Abstract
Relayed communications are typically used in body sensor networks (BSNs) to improve reliability and energy efficiency. Particularly, a biosensor communicates with the sink device through relay(s) to conserve its limited battery power. When the relay malfunctions, the biosensor may increase its transmission power to enable a direct communication with the sink. However, such a direct communication may last only a limited, uncertain time; the biosensor becomes isolated to the rest of the BSN when the sensor's remaining power depletes to the level insufficient to support the direct communication. This paper makes contributions by addressing realistic random isolation time (RIT) in reliability analysis of BSN systems. Moreover, competitions between the isolation and propagated failures originating from the biosensor are considered. Based on the total probability theorem, a combinatorial methodology is suggested for reliability analysis of BSNs subject to the competitions and RIT. The method is applicable to any types of time-to-failure and time-to-isolation distributions of biosensors. As demonstrated through a detailed case study with numerical results, RIT can significantly influence the BSN reliability and thus it is pivotal to consider this factor for accurate system reliability assessment. Correctness of the proposed method is verified using a continuous-time Markov chain-based method.