Study On Forest Fire Prediction And Key Algorithm Based On Wireless Sensor Network

Recently, the annual frequency of forest fires, a lot of devouring the life of the forest, and further deterioration of the ecological environment, a huge loss to the production of human life, to become one of the four major natural disasters. Therefore, the use of advanced scientific and technological means to predict the forest fire, reduce the incidence of forest fires, to further reduce or avoid the enormous loss of forest fires for mankind, is of great significance. Accompanied by the "Digital Earth", the launching of digital forestry are deepening, extensive forest fire prediction techniques from prediction methods gradually shift to the technology information. This paper will integrate wireless sensor network technology, control technology, computer simulation technology, wireless networking technology, and forest fire prediction methods, the establishment of forest fire prediction method based on wireless sensor networks to increase the protection, so as to enhance the level of forest fire prediction, the intensity of forest ecosystems, a new exploration and in-depth study on the theories and methods.Combine the characteristics of the forest fire prediction, according to meteorological data in front of the forest fire in Yichun in recent decades, a detailed analysis of the contribution of a single meteorological factors on forest fire occurred, the use of Logistic regression calculated for each meteorological factors affect the probability to determine the sensor factors of forest fire prediction, based on wireless sensor networks using multiple linear regression equation to establish an integrated regional meteorological index of forest fire insurance prediction methods. This method is based on a large number of meteorological data in the prediction regional forest fire before the derivation of multiple linear equations has a good geographical effectiveness and application value.Using ZigBee wireless networking technology, combined with GPRS wireless transmission mechanism to build a wireless sensor network for forest fire prediction, sensor nodes using the current energy consumption of small MSP430F5438MCU and DS18B20temperature sensor chip, the HR202humidity sensor chips combination of temperature and humidity sensing range to meet the projected demand of forest fires, efficient MCU with fast data processing capabilities, for the prediction of forest fire protection for wireless sensor networks.In order to accurately predict the location of forest fires occurred in the center scan positioning algorithm based on the improvement, the improved algorithm reduces the false positive rate of InToOut and OutToIn of neighbor nodes around the unknown node to solve the unknown node around the beacon node sparse problems and improve the coverage of the node, to reduce the error rate of node localization. Simulation results show that the unknown node location error rate gradually decreased with the increase in the density of beacon nodes; compared with the original center scan localization algorithm, the localization error rate of the improved center scan localization algorithm Tell by about40%in a sparse network, greatly enhance the localization accuracy of the network, and greater flexibility for the positioning of randomly distributed wireless sensor nodes, and more advantage than other localization algorithms.To extend for forest fire prediction the life cycle of wireless sensor network, this paper proposed a energy clustering algorithm based on ant colony(ECBAC), the algorithm in the circumstances to ensure that the network energy consumption of the smallest sub-cluster number use of the energy consumption of cluster head nodes in a balanced network of the main cluster heads and vice cluster head, integrated energy, distance and division of factors to select the master, the vice cluster head, to determine the cluster radius, the network is reasonable, energy consumption equilibrium points cluster division. In the main cluster head to the sink node during data transmission, combined with ant colony algorithm for data transmission using the shortest path distance between the node residual energy and the cluster head. The entire algorithm in the design process, and always to the energy consumption is the premise on the basis of the complexity of the algorithm does not increase greatly reduce network energy consumption, extend the network life cycle. Simulation results show that the algorithm uses the ECBAC two cluster head node cluster affairs division of labor, both to increase the continuity of the network, when compared with other algorithms to reduce the energy consumption of cluster head, the network’s total energy consumption compared to the LEACH algorithm decreased15%good to extend the life cycle of the network, wireless sensor network, energy saving algorithm has practical value.Forest fire prediction, localization algorithm and energy saving algorithm based on wireless sensor networks for the current frequent fires can play the role of prediction in advance, to prevent fires, and promote China’s forest fire prediction scientific and intelligent and Information Technology, for the protection of forest ecosystems to provide new ideas.