Research On Key Technologies Of Over-The-Air Test For 5G Multiple-Input Multiple-Output Devices

The fifth generation mobile communication(5G)is designed to significantly improve wireless system performance and achieve higher data rates,energy efficiency,and reliability than the fifth generation mobile communication(4G).With the application of new enabling technologies in 5G,such as massive multiple-input multiple-output(MIMO),large bandwidth in millimeter-wave(mmWave),3D beamforming,and antenna-in-package(AiP)with integrated design,the over-the-air(OTA)testing becomes critical for performance evaluation of 5G MIMO devices.Advances in 5G technology have posed significant challenges to 5G OTA test methods,e.g.significant increases in test cost and complexity.Faced with these problems,this dissertation explores and researches the key technologies of OTA testing for 5G MIMO devices.The main research work is summarized as follows:For the OTA performance testing of carrier aggregation(CA)enabled MIMO terminals,this dissertation paper proposes two test setups(A and B)of radiated two-stage(RTS)method to establish the required wireless cable connections.Setup A directly employs multiple sets of setup resources of a single carrier according to the number of component carriers.The calibration matrix is determiined for each component carrier and then implemented in the radio frequency(RF)domain of the channel emulator(CE).Setup B can implement the calibration matrix directly in the digital baseband of the CE.In this setup,the system cost and calibration time are greatly reduced,especially for a large number of component carriers.To validate RTS OTA testing for CA scenarios,the transfer matrix from the OTA antenna ports to the device-under-test(DUT)antenna ports in three types of RF shielded enclosures is experimentally investigated.The results show that the RTS method extended to CA scenario works best in an anechoic chamber,and the isolation level of more than 18dB can be achieved for each wireless cable connection.A small anechoic box can be used in practice to save the system cost.Theoretical derivation and analysis further reveal that it is necessary to adopt one calibration matrix for each component carrier when the frequency spacing of the component carriers is large(e.g.larger than 150 MHz).For the OTA performance testing of 5G adaptive antenna system with beam-steering features,this dissertation firstly investigates the beam probability metric in evaluating the channel emulation accuracy in the multi-probe anechoic chamber(MPAC)OTA test setup.Taking a standard two-dimensional(2D)spatial channel model as an example in the simulation results,the effects of the DUT size and the number of OTA antennas on the beam probability emulation accuracy are discussed.Two quantitative measures(beam peak distance and beam statistical distance)of beam probability emulation accuracy and their rationality are further investigated.The simulation results show that the beam statistical distance might be an more instructive quantitative measure of emulation accuracy for beam probability metric compared with the beam peak distance.For 5G massive MIMO base station(BS)performance testing,based on the MPAC method,this dissertation proposes a virtual OTA testing method using flexible virtual probes.The probe wall and probe switching circuit in the traditional 3D sectored MPAC setup can be replaced with the proposed virtual probes,which further reduces the complexity of actual probe installation and the limitation of multi-probe switching in traditional MPAC setups.To further validate the proposed virtual OTA test system,the channel emulation accuracy results are simulated in terms of rms error of spatial correlation error,total variation distance of PAS,and beam statistical distance with 0.093,0.062,and 0.053,respectively.For the the standardization work of 5G MIMO OTA testing,this dissertation finally summarizes the testing methods of 5G MPAC OTA in the Third Generation Partner Project(3GPP)standardization,and discusses their differences from 4G MPAC OTA test.Furthermore,the throughput measurement results of real 5G terminals under standard channel models in MPAC OTA test setups are demonstrated.The throughput testing results show that the downlink rate of 1.1 Gbps can be achieved in good signaling conditions.For the RF test of mm Wave phased array antenna in 5G OTA,this dissertation introduces the test and measurement methods of 5G mm Wave AiP.In order to effectively and conveniently evaluate the control accuracy of AiP antenna elements,and to reliably calibrate AiP antenna arrays,this dissertation first proposes an efficient and automated mm Wave phased array AiP experimental platform.The experimental platform has flexible amplitude and phase control of AiP elements excitation,which enables automated and efficient custom measurements,avoiding the heavy workload of controlling AiP in test and measurements.For example,it takes less than 0.7 s in total for the experimental platform to automatically complete one combination operation for the status update of AiP,RF parameter measurement,and data saving.Besides,the effectiveness of the experimental platform is demonstrated by several measurement campaigns where control accuracy,array calibration,and beam steering are extensively investigated with good performance.In order to accurately evaluate the element excitation control accuracy in the normal working mode of AiP,a novel all-on method is then proposed in this dissertation and is experimentally validated.The proposed method introduces a phase compensation technique to accurately measure the control accuracy of each element in the all-on mode of the AiP,and the error range within ±0.1 dB and ±1° can be observed for amplitude and phase,respectively.The experimental analysis shows that the AiP control accuracy results obtained in an all-on mode are more instructive and valuable in practical AiP than those in the on-off mode.For the difference between the two calibration modes of AiP,this dissertation finally experimentally compares the beam-steering performance of the AiP after the two calibration modes.The objective is to investigate whether the difference between them leads to different beam-steering performance of AiP.The experimental results show that the on-off calibration mode and the all-on calibration mode can achieve the same beam-steering performance within the small range of steering angles,although the results of element discrepancies obtained in these two calibration modes are different(e.g.up to deviation range of ±2.5 dB and ±20°for amplitude and phase,respectively).