This project addresses the development of a low cost approach for intrusion detection. The system concept is to create a packet radio network that would connect a group of sensor packages. Each sensor package will contain a set of physical sensors. Multisensor data fusion will be done by a microprocessor within each sensor package using data from the sensors within the package. Using radio communications, the sensor packages in a detection zone will perform zone-wide sensor data fusion which will provide the intrusion alarm. The project directly addresses the use of adaptive fuzzy systems for multisensor data fusion. Further, the selection and use of sensors that can provide means of automatic initialization and self-check are sought. Costs of system acquisition, initialization, and maintenance will be drastically reduced by the use of multiple cooperating sensor packages, deployment scenarios, and machine intelligent learning. A packet radio network with an artificially intelligent router will be part of the system. Spread spectrum radios are used for low probability-of-intercept operations. The possibility of using the spread spectrum radios as a bistatic radar sensor will be explored. The primary objective of Phase I is to demonstrate the feasibility of a systems approach to the selection of sensors, multisensor data fusion, and learning feedback scenarios.Anticipated Results /Potential Commercial Applications as described by the awardee:This project will result in a product that can provide low cost intrusion detection aids to guarded facilities, a subsystem component to augment security systems with more expensive equipment and site preparation, and a rapidly deployable and easily relocatable intrusion detection system.
