Energy Efficient Cooperative Multimodal Ambient Monitoring

Wireless Video Sensor Networks (WVNs) are lively interest in the research community as flexible means for monitoring isolated areas. WVN effectiveness can be augmented when coupled with a network of low-power, low-cost Pyroelectric InfraRed (PIR) detectors to form a multimodal surveillance system. Autonomy is a key issue as battery replacement is often impractical and even using energy scavenging techniques, such as solar harvester, ad-hoc power management policies are essential to extend network lifetime. In this paper we propose a cooperative policy to manage power consumption of a WVN powered by solar scavengers and supported by a network of PIR sensors that perform a coarse classification of movements. A cost function is calculated by each VSN according to its available energy and information from the PIR network. Such functions are used by a distributed energy aware policy that selects the best VSN to observe the activity. This VSN locally analyzes the image and detects whether or not it represents a person, and only this information is forwarded to users. The effectiveness of this technique is evaluated through simulation and compared to an approach presented in previous work. Results show an increase in system lifetime without any loss in people detection ratio.