Three Dimensional MIMO Spatial Channel Modeling Research Based On Radio Channel Measurement

Each new technology of mobile communication focus on dealing with the fading characteristics of wireless channel, and achieve effective, safe, and reliable communication. Nowadays the data rate keeps increasing. And the carrier frequency is up to the GHz, which incur the cell radius become much smaller. That makes the elevation dimension cannot be ignored. Meanwhile, the appearance of new tech such as3D MIMO which can fully use the elevation dimension by the3D beamforming, makes the current two dimension channel model not capable to fulfill the needs of wireless communication situation. Three dimensional channel model has becoming the hottest research topic, how to model the3D space channel model’s characteristics is become a serious problem.This paper chooses the three dimension channel model technology which based on channel measurement as the issue, followed by the standardization progressing step, main work includes:First:This paper summarize the early studies of three dimension channel model, from two aspects:Theoretical model and the experimental model based on measurement. Through the early study, the elevation angle of radio waves was confirmed by LEE in1973, and the value is within the range of [0,50°]. Then the elevation angle distribution is summarized, the early result show the distribution vary with different scenario and frequency. Also this paper analyze the model method of current MIMO channel model, so propose the idea to model the3D channel by extend the2D channel model to3D. At last, according to the idea, this paper categorizes the key parameter of3D channel, then divide them into different priority, and use these priority to direct following channel measurement and modeling.Second:This paper using Propsound channel measurement equipment and3D antenna array to measurement the UMA UMI021scenario. The results are following:all the three scenario elevation angle follows laplace distribution, and the global angular spread in these three situation are follow lognormal distribution and normal distribution. This paper also studies the mean value of angular spread, and this value is related to distance. And strong relation occurs when distance <50m. In this paper, the mean value of elevation angle of measurement is not along with LOS direction, the difference can be modeled as a function of distance. So the model method of elevation angle is modified to fit the measurement result. In the end, the clustering result and cross correlation coefficient is given. Through above results, this paper offers a prime framework of3D channel and the model method of different spatial parameters, our results is written as proposal deliver to3GPP for further discussion.