Neural Network Sensor Fusion for Security Applications

Abstract

Neural networks can detect patterns in information that they receive and have been used successfully in many classification problems. The research discussed in this paper was pursued to determine if neural networks could be successfully used for alarm classification for security systems operating in unstructured environments. The project investigated the usefulness of various types of neural networks for this application. The results from the research exceeded the original goals, promising sound application to security systems in unstructured environments.