Optimization Of Wireless Sensor Layout And Scheduling Based On Mutual Information

As an important means of obtaining information,wireless sensor networks have become an important monitoring measure for wide areas such as forests and pastures due to their advantages of convenient layout,self-organization,and strong environmental adaptability.This paper focuses on the sensor layout and scheduling of wireless sensor networks.First,according to the characteristics of node perception and the submodularity of the layout problem,a probabilistic layout model based on mutual information is established.In order to improve the robustness of the layout and reduce the monitoring effect loss caused by the uncertainty of the nodes,a submodular benefit model based on CVa R(Conditional Value at Risk)and OS(Ordered Searching)-LE(Lazy Evaluation)-Greedy algorithm are proposed.Then,to further optimizing the sensor layout under drone deployment error,this paper proposes an error compensation sensor layout model based on mutual information.SOG(Submodular-Online-Greedy)algorithm is also to proposed to reduce the impact of deployment errors to subsequent nodes.Finally,considering the sensor layout and scheduling,a balanced layoutscheduling model based on mutual information is proposed,and the traditional grouping algorithm is improved to propose the BG(Balanced-Grouping)algorithm to complete the simultaneous optimization of sensor layout and scheduling.The main work of this article is as follows:(1)A mutual information deployment model based on CVa R(Conditional Value at Risk)and OS(Ordered Searching)-LE(Lazy Evaluation Greedy)algorithm are proposed.In view of the problem of how to ensure the optimal effect of the sensor layout in the presence of uncertain nodes,this paper considers the uncertainty of the nodes when initially deploying the nodes,and designs a submodular benefit model based on CVa R to minimize the impact which uncertainty has on sensor deployment.At the same time,the traditional greedy algorithm is improved in order to quickly and efficiently obtain the optimal sensor layout under the model.According to the parameter ? in the model,the global optimal solution is searched in an orderly manner,and lazy evaluation is introduced to decrease time complexity.Experiments show that the mutual information obtained by OS-LE-Greedy algorithm is 14.2% and 68.4% higher than the traditional algorithm and the random deployment method,respectively.In the case of node death and node deviation from the expected position,the loss caused by the mutual information deployment model based on CVa R is also less than loss under the traditional submodular benefit model.(2)An error compensation sensor layout model based on mutual information and SOG algorithm are proposed.According to the error existed in sensor deployment when using drone scattering,an error compensation layout model based on mutual information is proposed.The actual deployment position of the node is the sum of the expected position after considering the error and the deviation.In order to obtain the expected deployment position after considering the error,the SOG algorithm is proposed,taking the covariance between the expected deployment position and the actual deployment position as input,and selecting the node that increases the mutual information amount in the network as the next deployment position each time.Experiments show that the amount of mutual information that the SOG algorithm can achieve in actual deployment reaches 133.34% and 174.36% of the traditional greedy algorithm and random deployment method,respectively,and the actual layout effect of the sensor under the compensation model is closer to the expected effect.(3)A balanced layout-scheduling model based on mutual information and BG algorithm are proposed.For the combination of sensor layout and scheduling,the idea of simultaneous layout-scheduling is adopted.At the beginning of the layout,the sensor networks are grouped,and a balanced layout-scheduling model based on mutual information is proposed to maximize the mutual effectiveness of the worst group in order to optimize the monitoring effect of each group.In order to complete balanced scheduling,the BG algorithm is proposed to redistribute each group under the set threshold,which effectively guarantees the monitoring quality of each group and ensures that each group after grouping can work independently on the premise of ensuring that the monitoring effect meets the requirements,so that the network lifetime extended.