Research And Verification Of Non-orthogonal Random Access Technology For Io T Network

With the dramatic increase in the number of internet of things(IoT)devices,the wireless communication technology based on orthogonal multiple access(OMA)is greatly restricted by limited spectrum resources,data conflict and loss due to massive IoT device access becomes unacceptable.It becomes especially important to consider new access technologies to accommodate the proliferation of IoT devices.This thesis is mainly focuses on the technical difficulties such as active user detection and system optimization faced by the application of non-orthogonal multiple access(NOMA)technology to IoT scenarios,researches are carried out from the aspects of uplink active user detection and system throughput optimization,and built a verification platform based on sparse code multiple access(SCMA)technology through universal software radio peripheral(USRP),it verifies the feasibility of SCMA technology in IoT scenarios.Above all,for different NOMA schemes,this thesis presents a unified activation user detection model,and a detection algorithm based on long short-term memory(LSTM)network is proposed to solve the problems of low accuracy and high complexity of traditional active user detection algorithms.The simulation results show that,compared with the orthogonal matching pursuit(OMP)algorithm and the sparse bayesian learning(SBL)algorithm,the optimal false alarm probability and missed detection probability of the activated user detection algorithm based on LSTM network are reduced from 0.051and 0.138 to 0.01 and3.326×10^-4,respectively,when the number of activated users is150.Moreover,the pilot designed for the SCMA scheme are as low as 0.006 and7.285×10^-5.In addition,the detection complexity based on LSTM network is reduced to o(n~2)compared to o(n~3)of OMP and SBL algorithms.Secondly,combined with NOMA technology,this thesis solves the problem that the traditional ALOHA scheme does not allow data collision,derives the system throughput expression for uplink multi-user access in IoT scenarios,and presents the algorithm flow for optimizing system throughput by combining node rate and activation probability.Simulation results show that with the number of potential nodes determined,with the increase of the number of active nodes,the system throughput will also increase,but the data rate of each node will decrease.In addition,the activation probability of the node corresponding to the maximum system throughput is also increasing.On the other hand,when the node activation probability is uncertain,the system throughput is not necessarily related to the total number of potential users.Finally,this thesis uses USRP to build a test platform for uplink and downlink SCMA solutions to verify the performance of actual NOMA technology transmission.Through the test of actual transmission data,it is found that the downlink SCMA scheme can accurately restore the transmitted data information,and the uplink SCMA scheme can also accurately restore the node data when the access delay difference of each node is less than 0.3?s.