Uncrewed Aerial Vehicles (UAVs) are increasingly used in civilian and industrial applications, making secure low-altitude operations crucial. In dense mmWave environments, accurately classifying low-altitude UAVs as either inside authorized or restricted airspaces remains challenging, requiring models that handle complex propagation and signal variability. This paper proposes a deep learning model, referred to as CoBA, which stands for integrated Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Attention which leverages Fifth Generation (5G) millimeter-wave (mmWave) radio measurements to classify UAV operations in authorized and restricted airspaces at low altitude. The proposed CoBA model integrates convolutional, bidirectional recurrent, and attention layers to capture both spatial and temporal patterns in UAV radio measurements. To validate the model, a dedicated dataset is collected using the 5G mmWave network at TalTech, with controlled low altitude UAV flights in authorized and restricted scenarios. The model is evaluated against conventional ML models and a fingerprinting-based benchmark. Experimental results show that CoBA achieves superior accuracy, significantly outperforming all baseline models and demonstrating its potential for reliable and regulated UAV airspace monitoring.

Unmanned Aerial Vehicles (UAVs) can be used as low altitude platforms in several applications. In this paper, we propose their use to localize a ground Radio Frequency (RF) emitter by collecting measures of the Received Signal Strength Indicator (RSSI) at different positions. The main contribution of the work consists in the definition of an experimental setup for the simultaneous measures of RSSI and receiver position. The RSSI is measured by an actual transceiver, the Adalm Pluto Software Defined Radio (SDR) development board, programmed with the open-source software GNU Radio. The position is provided by GPS and Inertial Measuring Unit (IMU) sensors on the drone. The measures are acquired in the 865MHz Short Range Device (SRD) and 2.4 GHz Industrial Scientific Medical (ISM) unlicensed frequency bands. Since the ISM measures can be affected by interference generated by different sources (e.g. Wi-Fi access points and UAV controller), the SRD band is exploited for collecting the RSSI measures with less interference. A maximum likelihood (ML) algorithm is applied to the collected data for estimating the transmitter location. For the considered setup we show that the mean absolute localization error is around 4m without interference and 5m with interference. A threshold-based technique is proposed to improve the accuracy in presence of interference.

The high mobility of Unmanned Aerial Vehicles (UAVs) and their capability to rapidly deploy Aerial Base Stations (ABS) in areas where the terrestrial network becomes unavailable is a key enabler for Public Safety Networks. In our work we introduce a model in order to identify Line of Sight (LoS) and Non-Line of Sight (NLoS) conditions for User Equipments (UEs) that attempt a connection to an ABS through the Physical Random Access Channel (PRACH) based on Convolutional Neural Networks (CNNs). Our method limits the number of antennas employed with respect to other methods that were developed for traditional approaches, while achieving higher than 80% accuracy for SNR of -20 dB. Finally, we study the impact of UAV’s height on the accuracy of our method and we compare it with typical computationally efficient methods based on the delay spread with and without the aid of beamforming.

In order to leverage the spectrum resources, several forms of wireless duplex have been introduced and investigated in recent years. In Partial Duplex (PD) schemes, part of the band is transmitted in Full-Duplex (FD) and the rest in Half-Duplex (HD); therefore, some transmitted symbols will be characterized, at the receiver, by high SNR (Signal-to-Noise Ratio) and others by low SNR because of the residual self-interference (SI) in the FD part. Combining properly the patterns of these high and low SNR symbols affects the performance of the encoding schemes used in the system; in order to overcome this issue, different encoding and allocation schemes can be adopted for achieving a satisfactory solution. This paper investigates the performance of Low-Density Parity-Check (LDPC), turbo, polar codes for wireless PD. Orthogonal Frequency Division Multiplexing (OFDM) is an efficient multicarrier modulation technique, used in 4G and in the upcoming 5G, and it can be exploited for realizing a proper symbol allocation according to the SNR on each subcarrier. In this context, performance of LDPC, polar, and turbo codes derived from existing specifications has been studied when the system faces a mixture of high and low SNRs on the bits and hence on the symbols coming from the same codeword and this unbalanced SNR distribution is known a-priori at the transmitter, a condition associated with a scheme in which part of the symbols is subject to FD interference.

The 5thgeneration of cellular networks (5G) will provide high speed and high-availability wireless links for communication between mobile users. The usage of aerial platforms as base stations has been recently proposed to meet the above requirements, especially in densely-packed urban areas. To make an accurate prediction of the performance in such a communication system the availability of suitable channel models is a fundamental requirement. Here, we concentrate on a simple path loss and shadow fading channel model that is commonly used to describe the propagation between an aerial base station and a user on the ground. A commercial 3D ray-tracing simulator is used to extract the main parameters used in the model and the Line of Sight/Non Line of Sight probabilities as a function of the transmitter height and elevation angle. We consider three reference scenarios: Suburban, Urban and Urban High Rise generated according to ITU-R specifications. As a novel contribution, we also show simulation results for the spatial correlation of the received signal in the three considered scenarios.

Accurate positioning of vehicles and pedestrians is crucial for enhancing road safety. In this paper, we propose and compare two implementations based on Unscented Kalman Filter (UKF) and Particle Filter (PF) to perform trajectory estimation with sensor fusion. For the latter, a novel soft map-matching technique is applied on top of a PF. The main benefit of our method is the possibility of detecting reliably critical situations, like vehicles skidding off the road. Moreover, we can reduce the positioning error by 45% w.r.t. prior art approaches. Our solution can be implemented as a cloud service in the 5G mobile radio network.

One of the key challenges in next generation 5G networks is to deliver Ultra-Reliable Low-Latency Communications (URLLC). Recent advances in information theory about principles that govern short packet transmissions pointed out that, for the URLLC typical short packet dimension, achieving higher reliabilities comes at the price of a lower maximum achievable rate, thus reducing spectral efficiency. Hence, retransmissions are used in LTE and planned for 5G, in order to achieve reliability with a better resource consumption, at the price of increased packet latency. Keeping in mind the stringent requirements for URLLC, in this paper we analyze the tradeoffs and limitations of retransmission strategies considered in the literature, either too demanding in terms of wireless resources and aggressive URLLC performance or the contrary. Then we propose a novel scheme, whose purpose is to match the URLLC requirements, minimizing the resource consumption. We evaluate the schemes through simulations and highlight the advantages of the proposed scheme, providing also insights on the performance HARQ techniques in URLLC scenarios.

In the context of fingerprinting (FP) applications, this paper investigates the reduction of quantization levels in the Received Signal Strength Indicator (RSSI) till to its binary representation. One of the common drawbacks of FP is the large data size and consequently the large search space and computational load as a result of either vastness of the positioning area or the finer resolution in the FP grid map. This complexity can be limited reducing the RSSI quantization till to a simple binary indicator at the expense of an increased number of reference points or beacons. This approach turns out to be advantageous for the deployment of FP systems based on diffused beacons equipped with inexpensive technologies, such as Bluetooth Low Energy (BLE) or other technologies for the Internet of Things (IoT). An appropriate quantization and design of RSSI signatures will make possible the deployment of FP in larger areas maintaining the same computational load and/or the desired localization performance. The experimental results confirm promising computational savings without a relevant impact on the localization performance.

Low-density parity-check (LDPC) code is one of the fundamental error correction techniques used in wireless networks for achieving low probabilities of error. Wireless full-duplex (FD), according to its inherent nature, faces a considerable self-interference, which cannot be cancelled completely at the FD receivers. LDPC codes are excellent candidates for enabling full duplex in the presence of residual self-interference. This paper studies performance of a class of LDPC codes when the system faces a mixture of high and low bit SNR in the same codeword, condition associated with a scheme in which part of the bits is subject to FD interference (partialduplex scheme). In partial duplex (PD) schemes, part of the band is transmitted in FD and the rest in half-duplex and, consequently, some transmitted bits will be characterized by high SNR and the others by low SNR according to a pattern which is known, a-priori, by the system. Combining properly the patterns of these high and low SNR bits affects the coding performance of the system. In this study, we show the performance achieved by different strategies for allocating low SNR bits in the codeword of LDPC codes derived and adapted by the DVB-S2 standard.

This paper investigates the role of quantization in the Received Signal Strength Indicator (RSSI) information used for fingerprinting (FP) applications. One of the common drawbacks of FP is the large data size and consequently the large search space and computational load as a result of either vastness of the positioning area or the finer resolution in the FP grid map: this limits the application of FP to small environments or scenarios with largely spaced grid points leading to poor localization performance. We show that the computational complexity can be limited adapting the RSSI quantization w.r.t. the variance of the measured RSSI error at the target. This approach turns out to be advantageous for the deployment of FP systems based on beacons equipped with inexpensive technologies, in which the measures precision loss could be compensated by larger numbers of beacons. An appropriate quantization and design of RSSI signatures will make possible the deployment of FP in larger areas maintaining the same computational load and/or the desired localization performance.

This paper considers the interference mitigation problem in multi-cell networks based on Orthogonal Frequency-Division Multiple Access (OFDMA). The approaches studied here rely on allocation algorithms exploiting base stations cooperation achieved by means of inter-cell interference information sharing. Our emphasis is concentrated on the terminals at the cell edge since, obviously, they are more affected and limited by interference. The paper presents results on several schemes of cooperation (centralized, semi-distributed, coordinated-distributed) combined with allocation algorithms characterized by the awareness of interference relations among users, achieved sharing channel state information among base stations in a typical multi-cell layout. The interference relations are revealed by processing particular graphs, either for partial frequency reuse or soft frequency reuse schemes. So the network realizes, in several forms, a dynamic fractional frequency reuse and the study highlights the main advantages of the best strategies in terms of cooperation and resource allocation.

Ultra-Wideband (UWB) is an emerging technology for short-range wireless communications. Due to the high bandwidth of UWB signals, accurate ranging is possible, which is one of many reasons why UWB is a candidate physical layer for wireless sensor networks (WSNs). Reliable position information, obtained with a reasonable complexity overhead and possibly low energy consumption, is often paramount for a successful operation of these networks. The paper concerns a novel approach for improving the trade-off between energy consumption and performance in a localization tracking process. The scenario of application is common: a set of fixed beacons is used for tracking the position of a target that is moving in a limited indoor environment. The principle behind the proposed approach is relatively simple: tracking a small target device is realized by mixing active signal transmissions, which allow using standard techniques for deriving distances and locations, as well as passive signal receptions, which exploit scattering caused by small targets during signal propagation. The entire tracking process exploits the combination of these two types of transmissions with the advantage of possibly saving energy in the target device. The algorithm performance is highlighted by computer simulations using channel models adopted by the IEEE 802.15.4a working group.

Interference mitigation and adaptive multi-user resource allocation are among the most promising technological breakthroughs that should improve capacity of future broadband wireless communications. In a multi-cell scenario, where co-channel interference is the performance limiting factor, discrete-rate resource allocation algorithms, selecting those channels with best Signal-to-Interference and Noise-Ratio (SINR), can reduce the transmission power in each base station, acting indirectly on the co-channel interference levels. In this paper we investigate multi-cell systems in which power adaptation acts as a positive factor for the interference reduction. We are interested here in the convergence properties of the dynamic model of the system and in the role of the number and position of SINR thresholds of the discrete-rate profiles. We find that, in multi-cell scenarios, the application of power reduction makes the spectral efficiency of the system relatively insensitive to the number of discrete-rate profiles, once fixed the minimum and maximum SINR thresholds.

Adaptive multi-user resource allocation is one of the key features towards high speed wireless network based on Orthogonal Frequency Division Multiplexing Access (OFDMA). According to IEEE 802.16e (WiMAX) standard radio, resource allocation problem has to be performed on a two-dimensional space (frequency and time) with the constraint that each transmitted burst should occupy a contiguous portion of the frame. The paper introduces two different strategies and it investigates the advantages, drawbacks and challenges of the radio resource allocation procedure in two dimensions.

Time hopping is the strategy employed in impulse radio systems to increase the communication reliability by spreading the multiple access interference (MAI) over numerous time intervals. This results in a non-stationary interference that makes the error probability performance to be dominated by the worst-case. In this paper we employ a collision model analysis for this time-varying interference that is ruled by the system loading in term of the corresponding MAI collision probability. The model proves to be simple and accurate to characterize the decision variable conditioned to the collision events. This provides a reliable framework for evaluating the error probability in multipath faded channels with RAKE receiver either for coded and uncoded systems. In this paper we derive closed form approximations of the bit-error-rate for multipath channels with exponential power delay profile and Nakagami-m fading. Validations by numerical analysis show that the collision model provides an accurate framework to evaluate also performance of realistic channel models.

In the context of UWB (ultra-wide band) impulse radio (IR) heterogeneous low data-rate networks, the need for coding/modulation schemes amenable of both coherent and noncoherent detection has emerged. The rationale is the potential coexistence of different classes of nodes, characterized by different complexity and performances. In such scenario, one of the main technical challenges is the design of efficient coding/modulation schemes incorporating forward error correction (FEC) and spreading capabilities, thus providing the necessary processing and coding gain, to guarantee robust communications. In this work, we introduce novel coding schemes exploiting the trellis structure of orthogonal convolutional codes for generating time-hopping sequences suitable for UWB-IR, while simultaneously mapping the information bits onto the transmitted waveforms. Namely, we present a new class of orthogonal convolutional modulation (OCM) schemes, highlighting the trade-offs in terms of coherent vs. non-coherent detection performance. Then we develop an adaptive OCM scheme (AOCM) for application to broadcast transmission in heterogeneous UWB-IR networks, illustrating advantages in terms of energy savings and potential drawbacks.

Adaptive multi-user resource allocation is one of the most promising technological breakthroughs that could improve performance of future broadband wireless communications. In a multi-cell scenario, where co-channel interference is the performance limiting factor, resource allocation algorithms select those channels with best Signal-to Interference and Noise-Ratio (SINR) and adapt transmission power, acting indirectly on the interference levels. The paper introduces an analytical procedure for evaluating how the distribution of co-channel interference is modified by different resource allocation algorithms. This procedure can be used for deriving a model of interference or for providing a deeper insight on the performance limits of the system.

IEEE 802.16 is an emerging technology for broadband cellular wireless communications. In the context of high capacity systems for mobile terminals, implementation issues like channel estimation become more challenging because of the sparsity of the pilots grid (w.r.t. the channel coherence bandwidth) and the time varying nature of the channel. This paper proposes a novel algorithm for channel estimation that is characterized by a good trade-off between complexity and performance. The algorithm, that is suitable for pilot-aided OFDM transmission, is applied here to IEEE 802.16e systems. Numerical results show that a meaningful performance gain (in terms of mean squared error and error probability) is provided by this estimation procedure, especially in the low SNR (Signal to Noise Ratio) region.

In this letter, we discuss the problem of positioning an unknown number of amplify-and-forward relays or point scatterers using a wireless network of wideband transceivers. We propose to model this positioning problem as an assignment problem. In doing so, we avoid the problems associated with an exhaustive gridsearch approach, which has recently been proposed to solve the problem. Numerical simulations show promising results in terms of performance and robustness.

In this invited paper, we discuss positioning of road-users in a traffic environment. Once accurate and reliable position information can be made available, many interesting and innovative telematics applications may be realized. Some of the problems facing radio-frequency (RF) based positioning systems today include blocked RF channels and multipath propagation, complicating distance estimation between network members and between network members and the infrastructure. In this work, we propose a novel estimator of range operating on an ultra-wideband (UWB) physical layer. This distance estimator has been designed specifically with a novel type of positioning algorithm in mind, based on the method of projections onto convex sets. Also, we investigate the feasibility of positioning users that are not members of the telematics network based on arbitrary transmitted RF signals. Numerical simulations, using channel models recently adopted by the IEEE 802.15.4a working group, confirm the robustness and accuracy of the proposed approach in the presence of both non-line-of-sight (NLOS) channels, and severe multipath propagation

IEEE 802.16 (WiMAX) is an emerging technology for medium-range wireless communications. Among the technological breakthroughs that could improve its performance, adaptive multi-user resource allocation is one of the most promising. The paper introduces a novel algorithm and provides an insight into some aspects and design issues related to this application. Numerical results rely on the channel models adopted by IEEE 802.16 and highlight the advantages and drawbacks of the selected strategies.

In presence of multiple access interference, the performance of impulse radio system is affected by collisions with other users. In this paper we evaluate the bit error probability by employing a collision model for describing the non-stationary interference caused by simultaneous transmissions in time hopping systems. The collision model provides a simple approximation for the error probability even in multipath channels since the decision variable can be conditioned to the collision events. Numerical validations of the error probability for several propagation settings (e.g., over Nakagami-m fading channels or realistic scenarios) employing the collision model show that the proposed method is simple and accurate for most practical applications.

In the context of UWB (ultra-wideband) impulse radio, one of the challenges, for both high and low data-rate applications, is the design of an efficient coding/modulation scheme incorporating forward error correction (FEC) and spreading capabilities. In this paper, we introduce encoding strategies exploiting the trellis structure of orthogonal convolutional codes for generating time-hopping sequences suitable for UWB impulse radio and, at the same time, for mapping the information bits onto the transmitted waveforms. It is shown that this solution improves the Euclidean distance among transmitted signals, without decreasing the corresponding information rate.

UWB (ultra wide band) is an emerging technology for short-range wireless communications, such as WPAN (wireless personal area networks). In the context of UWB impulse radio, one of the crucial challenges remains the acquisition of synchronization in terms of time, precision and number of operations. In this paper, we present a study on the acquisition time for hybrid strategies; the analysis relies on common assumptions on the fading statistics and the multipath intensity profile, while the simulations consider the complete channel models adopted by IEEE 802.15.4a. The analysis and simulations highlight the advantages and drawbacks of this type of search strategies

Ultrawideband (UWB) impulse radio is an emerging technology suitable for high-rate tactical wireless communications. One of the crucial challenges for a connecting station remains the initial code acquisition, in a hostile propagation environment (e.g., urban combat). In this paper, we address the coarse acquisition of pseudonoise (PN) codes and propose algorithms for speeding up the acquisition process and/or reducing the complexity of the acquisition algorithm itself. Also, in the case of energy detection, we show that the front-end sampling rate may be reduced. An in-depth analysis, supported by simulations in the presence of multipath is presented, and the results discussed.

Turbo equalization is an iterative equalization decoding scheme for detecting encoded data transmitted over channels that introduce intersymbol interference. In such a scheme, the complexity of the optimal soft-input soft-output (SISO) equalizer, which implements the BCJR a posteriori probability computation algorithm, is one of the major concerns. We extend the reduced state sequence detection (RSSD) algorithm to a BCJR-SISO equalizer matched to a non binary modulation. The resulting reduced-state SISO (RS-SISO) algorithms can take full advantage from the flexibility offered by RSSD in reducing the number of trellis states. However, extension of RSSD algorithm for obtaining the RS-SISO equalizers requires attention in the recombination procedure of the forward and backward metrics.

It is well known that in high density magnetic recording channels noise models should comprise the contributions due to the non-ideal transitions. A simplified model of 'transition noise' includes the first order Taylor's expansion of the function that represents the ideal response in order to include the effects of transition (position) and width jitters. In this work we present an analysis that provides an evaluation of the probability density function of the noise projection over a generic error event. This procedure is used to evaluate the probability density functions of the reliability measures obtainable by a soft detection as max-log-MAP (maximum a posteriori) or SOVA (soft output Viterbi algorithm) applied to a magnetic channel.

SISO (soft input soft output) algorithms are present in several decoding schemes, particularly in concatenated ones that admit an efficient iterative detection, e.g. turbo codes. These algorithms have been an interesting research subject, especially for the aspects related to complexity reduction. This paper presents the idea of partitioning the set of original states in a channel or a code, generating a scaled (or reduced) trellis. Merging two or more states produces a new super-state and a standard SISO algorithm can follow the system evolution in terms of the new description.

High-density magnetic recording systems require increasingly sophisticated signal processing techniques. In magnetic recording channels, the information bits are encoded by the concatenation of an outer nonbinary error correcting code (ECC) and an inner line code. Furthermore, the high intersymbol interference that characterizes the channel is controlled by "partial response" equalization. This paper presents a study on a better-integrated decoding procedure between the inner and outer codes. Inversion of the concatenated codes (reverse concatenation) allows the direct mapping of soft information from the partial response channel to the outer ECC. A simplified soft decoding technique, based on the use of erasures, is applied to two typical magnetic recording systems. Performance curves are obtained by an analysis of the distributions of the reliability measures associated with incorrect and correct symbols.