Sparse vector coding (SVC) and its variants convey information through sparse vectors (SVs), or sparse support patterns and detecting information by solving sparse recovery problems, making them attractive for short-packet ultra-reliable low-latency communications (URLLC). However, as the required information rate increases, conventional SVC schemes or the enhanced SVC (ESVC) schemes require a rapidly growing number of SVs, resulting in significant increase of complexity and receiver processing latency. On the other hand, when communicating in doubly selective channels, orthogonal frequency division multiplexing (OFDM) suffers from the Doppler spread induced inter-carrier interference, while affine frequency division multiplexing (AFDM) experiences the same peak-to-average power ratio (PAPR) problem as OFDM, although it allows diversity gain from channel's double selectivity. To address these issues, this paper proposes a constant-weight sparse vector coded augmented AFDM (CW-SVC/A$^{2}$FDM) scheme, along with a corresponding joint support and symbol detection algorithm. In this scheme, SVs are represented by constant-weight codewords that feature a fixed Hamming weight and a large Hamming distance, resulting in relatively small SV sets. Consequently, highly reliable sparse recovery is achievable. Meanwhile, because the primary information carrier shifts from SV candidates to channel-coded quadrature amplitude modulation (QAM) symbols, the proposed detection algorithm delivers both highly reliable information detection and low-latency receiver processing. Simulation results compare various SVC/ESVC schemes and diverse coding schemes employed by the CW-SVC/A$^{2}$FDM, including Hamming, BCH, RS, and polar codes, for QAM symbol protection. The results demonstrate that the proposed scheme provides a distinct reliability advantage over conventional SVC/ESVC schemes, achieving a gain of up to 14 dB. In the CW-SVC/A$^{2}$FDM system, errors are dominated by QAM demodulation, with polar codes offering the strongest protection among the evaluated options. Furthermore, the study indicates that the CWC's weight and minimum Hamming distance must be carefully selected to balance detection reliability, QAM modulation order, and sparse recovery feasibility, thereby maximizing the overall system's performance.
Vehicle-to-everything (V2X) is a communication technology that enables wireless communication among vehicles, roadside infrastructure, pedestrians, networks, and other entities, supporting safe autonomous driving and connected cars. Installing a large number of base stations (BSs) to achieve highly reliable V2X communication over a wide area increases installation and operational costs. Therefore, relay communication using the sidelink, which enables direct communication between terminals without passing through a BS, is a viable option for extending the communication range while minimizing the number of BSs. A multi-hop routing scheme is proposed for New Radio (NR)-V2X in fifth-generation mobile communication systems (5G) that uses BSs, roadside units, and vehicles as relay nodes. In the proposed method, a link metric is calculated based on the signal-to-noise power ratio between nodes and the number of information bits per orthogonal frequency-division multiplexing symbol in 5G NR. Relay nodes and communication paths are then determined to reduce radio resource requirements for end-to-end transmission from the source node to the destination node. Furthermore, the proposed method uses the frame formats and channels specified in 5G NR. Assuming a highway environment as defined by the 3rd Generation Partnership Project, computer simulations were used to evaluate the packet reception ratio (PRR) of the proposed method in a setting where multiple vehicles travel at high speeds. The results demonstrated that even when the number of BSs was reduced to one-sixth of that in a conventional configuration, the average PRR could be improved to a value close to one.
This paper presents an experimental characterization and comparison of the conducted emissions generated by two commercial Electric Vehicles (EVs) during Alternating Current (AC) and Direct Current (DC) charging in the 9-500 kHz frequency range under realistic low-voltage grid conditions. The results show that DC charging generates higher conducted emissions than AC charging in more than 94% of the analyzed frequency bins, although the spectral characteristics remain strongly dependent on both the EV model and the frequency range. In particular, ultra-narrowband and narrowband emissions dominate the 9-150 kHz band, whereas colored noise is observed in the 150-500 kHz band. Time-domain analysis reveals that the overall spectral characteristics remain largely stable over consecutive 50-s intervals within 600-s measurement periods, despite amplitude variations of up to 10 dB in some high-amplitude spectral components. Fast Fourier Transform (FFT) analysis shows that most emissions exhibit periodic variations synchronized with the mains cycle (20 ms), whereas emissions whose central frequency oscillates with time follow a distinct periodic pattern with a period of approximately 24.4 ms. For both EV models, the highest short-term variability is observed during DC charging. The transitions between charging (ON) and non-charging (OFF) states are also investigated. The results indicate that these transitions may be either abrupt or gradual, depending on the charging process under consideration rather than on the charging mode itself. Furthermore, some emissions exhibit delayed appearance and variations in their center frequency during transitions, highlighting the importance of considering temporal effects when characterizing conducted emissions generated by EV charging processes.
Learned car-following models are usually trained on one-step targets and accumulate errors when simulated in closed loop. Training through a differentiable vehicle update and adversarial imitation learning have both been proposed as remedies, but their contributions have not been separated for long rollouts at high sampling rates. We present a controlled study on the 30-second HighD subset of the FollowNet benchmark, where each event is simulated for 725 autoregressive steps at 25 Hz. With a common Transformer policy, we vary closed-loop unrolling, an adversarial term, a supervised acceleration anchor, an explicit spacing loss, and the training horizon. Every configuration is trained in multiple runs and compared with paired two-level bootstrap intervals and collision counts, and all models are evaluated without retraining on five further datasets (NGSIM I-80, Waymo, Lyft, and two SPMD subsets). Closed-loop unrolling is the most consistent factor: it lowered the spacing error in all twelve comparisons across the six datasets, significantly in ten, by 12 to 61%. The adversarial term never significantly improved accuracy on HighD, increased the error on three of the five other datasets, and, for the retained architecture at a 40-step horizon, was associated with collisions, whereas the closed-loop supervised models collided once in more than 90 000 simulated episodes. With a spacing loss, closed-loop supervised training had the lowest mean error and run-to-run spread on HighD ($26.4 \pm 0.2$ m$^{2}$; not significantly below training without it), compared with $31.5 \pm 3.7$ m$^{2}$ for the hybrid adversarial model and 62.8 m$^{2}$ for a calibrated Intelligent Driver Model. The effects of the spacing loss, the anchor, and the horizon depended on the dataset.
Multiple frequency ranges, including mmWave and Sub-6 GHz, open the door to a variety of applications in current and future wireless communications networks, such as Vehicle-to-Everything (V2X) networks. V2X networks support various use cases, contributing to a safe and efficient transportation system. These use cases require a flexible Transmission Control Protocol/Internet Protocol (TCP/IP) stack that enables layers within the stack (PS) to exchange information. This paper proposes a flexible cross-layer design for a hybrid mmWave and Sub-6 GHz system. The hybrid system relies on switching between mmWave and Sub-6 GHz spectrum based on the defined bandwidth threshold. The flexible cross-layer combines the Application, Facility, Medium Access Control (MAC), and Physical layers, denoted as (FC-AMP), to adjust the transmission parameters, accounting for the channel parameters and the rate requirements of the triggered use case. The FC-AMP relies on the current status of the V2X link characteristics, known at the receiver vehicle. Based on a use case's rate, including message rate and frequency, FC-AMP determines the physical-layer parameters, such as the frequency range and bandwidth, using the numerology selected and returned by the receiver as per channel parameters, including Delay and Doppler offsets. The simulations were conducted using Network Simulator 3 (NS-3). Extensive results have been reported, including a link-level evaluation of the Sub-6 GHz and mmWave spectrums, compared to the literature. The results demonstrate that the outage probability differs marginally between mmWave and Sub-6 GHz as velocity varies, suggesting that radio communications using mmWave and Sub-6 GHz maintain data rates within the outage ranges. The results also demonstrate that the throughput and packet Inter-reception ratio (PIR) of the FC-AMP exhibit a trade-off across different numerologies for a V2X use-case implementation, and an enhancement over the fixed numerology system when multiple use cases are triggered.
