A QoS based Heuristics for Clustering in Two-Tier Sensor Networks

Once sensors detect an event they always have to route the data to base station where the data is processed. Since sensors usually have concerns regarding coverage, energy, processing power and memory, etc achieving Quality of Service is hard in sensor networks. Therefore to deal with such issues of sensors and to maximize the Quality of Service, initially the two tiers Heterogeneous Sensor Networks approach is used to route the data. Second, the sensors are partitioned into clusters to increase the network coverage and to reduce transportation costs and energy utilization. Voronoi clustering and Tabu search meta-heuristics have been used for making such clusters. An Improved Tree Routing technique is applied to two-tier Heterogeneous Sensor Networks to route the data through cluster heads. This approach largely increases the performance of sensor networks. Through simulation results, we show that the Voronoi-Tabu based clustering technique when added to Improved Tree Routing has better Quality of Service than Directed Diffusion and Low Energy Adaptive Clustering Hierarchy routing protocols. Furthermore, empirical evaluations show that Voronoi-Tabu based clustering increases the throughput of the network, in addition to decreasing the energy utilization and network delays.