| The growth of crops is based on soil,it is very important for the production of crops to get the soil moisture content accurately and in time.Through sensors distributed in farmland,a large amount of soil moisture data(such as soil moisture,temperature,etc.)is collected to accurately obtain the soil moisture content and guide agricultural personnel to carry out agricultural activities,such as irrigation.Whether the sensor is effectively arranged has an important impact on the decision of irrigation system,as well as the construction cost of the system.This paper studied the optimal layout methods of soil moisture sensor for tea plantation,and mainly optimized the number of soil moisture sensor,data redundancy,network energy consumption and the path length.The optimization methods mainly include clustering algorithm,ant colony optimization(ACO)and modified ACO.(1)Affinity Propagation(AP clustering algorithm)based sensor layout Optimization.The tea plantation in the National High-tech Agricultural Park of An Hui Agricultural University was chosen for the experiment,25 points were randomly selected to arrange the sensors,to collect soil moisture data for constructing the similarity matrix of soil node moisture content according to the collection of soil moisture data of each node in the experimental area,and iteratively calculating the attraction and attribution values of each node,obtaining the cluster number and cluster center position,taking cluster center as the benchmark,eliminating redundant sensor nodes,and using cluster center and cluster number as the placement position of tea plantation sensors.The experimental results showed that the number of sensors was reduced from 25 to 2,which reduced the number of the sensors and the redundancy of soil moisture data was also reduced.(2)ACO based tea plantation wireless sensor network(WSN)sensor layout optimization.After the optimizing sensor layout by AP clustering algorithm,the data redundancy was reduced and the number of sensors as well,but the distance between nodes was beyond the communication range and the data transmission could not be completed.Considering the sensor transmission distance,we proposed ACO to optimize the sensor layout,the distance model between sensor nodes and data redundancy model were introduced into the ACO probability transfer formula,to calculate the transfer probability,and select the next visited sensor node.The experimental results showed that optimized sensor layout by ACO required 9 sensors,which could realize data transfer,and the redundancy of soil moisture data was low,but higher than that in the sensor layout optimized by AP clustering algorithm.(3)Modified ACO based tea plantation wireless sensor network(WSN)sensor layout optimization.Energy loss was an important factor affecting the network life cycle in wireless sensor networks.Therefore,modified ACO was proposed to optimize the sensor layout,to introduce network energy loss model in the heuristic function,so as to enlighten ants to choose paths with small energy loss and save network energy;to introduce trend guide in probability transfer formula to guide ants to select the next sensor node from the starting point along the end point,prevent detour and reduce path length.The experimental results showed that the modified ACO optimized sensor layout required 6 sensors,which saved 3 sensors compared to the ACO,and the network energy consumption and data redundancy were also lower than the ACO,as well as the the shortest path length. |
