Probabilistic analysis of a calculus for wireless sensor networks

The CWQ Calculus (a Calculus for Wireless sensor networks from Quality perspective) was recently proposed for modeling and reasoning about Wireless Sensor Networks (WSNs). It has the flexibility that not all input data in a binder need to be received in order for the process to continue. Meanwhile, it has the unique and important characteristic that, in order for the decision of a system of a WSN to be of high trustworthiness, the decision is expected to be made by considering all data from all network nodes in the WSN. Consequently, decisions of a system may have different trustworthiness depending on which input data have actually been received. In this paper, we propose a data-driven probabilistic trust analysis of the CWQ Calculus for WSNs. We assume that data received from a channel have trust values that follow a probability distribution; that is, the trust value of a data represents the trust of the decision of a system made solely based on that data. Thus, we decouple the probability of receiving input data from the probability of data trustworthiness. The overall trustworthiness of the decision of a system is determined by performing a relational analysis to combine these probability distributions.