Study On Data Collection Of Heterogeneous Wireless Sensor Networks For Farmland Monitoring

Farmland monitoring environment is characteristic of gradient. The features of farmland monitoring environment are large scale, long monitoring period, plants grow with time, machinery operation and complex environment factors. How to monitor the environment parameters of large-scale farmland in order to improve the productivity of farmland is the key issue in precision agriculture. The environment information which gathered by WSN could guide the agricultural production and lead to great economic benefit.In recent years, achievements have been made in data collection on homogeneous wireless sensor networks. However, there are still some problems in research of heterogeneous WSN. These problems include uniformly deployed problem of heterogeneous network with solar and battery nodes; instability of WSN routing protocol caused by weather factors; instability in the transmission channel and data collection caused by heterogeneous deployment of the ground and underground sensor nodes. To solve all these problems, this paper studies the heterogeneous WSN, focusing on the channel modeling, node deployment, routing algorithm and mobile sink data collected.The main contributions of this dissertation are as follows:(1)According to the changes of radio propagation channel causes by crop growth, environment correlation factor optimizationof large-scale and multi-scale fading model is studied. The large-scale logarithm fading model and multi-scale statistical fading distribution model is carried on the preliminary fitting of the experiment data. The environmental correlation factors such as leaf area index, leaf density index, antenna height, occlusion height are also introduced in the data fitting. As a result, the large-scale and multi-scale fading model based on environmental factors and signal transmission is acquired. In addition, the channel fading of underground wireless channel signal is also studied based on the experimental data. A two path model of wireless signal in maize field is established. The experimental results show that the model has good fitting performance, which can accurately describe the dynamic changes of the channel transmission.(2)Aiming at the problem of the multipath propagation of wireless signal and the energy heterogeneity of nodes in the complex environment of farmland, a new deployment method of heterogeneous WSN nodes based on path loss is studied. According to the crop growth trend, using square segmentation method of structure deployment, on the basis of the regional path loss, random deployment of nodes is drawn. Nodes positions are adjusted based on the virtual collision force method. Meanwhile, solar power nodes are introduced and the locations are selected by the network load factor. Simulation results show that the method not only ensures the network coverage, optimizes the load balance between nodes, but also improves the utilization efficiency of solar energy node.(3)Aiming at the unstability of clustering routing caused by the solar energy limitation, this paper proposes an energy aware routing algorithm based on heterogeneous WSN. By using the method of fuzzy set evaluation factor of solar charging capacity of the node, the node residual energy factor and other factors is considered the backbone nodes for the selection. The fuzzy soft set and clustering methodare used to elect the backbone nodes and the cluster heads. In accordance with the improved hierarchical routing, the extended backbone nodes introduced as relaying routing. Simulation results show that this algorithm optimizes the data collection routing, ensures the stability of network operation, improves the utilization efficiency of solar energy of nodes and prolongs the network lifetime.(4)Aiming at the problem of packet loss rate, high energy consumption and transmission instability of mobile sink data collection in large-scale farmland environment, a new mobile sink method of heterogeneous WSN data collection is proposed. In the initial stage, the square structured incremental is deployed. The virtual RP(Rendezvous Point) location is designed, and all the real node information is collected while mobile sink traversing all the RP nodes. Then, according to the time constraint and the priority of the key nodes, RP nodes are selected and the path of mobile sink is optimized. At the same time, the mechanism of the environment of the isolated node is designed, and the stability mechanism of the packet loss process is designed. Simulation results show that this method optimizes the sink movement path, ensures the effective collection of underground node data under different humidity changes, reduces the packet loss rate and stabilizes the network data collection.