Complex situation modeling in distributed sensor networks

Surveillance typically involves monitoring humans, buildings, and other mobile objects to detect abnormal behavior; primarily to sense and detect any anomalies in real time. Conventionally this has been done manually, but with a growing demand for day-to-day surveillance and the need for intense monitoring; decision support has proven to improve the overall system performance. Decision support also called situation assessment from a surveillance perspective, takes form of a higher order pattern recognition problem involving complex reasoning and inference. A distributed approach to knowledge modeling and inference is proposed here for effective representation of the domain knowledge. In addition optimal local and global rules to combine situation level information have been developed.