The Research And Application Of Energy-efficiency In WSN For Debris Flow Warning

Due to the complexity and harsh environments, and in wide-area monitoring landslides occurred in this area has always been the lack of suitable methods. The wireless sensor network has low power, low cost, fast response characteristics, can be applied to just the right among the debris flow disaster monitoring, which is to protect the safety and wealth under the slope of the residents are very necessary. But there is an important issue in today’s wireless sensor network-node energy consumption problem. Since the nodes are battery powered, the various types of sensor node integrated energy is large, so if the node is in continuous operation so that the state of the case, the energy of the battery will be exhausted in a few days. This is contrary to the very far nodes need several months of work in the field of requirements.Currently, wireless sensor networks for energy conservation research program is divided into three main methods:a topology control protocol.2. The routing control protocol.3. Sleep control management. The main focus of previous scholars As topology control protocols and routing control protocol. The third method for the study of control and management of less dormant. However, in the application of wireless sensor networks is now in energy consumption compared to the energy consumption of the sensor nodes communicate almost the same, even more need for sleep control management node-depth research.In this article, we propose a wavelet neural network based on nodes sleep conduct scientific management control program. Specifically, it has two advantages:1) when the node is in sleep mode in which the energy consumption of nodes is negligible. So let node intelligent system switch back and forth between the operating mode and sleep mode, as the cycle of work; 2) node sampling rate will vary according to changes in the environment. Combination of these two advantages can make the nodes at the time of the frequent need to work to work in a timely manner when needed saving hibernation, greatly extending the life of wireless sensor networks and did not affect the reliability of the network. Due to the special nature of the outbreak when the debris flow monitoring environment, we did not find a suitable low-power wake to wake up the sleeping sensor nodes, we propose the use of wavelet neural network to predict the next point in time to predict the performance of each sensor data possibility to adjust the sampling rate and thus the node based on timely forecast data, to delay the network lifetime without cuts destination network reliability. As we all know, even after repeated training the neural network learning will exist error, not a very accurate prediction of the data we want, so in this paper we use the convergence time and precision are more excellent convergence particle swarm optimization. But in the context of multi-network forecasting the impact of time, as much as possible to improve the prediction accuracy of wavelet neural network. Through the above solution, we can make life a node extended to a year and a half ago, but also to make the entire network efficiently and reliably.