1 A Near Optimal k-Coverage Algorithm For Large-Scale Sensor Networks Majid Bagheri, Mohamed Hefeeda, and Hossein Ahmadi School of Computing Science Simon Fraser University Surrey, Canada {mbagheri,mhefeeda,hahmadi}@cs.sfu.ca Technical Report: TR 2006-10 Abstract In sensor networks, k-coverage means that each point in the surveillance area must be within the sensing range of at least k sensors. Finding the minimum number of sensors to achieve k-coverage is NP- hard [1], [2]. We propose an efficient approximation algorithm for k-coverage problem which achieves a solution of size within a logarithmic factor of the optimal solution. We prove that our algorithm is correct and analyze its complexity. Furthermore, we show how our algorithm can: (i) cover arbitrarily- shaped areas, (ii) provide different degrees of coverage at different spots, and (iii) tolerate deployment inaccuracies. Most of these scenarios can not be handled by previous algorithms. We also compare our implemented algorithm against others in the literature. Our experimental results indicate that the algorithm is very scalable: Coverage for an area of size 100km 2 with sensors of range 15m is obtained within a minute on a commodity PC. Our results also show that the logarithmic factor is only a worst case upper bound and the solution is within a constant factor of the optimal in most cases. I. INTRODUCTION Mass production of sensors with low cost enables the deployment of large-scale sensor networks for real-life applications such as forest fire detection and vehicle traffic monitoring. A fundamental issue in This work is partially supported by a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) and a Simon Fraser University President’s Research Grant.