Federated learning is envisioned to be a key enabler of network functionalities based on artificial intelligence. Multiple access mechanisms supporting the learning task must then be designed, in order to provide an efficient interplay between the communication and computation resources. This work considers thus a multi-level slotted random access scheme autonomously optimised by each node. Due to their mutual coupling, the nodes’ interaction is an instance of the Best Response Dynamics (BRD) of a Generalised Nash Equilibrium Problem (GNEP). Within this framework, levers are identified, guaranteeing the convergence of the interactions to an equilibrium point at which the federated learning task is supported. These levers, on which the network manager can act, are validated by numerical simulations. These latter moreover show that the performance loss due to the autonomous character of the nodes is negligible with respect to the result of a centralised optimisation. On a broader mathematical level, this work defines a class of GNEPs for which sufficient convergence conditions for the totally asynchronous BRD are obtained. The considered class, named the GNEPs with polyhedral strategy sets and variable right-hand sides, encompasses a wide variety of GNEPs, and in particular GNEPs which are neither jointly convex nor generalised potential games. The obtained conditions depend on the first and second derivatives of the objective and constraint functions, and they constitute thus an off-the-shelf framework to study the BRD of GNEPs belonging to the identified class.

The use of deep learning (DL) has become more common in applications that require high accuracy, but the high computational demand of DL makes it challenging to execute on a single resource-constrained end device (ED) due to low computing capability and limited energy. One way to execute DL is to use the mobile cloud computing concept, but it results in intolerable latency, network congestion and raises privacy concerns. In the absence of powerful servers, another solution could be to form a collaboration among multiple EDs to execute DL computation. This paper proposes a distributed collaborative on-device convolutional neural network (CNN) execution scheme using model parallelism for latency-critical applications. In the proposed scheme, the ED that owns the input data uses its communication capabilities to get additional computing capabilities from other EDs. A convex optimization problem is formulated to minimize the energy consumption of all EDs by jointly optimizing the communication and computation parameters along with the number of filters assignment in each convolutional layer in the CNN model to multiple EDs. The problem is then decomposed into two sub-problems to obtain analytical expressions for the optimizing parameters and make the optimization distributed among multiple EDs. The simulation results show the importance of optimizing the communication-computation trade-off and the advantages of collaborative computing.

Wireless power transfer (WPT) is a possible alternative to batteries for supplying smart sensors. It is often described as “greener" because it avoids the use of batteries. However, this claim usually lacks the proper assessment of the environmental impacts of WPT and is particularly questionable given the poor efficiency of wirelessly transmitting power over a few meters of distance. In this paper, we design and thoroughly optimize a complete state-of-the-art WPT system and compare it to an equivalent battery-powered solution to assess whether it effectively represents an eco-friendly alternative. The proposed WPT system is a 2.45-GHz simultaneous wireless information and power transfer (SWIPT) system with beamforming and a high peak-to-average power ratio waveform for supplying a battery-less passive infrared sensor for room occupancy tracking. The final prototype of the smart sensor consumes 4and can operate in steady state at 5from the remote power head with an overall power transfer efficiency of 17/and can cold-start at 3.5. We carry out a life-cycle assessment (LCA) of the environmental impacts through four indicators, i.e., primary energy demand, global warming potential, terrestrial ecotoxicity, and freshwater consumption. Results show that for a 10-year lifetime, the WPT alternative has at least 5.5 to 10.3× higher environmental impacts than its battery-powered equivalent. This demonstrates the importance of LCA during the design of IoT smart sensors and shows that the use of meter-range 2.45¯/GHz SWIPT should be limited to applications where conventional battery-powered systems cannot be used.

Estimation of User Terminals’ (UTs’) Angle of Arrival (AoA) plays a significant role in the next generation of wireless systems. Due to high demands, energy efficiency concerns, and scarcity of available resources, it is pivotal how these resources are used. Installed antennas and their corresponding hardware at the Base Station (BS) are of these resources. In this paper, we address the problem of antenna selection to minimize Cramer-Rao Lower Bound (CRLB) of a planar antenna array when fewer antennas than total available antennas have to be used for a UT. First, the optimal antenna selection strategy to minimize the expected CRLB is proposed. Then, using this strategy as a preliminary step, we present a two-stage greedy antenna selection method whose goal is to minimize the instantaneous CRLB. The optimal start point of the greedy algorithm is presented alongside some methods to reduce the algorithm’s computational complexity. Numerical results confirm the accuracy of proposed solutions. They demonstrate that the proposed antenna selection method only requires a small proportion of the total available antennas to accomplish a significant amount of the total performance, enhancing hardware utilization efficiency. Also, it is shown that the presented algorithm has a high error tolerance.

In the next generation of wireless systems, massive Multiple Input Multiple Output (MIMO) offers high angular resolution for localization. By virtue of large number of antennas, Angle of Arrival (AoA) of User Terminals (UTs) can be estimated accurately. Measurements of Dense Multipath Channel (DMC) indicate multipath signals contribute up to 95% of total link power. This fact urges the necessity of studying the contribution of multipath signals accompanying dominant path in AoA estimation. We obtain a deterministic form for Cramer-Rao Lower Bound ($CRLB$) in multi-user scenario when contribution of multipath signals is considered. We do this when the multipath signals are independent and identically distributed (i.i.d) with arbitrary distribution. Then, we redefine a localization efficiency function for multi-user scenario and optimize it with respect to (w.r.t) the number of antennas. When only a subset of available antennas is used, we prove that$CRLB$can be minimized w.r.t which set is used. An antenna selection strategy that minimizes$CRLB$is proposed. As a benchmark, we apply the proposed antenna selection to MUltiple SIgnal Classification (MUSIC) algorithm and study its efficiency. Numerical results validate the accuracy of our analysis and show significant improvement in efficiency when the proposed antenna selection is employed.

Massive MIMO systems provide high angular resolution in the next generation of wireless systems. This opportunity can be used to estimate the location of user terminals (UTs) accurately. In this paper, we analyze the Cramer-Rao lower bound (CRLB) of planar antenna arrays in Massive MIMO systems for Angle of Arrival (AoA) estimation. With the help of Random Matrix Theory, we prove that for Massive antenna arrays with independent and identically distributed (i.i.d) multipath signals, instantaneous CRLB for AoA estimation converges toward a deterministic value, regardless of channel distribution. In this scenario, CRLB is a function of channel variance instead of instantaneous realizations. Then, antenna selection is studied, and it is shown that using different subsets significantly affects the CRLB of a planar array. Numerical results confirm the convergence of deterministic results and indicate the benefits of antenna selection and the importance of the selection strategy.

In this paper we analyze the performance of a mobile edge computing (MEC) system in the presence of spatially correlated shadowing. To achieve this goal, obstacles like walls are modelled by a Manhattan Poisson line Process (MPLP). Simultaneous wireless information and power transfer (SWIPT) is used to send command messages as well as to deliver the power to low power devices (LPDs). The energy harvested by the LPDs in the downlink phase is then used to either locally execute the assigned computation task or to offload it to the MEC server. By leveraging stochastic geometry, we derive analytical expressions for the success probability of the system which is defined as the joint probability of SINR downlink coverage and of harvesting enough energy to execute the computation task. Moreover, the impact of different system parameters on the trade-off between the harvested energy and the computational task admission rate is analyzed.

In this paper, we propose the use of simultaneous wireless information and power transfer (SWIPT) to control distributed computation process while delivering power to perform the computation tasks requested. A real-time mobile computing (MC) system is considered, meaning that the trade-off between the information rate and the energy harvested must be carefully chosen to guarantee that the central processing unit (CPU) may perform tasks of given complexity before receiving a new control signal. In order to provide a system-level perspective on the performance of such a SWIPT-MC network, we make use of stochastic geometry to characterise the rate-energy trade-off of the system. The resulting achievable performance region is then put in relation with the CPU energy consumption to investigate the operating conditions of real-time computing systems.

Simultaneous wireless information and power transfer (SWIPT) is characterized by the ambiguous role of multi-user interference. In short, the beneficial effect of multi-user interference on RF energy harvesting is obtained at the price of a reduced link capacity, thus originating nontrivial tradeoffs between the achievable information rate and the harvestable energy. Arguably, in indoor environments, this tradeoff might be affected by the propagation loss due to blockage objects like walls. Hence, a couple of fundamental questions arise. How much must the network elements be densified to counteract the blockage attenuation? Is blockage always detrimental on the achievable rate-energy tradeoff? In this paper, we analyze the performance of an indoor multiple-input multiple-output SWIPT-enabled network in the attempt to shed a light of those questions. The effects of the obstacles are examined with the help of a stochastic geometry approach in which the locations of energy sources (also referred to as power heads) are distributed by using a Poisson point process and walls are generated through a Manhattan Poisson line process. The stochastic behavior of the signal attenuation and the multi-user interference is studied to obtain the joint complementary cumulative distribution function of information rate and harvested power. Theoretical results are validated through Monte Carlo simulations. Eventually, the rate-energy tradeoff is presented as a function of the density of walls to emphasize the cross-dependences between the deployment of the network elements and the macro parameters characterizing the topology of the venue.

In this paper, we consider a fog computing system consisting of a multi-antenna access point (AP), an ultra-low power (ULP) single antenna device and a fog server. The ULP device is assumed to be capable of both energy harvesting (EH) and information decoding (ID) using a time-switching simultaneous wireless information and power transfer (SWIPT) scheme. The ULP device deploys the harvested energy for ID and either local computing or offloading the computations to the fog server depending on which strategy is most energy efficient. In this scenario, we optimize the time slots devoted to EH, ID and local computation as well as the time slot and power required for the offloading to minimize the energy cost of the ULP device. Numerical results are provided to study the effectiveness of the optimized fog computing system and the relevant challenges.

Equilibrium problems with vector objective functions may arise whenever uncoordinated wireless devices have to consider multiple conflicting requirements. In this paper we propose the use of a competitive optimality criterion to provide a systematic study of the relationship between the theoretical limits of information rate and energy efficiency for a group of transmitter-receiver pairs (TRPs) with local channel state information (CSI) only. To achieve this, we modeled the power control problem in a Gaussian interference channel as a competitive multiobjective game. After having provided the reader with some essential results on multiobjective games, we propose a game scalarisation based on the Quasi Variational Inequality (QVI) framework. This provides the mathematical tools to study the characteristics of the equilibrium points and to derive a power control distributed algorithm. Eventually, the proposed solution is validated by means of numerical results.

This paper investigates the optimal design of multiple-input-multiple-output (MIMO) precoders for orthogonal frequency-division multiplexing (OFDM) systems in the presence of nonlinear distortions due to the high power amplifier (HPA). In traditional approaches, the designer lets the input backoff (IBO) of the HPA be large enough to neglect the unavoidable nonlinear effects of the amplifier. However, this hypothesis may severely hinder the efficiency of the HPA and, consequently, the global energy efficiency of the link. On the other hand, making full use of the available power gives rise to unbearable nonlinear effects with consequent strong rate degradations. In this work, we take advantage of the Bussgang theorem to unveil the impact of the nonlinear distortion noise on the optimal precoding and to derive a power allocation algorithm that achieves the optimal tradeoff between HPA efficiency and rate degradation. We also provide a sufficient condition for the concavity of the information rate objective function in this nonlinear scenario. Finally, numerical results shows that the proposed algorithm outperforms conventional fixed-IBO precoding strategies.

This paper studies an energy efficient design of precoders for point-to-point multiple-input-multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) systems. Differently from traditional approaches, the optimal power allocation strategy is studied by modelling the circuit power as a rate-dependent function. We show that if the circuit power is a constant plus an increasing and convex function of the transmission rate, the problem of minimizing the consumed energy per bit received can be reformulated as a convex fractional program and solved by means of a bisection algorithm. The impact of the some system parameters is investigated either analytically or by means of computational results.

This paper presents an energy-efficient power allocation for relay-aided heterogeneous networks subject to coupling convex constraints, that make the problem at hand a generalized Nash equilibrium problem. The solution to the resource allocation problem is derived using a sequential penalty approach based on the advanced theory of quasi variational inequality, which allows the network to converge to its generalized Nash equilibrium in a distributed manner. The main feature of the proposed approach is its decomposability, which leads to a two-layer distributed algorithm with provable convergence.

In this work, we investigate the power allocation (PA) problem aimed at minimizing the users' packet error rate (PER) over a noncooperative link, i.e., a link where the set of users, employing packet-oriented bit-interleaved coded (BIC) orthogonal frequency division multiplexing (OFDM) systems, compete for the same bandwidth. For these kind of systems, the PER is not available in closed-form, but a very efficient solution is offered by the effective SNR mapping (ESM) technique. This method allows each user to evaluate a single scalar value, the effective SNR (ESNR), accounting for all the SNIRs experienced over the active subcarriers, and to univocally map it into a PER value. Thus, in order to derive a decentralized strategy allowing each user to minimize its own PER, the problem is described as a strategic game, called min-PER game, with the set of player, utilities and strategies represented by the competitive users, the ESNRs and the set of feasible power allocations, respectively. We will show both the existence of at least one Nash Equilibrium (NE) for the min-PER game and its equivalence with a Nonlinear Variational Inequality (NVI) problem. Finally, relying on the theory of contraction mappings, we will derive a distributed algorithm to reach the NE of the game.

This paper compares the energy efficiency of some cooperative MIMO schemes in wireless networks. By energy efficiency we denote the spectral efficiency seen at the receiver normalized by the total energy consumption, which includes the circuitry, the efficiency of the power amplifier, and the transmission rate. We focus on transmit antenna selection (TAS) and switch and stay combining (SSC) at the receiver. The performance of TAS+SSC is compared to that of TAS and maximal ratio combining (MRC), and to that of transmit/receive beamforming using a singular value decomposition (SVD) technique. We derive closed-form outage probability expressions, and analyze the effect of selecting the antenna from the source to optimize the communication with the relay or with the destination. Our results show that selecting the antennas with respect to the destination is in general a better option when the energy consumption is accounted. Moreover, some power allocation strategies are described and the comparison of the schemes in terms of energy efficiency reveals a considerable improvement with the use of TAS+SSC for low to moderate spectral efficiency, while beamforming outperforms the other schemes for high spectral efficiency when the number of antennas at each node is small.

In this paper, an approximate analytical expression for the peak-to-average power ratio (PAPR) of a Multi-Carrier Code Division Multiple Access (MC-CDMA) signal is derived. Then, it is demonstrated that the PAPR of a MC-CDMA system employing Walsh-Hadamard (WH) codes can be suitably reduced by resorting to a judicious strategy for the allocation of the spreading signature codes. Eventually, a low-complexity implementation of the proposed strategy is presented and the relevant benefits, in terms of peak signal occurrence and robustness against nonlinear distortions, is numerically assessed over conventional random allocation strategy.

An efficient and flexible air interface is the necessary condition for enabling high data rates next generation wireless systems. A practical answer to this need consists in exploiting channel state information at the transmitter for an adaptive and efficient use of the available radio link resources. This paper contributes with a novel strategy for the allocation of the sub-carrier transmit power in the case of a wireless system featuring: i) bit-interleaved coded modulation (BICM); ii) orthogonal frequency division multiplexing (OFDM); iii) multiple antenna profile (MIMO). The proposed power-allocation algorithm aims at maximizing the so-called goodput performance metric for packet transmissions employing automatic repeat request (ARQ) schemes. In addition, it is also proven to be near-optimum in maximizing the input/output mutual information of the MIMO-OFDM subchannels. Numerical simulations over typical wireless indoor scenarios confirm the effectiveness of the procedure when applied in the context of MIMO-BICM-OFDM systems.

In this paper, a novel pragmatic bit and power allocation algorithm for Non Orthogonal Frequency Division Multiplexing (NOFDM) systems is presented. It is based on the well-known Campello approach and takes into account the lack of orthogonality between adjacent waveforms. Simulation results in a typical NOFDM wireless applicative scenario show the effectiveness of the proposed algorithms.

Coded multi-antenna and multi-carrier techniques combined together with link resources adaptation algorithms are the key technologies toward efficient high-data-rate communications over wireless fading channels. The practicability of this concept, however, requires that the transmitter can perform accurate and simple evaluation of the actual link performance. This paper contributes with a novel method specifically developed to predict the link performance of bit-interleaved coded MIMO- OFDM links, which offers improved accuracy at the price of lower complexity when compared with conventional techniques. Its effectiveness is confirmed through extensive simulation results obtained over typical wireless channel environments.

In multicarrier (MC) broadband wireless networks, nonlinear power amplification is one of the main performance limiting factor, and therefore its effects have to be properly quantified during the design phase of the radio links. This letter outlines an analytical framework for the forward-link of a bandlimited MC code division multiple access that intends to fulfill this need while improving the scarce efficiency of conventional simulation-based-only design procedures. Through the extended Bussgang theorem, the approach presented herein yields a theoretical description of several time- and frequency-domain system features, such as transmitted signal power spectrum, constellation warping, nonlinear distortion noise and BER metric. Extensive simulation results verify the accuracy of the proposed analysis under typical multiple-access operating conditions.

This paper deals with the power allocation problem for coded multicarrier transmission. Specifically, we focus on a bit interleaved coded modulation (BICM) packet transmission implemented with orthogonal frequency division multiplexing (OFDM) and in the presence of automatic repeat request (ARQ) protocol. Capitalizing on the binary-input output-symmetric (BIOS) nature of the BICM channel it is provided a simple upper-bound of the rate of information bits received without any error, the so called goodput. Based on this theoretical characterization, we develop a power allocation strategy among the different subcarriers so that the system goodput performance metric is maximized. The effectiveness of the proposed method is numerically testified for BICM-OFDM transmission in the context of the typical WLAN scenario.

This paper contributes with an adaptive modulation scheme intended for bit interleaved coded (BIC) orthogonal frequency division multiplexing (OFDM) packet transmission in the presence of automatic repeat request (ARQ) protocol. The basic idea consists in properly combining the power allocation strategy with a bit-loading procedure among the different subcarriers so that the rate of information bits received without any error is maximized. Numerical simulations corroborate the effectiveness of the proposed scheme when operating over typical wireless environments.

An efficient and flexible air interface is the necessary condition to achieve the ambitious goals of next generation broadband wireless networks. Multicarrier code division multiple access (MC-CDMA) is a practical answer to this need, but the considerable vulnerability to nonlinear distortions induced by high power amplifiers at the transmitter side hinders its appealing features, thereby demanding for adequate countermeasures. The current paper contributes to this issue with a novel multiuser data predistortion (MUDP) procedure aimed at mitigating the performance degradation of a MC-CDMA forward-link in the presence of a nonlinear channel. The proposed MUDP resides entirely at the transmitter of the base station, and simply requires a set of modified constellations, which are jointly and iteratively optimized so that the distortions experienced by each active user are significantly reduced. The required computational complexity is then kept at affordable levels by resorting to a suboptimal yet effective scheme based on the identification of the strongest interfering users combined with the set-partitioning and the symmetry properties of their constellations. Numerical simulation results corroborate the effectiveness of the MUDP in improving the link performance under typical demanding multiple-access operating conditions.

This paper presents an analytical framework for the theoretical characterization and performance evaluation of an MC-CDMA forward-link in the presence of nonlinear distortions induced by transmitter's high-power amplifiers. The analytical characterization of the signal features, both in the time and frequency domains, is carried out through the application of the Bussgang theorem. Simulation results are derived to validate the accuracy of the proposed method under typical operating conditions