Mobile application classification is essential for advanced network management and application-based QoS policy enforcement in future, AI-enhanced, beyond 5G and 6G mobile networks. This article proposes to use AI methods to categorize applications as functional types (e.g., Video, Audio, Browsing) despite encryption and limited labeled data. We tackle these challenges through unsupervised representation learning, which maximizes the use of abundant unlabeled data in mobile networks. Due to the distributed nature of beyond 5G and 6G networks, we use this method in federated learning scenarios and compare it to the centralized ones. Our findings highlight that unsupervised learning improves model performance, especially with scarce labeled data. Additionally, federated learning provides effective results as compared to centralized methods.

Mobile networks are rapidly expanding, and there is an increasing demand for seamless connectivity. Detecting whether a user is indoors or outdoors is pivotal in optimizing network performance and enhancing user experience. This paper proposes a semi-supervised learning method using a Long Short-Term Memory Autoencoder (LSTM-AE) that detects the user's environment. It uses mobile network radio signal data of real users. The LSTM Autoencoder learns to capture the underlying structure of the data and identify patterns that distinguish indoor from outdoor environments. Three key features are used to train the model: Reference Signal Received Power (RSRP), Channel Quality Indicator (CQI), and Timing Advance (TA). Results show that the LSTM-AE model achieves a high accuracy of 84% and an F1 score of 89%. In our approach, we achieve a substantial reduction of 34.14% in the requirement for labeled data compared to traditional methods that primarily rely on fully supervised learning. By diminishing the dependence on labour-intensive and time-consuming data labelling processes, this improvement significantly enhances the overall efficiency of the machine-learning process.

Resilience is defined as the ability of a network to resist, adapt to and quickly bounce back from disruptions, to continue keeping an acceptable level of service from users’ perspective. To characterize and analyse the resilience of a 5G Radio Access Network, it is vital to be able to measure current and prospective resiliency levels using relevant metrics. In this work, we perform an analysis and a quantification of 5G-RAN resilience based on radio coverage indicator and one of the Quality of Service indicators, namely the service integrity. They are considered as the main performance indicators for network planners and operators. In this objective, we model network states using the performance indicators, Reference Signal Received Power and Received User Equipment Throughput. Then, we use them to build Markovian models and to define resilience metrics. By means of simulations, we evaluate the ability of our analytical model to accurately capture network resiliency levels. We discuss numerical results from multiple usage scenarios to show how the resilience metrics can be used to identify potential outages and to help network planners or operators deciding on which adequate adaptive resilient mechanisms to deploy.

In sensing-enabled mobile infrastructure, the network itself acts as a whole sensor by leveraging radio data or signals collected within Base Stations (BSs). This data is exploited for the development of data-driven machine learning solutions to augment network’s capabilities. Nevertheless, large-scale qualitative data is required for achieving high accuracy learning. However, their training phase leads to prohibitive cost and heavy constraints on data collection and storage that are not desirable for network. To overcome this problem, we propose to use synthetic data instead of real data for training machine learning models to avoid high cost data sharing/storage. In this paper, we are interested in real-life Environment Sensing Network in a context of limited data amount sharing. We focus on Indoor-Outdoor Detection (IOD) using unsupervised machine learning classification models. For this purpose, experiments are conducted following the paradigm of Training on Synthetic data and Testing on Real Data (TSTR). We conduct a comparative study of four well-known generative models, that are able to generate synthetic 3GPP radio data with similar distribution than the source data. We investigate the quality of these synthetic generated radio data according to three dimensions: distribution similarity, data variability and detection capability. The classification models trained with synthetic generated data are tested in real-life context to infer whether a user connected to the network is inside or outside a building. The study shows convincing results with an Indoor/Outdoor unsupervised classification performance up to 80% of F1 − score like in real-life data training scenarios.

This paper demonstrates the feasibility of an hybrid/semi-supervised classification method for detecting the environment of an active mobile phone, based on both labeled and unlabeled cellular radio data. Precisely, we provide answers to the following question: what is the environment of the mobile user when it is/was experiencing a mobile service/application: indoor or outdoor? Implementing this method within the mobile network is interesting for mobile operators since it has low complexity, is less human intrusive (minimal intervention of mobile users) and more accurate. The semi-supervised classification algorithm learns to identify the environment using large and real collected 3GPP signals measurements. As compared to existing work, in addition to existing parameters used for classification, we propose to also use a radio metric called Timing Advance. It is computed within the mobile network. We empirically validate the innovative semi-supervised algorithm using new real-time radio measurements, with partial ground truth information, gathered daily, weekly, monthly, from indoor and outdoor locations and from multiple typical and diversified environments crossed by mobile users. The study confirms the effectiveness of the proposed scheme compared to the existing supervised classification methods including SVM and Deep Learning.

Future mobile networks promise to be more intelligent and to guarantee a better service for users. This intelligence can be highly accentuated by cognition of mobile users' behavior and conditions. The user speed is an important element of user profile. We are interested in real time speed profiling by detecting speed range of an active user. We refer to this as Mobile Speed Profiling (MSP) of users. Indeed, performing MSP can notably improve the self-adaptation and self-optimization capabilities of these networks. It can help mobile network in resource management and handover management. This paper introduces a Deep Learning based solution to automatically construct the MSP of a mobile user. We empirically evaluate the effectiveness of our approach using real-time and highly representative radio data that best captures the real daily movements of users. This data includes ground truth information and the whole dataset has been gathered massively from many diversified mobility situations. Results show that the profiling of UE's speed with fine granularity or on multiple ranges can be achieved with high accuracy on real data measured in heterogeneous deployments for 5G networks.

Detecting whether a mobile user is indoor or outdoor is an important issue which significantly impacts user behavior contextualization and mobile network resource management. Indoor Outdoor Detection (IOD) can be performed within mobile networks or in user terminals. Implementing IOD in mobile networks is interesting for operators as it is less costly, easier to deploy, and more energy efficient with centralized computing. This paper investigates hybrid/semi-supervised Deep Learning-based methods for detecting the environment of an active mobile phone user. They are based on both labeled and unlabeled large real radio data obtained from inside the network and from 3GPP signal measurements. We empirically evaluate the effectiveness of the semi-supervised learning methods using new real-time radio data, with partial ground truth information, gathered massively from multiple typical and diversified locations (indoor and outdoor) of mobile users. We also present an analysis of such schemes as compared to the existing supervised classification methods including SVM and Deep Learning.

Cognition of user behavior can be seen as an efficient tool for automation of future mobile networks. As a matter of fact, it amplifies the intelligence of autonomic networks in a sense that the network is more aware of the operational context. However, predicting the context of mobile users is a prerequisite for inferring the user behavior. This paper deals with the user behaviour modeling. The model includes the prediction of two main features related to mobile user context: the environment and the mobility. Practically, the question is how and where the mobile user consumes the mobile services. We investigate Deep Learning based methods for simultaneously detecting the environment and the mobility state. We empirically evaluate the effectiveness of the proposed methods using real-time radio data, which has been massively gathered from multiple diversified situations of mobile users.

This paper presents a novel example-based single-image superresolution procedure that upscales to high-resolution (HR) a given low-resolution (LR) input image without relying on an external dictionary of image examples. The dictionary instead is built from the LR input image itself, by generating a double pyramid of recursively scaled, and subsequently interpolated, images, from which self-examples are extracted. The upscaling procedure is multipass, i.e., the output image is constructed by means of gradual increases, and consists in learning special linear mapping functions on this double pyramid, as many as the number of patches in the current image to upscale. More precisely, for each LR patch, similar self-examples are found, and, because of them, a linear function is learned to directly map it into its HR version. Iterative back projection is also employed to ensure consistency at each pass of the procedure. Extensive experiments and comparisons with other state-of-the-art methods, based both on external and internal dictionaries, show that our algorithm can produce visually pleasant upscalings, with sharp edges and well reconstructed details. Moreover, when considering objective metrics, such as Peak signal-to-noise ratio and Structural similarity, our method turns out to give the best performance.

This paper presents a new method to construct a dictionary for example-based super-resolution (SR) algorithms. Example-based SR relies on a dictionary of correspondences of low-resolution (LR) and high-resolution (HR) patches. Having a fixed, prebuilt, dictionary, allows to speed up the SR process; however, in order to perform well in most cases, we need to have big dictionaries with a large variety of patches. Moreover, LR and HR patches often are not coherent, i.e. local LR neighborhoods are not preserved in the HR space. Our designed dictionary learning method takes as input a large dictionary and gives as an output a dictionary with a “sustainable” size, yet presenting comparable or even better performance. It firstly consists of a partitioning process, done according to a joint k-means procedure, which enforces the coherence between LR and HR patches by discarding those pairs for which we do not find a common cluster. Secondly, the clustered dictionary is used to extract some salient patches that will form the output set.

This paper presents a new nonnegative dictionary learning method, to decompose an input data matrix into a dictionary of nonnegative atoms, and a representation matrix with a strict ℓ0-sparsity constraint. This constraint makes each input vector representable by a limited combination of atoms. The proposed method consists of two steps which are alternatively iterated: a sparse coding and a dictionary update stage. As for the dictionary update, an original method is proposed, which we call K-WEB, as it involves the computation of k WEighted Barycenters. The so designed algorithm is shown to outperform other methods in the literature that address the same learning problem, in different applications, and both with synthetic and “real” data, i.e. coming from natural images.

In this paper we present a super-resolution (SR) method for upscaling low-resolution (LR) video sequences, that relies on the presence of periodic high-resolution (HR) key frames, and validate it in the context of video compression. For a given LR intermediate frame, the HR details are retrieved patch-by-patch by taking sparse linear combinations of patches found in the neighbor key frames. The performance of the video SR algorithm is assessed in a scheme where only some key frames from an original HR sequence are directly encoded; the remaining intermediate frames are down-sampled to LR and encoded as well, with a possibly different quantization parameter. SR is then finally employed to upscale these frames. For comparison, we consider the best case where the whole original HR sequence is encoded. With respect to this case, our SR-based approach is shown to bring a certain gain for low bit-rates (consistent when all frames are encoded independently), i.e. when a poor encoding can actually benefit of the special processing of the intermediate frames, so proving that video SR can be an useful tool in realistic scenarios.

International audience

This paper describes a novel method for single-image super-resolution (SR) based on a neighbor embedding technique which uses Semi-Nonnegative Matrix Factorization (SNMF). Each low-resolution (LR) input patch is approximated by a linear combination of nearest neighbors taken from a dictionary. This dictionary stores low-resolution and corresponding high-resolution (HR) patches taken from natural images and is thus used to infer the HR details of the super-resolved image. The entire neighbor embedding procedure is carried out in a feature space. Features which are either the gradient values of the pixels or the mean-subtracted luminance values are extracted from the LR input patches, and from the LR and HR patches stored in the dictionary. The algorithm thus searches for the K nearest neighbors of the feature vector of the LR input patch and then computes the weights for approximating the input feature vector. The use of SNMF for computing the weights of the linear approximation is shown to have a more stable behavior than the use of LLE and lead to significantly higher PSNR values for the super-resolved images.

The Digital Video Broadcasting Satellite services to Handhelds (DVB-SH) is a hybrid satellite/terrestrial system for the broadcasting of multimedia services to mobile receivers. To mitigate deep signal fading events occurring mainly in the land-mobile satellite channel, a Multi-Protocol-Encapsulation Inter-burst Forward Error Correction (MPE-IFEC) technique has been recently introduced. It is based on a clever organization of data and parities offering long interleaving. Two decoding modes have been proposed: the late decoding mode and the early decoding mode. In this paper, we propose to discuss and to analyse the impact of these decoding strate-gies on the channel change delay, known as the key issue of DVB-SH system. Based on an end-to-end DVB-SH network simulator, two criteria: the channel change delay and the cu-mulative freeze time, are computed by considering different MPE-IFEC parameters and typical DVB-SH propagation environments. These two criteria are sufficient to give us the subjective video quality of the newly joined channel.

The DVB-SH system relies on a hybrid satellite/terrestrial infrastructure. A complementary solution using the interworking with a bi-directional path offered by cellular network is under investigation in DVB-Convergence of Broadcast and Mobile Services Group. This aims to provide good QoS of DVB-SH service anywhere anytime. Due to the strict reception conditions and the severe satellite channel propagation, many time-sliced bursts are lost impacting negatively the subjective video quality because the new link layer : inter-burst FEC, fails to recover the burst losses even with an adequate coding rate. In this paper, we propose a repair mechanism and a suitable architecture for real-time streaming service error handling. By recovering only some specific lost information of the broadcasted stream, the quality of TV service is improved.

The improved convergence speed and tracking properties of fractionally-spaced equalizers are analyzed. We consider in particular the effect of a frequency offset between the transmitter baud rate and the receiver sampling clock that induces important time-variations. We show that a fractionally-spaced equalizer can handle the intersymbol interference (ISI) induced when the propagation channel does not introduce too much ISI.