A Research Of Sleep Scheduling Algorithm Based On Energy In WSN

Wireless sensor network(WSN) is a multi-hop self-organizing wireless network system formed by a variety of sensor nodes via wireless communication. WSN is a kind of wireless network without infrastructures that each node is equipped with very little energy, but it is impossible to charge the node in the practical application, therefore, energy has become the vital element that restricts its large-scale application. There are many sensor nodes deployed in WSN. If all sensor nodes work at the same time, WSN will produce a large number of redundant data, which not only waste energy, but also increase the difficulty of data fusion.Taking advantage of the sleep scheduling method, we let some nodes work and others stay the state of dormant. Working nodes are responsible for data acquisition. Then awake the rest nodes to replace the invalid nodes in order to ensure the normal running of the network when the network topology changes. Efficient sleep scheduling algorithm can prolong the network life cycle, it also can improve the efficiency of data collection. Therefore, designing an efficient sleep scheduling algorithm is the primary goal of the WSN.Considering that existing sleep scheduling algorithms generate working nodes with many overlapping area, a new sleep scheduling algorithm which finds the minimum working set(FMWS) of nodes is proposed. The algorithm adopts an iterative loop mode, which first uses classic cover algorithm to get the initial cover set, and then reduces the overlapping area between nodes by means of consultation, so that the coverage area of each node is maximized.In the first phase, this paper employs the CCP algorithm to get the initial working node set. In the second phase, the pre-consultation mechanism is introduced, which conducts the consultation between nodes by simulating the sleeping nodes as working nodes, and to find more working nodes which can be transformed to sleeping nodes and transformed them,thereby working nodes are reduced further. The location factor is included in the negotiation content, so that the overlapping area between the nodes is further reduced, which saves the network energy and extends the network life cycle. To balance the load of network energy and avoid state transitions conflict of nodes, a fallback mechanism is introduced, in which fallback time and the residual energy is related to ensure the more residual energy of nodes the greaterthe probability of them into an working state. In order to reduce the complexity of the CCP algorithm, a temporary state is introduced, which limits the involvement of the current node in redundant computation of other nodes, thereby cuts down the number of neighboring working nodes during node redundant computation, so that the complexity is reduced to O(n).In most cases of WSN applications, it has no requirement for the full nodes’ coverage,and allows some coverage blinds in the monitoring area to reduce network energy consumption. We proposed a distributed partial coverage sleep scheduling algorithm named DSSAP, using two rounds of iterative work mode. The first stage of it divided nodes’ sense area into various small squares and then used the ratio of numbers of nodes to replace the ratio of the areas of nodes. And set thread to ensure the redundant node. It helps to distribute the continuous node area and then it reduces the computation and improve the accurately. The second stage of it adopted the consulting method based on the working mechanism to reduce the redundant nodes. Considering that there may be "isolated" node in the network, GAF algorithm theory is introduced. Monitoring area is divided into a plurality of small areas, and all isolated sub-regions are searched by applying node negotiation method, finally, network connectivity is ensured by waking up some sleeping nodes in isolated area.Simulation is carried out in the NS-2 simulation platform, and we compared the performance of the proposed algorithm with other algorithms. The experimental results showed that the FMWS algorithm could keep away from blind spots successfully and has a better performance regarding to the numbers of working nodes and the average coverage.More initial nodes means less working nodes in the proposed algorithm. Its number of working nodes is about 10% less than the CCP algorithm. The number of working nodes generated by DSSAP algorithm is less than the full coverage algorithm. With the network coverage ratio decreasing, the number of working nodes also continuously reduced. In general,high coverage ratio means high quality service. Hence, it is necessary to set up a suitable coverage ratio according to the different requirements of the application in actual application.Experiments showed that the network could achieve a good balance when the coverage ratio was 90%.