Research On Intelligent Tracking Method Of Heterogeneous Sensing Based On Machine Learning

With the rapid development of communication,computing and control technology,wireless sensing system has been more and more widely used,and now it has become an important part of China’s "new infrastructure" to promote the development of digital economy.As an important application area of wireless sensing system,multi-source heterogeneous target tracking has a broad market and good prospect.Energy consumption and tracking accuracy problems are fundamental to the development of wireless sensing target tracking systems,and there is a contradiction between energy consumption and tracking accuracy of the sensing system.The energy consumption problem is mainly reflected in the cluster head dying too fast.To improve the tracking accuracy of heterogeneous wireless sensing system tracking accuracy needs to increase the amount of data collection,and the transmission of a large amount of data and intensify the energy loss of the cluster head of heterogeneous wireless sensing system.However,the existing node scheduling methods are not good enough to solve this problem.Node scheduling is one of the important means to solve the problem of energy consumption and tracking accuracy of wireless sensing target tracking system,which can be divided into communication scheduling and functional scheduling.Communication scheduling means to reduce the energy consumption of the network and extend the lifetime of the intelligent sensing system by improving the network topology control optimization.Functional scheduling means to optimize the energy consumption and tracking accuracy of wireless sensing target tracking system by improving the nodes’ sleep-wake method.In order to solve this problem,this paper designs the attention-shifting mechanism of cluster head energy consumption and the sleep-wake scheduling method of heterogeneous nodes based on trajectory prediction to realize the transfer of cluster head energy consumption,reduce the large amount of irrelevant data generated by the traditional sleep-wake method of nodes,and ensure the target tracking accuracy.The details of the study are as follows:Optimization of energy consumption control of cluster head for wireless sensing target tracking system based on energy consumption attention shifting.In this paper,Attention-shifting Mechanism Based on Hybrid Energy-Efficient Distributed Clustering is proposed to control the cluster head energy consumption of wireless sensing target tracking system.Cluster head screening is performed by calculating the communication loss in a competitive manner,and then the energy consumption of cluster heads is reduced by transferring the data reception and fusion energy consumption of cluster heads to other cluster member nodes(Agents)within their own competitive radius.The selection of agents is performed by communication energy consumption comparison,while ensuring the rationality of cluster head and agent positions and reducing the communication energy consumption for re-clustering.In addition,this paper optimizes the traditional energy consumption model of sensing systems to adapt to the needs of adapting heterogeneous sensing systems.The experimental results show that the proposed method can maximize the balance of energy consumption of cluster heads and common nodes,maintain the integrity of the network,and extend the lifetime of the intelligent sensing system.Trajectory prediction-based optimization of sleep-wake scheduling for wireless sensing target tracking systems.In this paper,an error analysis-aware node pre-scheduling method based on Sparse Graph Convolution Network(SGCN)pedestrian trajectory prediction is used.First,the SGCN pedestrian trajectory prediction model is established,and the sparse graph is used to learn the movement trend and interaction characteristics of pedestrians.Then,the error analysis is performed between the historical predicted trajectory and the actual trajectory,and the pre-scheduling radius proposed in this paper is calculated.Finally,the heterogeneous sensing nodes within the pre-scheduled radius are pre-scheduled with the predicted coordinates as the circle center,and the relationship between the actual set of scheduled nodes and the set of pre-scheduled nodes is compared.This model combined with proven target tracking technology can achieve accurate and energy-efficient dispatch tracking.The experimental results show that the heterogeneous sensing node pre-scheduling method based on pedestrian trajectory prediction can achieve a larger range of node pre-scheduling and reduce the energy consumption of the target tracking sensing system while ensuring the target tracking accuracy.