerant technology for sensor networks shown to be feasible in the SBIR Phase I study in response to topic N01- 080, "Fault Location in an Intelligent Open Sensor Network." The proposed approach considers that for Conditional-Based Maintenance purposes an entire ship can be classified as a hierarchy of subsystems, which conceptually leads to a hierarchy of sensor networks that govern/monitor an entire ship. This SBIR Phase II effort includes (1) fault-tolerant implementation of the IEEE 1451.2 Standard, (2) implementation of sensor networking technology using COTS-based technology, (3) implementation of the physics-based fault tolerance approach for a network of smart sensors, and (4) implementation of a hierarchical approach corresponding to the hierarchical abstraction of a physical system of interest (i.e., an entire ship). In addition, technology for interfacing the proposed software products with products that are already in use in naval vessels will be developed as part of this effort. Demonstration of the fault- tolerant technology to be developed with suitable sensor networks, and further refinement of the technology based on feedback will occur in stages during the Phase II effort and will continue, with full demonstrations, through the Phase II Option.BENEFITS: Continuous modernization of Naval ships as well as decreasing reliance on on-board military personnel calls for the building of autonomous systems. The use of fault-tolerance enabled computer-sensor networks would be valuable in achieving this objective, since sensor faults would be indicated separately from component failures. Additionally, an integrated computer-sensor network with fault-tolerance capabilities would offer end-users in defense, manufacturing, and processing industries the ability to remotely monitor and diagnose any sensor on a network. The proposed physics-based fault-tolerant technology would also be useful in monitoring and controlling various sub-systems in automobiles. Aerospace industry and NASA are other potential customers. Use of this technology for large-scale energy management in office buildings and other industrial settings is also a potential market.