Research On Congestion Control Of Wireless Sensor Networks Based On L_1/2 Regularization

As a distributed network,wireless sensor networks have the advantages of wide monitoring range,wireless communication,and low energy consumption.It is widely used in military,industrial,medical and other fields.However,wireless sensor networks transmit data in many-to-one mode,which is easy to lead to network congestion.Network congestion can cause problems such as communication delay,packet loss,and throughput decline.These problems can affect the quality of data transmission and the quality of network services.In this thesis,the congestion of wireless sensor networks is studied.Combining L_1/2regularization with control methods,two new congestion control algorithms are proposed to achieve link level congestion and node level congestion of the network.The feasibility of the algorithm is verified by simulation experiments.The main contribution includes:?1?In the data transmission process of the wireless sensor network,redundant information is easily generated.In order to solve the above problem,at the transmitting end,the transmission data is compressed and observed.And at the receiving end,the compressed signal is reconstructed by using an iterative half threshold algorithm to solve the L_1/2 regularization.L_1/2 regularization is the non-convex optimization problem.This algorithm can reconstruct compressed data with high precision.The simulation results show that L_1/2 regularization half threshold iterative algorithm can reconstruct the compressed signals in wireless sensor networks accurately.?2?For the problem of link level congestion in wireless sensor networks,the fuzzy neural network congestion control algorithm based on L_1/2 regularization is proposed.The network congestion is difficult to describe with accurate mathematical models.In order to enhance the adaptability of the algorithm,the fuzzy neural network is used to automatically adjust the compression observation matrix dimension.At the receiving end,L_1/2regularization is used to reconstruct the compressed transmission data with high precision.The simulation results show that the algorithm can alleviate the link level congestion problem of wireless sensor networks and the effect is obvious.?3?In order to solve the node level congestion problem of wireless sensor networks,fuzzy PID queue management congestion control algorithm based on L_1/2 regularization is proposed.This algorithm compresses the data and reduces the amount of data transferred.The fuzzy control is used to optimize and adjust the parameters in the PID algorithm in real time,so that the length of the node queue can maintain the expected value more stably.At the link level,compressed data is reconstructed using an iterative half threshold algorithm to solve the L_1/2regularization.Simulation experiments show that under different congestion conditions,the algorithm can improve network transmission quality and achieve control of network node level congestion.