A New Mixed Strategy to Cover Continuous Region

Wireless sensor networks (WSNs) is now a developing area. The coverage problem is an important part of that. WSNs contain sensors which cover a three or two-dimensional bounded convex set. In general, the sensors are deployed randomly from the air. Sometimes sensors are placed by robot(s) or sensor can move to placed themselves on some points of a pre-assigned region, known as Region of Interest (ROI). Since there is a randomness of placing the sensors, they may not be placed on the pre-assigned points. Hence, ROI may not completely cover by the sensors. So, the problem is, how one decrease the uncovered area? Most of the situations, extra sensors are deployed on some randomly chosen points to cover the ROI. In one of our previous work, we reduced the distance between two adjacent vertices and drop exactly one sensor on each vertex. We compare the uncovered area using two strategies (the general one and our previous one), for normal and uniform distributions, and for a different percentages of extra sensors. We observed that our strategy is better for low variance-distribution, but not for high variance-distribution. So, we combined the two strategies in this paper.We partitioned the ROI in regular hexagons and develop a new mixed strategy, for dropping the sensors. The extra sensors are divides in two groups. One group is used for reducing the distance between the two adjacent vertices and other parts is used for deploying an extra sensor on randomly selected points. We simulate with a non parametric distribution also along with above two. Simulation shows the optimal trade-off between the percentage of these two groups, which change with the variance of randomness.