Research on multi-agent based remote distributed video monitoring

In order to overcome the disadvantages of lacking flexibility, robustness and scalability of the traditional remote video monitoring system (RVMS) for substations, we propose a model of distributed RVMS based on multi-agent techniques. In this model, complicated control of traditional monitoring system can be simplified and decentralized based on the features of independence and cooperativeness of agent work groups. The architecture and operation principle of the model is analyzed in detail and the design and implementation of universal agent modules are also presented. Furthermore, two algorithms based on the new H.264/AVC coding standard, namely, the controllable predictive fast search algorithm for block matching motion estimation and the motion vector recovery algorithm are adopted respectively in two important agent work groups, DCG and ECG. Simulation results show that real-time video transmission via error prone channels can be realized by cooperative communications of agent groups, the strategies of video compressing and communications can be dynamically changed according to correspond situation. Therefore, efficient video monitoring of substations can be achieved by using this model.