Sensor Deployment under Probabilistic Sensing Model

In recent years, with the rapid development of wireless sensor networks (WSNs), sensors are widely used to monitor a region of interests (ROI). Therefore sensor deployment becomes one of the important issues that need to be solved because it determines the cost of constructing the WSN and affects how well the ROI is monitored. Sensors can be deployed in a pre-planned manner or in an ad-hoc manner. Moreover, the sensing model of a WSN can be binary disk model or probabilistic sensing model. Most of previous researches focus on binary disk model, which assumes that sensors can accurately detect targets within their sensing ranges. Recently, probabilistic sensing model has been proposed; in this model, the probability for a sensor to detect a target decays with the distance between the sensor and the target. Probabilistic sensing model is therefore more realistic than the binary disk model. In [2], the sensor deployment problem in a pre-planned manner under probabilistic sensing model is solved by using a transformation from probabilistic sensing model to binary disk model. We find that such a transformation wastes too many sensors, and it can be avoided. In this paper, we solve the same problem (i.e., pre-planned deployment under probabilistic sensing model) by using a "direct" method. We show that even under a simplified probabilistic sensing model, there are examples such that more than 86% of sensors used in [2] can be saved.