Research And Implementation Of Field Environment Monitoring System Based On Wireless Sensor Networks

As one of the internet of things systems, wireless sensor networks is the important channel for changing agriculture production mode, promoting the sustainable development of agriculture in China and efficient utilization of rural resources. As a wireless information acquisition and processing technologies,it is now attracting widespread attention in the field of agricultural production.WSN is a dynamic network. It has dynamic topology organization function to achieve full coverage of the monitoring area of farmland and guarantee the connectivity of the whole network. Farmland environment is a vast territory, keeping a large amount of information. Therefore, reducing the energy consumption of the system, maximizing the network lifetime,accurate positioning and accurate information acquisition and transmission are important performance indexes for designing field environment monitoring system which based on wireless sensor networks.Currently, there are large number of sensor nodes deployed in the farmland environment, which are powered by battery capacity. And the node’s energy consumption is considerable. The majority of scholars have conducted research on a variety of energy-saving measures and some progress had been made. But there are fewer energy saving strategy for node deployment strategies and routing protocol study in terms of the actual needs of farmland.In addition, the node location information in farmland wireless sensor networks is also an important part of the research. There is a large amount of localization algorithm in wireless sensor networks, among which RSSI localization algorithm is greatly influenced by the actual farmland environment. And thus as for the actual situation of farmland, the precise distance measurement model still needs to be perfect.Moreover, the accuracy of sensor itself and the nonlinear error of sensor node and other factors will lead to measurement data inaccuracy. And the reasonable compensation can improve the accuracy of the system measurement. The continuous acquisition of multiple sensor nodes will produce large amounts of data and can reduce the amount of upload gateway data through effective data fusion, so as to save the system energy consumption and improve network robustness.In response to the above problems, the article studies the key technical issues of how the wireless sensor networks was applied in field monitoring on the basis of analyzing the domestic and abroad research in this field. The main research work was as follows:(1) The experimental study was mainly about nodes deployment properties of homogeneous and heterogeneous network. It analyzed the coverage rate in homogeneous network in three conditions.That was, regular hexagon, regular quadrilateral and random deployments, in which when the distance between nodes in regular hexagon was√3times the length of the sensing radius, and quadrilateral relationship was√2times, the network reaches full coverage and the coverage rate was1.The study and comparison on the relationship among node sensor radius, the distance between nodes and power loss in the two kinds of regular deployments showed that, the energy losses in regular hexagons were relatively small. On this basis, NS2simulation software analyzed three kinds of network performance---the network life cycle, end-to-end delay and packet loss---in the three ways of deploying of homogeneous network. Simulation results shown that network performance of the regular hexagon was the best.At the same time, the network performance in heterogeneous network was studied, the energetic nodes and ordinary nodes in the deployment of hexagonal and square way to reach complete coverage the position and the number relationship comparative analysis and simulation experiment. Simulation analysis found in the first way, Hexagon deployment forms network life cycle was the longest, and in the second way, the quadrilateral deployment life cycle was the longest in the low density deployment. While in the high-density deployment, the random deployment of network life cycle was longer in the first way, and the time when death node appeared in Hexagon deployment way was more advanced in the second way, but the overall network survival time was less than random way. The result showed the heterogeneous way can better extend the life cycle of network.(2) The test experiment was designed for heterogeneous network performance of2.4GHz of ZigBee nodes. The effective transmission range of ordinary node and high energy node was tested in obstacle and non-obstacle as well as in three different heights. Thus the distance between nodes and node height were decided in the farmland node deployment. On this basis, a comparative analysis was made for the number of nodes placement under the regular hexagon and quadrilateral heterogeneous deployment in the same transmission distance and the same coverage rate. The result indicated that the number of nodes placement under the regular hexagon was more than the one under regular quadrilateral in the same transmission distance and the number of nodes placement in quadrilateral-shaped place was more than the one in regular hexagons under the same coverage rate. At last, a small amount of nodes were placed in the farmland to make an experiment on the remaining energy in the same transmission distance and the same coverage rate. Regular hexagons node voltage was reduced more slowly in the same distance, and quadrilateral node voltage was reduced more slowly in the same coverage rate.(3) When the localization algorithm of RSSI wireless sensor networks was applied to the farmland information monitoring system, the actual RSSI value was affected by the distance between nodes,antenna height,crop growth and other factors.The signal strength experiment of the2.4GHz sensor nodes was designed under various environmental conditions such as having obstacles or not in the surrounding farmland, at different heights and in different distance. The research analyzed the attenuation relation between RSSI value and distance and height between nodes and determines the radio signal attenuation model by multiple linear regression method. On the basis of this model, distance measuring experiment was conducted. Then the positioning of an unknown node was achieved on the basis of Gaussian mixture and maximum likelihood methods.Select a8x8m2area and deploy4beacon nodes and2unknown nodes to make location experiment. The results showed that the system can fulfill better positioning of the two location nodes and the average positioning accuracy was up to1.01m in4beacon nodes.(4) A hierarchical cellular Ant Colony Optimization routing algorithm was proposed on the basis of analysis of principle and advantage of Ant Colony algorithm. The sensor nodes in the actual farmland environment were mapped to cellular nodes. Store the remaining energy of neighboring nodes, node distance, pheromones and other information. Select the next node by calculating the pheromone probability. Keep up to update pheromones in the various lines until reach the sink node. The algorithm optimized net routing by improved cellular node hibernate conversion mechanisms and pheromone updating rule. Experiments showed that compared with the traditional LEACH algorithms, survival time increases by15.9%and more balances network energy.(5) When there was farmland information acquisition, multiple sensors would be placed, such as soil temperature and humidity, light intensity. Based on multi-parameter, a two-level information fusion and calibration method was presented. After analysis of main factors affecting soil temperature and humidity and light intensity sensor measurement, a measurement of30samples was carried out. There was a calibration of individual sensors measurements through BP neural networks and least squares fitting. Eliminate nonlinear error of sensor measurements and incorrect data. Experimental results showed that the neural network calibration works better and the calibration error of two kinds of sensors was within±2%. And then the fusion of multiple sensor measurement data through the measurement of Adaptive weighted fusion algorithm had reasoning results and get the current state of the tested object. Fusion experiments showed that Adaptive fusion algorithm can achieve a better integration. By a two-level fusion approach not only was a single-sensor data fusion methods optimization achieved, but also multiple sensors data fusion processing was fulfilled. Thus the reliability and accuracy of measurement and control system as a whole were increased.(6) According to farmland environmental characteristics, completed the selection and a modular design of the four kinds of sensors such as the soil humidity, light intensity, temperature and humidity and design the sensors node, gateway node and power control circuit and then NET remote monitoring platform and mobile terminals Android monitoring platform. Farm monitoring system acquire and process the humidity, temperature, lighting, soil humidity and video information. The stable and reliable system operation met the needs of field monitoring.