Analysis and Optimization of 5G Coverage Predictions Using a Beamforming Antenna Model and Real Drive Test Measurements

Abstract

The ability to estimate radio coverage accurately is fundamental for planning and optimizing any wireless network, notably when a new generation, as the 5 th Generation (5G), is in an early deployment phase. The knowledge acquired from radio planning of previous generations must be revisited, particularly the used path loss and antennas models, as the 5G propagation is intrinsically distinct. This paper analyses a new beamforming antenna model and distinct path loss models -3 rd Generation Partnership Project (3GPP) and Millimetre-Wave Based Mobile Radio Access Network for Fifth Generation Integrated Communications (mmMAGIC) -applying them to evaluate 5G coverage in 3-Dimensional (3D) synthetic and real scenarios, for outdoor and indoor environments. Further, real 5G Drive Tests (DTs) were used to evaluate the 3GPP path loss model accuracy in Urban Macro (UMa) scenarios. For the new antenna model, it is shown that the use of beamforming with multiple vertical beams is advantageous when the Base Station (BS) is placed below the surrounding buildings; in regular UMa surroundings, one vertical beam provides adequate indoor coverage and a maximized outdoor coverage after antenna tilt optimization. The 3GPP path loss model exhibited a Mean Absolute Error (MAE) of 21.05 dB for Line-of-Sight (LoS) and 14.48 dB for Non-Line-of-Sight (NLoS), compared with real measurements. After calibration, the MAE for LoS and NLoS decreased to 5.45 dB and 7.51 dB, respectively. Moreover, the non-calibrated 3GPP path loss model led to overestimations of the 5G coverage and user throughput up to 25% and 163%, respectively, when compared to the calibrated model predictions. The use of Machine Learning (ML) algorithms resulted in path loss MAEs within the range of 4.58 dB to 5.38 dB, for LoS, and within the range of 3.70 dB to 5.96 dB, for NLoS, with the Random Forest (RF) algorithm attaining the lowest error.

An Ubiquitous 2.6 GHz Radio Propagation Model for Wireless Networks Using Self-Supervised Learning From Satellite Images

“…The LoS classification was performed deterministically, using terrain and 3D building information [53]. Afterwards, the 3GPP model was applied, and an RMSE of 20.96 dB was obtained, which is within the values reported in [8]. The 3D building data was limited to the train and validation areas, preventing an evaluation of the 3GPP PL model on the locations of the generalization DT data.…”

“…The goal is to achieve higher prediction accuracy than conventional (empirical) PL models without introducing excessive computational complexity or requiring extensive environment data. Nonetheless, the generalization capability -i.e., the ability to learn from a limited volume of data and perform similarly in an out-of-distribution data-of ML or deep learning-based models is still being investigated [8]. Moreover, PL models, when calibrated with DT measurements, generally require a specific calibration for each propagation environments [9].…”

Mobile Network Coverage Prediction Based on Supervised Machine Learning Algorithms

“…As shown in ( 1) and ( 2), two additional parameters were derived from the above-generated parameters, i.e., the 3D separation distance between eNB and UE (D 3 ) [22], [25], [26] and the height ratio between eNB antenna and EU (H R ) [28].…”

“…The work was based on 2,558 datasets with ratios of 53.6% and 46.4% for model training and testing, respectively. Similarly, in [25], the work was conducted in an urban environment in Lisbon, Portugal, at 3.7 GHz and 26 GHz frequency bands, using a real 5G network. The input parameters and response variables used are almost the same as [27].…”

“…• Performance analysis of ML-based model against previous work done by [22], [25], but considering an extension of scenarios and other improvements in methodology, as shown in Table 1.…”