Multi-view 3D object detection has garnered increasing attention, particularly due to its success in autonomous driving systems. Although multi-view systems possess rich semantic information, their spatial-geometric reasoning capabilities remain limited. Recent studies employ simulated point cloud generation mechanisms to facilitate LiDAR-camera multi-modal knowledge distillation, achieving formal structural consistency. Despite advancements, these methods still face two main issues: i) alignment challenges caused by discrepancies between LiDAR and camera data, and ii) prediction errors from simulated point clouds that compromise the semantic information extracted from images during fusion. To address these problems, we propose adaptive-smooth distillation to optimize alignment granularity based on feature discrepancies for improved LiDAR-camera knowledge distillation. Specifically, this work considers both LIDAR-to-camera cross-modal distillation and LiDAR-camera fusion to simulated point cloud-camera fusion multi-modal distillation. Then, we introduce a heterogeneous fusion module to strategically bias the fusion process toward the extracted camera features, thereby enhancing the robustness of the fusion feature. Additionally, soft-weighted response distillation is proposed to facilitate the student model to selectively mimic the high-quality output of the teacher model. Extensive experiments have demonstrated the superiority of our method, achieving statistically significant improvements of 4.9% in mean Average Precision (mAP) and 4.5% in NuScenes Detection Score (NDS) over the benchmark.

This paper investigates a heterogeneous multi-vehicle, multi-modal sensing (H-MVMM) aided online precoding problem. The proposed H-MVMM scheme utilizes a vertical federated learning (VFL) framework to minimize pilot sequence length and optimize the sum rate. This offers a promising solution for reducing latency in frequency division duplexing systems. To achieve this, three preprocessing modules are designed to transform raw sensory data into informative representations relevant to precoding. The approach effectively addresses local data heterogeneity arising from diverse on-board sensor configurations through a well-structured VFL training procedure. Additionally, a label-free online model updating strategy is introduced, enabling the H-MVMM scheme to adapt its weights flexibly. This strategy features a pseudo downlink channel state information label simulator (PCSI-Simulator), which is trained using a semi-supervised learning (SSL) approach alongside an online loss function. Numerical results show that the proposed method can closely approximate the performance of traditional optimization techniques with perfect channel state information, achieving a significant 90.6% reduction in pilot sequence length.

Deep learning-based channel state information (CSI) feedback schemes offer strong compression but are typically confined to fixed system configurations, limiting their generalizability and flexibility. To address this challenge, this work proposes WiFo-CF, a novel wireless foundation model tailored for CSI feedback. WiFo-CF uniquely accommodates heterogeneous configurations, including varying channel dimensions, feedback rates, and data distributions, within a unified framework through two key innovations: a multi-user, multi-rate self-supervised pre-training strategy and a Mixture of Shared and Routed Experts (S-R MoE) architecture. To support its large-scale pre-training, we introduce the first heterogeneous channel feedback dataset; its diverse patterns enable WiFo-CF to achieve superior performance on both in-distribution and out-of-distribution data across simulated and real-world scenarios. Furthermore, the learned representations effectively facilitate adaptation to downstream tasks such as CSI-based indoor localization, validating the model’s scalability and deployment potential.

Integrated sensing and communication (ISAC) at sub-THz frequencies is crucial for future air-ground networks. However, optimizing ISAC performance while managing operational latency is challenging due to unique propagation characteristics and hardware limitations. This paper introduces a multi-modal sensing fusion framework inspired by synesthesia of machine (SoM) to enhance sub-THz ISAC transmission. By exploiting inherent degrees of freedom in sub-THz hardware and channels, the framework succeeds in tuning the radio-frequency environment. It features squint-aware beam management to improve air-ground network adaptability, enabling dynamic three-dimensional ISAC links. By leveraging multi-modal information, the framework enhances ISAC performance and reduces latency. Visual data is used to rapidly localize users and targets, while a customized multi-modal learning algorithm optimizes the hybrid precoder. A new metric is proposed for comprehensive performance evaluation. Extensive experiments demonstrate that the proposed scheme significantly improves ISAC efficiency.

This paper presents our indoor pathloss prediction solution to ICASSP 2025 Signal Process Grand Challenge: First Indoor Path Loss Prediction Challenge. The proposed U-Net-based network incorporates dedicated asymmetric convolutions and spatial pyramid pooling to enhance reconstruction quality. Our approach achieves a weighted root mean squared error (RMSE) of 9.411 dB on the final test set, securing the 1st place in the challenge.

High complexity in precoding design for frequency division duplex systems necessitates streamlined solutions. Guided by Synesthesia of Machines (SoM), this paper introduces a heterogeneous multi-vehicle, multi-modal sensing aided precoding scheme within a vertical federated learning (VFL) framework, which significantly minimizes pilot sequence length while optimizing the system's sum rate. We address the challenges posed by local data heterogeneity due to varying on-board sensor configurations through a meticulously designed VFL training procedure. To extract valuable channel features from multimodal sensing, we employ three distinct data preprocessing methods that convert raw data into informative representations relevant for precoding. Additionally, we propose an online training strategy based on VFL framework, enabling the scheme to adapt dynamically to fluctuations in user numbers. Numerical results indicate that our approach, utilizing short pilot sequences, closely approximates the performance of traditional optimization methods with perfect channel state information.

This paper presents a novel and robust target-to-user (T2U) association framework to support reliable vehicle-to-infrastructure (V2I) networks that potentially operate within the hybrid field (near-field and far-field). To address the challenges posed by complex vehicle maneuvers and user association ambiguity, an interacting multiple-model filtering scheme is developed, which combines coordinated turn and constant velocity models for predictive beamforming. Building upon this foundation, a lightweight association scheme leverages user-specific integrated sensing and communication (ISAC) signaling while employing probabilistic data association to manage clutter measurements in dense traffic. Numerical results validate that the proposed framework significantly outperforms conventional methods in terms of both tracking accuracy and association reliability.

Integrated sensing and communication (ISAC) technology is vital for vehicular networks, yet the time-varying communication channels and rapid movement of targets present significant challenges for real-time precoding design. Traditional optimization-based methods are computationally complex and strongly depend on perfect prior information, which is often unavailable in double-dynamic scenarios. In this paper, we propose a synesthesia of machine (SoM)-enhanced precoding paradigm that leverages modalities such as positioning and initial channel information to adapt to these dynamics. Utilizing a deep reinforcement learning (DRL) framework, our approach pushes ISAC performance boundaries. We also introduce a parameter-shared actor-critic architecture to accelerate training in complex state and action spaces. Extensive experiments validate the superiority of our method over existing approaches.

Integrated sensing and communication (ISAC) emerges as a promising technology for 6G, particularly in the millimeter-wave (mmWave) band. However, the widely utilized hybrid architecture in mmWave systems compromises multiplexing gain due to the constraints of limited radio-frequency (RF) chains. Moreover, additional sensing functionalities exacerbate the impairment of spectrum efficiency (SE). In this paper, we present an optimized beam pattern modulation-embedded ISAC (BPM-ISAC) transceiver design, which spares one RF chain for sensing and uses the remaining ones for communication. To compensate for the reduced SE, index modulation across communication beams is applied. We formulate an optimization problem aimed at minimizing the mean squared error (MSE) of the sensing beampattern, subject to a symbol MSE constraint. This problem is then solved by sequentially optimizing the analog and digital parts. Both the multi-aperture structure (MAS) and the multi-beam structure (MBS) are considered in the analog part. We conduct theoretical analysis on the asymptotic pairwise error probability (APEP) and the Cramér-Rao bound (CRB) of direction of arrival (DoA) estimation. Numerical simulations validate the overall enhanced ISAC performance over existing alternatives.

This work aims to tackle the labor-intensive and resource-consuming task of indoor positioning by proposing an efficient approach. The proposed approach involves the introduction of a semi-supervised learning (SSL) with a biased teacher (SSLB) algorithm, which effectively utilizes both labeled and un-labeled channel data. To reduce measurement expenses, unlabeled data is generated using an updated channel simulator (UCHS), and then weighted by adaptive confidence values to simplify the tuning of hyperparameters. Simulation results demonstrate that the proposed strategy achieves superior performance while minimizing measurement overhead and training expense compared to existing benchmarks, offering a valuable and practical solution for indoor positioning.

Integrated sensing and communication (ISAC) technology is essential for supporting vehicular networks. However, the communication channel in this scenario exhibits time variations, and the potential targets may move rapidly, resulting in double dynamics. This nature poses a challenge for real-time precoder design. While optimization-based solutions are widely researched, they are complex and heavily rely on perfect channel-related information, which is impractical in double dynamics. To address this challenge, we propose using constrained deep reinforcement learning to facilitate dynamic updates to the ISAC precoder. Additionally, the primal dual-deep deterministic policy gradient and Wolpertinger architecture are tailored to efficiently train the algorithm under complex constraints and varying numbers of users. The proposed scheme not only adapts to the dynamics based on observations but also leverages environmental information to enhance performance and reduce complexity. Its superiority over existing candidates has been validated through experiments.

Integrated Sensing and Communication (ISAC) emerges as a promising technology for BSG/6G, particularly in the millimeter-wave (mmWave) band. However, the widespread adoption of hybrid architecture in mmWave systems compromises multiplexing gain due to limited radio-frequency chains, resulting in mediocre performance when embedding sensing functionality. To avoid sacrificing the spectrum efficiency in hybrid structures while addressing performance bottlenecks in its extension to ISAC, we present an optimized beam pattern modulation-embedded ISAC (BPM-ISAC). BPM-ISAC applies index modulation over beamspace by selectively activating communication beams, aiming to minimize sensing beampattern mean squared error (MSE) under communication MSE constraints through dedicated hybrid transceiver design. Optimization involves the analog part through a min-MSE-based beam selection algorithm, followed by the digital part using an alternating optimization algorithm. Convergence and asymptotic pairwise error probability (APEP) analyses accompany numerical simulations, validating its overall enhanced ISAC performance over existing alternatives.

The future of vehicular communication networks relies on mmWave massive multi-input-multi-output antenna arrays for intensive data transfer and massive vehicle access. However, reliable vehicle-to-infrastructure links require exact alignment between the narrow beams, which traditionally involves excessive signaling overhead. To address this issue, we propose a novel proactive beamforming scheme that integrates multi-modal sensing and communications via Multi-Modal Feature Fusion Network (MMFF-Net), which is composed of multiple neural network components with distinct functions. Unlike existing methods that rely solely on communication processing, our approach obtains comprehensive environmental features to improve beam alignment accuracy. We verify our scheme on the Vision-Wireless (ViWi) dataset, which we enriched with realistic vehicle drifting behavior. Our proposed MMFF-Net achieves more accurate and stable angle prediction, which in turn increases the achievable rates and reduces the communication system outage probability. Even in complex dynamic scenarios with adverse environment conditions, robust prediction results can be guaranteed, demonstrating the feasibility and practicality of the proposed proactive beamforming approach.

Meta-learning owns unique effectiveness and swiftness in tackling emerging tasks with limited data. Its broad applicability is revealed by viewing it as a bi-level optimization problem. The resultant algorithmic viewpoint however, faces scalability issues when the inner-level optimization relies on gradient-based iterations. Implicit differentiation has been considered to alleviate this challenge, but it is restricted to an isotropic Gaussian prior, and only favors deterministic meta-learning approaches. This work markedly mitigates the scalability bottleneck by cross-fertilizing the benefits of implicit differentiation to probabilistic Bayesian meta-learning. The novel implicit Bayesian meta-learning (iBaML) method not only broadens the scope of learnable priors, but also quantifies the associated uncertainty. Furthermore, the ultimate complexity is well controlled regardless of the inner-level optimization trajectory. Analytical error bounds are established to demonstrate the precision and efficiency of the generalized implicit gradient over the explicit one. Extensive numerical tests are also carried out to empirically validate the performance of the proposed method.

Over-the-air federated learning (OTA-FL) is a communication-effective approach for achieving distributed learning tasks. In this paper, we aim to enhance OTA-FL by seamlessly combining sensing into the communication-computation integrated system. Our research reveals that the wireless waveform used to convey OTA-FL parameters possesses inherent properties that make it well-suited for sensing, thanks to its remarkable auto-correlation characteristics. By leveraging the OTA-FL learning statistics, i.e., means and variances of local gradients in each training round, the sensing results can be embedded therein without the need for additional time or frequency resources. Finally, by considering the imperfections of learning statistics that are neglected in the prior works, we end up with an optimized the transceiver design to maximize the OTA-FL performance. Simulations validate that the proposed method not only achieves outstanding sensing performance but also significantly lowers the learning error bound.

Wireless multi-casting is an efficient transmission modality for downlink network control and content sharing. Conventional multi-casting amounts to finding an optimized beamforming vector under a specific performance metric and without the need of a dedicated treatment at the receiver. However, the so-calledone-fit-allstrategy is not directly applicable to mmWave systems equipped with 1-bit digital-to-analog converters (DAC), prompting us to develop a new multi-casting framework. The overarching design is built upon a vector-based instead of a scalar-based modulation, with which the multi-casting exhibits special features on both the transmitter and receiver. Specifically, we start by revealing the theoretical fundamentals for multi-casting under different 1-bit setups. Then, we design an iterative scheme to generate high-order constellations by establishing the equivalence between the codeword basis and low-order constellations. Finally, effective detection solutions are individually proposed for low-order and high-order constellations per their specific signal structures. Extensive analyses and simulations have been carried out to corroborate the decency of the proposed multi-casting strategy.

With the unprecedented development of smart vehicles and roadside units equipped with wireless connectivity, the transportation system is undergoing revolutionary changes in the past decade or two. Bearing safety and efficiency as the utmost objectives, the vehicular environments are witnessing explosive increase of various sensors onboard vehicles and equipped at transportation infrastructures. On the one hand, these sensors are destined to be wirelessly connected to provide more comprehensive situational awareness for transportation purposes. On the other hand, the abundance of sensor data of the environment can potentially shed light on the channel propagation characteristics that lie at the core of any communications system design. The integrated sensing and communications (ISACs) is henceforth both necessary and natural in vehicular communications networks (VCN). Different from existing ISAC works that target generic environments but are limited to dual-function radar-communications (DFRC), in this article we focus on transportation scenarios and applications but take a wholistic view of ISAC possibilities. First, we argue that, even though many sensors in transportation settings are nonradio-frequency (RF)-based, functional ISAC (fISAC) is feasible and necessary, in both communication-centric (CC) or sensing-centric (SC) modes. To facilitate this, the concept of synesthesia is introduced to ISAC to accommodate “machine senses” in the RF and non-RF formats. We then zoom in to RF-based sensors and propose the so-termed signaling ISAC (sISAC), with either unified-hardware (UH) or separate-hardware (SH) platforms, and delineate the unique issues arising in transportation settings. Several transportation-specific case studies are included to demonstrate these various ISAC regimes. Toward the end, the relationships of these ISAC subcategories are discussed with a roadmap laid out.

To enable next-generation mmWave cellular, it is vital to design high-performance precoding schemes for wideband multi-user (wMU) mmWave massive MIMO (mMIMO). As existing approaches are mostly ad-hoc, thereby lacking performance guarantee, we will tailor an enhanced transceiver design explicitly for wMU mmWave mMIMO, with the goal of maximizing mutual information (MI). The proposed scheme follows the prevalent hybrid block diagonalization (HBD-)based framework that is well-known for balancing the transmitter-end processing flexibility and the user-end detection complexity. In this paper, we for the first time prove that HBD is optimal in the sense of MI. In terms of the transceiver design, we start by decoupling the hybrid processing into a two-stage analog and digital processing, and then derive the MI bounds associated with HBD. By optimizing the tight MI bound, excellent HBD transceivers are devised for both the multi-aperture structure (MAS) and the multi-beam structure (MBS). The proposed HBD technique does not rely on substantial computational complexity, striking channel sparsity, or high-resolution analog beamformers, and can achieve a superb MI performance even with inferior hardware configurations.

Referring to the system equipped with single-bit converters, 1-bit mmWave communications is gaining increasing attention for its superb cost efficiency. However, the inherent non-linear distortion renders the detectors designed for classical transparent communications inapplicable, leading to an urgent need for novel detecting solutions dedicated to 1-bit systems. Although a few endeavours have been made towards learning-based (as opposed to the traditional model-based) detectors for multi-user (MU) 1-bit systems, they are exclusively limited to narrowband channels and fail to cope with the multi-path effects inevitable to mmWave systems. In this paper, we first design a learning-based detector (LeaD) for general wideband multi-user (wMU) scenarios. Though stemming from block-based detection, the classic workhorse for transparent systems, LeaD faces either unaffordable complexity or unacceptable data rate in 1-bit systems. Given the impracticability of block-based detection, we resort to the serial detection mechanism and henceforth devise a so-termed model-enhanced (Me-)LeaD by utilizing the channel delay-domain information. Me-LeaD can be further augmented by exploiting the channel angular-domain information. Underpinned by a judiciously tailored method for extracting tbe model information, the proposed Me-LeaD demonstrates a decent overall performance in general 1-bit wMU scenarios.

This paper investigates the transceiver design for downlink hybrid mmWave multi-user multi-carrier massive MIMO systems. In order to balance the processing complexity and the design flexibility, we adopt a prevalent hybrid precoding technique named hybrid block diagonalization (HBD) for downlink multi-user transmission. Aimed at maximizing the end-to-end mutual information (EEMI), a novel virtual EEMI assisted two-stage HBD scheme is judiciously devised. Apart from a low implementing complexity, the developed scheme is a more generic HBD solution, as it not only takes the frequency selectivity into account, but also removes the reliance on the high-resolution analog network. Simulations show that, even when applied with an inferior hardware configuration, the proposed HBD could still remarkably outperform its counterparts in terms of the EEMI performance at different levels of channel sparsity.

In mmWave massive multiple-input multiple-output (mMIMO) systems, hybrid (digital/analog) structure has been a prevalent option to balance system cost and performance. To facilitate transceiver design in hybrid mmWave mMIMO, acquiring an accurate channel state information is critical. To this end, a novel doubly-sparse approach is proposed to estimate doubly-selective mmWave channels under hybrid mMIMO. Via the judiciously designed training pattern, the well-utilized beamspace sparsity alongside the under-investigated delay-domain sparsity that mmWave channels exhibit can be jointly exploited to assist channel estimation. Thanks to our careful two-stage (random-probing and steering-probing) design, the proposed channel estimator possesses strong robustness against the double (frequency and time) selectivity whilst enjoying the benefits brought by the exploitation of double sparsity. Compared with existing alternatives, our proposed mmWave channel estimator not only works in doubly-selective channels, but also largely reduces the training overhead, storage demand as well as computational complexity.

The widespread deployment of mmWave communication systems is an unstoppable trend, but its success will heavily rely on well-designed transceivers to combat the severe propagation loss. To acquire the accurate channel state information (CSI) so that the transceiver design can be facilitated, in this paper, we provide a new time-domain channel estimation scheme for hybrid mmWave massive multiple-input multiple-output (mMIMO) systems. In addition to utilizing the well-known angular sparsity, the delay-domain sparsity will also be exploited to accomplish the channel estimation with satisfactory accuracy yet affordable complexity. The successful combination of these two types of sparsity is attributed to a judiciously designed training pattern. Thanks to our innovative exploitation of the double sparsity, a satisfactory performance can be achieved together with largely reduced training overhead, storage demand, and computational complexity.

Millimeter wave (mmWave) massive multiple-input multiple-output (mMIMO) has been recognized as a promising candidate for 5G communications for its capability of supporting Gb/s transmission. However, it is a common exercise to deploy a limited number of radio-frequency (RF) chains at mmWave mMIMO transceivers due to hardware complexity and cost. As a result, the potential multiplexing gain (MG), which is restricted by the smaller number of RF chains at the transmitter and receiver, is markedly compromised. In order to boost the MG and spectral efficiency (SE), we innovatively develop a novel index modulation termed as the generalized beamspace modulation (GBM). The acquisition of (sub-)beamspace is owing to a natural exploitation of the unique features of mmWave mMIMO. Based on the (sub-)beamspace, a complete GBM transceiver is designed and optimized. Unlike existing alternatives that are largely digital based, our GBM is tailored for the hybrid structure of mmWave mMIMO and can, thereby, realize efficient spatial multiplexing despite the limited RF chains. Extensive analyses and simulations have demonstrated remarkable superiority of GBM over existing counterparts in terms of the error performance and SE.

As a recently emerging technology, index-based modulation (IBM) has been attracting increasing research interests for its improved bit error rate (BER) performance and power efficiency. At present, the applications of two typical schemes named spatial modulation (SM) and subcarrier index modulation (IM), as well as their variants are introduced to microwave systems. To make IBM applicable in mmWave systems, the special properties of channel environments and system architectures should be taken into account. In this paper, we present a novel IBM scheme termed as generalized beamspace modulation (GBM) for mmWave beamspace multiple-input multiple-output (MIMO) systems. Unlike the frequency or spatial domain in which the existing IBM schemes are typically performed, GBM is implemented in the beamspace. To achieve near- optimal BER performance in GBM systems, a general effective beamspace channel (EBC) optimization method is derived based on the minimum asymptotic pairwise error probability (APEP) criterion. The optimal maximum-likelihood (ML) detector and the lowcomplexity detector are both provided. Thanks to our proposed GBM scheme, the BER performance can be noticeably enhanced compared to plain mmWave systems, with a smaller number of active frequency chains are used during transmission.

In this paper, we propose a novel digital filter-and-forward full-duplex (FF-FD) relaying system to exploit the loop-back signal (LBS). Unlike treating the LBS as a generic interference, we utilize the fact that the LBS actually conveys redundant information from the source to the destination in the relaying communication scenarios. Thus, the LBS could potentially be exploited instead of being cancelled. We prove that the FF-FD can achieve higher system achievable rate (SAR) than the amplify-and-forward full-duplex (AF-FD) system. To maintain the system linearity, a digital FF-FD relay is proposed without non-linear distortion induced in the digital domain. The relay can be modeled as a linear digital filter to exploit the LBS via linear filtering. The design challenge at the relay is accordingly shifted to optimizing its discrete frequency response (DFR) rather than the complicated interference cancellation design. Based on the metric of maximum SAR, the DFR is efficiently optimized with and without CSI at the source. A low-complexity yet near-optimal alternative is also provided to reduce the computation cost for large multi-carrier systems. Both theoretical analyses and simulations validate the considerable advantages of the proposed FF-FD over the conventional AF-FD, making FF-FD an appealing candidate for future relaying communications.

Index modulated orthogonal frequency division multiplexing (IM-OFDM) is a novel multicarrier transmission scheme, which provides considerable performance improvement compared with classical OFDM by conveying information via the active subcarrier indices in conjugation with the constellation symbols. In this paper, we extend the idea of IM-OFDM to multi-input multi-output (MIMO) systems and propose precoded MIMO-OFDM (PIM-MIMO-OFDM). Based on the channel state information at the transmitter, PIM-MIMO-OFDM selects the active elements of the receiver-side space-frequency subblocks via linear precoding. The spectral efficiency enhancement method of in-phase/quadrature index modulation is also employed to construct PIM-MIMO-OFDM with in-phase quadrature modulation by selecting the active space-frequency elements in the in-phase and quadrature components of the constellation symbol. Thanks to the carefully designed precoding, the co-channel interference among received antennas can be completely eliminated. Consequently, low-complexity maximum likelihood and suboptimal detectors are devised with linear detection complexity. Both analytical and numerical results show that, with the help of precoding, PIM-MIMO-OFDM can achieve better bit error rateperformance than traditional precoded MIMO-OFDM (P-MIMO-OFDM) in various system configurations. In addition, the spectral efficiency can be also enhanced compared with the P-MIMO-OFDM under certain configurations.