Massive Connectivity Asynchronous Access Detection Algorithm For Internet Of Things

With the rapid development of Internet of Things(Io T),massive machine type communication(m MTC)has gradually become one of the most important communication modes in cellular networks.The traditional algorithms for joint detection of device activity and channel response require that the length of pilot sequence is proportional to the total number of users in the system.The traditional methods require a large amount of pilot resources in the m MTC system,which makes it difficult to achieve them.To address the problems effectively,considering the sporadic transmission characteristics of the m MTC system,the application of compressed sensing(CS)technology to large-scale access scenarios has become a current research hotspot.However,if the frequency offset and time delay are not effectively processed,conventional CS reconstruction algorithms will seriously affect the performance of device activity detection and channel estimation in large-scale access scenarios of asynchronous Internet of Things.Therefore,this thesis focus on the largescale access detection algorithm of asynchronous Internet of Things.The main research contents of this thesis are summarized as follows:First of all,the conventional compressed sensing algorithms are studied,and the typical greedy reconstruction algorithms are simulated and compared from the perspective of reconstruction performance.Using the sporadic transmission characteristics of m MTC,system model of joint device activity and channel estimation is established.Secondly,considering the time delay in large-scale access scenarios,a system model of asynchronous transmission is established.Then,in order to reduce the effect of time delay on the reconstruction performance,this thesis derives an optimization model for joint delay estimation,device activity detection and channel estimation to synchronize the received signal.aiming to reduce the detection error rate of active users and the mean square error of channel estimation,an active user detection and channel estimation algorithm based on greedy recovery and convex optimization solution are designed.The simulation compares the detection error rate and channel estimation performance of the OMP algorithm based on successive interference cancellation and the optimization algorithm based on Lasso.The results show that the performance of the greedy recovery algorithm has better performance.Lastly,a system model of large-scale access and an optimized model with both time delay and frequency offset are established.Because of the influence of time delay and frequency offset on the detection system,the large-scale optimization is decomposed into a series of small-scale sub-problems,and then each sub-problem is solved in turn.And then,an algorithm based on sparse constraints is designed according to the decomposed optimization sub-problem.The algorithm uses a parallel method to perform a two-dimensional search for time delay and frequency offset using correlation values.And particle swarm optimization algorithm with fast convergence speed is used to obtain the optimal frequency offset value.The simulation results verify the feasibility of the sparse constraint algorithm,and describe the influence of the pilot length,the number of devices,and the sparse level on the error rate of device detection in asynchronous scenarios.