Minimizing the operational cost of chemical sensor networks

Minimizing the sensing activity during the pre-attack phase while meeting the performance requirements is critical for reducing the operational costs of chemical sensor networks. In this paper, we present a formal problem formulation for this optimization problem, and propose a decentralized, dynamic sensor activation protocol for meeting this goal. We analyze the behavior of the activation protocol using an epidemiology-based approach and derive analytical expressions for the system performance metrics, namely, operational cost, detection delay, probability of false alarms and probability of detection. We analyze the interdependency between the performance metrics.