Research On Node Location Prediction Method In Mobile Wireless Sensor Networks

The mobile wireless sensor network is a medium that connects human society with nature.It is widely used in various fields,such as the monitoring of forest vegetation coverage,the monitoring of animal habits in deep areas of pastoral areas,and the rescue of mine disasters,etc.Accurately obtaining the location information of the monitoring node in the application scenario becomes an extremely critical and difficult problem to solve.It not only has theoretical research significance,but also has a certain commercial value.This article summarizes the progress of previous researches on mobile node location prediction methods,and analyzes their deficiencies,such as the low accuracy of location prediction;it requires a large number of historical trajectories to build prediction models and the robustness of the prediction models is not very good.problem.To solve these problems,first,a node position prediction algorithm is proposed that can predict the position of the mobile node at the next moment in time even in the case of small sample data sets.The algorithm uses uncertain support vector machines to handle the problem.The ability to determine information is combined with the heuristic genetic optimization algorithm to construct a good position prediction model.At the same time,the calculation of the yaw direction of the mobile node is used to narrow the range of prediction values,thereby improving the accuracy of position prediction.Secondly,on the basis of the last simulation experiment,consider that in the actual monitoring environment,nodes may die due to node energy exhaustion,fire or earthquake,and other external forces,resulting in a sparse mobile sensor network with sparse sensing.Because of the low accuracy of mobile node location prediction in the network,this paper also proposes a sparse sensor network mobile node location prediction algorithm based on deep belief network.This method makes full use of the unique unsupervised learning ability of the deep belief network and the signal intensity distribution characteristics under the autonomous learning complex environment.Based on this,it trains and builds a distance prediction model,and uses the constructed distance prediction model todetermine the unknown node and its neighboring nodes.The distance between them to determine the location of the area,and then to achieve the purpose of predicting the location of unknown nodes.Finally,the experimental simulation results show that the proposed method based on the uncertainty support vector machine(USVM)proposed in this paper is suitable for small samples while maintaining good prediction accuracy,and the calculation is relatively small and suitable for energy-limited.Mobile node location prediction.The position prediction algorithm for mobile nodes based on deep belief networks is superior to other methods for predicting the position of nodes in sparse sensor networks.It also improves prediction accuracy and robustness,and is suitable for sparse sensor networks.Perform node position prediction.