Nikos Perpinias

Measurements of Shadow Correlations in a Suburban Environment on the 485 MHz Band

Correlations in shadow fading play a key role in the performance of wireless networks. However, despite of this importance, relatively few comprehensive measurement studies have been made on shadow correlations, and the methodology followed in the existing studies is often poorly documented. In this paper we report results from an extensive measurement campaign on shadow correlation in the 485 MHz band in a suburban environment. In particular, we study carefully the influence of fast fading and different data processing choices on the results, as well as describe in detail the design and verification of the measurement set up. Our results show that shadow correlations decay exponentially as a function of distance also on this band, but that the details of the arising correlation model are highly dependent on the steps taken to process the measurement data. In particular, we show that unless fast fading is properly averaged over, highly incorrect conclusions would be drawn with respect to both marginal variance and the correlation distance of shadow fading.

60 GHz outdoor urban measurement study of the feasibility of multi-Gbps mm-wave cellular networks

Future 5G cellular networks are expected to exploit the abundant spectrum resources of the millimeter-wave (mm-wave) bands to satisfy demand for multi-Gbps mobile links anticipated by exponential data traffic growth. However, given the directional nature of mm-wave links, the feasibility of mm-wave mobile networks is critically dependent on efficient antenna beamsteering and a rich inventory of strong LOS (line-of-sight) and NLOS (non line-of-sight) paths from effective reflectors in the urban environment. In this paper we report results from detailed angular measurements of 60 GHz links at an example outdoor pico-cellular site in a mixed-use urban environment typical of European cities. Our work is the first to systematically analyze the beamsteering requirements of future mm-wave cellular networks based on real measurements. Our results reveal that the urban environment provides substantial opportunities for multi-Gbps mm-wave connectivity, but that the availability of strong LOS/NLOS links is highly location and orientation-specific. Our results also show that high speed mm-wave links are very sensitive to beam misalignment. This has important implications for practical mm-wave cellular network design: (i) high-precision beamsteering is required to maintain stable data rates even for quasi-stationary users; and (ii) providing seamless high speed service in mobility scenarios will be extremely challenging. Our results thus cast doubt on whether outdoor mm-wave cellular deployments will be feasible in practice, given the high network control overhead of meeting such stringent beamsteering requirements.

Empirical characterization of mm-wave communication links in realistic indoor scenarios

The large contiguous bandwidths available in the mm-wave frequency bands have recently started to gain significant attention due to the growing belief that it can alleviate the spectrum shortage at lower frequencies and meet the increasing high data rate demands. While some commercially available mm-wave transceivers in the 60 GHz and 70 GHz bands have been developed, only a few studies exist on the achievable data rates in realistic deployment scenarios. Moreover, practical issues such as interference and coexistence of heterogeneous applications have not been investigated. In this paper, we present our empirical study on the bit error ratio and throughput for realistic indoor application scenarios in 60 GHz and 70GHz using orthogonal frequency division multiplexing based physical layer on various commercially available transceivers. Furthermore, we analyze the effects of interference on the performance characteristics of the mm-wave frequency links in different deployment setups.

A measurement-based study on the use of spatial interpolation for propagation estimation

In this paper, we provide a first comprehensive comparison on the performance of two spatial estimation approaches used for propagation estimation. The first estimates the shadowing distribution with considerably low number of measurements and the second estimates the received power with no information of the transmitter location. We discuss in details the relative advantages of these techniques and study particularly their performance when using different number of measurement points. Our study is based on data from a very carefully conducted large-scale measurement campaign with a high spatial density. The robustness of our results is verified in different propagation environments and different frequencies. In particular, we show that for very small number of samples, prediction based on shadowing results in better performance. However, with small increase in the number of available measurements, this more complex approach becomes unnecessary and direct prediction of the received power yields very good results.

Studying and mitigating the impact of GPS localization error on radio environment map construction

Localization errors can play an important role in measurements of radio propagation or spectrum utilization, especially in urban environment where GPS localization accuracy is highly varying. In this paper we present a low cost and simple method for mitigating the impact of localization errors, based on multiple GPS receivers. The presented results are based primarily on performance evaluation of multiple GPS receivers based method under realistic conditions. We show how to reduce the time needed to estimate the statistics of the spatial radio environment by combining the input of multiple GPS sensors. We also present a simple scenario for small cell size networks based on real measurements in three distinct environments, where the localization errors are the determining factor of the performance of such systems.

On the Study of Shadow Correlations of Two Radio Frequencies in an Urban Environment

We present the results of an extensive, high spatial density measurement campaign made for investigating the shadowing correlations on two different frequency bands, namely 485MHz and 2600MHz in an urban propagation environment. In particular, we report results using different analysis procedures and tools. Our results show that the signal detector has only a small impact on the extracted data, given that the fast fading is correctly averaged out. Moreover, we extract the main parameters of shadow correlations. Finally, we report a preliminary comparison between a suburban and an urban propagation environments by comparing the parameters of the fitted variograms.

Impact of the path loss model on the spatial structure of shadow fading

In this paper we describe the effects of the radio propagation model on the spatial characteristics of the radio environment. For this study we have conducted a dedicated propagation measurement campaign focusing on the three different frequency bands of 485 MHz, 1700 MHz and 2600 MHz in order to capture the impact of the used frequency on the radio propagation. The presented results are coming from a high number of samples of very high spatial resolution. We quantify how much the propagation modelling affects the calculation of spatial statistics of radio propagation and in particular shadow fading for the three different frequency bands. Our results show that while the used propagation model does play a role in both the marginal distribution and the spatial correlation characteristics derived, the impact of the precise propagation model chosen is small. This indicates that correlation structure of shadow fading can be accurately derived also using predefined propagation models not directly fitted to the data and radio environment at hand.

Slicing the shadowing uncertainty: A measurement-based critical study of spatially correlated Gaussian models of shadow fading

We report results from extensive measurements based study to understand the shadow fading correlations. We have particularly studied and analyzed common assumptions done in literature for shadow fading, which is typically modeled as a spatially correlated Gaussian random process in the logarithmic scale. We show that these assumptions, starting from the Gaussianity of the radio propagation environment, do not always hold. We then discuss the assumptions regarding spatial correlations, usually taken to depend only on the distance between the points and assuming that these correlations remain the same throughout the area under consideration. This is done by comparing received power measurements for two different propagation environments, an urban and a sub-urban one, against simulated values based on typical theoretical assumptions. We conclude by explaining how these assumptions should be carefully taken into account and how they can affect simulation results and the performance of applications.

Impact of Model Uncertainties on the Accuracy of Spatial Interpolation Based Coverage Estimation

Improved propagation prediction has become important due to emerging 5G and ultra-dense wireless networks. One important aspect in this domain is to understand and model better the spatial correlations of the underlying shadowing fields. In this work, we investigate the sources of uncertainty on the predictions made by employing the optimal linear predictor, namely \emph{kriging}. We show that kriging is robust to the estimation errors of the spatial correlation structure. Moreover, we show that the amount of the training data (known data points) has great impact on the prediction error. Our results help to quantify the trade off between number of data points collected and accuracy that can be reached by interpolation. The results especially help to design and optimize expensive measurement and test drive campaigns. Another contribution of the paper is to explicitly show the capabilities of kriging for coverage prediction.

Measurement-based study on the influence of localization errors on estimated shadow correlations

In recent years, novel spectrum access schemes have increased the need to further explore and exploit radio propagation dynamics. One proposed approach is to use spectrum sensing nodes for propagation environment estimation, and particularly to allow shadowing field extraction. In general there is a tendency to develop techniques that allow the automated and localized estimation of the spatial correlation structure of the shadowing field. Key issue when using such data is to take into account inherent localization errors and the impact that it has on these estimates. In this paper we use real measurement data acquired with extremely high localization accuracy to demonstrate and study the impact of localization errors. We present results for the propagation estimation, the extracted shadow field, and estimated shadow correlations. In particular, we show that spatial correlation metrics such as semivariograms are robust against localization errors higher than arising from the typical embedded GPS chipsets, up to approximately 20 m. Moreover, we have considered various probability distributions of localization errors, exploring their impact into our analysis. The reported results are highly relevant to development of measurement-based coverage estimation techniques and planning of drive tests or understanding limits of crowd-sourced data from user devices.