In this paper, trellis coded modulation (TCM) is optimized to achieve better bit error ratio (BER) performance by training a neural network-based encoder. Specifically, two blocks, i.e., the mapper and constellation modulator, are jointly optimized, while convolutional coding and Viterbi decoder are maintained in the traditional manner. For the purpose of numerical optimization, a BER upper bound expression is derived as a function of Euclidean distance, and a loss function is designed based on the upper bound expression. The optimized TCM system can achieve higher coding gain and shaping gain, and can be implemented using the traditional encoder and decoder structure. Notably, the proposed scheme supports arbitrary code lengths with guaranteed scalability and manageable computational complexity. Simulation results demonstrate that the proposed scheme outperforms not only the traditional TCM with symmetric constellations, but also other TCM optimization schemes with asymmetric constellations.

Full-duplex technology can improve bandwidth and energy efficiency but has serious self-interference. In this paper, we introduce a novel full-duplex communication scheme that leverages the Doppler effect. By rotating the circular antenna array, the Doppler frequency shift is introduced into the received uplink (UL) signal, and the UL signal can be separated from the downlink (DL) signal in the frequency domain, thus eliminating self-interference. In order to maximize the interference-free bandwidth, an antenna switching criterion to maximum frequency offset interval is proposed for antenna switching control. Moreover, an antenna switching module based on the above criterion and a Doppler frequency shift compensation module are designed. Simulation results show that the proposed scheme can effectively avoid interference between UL and DL signals and achieves a substantial capacity improvement compared to conventional systems.

In this letter, we introduce a novel transmission scheme to a multi-input multi-output (MIMO) system employing an M × M antenna structure. The conventional MIMO loads parallel signals onto all transmit antennas to realize essentially M separate channels to achieve high channel capacity. In contrast, the proposed method can dynamically activate a subset of the transmit antennas at each transmission while maintaining the same channel capacity as conventional MIMO. The dynamic activation of the transmit antennas offers a statistically similar advantage to spatial modulation, leading to a reduced number of active RF chains and power saving as well. The theoretical analysis is conducted through the proposed transformation matrix that operates from the signal source to the transmit signal. The effectiveness of this approach is investigated regarding capacity issues and antenna utilization, which are confirmed by simulation results.

This paper proposes a perfect secrecy-achieving scheme for wireless communication systems. Most existing physical layer security schemes can achieve only weak or strong secrecy. One of the main reasons is the uncertainty of the eavesdropping channels. In order to achieve perfect secrecy, we design the artificial noise (AN) as a strictly typical sequence to guarantee that the eavesdropping channel is degraded with probability 1. In addition, a cascaded code is also designed to achieve both reliability and security. Theoretical analysis proves that the proposed scheme can transmit private message with finite code length without any information leakage. Finally, security performance of the proposed scheme is measured and compared with theoretical bounds, and a tradeoff between AN power consumption and decoding threshold is also analyzed numerically.

To promote energy and spectrum efficient communications in multiantenna channels, broadcast can provide a substantial gain in system throughput. However, the hardware constraints and strong Line of Sight (LoS) limit the implementation and performance of multiantenna broadcast in the Internet of unmanned aerial vehicles (UAVs). Based on the pseudo-Doppler principle, we propose a joint precoding and antenna array design to reduce the number of radio frequency (RF) chains required by the broadcast and free the LoS path from the interstream interference. The reduction of RF chains is realized by designing the precoding matrix that makes at least one of the transmit antennas have null inputs during any broadcast and, moreover, the LoS path is formed to match the obtained precoding matrix through antenna array design at the broadcasting UAV. The algorithms with low computational complexity for optimizing this design are developed to minimize the transmit power within the UAV broadcast paradigms. Theoretical formulation and numerical results in the metrics of sum data rate and bit error rate substantiate the validity of our proposed design, specifically in the Internet of UAVs with strong LoS.

For achieving the high spectral efficiency of the co-frequency co-time full duplex (CCFD), the self-interference (SI) should be cancelled drastically using the signal processing techniques. The radios frequency (RF) canceler uses the reconstructed SI symbol to cancel that arriving from the air interface, where the accurate time-alignment of the two symbols is required so that the cancellation is made at per entire symbol basis. Though using the variable fractional delay (VFD) method allows the exact time-alignment, it deforms inevitably the waveform of the reconstructed SI symbols in the practical application. Actually, the waveform deformation of the reconstructed SI symbol can impair the quality of the cancellation performance. To address this problem, we propose a new cost function that converts the mathematical error of the conventional VFD to the residual SI power directly. The sufficiency of the new cost function is proved and the superior performance is conformed by the simulation results.

For wireless communication systems with a long distance or severe interference, the insufficient transmit power limits the system performance. In this case, the maximum transmit power depends on the nonlinearity and the saturation region of the power amplifier (PA), which is referred to as a nonlinearity-constrained problem in this paper. To increase the transmit power as high as possible in a nonlinearity-constrained system, this paper proposes an autoencoder-based system to jointly optimize the modulation scheme and transmit power. The optimal solution can achieve a tradeoff between increasing the transmit power and reducing the nonlinear distortion. Meanwhile, the optimized signal constellation and the neural network-based receiver can effectively improve the capacity against nonlinear distortion. The simulation results indicate that the proposed method outperforms conventional methods in terms of symbol error rate (SER) and transmit power, and the SER of the proposed method is close to the SER lower bound of the nonlinear PA.

The quick evolution and widespread applicability of machine learning and artificial intelligence have fundamentally shaped and transcended modern life. Three key players stand behind such a ubiquitous emergence: big data, growing computing power, and improved algorithms. The need for distributed storage and processing arises from this ``data deluge'' that can flood any powerful machine, from the dispersively available data that are prohibitively costly to transfer to a central unit for further processing, and from the prevalent Internet-of-Things (IoT) devices that require real-time response as well as respect of privacy. Modern machine learning algorithms built to exploit such huge amounts of data are often computationally ``hungry'' and their ``appetite'' for computing power increases rapidly at a pace unmatched by the development of computing hardware. All these considerations justify the pressing need for distributed optimization algorithms that are scalable yet flexible to adapt to various configurations of networked computing nodes. To cope with these challenges, the present thesis first introduces a novel ADMM based approach (termed hybrid ADMM) for efficient decentralized optimization. By modeling the underlying communication patterns as hypergraphs, it provides a unifying framework that subsumes both centralized and fully decentralized counterparts as special cases, and allows nodes to communicate in centralized and decentralized ways at the same time. Leveraging the expressiveness of hypergraph models, a technique termed ``in-network acceleration'' is introduced enabling ``almost free'' performance gain by exploiting local graph topology. To account for heterogeneity of nodes and edges, a diagonal scaling based approach is proposed to tackle weighted updates, where proper edge weights are identified through solving a preconditioning problem. By assigning larger weights to critical edges, the proposed algorithm achieves higher efficiency and becomes more robust to perturbations. Finally, to boost the efficiency of the whole system to its full potential, an asynchronous method is introduced to mitigate the straggler problem so that nodes with different processing power can run at full speed. Convergence analysis for proposed algorithms is provided which reveals the connection between convergence rate and spectral properties of the communication graph. Numerical tests of several common tasks on different graphs are carried out to demonstrate the effectiveness of proposed algorithms.

As the digital version of a continuous-time delay, the concept of fractional delay (FD) is exploited to approximate a desired delay that is not a multiple of the sampling interval. However, in FD filters, there is always a severe distortion at the beginning of delayed signals, referred to as head distortion. This letter identifies the cause of head distortion and proposes a solution to this problem for reducing the overall distortion in FD filters. For the purpose of performance evaluation, relative root-mean-square (RMS) error is formulated as a metric to quantify the overall difference between the frequency-domain response of an FD filter and the ideal one. Moreover, illustrative numerical results on the proposed scheme applied in FD filters with classical sinc, Farrow and Lagrange interpolation substantiate the validity and feasibility of our solution.

In this letter, an uplink space division multiple access scheme based on the Doppler effect is proposed for wireless multi-user communication systems. By rotating the receive antenna array or using a circular sampling method, different Doppler frequency shifts are introduced to the received signals from different directions, and then different user’s signal can be separated in frequency domain. To maximize the inter-user interference (IUI)-free bandwidth, a max-min criterion is proposed for antenna switching control. The main advantage of the proposed scheme over the conventional space division multiple access (SDMA) schemes is that it needs less number of analog receive chain. Then simulations are carried out to evaluate the performance of the proposed scheme in terms of channel capacity. Simulation results show that the proposed scheme achieves a significant capacity gain, and its capacity is very close to that of a single user system without IUI.

Dual-connectivity technology enables a base station to assign multiple carriers from various bands to a mobile station (MS), thus increasing its bandwidth and data rate. However, when the downlink frequency assigned to the MS is approximately twice its uplink frequency, the MS’s receiver will be seriously interfered by the nonlinear self-interference from its own transmitter. This paper addresses the problem of nonlinear self-interference cancelation for MSs operating in the dual-connectivity mode. Compared with conventional systems, this scenario faces some new challenges because of the wide variety of nonlinear interference components, including not only harmonics but also intermodulation products, and the more complicated interference channels, including both nonlinear and linear devices. In addition, the frequency, bandwidth and frequency-selective channel parameters of the interference are influenced by the uplink resource block allocation. To solve these problems, a two-part nonlinear self-interference canceler is proposed, where one part is designed as a neural network to capture the nonlinear characteristics, and the other part is designed as a linear filter to capture the linear characteristics. Furthermore, a low-complexity two-step training scheme is proposed to approximate the interference channel in the entire system bandwidth. Finally, a hardware prototype is implemented to verify the effectiveness of the proposed scheme. The experimental results show that the proposed scheme achieves more than 20 dB interference cancelation, and significantly outperforms the conventional polynomial-based and pure neural-network cancelation schemes.

Providing security guarantee is a critical concern in the ad-hoc networks relying on multi-hop channels, since their flexible topology is vulnerable to security attacks. To enhance the security of a spatial modulation (SM) assisted wireless network, various SM mapping patterns are activated by random channel quality indicator (CQI) patterns over the legitimate link, as a physical-layer secret key. The SM signals are encrypted by random mapping patterns to prevent eavesdroppers from correctly demapping their detections. This secret key is developed for multi-hop wiretap ad-hoc networks, where eavesdroppers might monitor all the transmitting nodes of a legitimate link. We substantially characterise the multi-hop wiretap model with receiver diversity techniques adopted by eavesdroppers. The security performance of the conceived scheme is evaluated in the scenarios where eavesdroppers attempt to detect their received signals using maximal-ratio combining or maximum-gain selection. The achievable data rates of both legitimate and wiretapper links are formulated with the objective of quantifying the secrecy rates for both Gaussian-distributed and finite-alphabet inputs. Illustrative numerical results are provided for the metrics of ergodic secrecy rate and secrecy outage probability, which substantiate the compelling benefits of the physical-layer secret key generation via CQI-mapped SM.

The best coding scheme for Gaussian interference channel (GIC) is still an open problem so far, and has attracted much attention in the field of wireless communication due to it plays an important role in combating co-channel interference. In this paper, an adaptive power allocation and coding scheme based on time sharing (TS) strategy is proposed for the twouser GIC to coordinate the interference and achieve a higher sum-rate. In the proposed scheme, the codewords are divided into some segments, and the power of each segment is jointly optimized for the two users to maximize the sum-rate. To solve the power allocation problem, a heuristic search algorithm based on the optimal path planning algorithm is proposed. Simulation results show that the proposed scheme can achieve a higher sum-rate compared with the conventional coding schemes.

In this letter, a linear precoding scheme is proposed for downlink (DL) multi-user (MU) generalized spatial modulation (GSM) systems. The proposed precoding scheme achieves, not only inter-user interference (IUI) free, but also maximum signal-to-noise ratio (SNR). The former achievement leads to higher multiplexing gain, while the later results in higher diversity gain. By formulating the precoder design as a constrained optimization problem, the optimal precoding vectors are derived by using singular value decomposition (SVD). Outage and ergodic capacities of the proposed scheme are then analyzed, and compared to that of the conventional GSM scheme and the MU zero-forcing (ZF) transmit beamforming scheme by simulations. Simulation results indicate that the proposed scheme can effectively eliminate the IUI, and achieve significant diversity gain.

Radio FD has emerged as an attractive technique capable of doubling the spectral efficiency over half duplex. However, for signal reception, an FD node needs to suppress its transmitter's signals quite significantly. In point to point communication systems, these transmitter signals are termed self-interference. When working with an FD mobile network, the self-interference problem becomes much more complicated because the receiver of an FD base station (BS) receives interference not only from its BS transmitter in its cell, but also from those in the surrounding cells. For the UL channel, self-interference extends to the problem of multiple interference. And, a similar interference problem can be found among the MSs over a DL channel. In both cases, the interference owing to the FD implementation spreads beyond the scope of the self-interference. This article describes the development of FD BSs that use antenna arrays to deal with the BSs' interference, and thus enable FD communication over the UL channel, where the theoretical focus is placed on how to use the antenna array to nullify the multiple interference and receive the signals of the desired MSs simultaneously. To complete the system construction, FD MSs have also been developed to enable DL transmission. A prototype system is described for the scenario of two cells and one FD MS for tests of FD communication over UL channels and DL channels in terms of video performance. Good video quality is demonstrated at both the BS and MS.

We consider a full-duplex (FD) multiuser multiple-input multiple-output (MU-MIMO) system, where a FD base station (BS) with multiple antennas serves multiple half-duplex (HD) user equipments (UEs) in both uplink (UL) and downlink (DL) via the same time-frequency resources. UE scheduling is in demand to manage the UL-to-DL interference (UDI) incurred by the FD operation. Existing scheduling algorithms require UDI channel state information between each pair of the candidate UEs, which incurs a significant amount of overhead as the number of UEs grows. To reduce the overhead, we utilize channel reciprocity and UL-DL duality to design a DL-centric scheduling. By selecting UEs only based on their DL channels and received UDI strength, the proposed scheme no longer requires the massive UDI CSI. Numerical results demonstrate the proposed algorithm can achieve a near optimal performance without knowing any UDI CSI.

In a wideband full-duplex system, the residual self-interference (SI) channel exhibits significant frequency selective fading characteristics after radio frequency SI cancellation. Resource allocation using the difference between the signal channel and the residual SI channel is a new method of suppressing residual SI and increasing system capacity. In this work, the residual SI channel and the user uplink channel are modelled as multi-path frequency selective fading channels that satisfy the Rayleigh distribution. To improve system capacity, the authors propose to allocate the spectrum resource avoiding occupation of subcarriers with high residual SI, while guaranteeing a fair subcarrier assignment (SA) and power allocation (PA). A sum rate of uplink maximisation is formulated, and an algorithmic solution is developed to effectively conduct SA and PA. Numerical results demonstrate the effectiveness of the scheme in increasing system capacity and suppressing residual SI.

Flexible duplex is a promising technique, which provides access to flexible resource allocation and asymmetric traffic configuration. Inter-cell interference (ICI) is one of the key problems in flexible duplex cellular system. In this paper, a hybrid beamforming scheme is proposed to deal with the ICI problem. In particular, in designing the analog beamforming weights, a coordinate transformation method is introduced to transform the non-convex optimization problem into a convex one. Moreover, an iterative algorithm is proposed to find the global optimal weights of the analog beamformer with lower computational complexity. Simulation and hardware experimental results demonstrate that significant ICI suppression can be achieved by using the proposed scheme.

In this paper, we consider an Internet of Things (IoT) wireless network using Long Term Evolution (LTE) cellular system as backhaul. To provide high throughput, by using dual-connectivity technique, the IoT gateway simultaneously connects to two evolved Node Bs (eNBs) on two carriers, one for downlink and the other for uplink. As a result, the receive link will be severely interfered by the harmonic interference (HI) and inter-modulation (IM) components caused by the imperfections of power amplifier (PA) and in-phase/quadrature (I/Q) modulator. To solve this problem, in this paper, an neural-network (NN)based non-linear interference cancellation scheme is proposed for dual-connectivity IoT gateway. In the proposed scheme, the nonlinear interference is first reconstructed by using the transmit signal and the trained NN in baseband, and then subtracted from the received signal in digital domain at receiver. The NN precisely models the link behavior from the baseband transmitter to the baseband receiver, including all the linear and non-linear effect. Additionally, the NN can be used to reconstruct and cancel not only the HI, but also the IM components of the mirror-frequency interference (MFI) caused by I/Q imbalance, and direct current (DC) bias caused by local oscillator (LO) leakage. To evaluate the performance of the proposed scheme, a hardware prototype is designed and implemented. Experimental results show that the proposed scheme has a superior performance in dual-connectivity system compared with the traditional non-linear interference cancellation scheme using polynomial (PM) model.

Due to the severity of self-interferences, analog interference canceller is an essential part of a full duplex radio system. As the self-interference can potentially be even much stronger than the received signal, it must be meritoriously suppressed before analog-to-digital conversion. This paper is aimed to provide an optimal radio frequency domain multi-tap interference cancellation design. To this end, the following two key problems are addressed: the hardware design problem and the adaptive weight adjusting problem. For optimal hardware design, a trade-off between widening effective delay range and maximizing the minimum cancellation capability is formulated. The principles and guidelines for optimal solution and hardware parameter design are derived. We also formulate adaptive weight adjusting as a multi-dimensional optimization problem, which is proved to be non-convex while adjusting the attenuators and phase shifters independently. By using a coordinate transformation technique, it is then converted into a convex optimization problem, and a convergent iterative algorithm based on the gradient descent method is applied to obtain the global optimum. The hardware implementation and experimental results substantiated the effectiveness of the proposed principles and algorithms.