Research On The Fusion Algorithm Of Farmland Sensor Data

Wireless Sensor Network (WSN), an interdisplinary and newly developed research area, is a trend of the past few years. It utilizes integrated micro sensor to synergistically perceive, collect and obtain relevant data in real time from targets. Microprocessor is used for data processing. Then WSN transmits to aggregation nodes through routing protocols of wireless Ad hoc network by using multi-hop trunking scheme, combines information acquisition and data fusion together, ultimately provide users with useful information. Using WSN to get real time and accurate field information has been a hot issue for modern agriculture. It emphasizes on reduce the net consumption and extend service life, since it has the following disadvantages like disperse distribution, high volume data, and single node with limited resource. Some research has showed that the consumption of calculation is much less than communication. Only fuse the data of adjacent nodes, much energy can be saved. This work was supported by the doctoral program of higher education (specialized research fund, No.20114320110001) and Hunan Province important and special project, namely’Key technology study and display of Camellia Oleifera breeding and ecological cultivation’, the fourth special’The construction of informationalized service platform in Camellia Oleifera industry’(2013FJ1006-4). The main research fields are as follows.This paper gives a classification and comparison among common data fusion algorithms, especially in routing protocol. It sketches out and analyses research actuality and expectation of fusion strategies and algorithms in WSN both domestic and overseas.This paper also puts forward the L-GB algorithm based on genetic algorithm and neural network. This data fusion model is suitable for the perception of crops system. It includes a further refinement of LEACH routing protocol algorithm, and combines with neural network hierarchical structure, then optimizes sensor cluster by genetic algorithm, fuses data which collected by sensor nodes, ultimately transmits the fusion result to connection nodes only. Simulation testing showed that this algorithm can effectively eliminate the redundancy of data, save the energy of sensor network and extend service life.This paper proposes simulation test by using the modified LEACH protocol algorithm through MATLAB. Firstly, it proves the effectiveness of modified algorithm with the comparation of original one. Secondly, it demonstrates L-GB algorithm has data fusion after doing a series of simulation test in different scales. Then it compares energy consumption among three algorithms. Lastly, compared with the ant algorithm in convergenge,L-GB algorithm is better than others.