In this paper we propose a highly efficient and very accurate deep learning method for estimating the propagation pathloss from point x (transmitter location) to any y on planar domain. For applications such as user-cell site association device-to-device link scheduling, an knowledge of function all pairs transmitter-receiver locations is important. Commonly used statistical models approximate decaying distance between transmitter receiver. However, in realistic environments characterized by...

Global Navigation Satellite Systems typically perform poorly in urban environments, where the likelihood of line-of-sight conditions between devices and satellites is low. Therefore, alternative location methods are required to achieve good accuracy. We present LocUNet: A convolutional, end-to-end trained neural network (NN) for localization task, which able estimate position a user from received signal strength (RSS) small number Base Stations (BS). Using estimations pathloss radio maps BSs...

We consider a cache-aided wireless device-to-device (D2D) network of the type introduced by Ji et al., where placement phase is orchestrated central server. assume that devices' caches are filled with uncoded data, and whole content database contained in collection caches. After cache phase, files requested users serviced inter-device multicast communication. For such system setting, we provide exact characterization optimal load-memory trade-off under assumptions one-shot delivery. In...

To encourage further research and to facilitate fair comparisons in the development of deep learning-based radio propagation models, less explored case directional signal emissions indoor environments, we have launched ICASSP 2025 First Indoor Pathloss Radio Map Prediction Challenge. This overview paper describes path loss prediction problem, datasets used, Challenge tasks, evaluation methodology. Finally, results a summary submitted methods are presented.

In this paper we propose a highly efficient and very accurate method for estimating the propagation pathloss from point x to all points y on 2D plane. Our method, termed RadioUNet, is deep neural network. For applications such as user-cell site association device-to-device (D2D) link scheduling, an knowledge of function pairs locations important. Commonly used statistical models approximate decaying distance between points. However, in realistic environments characterized by presence...

This paper deals with the problem of localization in a cellular network dense urban scenario. Global Navigation Satellite Systems (GNSS) typically perform poorly environments, where likelihood line-of-sight conditions is low, and thus alternative methods are required for good accuracy. We present LocUNet: A deep learning method localization, based merely on Received Signal Strength (RSS) from Base Stations (BSs), which does not require any increase computation complexity at user devices...

In this article, we present a collection of radio map datasets in dense urban setting, which generated and made publicly available. The include simulated pathloss/received signal strength (RSS) time arrival (ToA) maps over large realistic setting real city maps. two main applications the presented dataset are 1) learning methods that predict pathloss from input (namely, deep learning-based simulations), and, 2) wireless localization. fact RSS ToA computed by same simulations allows for fair...

We consider an extension of the information bottleneck problem where underlying Markov Chain is X-o-(Y, S)-o-Z, and P <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">X</sub> ,S,Y = xmlns:xlink="http://www.w3.org/1999/xlink">S</sub> xmlns:xlink="http://www.w3.org/1999/xlink">Y|X,S</sub> joint distribution a source X, channel state S independent source, output Y state-dependent channel. For case SX + N with Gaussian circularly symmetric, we...

This paper considers the problem of recovering a k-sparse, N-dimensional complex signal from Fourier magnitude measurements. It proposes optics setup such that recovery up to global phase factor is possible with very high probability whenever M ≳ 4k log <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> (N/k) random intensity measurements are available. The proposed algorithm comprised two stages: An algebraic retrieval stage and...

This paper considers the problem of signal recovery from magnitude measurements for signals in modulation invariant spaces. It proposes a measurement setup such that almost every space can be reconstructed its amplitude up to global constant phase and with sampling rate four times innovation space. The applicability proposed scheme under noise is demonstrated by computer simulations.

We apply the coded caching scheme proposed by Maddah-Ali and Niesen to a multipoint multicasting video paradigm. Partially files on wireless devices provides an opportunity decrease data traffic load in peak hours via sending multicast messages users. In this paper, we propose two-hop network for multicasting, where common message is transmitted through different single antenna Edge Nodes (ENs) multiple Each user can decide decode any EN using zero forcing receiver. Motivated Scalable Video...

This paper deals with the problem of localization in a cellular network dense urban scenario. Global Navigation Satellite System typically performs poorly environments when there is no line-of-sight between devices and satellites, thus alternative methods are often required. We present simple yet effective method for based on pathloss. In our approach, user to be localized reports received signal strength from set base stations known locations. For each station we have good approximation...

The efficient deployment and operation of any wireless communication ecosystem rely on knowledge the received signal quality over target coverage area. This is typically acquired through radio propagation solvers, which however suffer from intrinsic well-known performance limitations. article provides a primer how integrating deep learning conventional modeling techniques can enhance multiple vital facets network operation, yield benefits in terms efficiency reliability. By highlighting...

Pathloss quantifies the reduction in power density of a signal radiated from transmitter. The attenuation is due to large-scale effects such as free-space propagation loss and interactions (e.g., penetration, reflection, diffraction) with objects buildings, vehicles, trees, pedestrians environment. Many current or planned wireless communications applications require knowledge (or reliable approximation) pathloss on dense grid (radio map) environment interest. Deterministic simulation methods...

To foster research and facilitate fair comparisons among recently proposed pathloss radio map prediction methods, we have launched the ICASSP 2023 First Pathloss Radio Map Prediction Challenge. In this short overview paper, briefly describe problem, provided datasets, challenge task evaluation methodology. Finally, present results of challenge.

In this paper, we study the localization problem in dense urban settings. such environments, Global Navigation Satellite Systems fail to provide good accuracy due low likelihood of line-of-sight (LOS) links between receiver (Rx) be located and satellites, presence obstacles like buildings. Thus, one has resort other technologies, which can reliably operate under non-line-of-sight (NLOS) conditions. Recently, proposed a Received Signal Strength (RSS) fingerprint convolutional neural...

Global Navigation Satellite Systems typically perform poorly in urban environments, where the likelihood of line-of-sight conditions between devices and satellites is low. Therefore, alternative location methods are required to achieve good accuracy. We present LocUNet: A convolutional, end-to-end trained neural network (NN) for localization task, which able estimate position a user from received signal strength (RSS) small number Base Stations (BS). Using estimations pathloss radio maps BSs...

We apply the coded caching scheme proposed by Maddah-Ali and Niesen to a multipoint multicasting video paradigm. Partially files on wireless devices provides an opportunity decrease data traffic load in peak hours via sending multicast messages users. In this paper, we propose two-hop network for multicasting, where common message is transmitted through different single antenna Edge Nodes (ENs) multiple Each user can decide decode any EN using zero forcing receiver. Motivated Scalable Video...

This contribution proposes a two stage strategy to allow for phase retrieval in state of the art sub-Nyquist sampling schemes sparse multiband signals. The proposed is based on data acquisition via modulated wideband converters known from sampling. paper describes how modulators have be modified such that signal recovery amplitude samples becomes possible and corresponding algorithm given which computational efficient. In addition, fairly general, allowing several constructions algorithms.