Petri Mähönen

Channel Selection Algorithm for Cognitive Radio Networks with Heavy-Tailed Idle Times

We consider a multichannel Cognitive Radio Network (CRN), where secondary users sequentially sense channels for opportunistic spectrum access. In this scenario, the Channel Selection Algorithm (CSA) allows secondary users to find a vacant channel with the minimal number of channel switches. Most of the existing CSA literature assumes exponential ON-OFF time distribution for primary users (PU) channel occupancy pattern. This exponential assumption might be helpful to get performance bounds; but not useful to evaluate the performance of CSA under realistic conditions. An in-depth analysis of independent spectrum measurement traces reveals that wireless channels have typically heavy-tailed PU OFF times. In this paper, we propose an extension to the Predictive CSA framework and its generalization for heavy tailed PU OFF time distribution, which represents realistic scenarios. In particular, we calculate the probability of channel being idle for hyper-exponential OFF times to use in CSA. We implement our proposed CSA framework in a wireless test-bed and comprehensively evaluate its performance by recreating the realistic PU channel occupancy patterns. The proposed CSA shows significant reduction in channel switches and energy consumption as compared to Predictive CSA which always assumes exponential PU ON-OFF times. Through our work, we show the impact of the PU channel occupancy pattern on the performance of CSA in multichannel CRN.

What will interference be like in 5G HetNets

In this paper we discuss challenges in interference modeling for performance analysis of future wireless networks. We show through detailed numerical and simulation case studies as well as through measurements that many of the commonly used models result in potentially highly inaccurate predictions of interference and performance. In particular, we identify correlations in node locations, three-dimensional structure of future network deployments, and complexity of in-building and inter-building radio propagation as key domains where further research is needed. We also discuss in detail potential approaches to be taken as starting points for new research in these domains.

Generic Multiuser Coordinated Beamforming for Underlay Spectrum Sharing

The beamforming techniques have been recently studied as possible enablers for underlay spectrum sharing. The existing beamforming techniques have several common limitations: they are usually system model specific, cannot operate with arbitrary number of transmit/receive antennas, and cannot serve arbitrary number of users. Moreover, the beamforming techniques for underlay spectrum sharing do not consider the interference originating from the incumbent primary system. This paper extends the common underlay sharing model by incorporating the interference originating from the incumbent system into generic combined beamforming design that can be applied on interference, broadcast, or multiple access channels. This paper proposes two novel multiuser beamforming algorithms for user fairness and sum rate maximization, utilizing newly derived convex optimization problems for transmit and receive beamformers calculation in a recursive optimization. Both beamforming algorithms provide efficient operation for the interference, broadcast, and multiple access channels, as well as for arbitrary number of antennas and secondary users in the system. Furthermore, this paper proposes a successive transmit/receive optimization approach that reduces the computational complexity of the proposed recursive algorithms. The results show that the proposed complexity reduction significantly improves the convergence rates and can facilitate their operation in scenarios, which require agile beamformer computation.

On the Stability of a Full-Duplex Aloha Network

This letter offers the first characterization of the stability for full-duplex (FD) large size networks. Through stochastic geometry tools, key performance metrics such as delay and maximum stable arrival rate are derived for the nonsaturated system and compared with a half-duplex counterpart, also accounting for imperfect self-interference cancellation. This analysis better identifies that the FD advantage is prominent for sparse networks, whereas in dense topologies, residual self-interference may hinder achievable gains.

LTE in Unlicensed Bands Is Neither Friend nor Foe to Wi-Fi

Proponents of deploying LTE in the 5 GHz band for providing additional cellular network capacity have claimed that LTE would be a better neighbour to Wi-Fi in the unlicensed band, than Wi-Fi is to itself. On the other side of the debate, the Wi-Fi community has objected that LTE would be highly detrimental to Wi-Fi network performance. However, there is a lack of transparent and systematic engineering evidence supporting the contradicting claims of the two camps, which is essential for ascertaining whether regulatory intervention is in fact required to protect the Wi-Fi incumbent from the new LTE entrant. To this end, we present a comprehensive coexistence study of Wi-Fi and LTE-in-unlicensed, surveying a large parameter space of coexistence mechanisms and a range of representative network densities and deployment scenarios. Our results show that, typically, harmonious coexistence between Wi-Fi and LTE is ensured by the large number of 5 GHz channels. For the worst-case scenario of forced co-channel operation, LTE is sometimes a better neighbour to Wi-Fi-when effective node density is low-but sometimes worse-when density is high. We find that distributed interference coordination is only necessary to prevent a “tragedy of the commons” in regimes where interference is very likely. We also show that in practice it does not make a difference to the incumbent what kind of coexistence mechanism is added to LTE-in-unlicensed, as long as one is in place. We therefore conclude that LTE is neither friend nor foe to Wi-Fi in the unlicensed bands in general. We submit that the systematic engineering analysis exemplified by our case study is a best-practice approach for supporting evidence-based rulemaking by the regulator.

Machine Learning for Performance Prediction in Mobile Cellular Networks

In this paper, we discuss the application of machine learning techniques for performance prediction problems in wireless networks. These problems often involve using existing measurement data to predict network performance where direct measurements are not available. We explore the performance of existing machine learning algorithms for these problems and propose a simple taxonomy of main problem categories. As an example, we use an extensive real-world drive test data set to show that classical machine learning methods such as Gaussian process regression, exponential smoothing of time series, and random forests can yield excellent prediction results. Applying these methods to the management of wireless mobile networks has the potential to significantly reduce operational costs while simultaneously improving user experience. We also discuss key challenges for future work, especially with the focus on practical deployment of machine learning techniques for performance prediction in mobile wireless networks.In this paper, we discuss the application of machine learning techniques for performance prediction problems in wireless networks. These problems often involve using existing measurement data to predict network performance where direct measurements are not available. We explore the performance of existing machine learning algorithms for these problems and propose a simple taxonomy of main problem categories. As an example, we use an extensive real-world drive test data set to show that classical machine learning methods such as Gaussian process regression, exponential smoothing of time series, and random forests can yield excellent prediction results. Applying these methods to the management of wireless mobile networks has the potential to significantly reduce operational costs while simultaneously improving user experience. We also discuss key challenges for future work, especially with the focus on practical deployment of machine learning techniques for performance prediction in mobile wireless networks.

Measurement procedures for design and enforcement of harm claim thresholds

Harm claim thresholds (HCTs) are a promising approach for regulators to specify interference limits in a technology-neutral fashion, and a useful parameter spectrum access systems can use to manage the aggregate interference caused by transmitters they control. However, existing literature provides very little guidance how HCTs should be set and enforced. In this paper we propose a detailed regulatory framework for gathering and processing of measurement data for enforcing and setting harm claim thresholds. We introduce the central concepts of stratification and weighting of measurement data, and show their importance in ensuring representativeness of measurements and enabling robust estimation of statistical confidence on results. For deriving HCT thresholds from measurements, we propose additional representativeness criteria that a regulator should apply to avoid underestimation of field strength levels related to existing wireless services. We demonstrate application of our proposed framework using an extensive drive test data set, and show that the chosen HCT percentile is critical in determining how much data needs to be gathered for enforcement. We also discuss the various design choices and parameters needed by our framework, and show through examples how they can be derived in cooperation with the different stakeholders. The developed framework has several applications beyond HCT enforcement, some of which are briefly described in the paper.

A Mean Field Analysis of CSMA/CA Throughput

Due to the fact that Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocols are extensively used in commercial wireless networks, the analysis of this family of medium access control protocols has been a topic of great interest for the wireless networking community. In this paper, we present a mean field analysis for the throughput of a slotted CSMA/CA scheme in single-hop networks. The concept of mean field approximation originates in statistical mechanics and it is a widely used approximation method employed to address systems of many interacting particles that cannot be solved exactly. The idea is to simplify the treatment of a many-body system by describing the interaction of one particle with the other interacting particles by an average potential, the so-called mean field. In our analysis, the nodes of a wireless network constitute a set of interacting particles; the interaction of a certain node with the other nodes in the network is reduced to an averaged channel condition. We estimate the amount of busy slots and the network throughput with respect to the offered traffic. Our analysis is of a recursive form, in which at each step the amount of traffic is increased by a specific amount. Also, our model takes into accounts retransmissions of collided packets.

When the whispers become noise: A contemporary look at radio noise levels

In this paper we present our results from our ongoing measurement campaign, which is targeting to provide information on the current radio noise levels. Our focus is to understand if the increasing number of users and devices with inbuilt transceivers has increased the noise levels considerably. In the literature it is almost universally assumed that the radio noise is a white Gaussian stochastic process and we are testing also how often deviations from this baseline are found. Our measurement approach is able to capture frequency and time domain data at very high accuracy. In this paper we report the first results from several different spatially separated measurement areas, enabling the study of noise levels at diverse indoor and outdoor locations. Our results show that the noise levels have indeed been increasing and we present a specific example on how these increased noise levels can affect todays telecommunications systems.

Large-Scale Cellular Network Modeling From Population Data: An Empirical Analysis

Accurate estimates of the spatial distribution of base stations (BSs) are crucial for the analysis of core performance and connectivity metrics in cellular networks. The appropriate tuning of these models, particularly through the fitting of point processes (PPs), enables practical simulations and theoretical analyses with realistic network configurations. However, the commonly made assumption of stationarity in model parameters does not hold for larger networks, with especially population inhomogeneities causing large network structure variations. We use BS location information from different operators in Germany, to explore the efficacy and limitations of using population data as a covariate to approximate cellular deployments. Our analysis shows that the overall network density is highly correlated already for small areas. Furthermore, we find that for moderately populated areas, purely population-driven PPs yield a good statistical match to our real data. The validity boundaries discussed in this letter provide a useful reference for determining at which length scales, more complex, interaction-based, PPs are necessary.