U.S. flag

An official website of the United States government, Department of Justice.

Development of an Automated Assessment Signal Processor (ASP) for Perimeter Intrusion Detection Systems (From Carnahan Conference on Security Technology, 1989: Crime Countermeasures, P 43-47, 1989, R. William De Vore, ed. -- See NCJ-117867)

NCJ Number
A R Hunt
Date Published
5 pages
General Research Corporation has developed and field tested a new product for automatically assessing the alarms in perimeter intrusion detection systems.
The hardware uses signal processing based on machine intelligence to learn the characteristics of sensor signals and to separate nuisance signals from intruders. The system uses signals generated by sensors commonly used in intrusion detection systems. It operates with one to four sensors for each zone. Signal features are extracted from the waveforms of each sensor and presented to a fusion and signal classification network based on machine intelligence. The network decides in real time whether the source of the disturbance is a nuisance, an intruder, or something it has not seen before. It saves all alarm data so that new classification algorithms can be produced to improve the performance of the system. The system's advantages include its ability to learn from its environment to adapt to nuisances specific to the site, its ability to correlate the inputs of multiple sensors installed at existing sites, and its ability to recognize when environmental or sensor conditions are changing. Thus, it reduces false alarms and opens a new dimension in the operation and economics of security systems. Figures, table, and 7 references. (Author abstract modified)