Multi-Agent Surveillance and Threat Evaluation for Indoor Environment

Abstract

This paper discusses a hierarchical multi-agent architecture for implementing semi-automated surveillance in the indoor environment. Specifically, a scientific lab environment is considered for demonstration purposes. A three-layered multiagent architecture of the surveillance system is presented with local intelligence, global intelligence, and human supervision. The internal architectures, models, and algorithms for local and global intelligence agents are presented and discussed in the context of surveillance of a scientific laboratory. Possible issues in practical implementation, scaling, and advancement of the proposed approach are also discussed in the paper.