Movie dubbing seeks to synthesize speech from a given script using a specific voice, while ensuring accurate lip synchronization and emotion-prosody alignment with the character’s visual performance. However, existing alignment approaches based on visual features face two key limitations: (1) they rely on complex, handcrafted visual preprocessing pipelines, including facial landmark detection and feature extraction; and (2) they generalize poorly to unseen visual domains, often resulting in degraded alignment and dubbing quality. To address these issues, we propose InstructDubber, a novel instruction-based alignment dubbing method for both robust in-domain and zero-shot movie dubbing. Specifically, we first feed the video, script, and corresponding prompts into a multimodal large language model to generate natural language dubbing instructions regarding the speaking rate and emotion state depicted in the video, which is robust to visual domain variations. Second, we design an instructed duration distilling module to mine discriminative duration cues from speaking rate instructions to predict lip-aligned phoneme-level pronunciation duration. Third, for emotion-prosody alignment, we devise an instructed emotion calibrating module, which fine-tunes an LLM-based instruction analyzer using ground truth dubbing emotion as supervision and predicts prosody based on the calibrated emotion analysis. Finally, the predicted duration and prosody, together with the script, are fed into the audio decoder to generate video-aligned dubbing. Extensive experiments on three major benchmarks demonstrate that InstructDubber outperforms state‑of‑the‑art approaches across both in‑domain and zero‑shot scenarios.

Integrated sensing, communication, and computation (ISC2) has been increasingly studied to support high-precision sensing and low-latency computing. Fairness in ISC2remains unaddressed due to the non-convexity introduced by local computing, task offloading, result delivery, and sensing. This paper develops a new approach to max-min fair task computing while ensuring target sensing demands in ISC2systems, by holistically optimizing the beamformers of the base station (BS), the transmit powers and CPU frequencies of users (UEs), and the uplink and downlink durations. To solve this non-convex problem, we develop a new algorithm that judiciously decouples the problem into manageable steps, leveraging Rayleigh quotient maximization, semidefinite relaxation (SDR), and successive convex approximation. We rigorously prove the (local) optimality of the algorithm by proving the tightness of the SDR, i.e., a rank-one solution for the dual-purpose beamforming of the BS for both communication and sensing, and the corresponding sensing-only beamforming, can always be constructed from a solution without rank constraint. Moreover, we reveal the optimal structure of the transmit powers of the BS and UEs, the CPU frequencies of the UEs, and the attainability of absolute fairness. As corroborated numerically, our algorithm substantially outperforms its alternatives in task completion and fairness.

Movie dubbing describes the process of transforming a script into speech that aligns temporally and emotionally with a given movie clip while exemplifying the speaker’s voice demonstrated in a short reference audio clip. This task demands the model bridge character performances and complicated prosody structures to build a high-quality video-synchronized dubbing track. The limited scale of movie dubbing datasets, along with the background noise inherent in audio data, hinder the acoustic modeling performance of trained models. To address these issues, we propose an acoustic-prosody disentangled two-stage method to achieve high-quality dubbing generation with precise prosody alignment. First, we propose a prosody-enhanced acoustic pre-training to develop robust acoustic modeling capabilities. Then, we freeze the pre-trained acoustic system and design an acoustic-disentangled framework to model prosodic text features and dubbing style while maintaining acoustic quality. Additionally, we incorporate an in-domain emotion analysis module to reduce the impact of visual domain shifts across different movies, thereby enhancing emotion-prosody alignment. Extensive experiments show that our method performs favorably against the state-of-the-art models on two primary benchmarks. The project is available at https://zzdoog.github.io/ProDubber/.

The rapid expansion of marine-related industries has intensified the need for robust maritime data transmission technologies. Traditionally, high-speed millimeter wave (mmWave) systems have been deemed unsuitable for marine environments due to significant propagation path losses (PLs). This study investigates the potential of marine evaporation ducts, a natural phenomena at the air-sea interface, to enhance mmWave frequency propagation by reducing signal attenuation. To examine the large-scale propagation characteristics of mmWave frequencies in environments featuring evaporation ducts, this research employed numerical simulations using ray tracing and parabolic equation methods. Additionally, a detailed field experiment was conducted to measure the PLs for a 22 GHz frequency over a 20 km shore-to-shore, over-the-horizon link with horizontal to horizontal polarization. The findings from both the numerical simulations and field tests reveal substantial multi-path effects caused by the evaporation duct phenomenon. These results indicate the necessity for orthogonal time frequency space technology to support high-speed, low-latency marine communications at mmWave frequencies.

Integrated Sensing, Communication, and Computing (ISCC) combines sensing, communication, and computing functions to improve spectrum efficiency and reduce hardware costs. However, poor link quality in the presence of obstructions leads to high offloading latency. This paper explores the use of intelligent reflecting surface and non-orthogonal multiple access in ISCC systems to enhance link reliability and improve computation offloading efficiency. A latency minimization problem is formulated by jointly optimizing computing, sensing, and communication parameters. Due to the strong coupling of the optimization variables, the problem is decomposed into two modules: computational design and sensing/communication design, which are optimized alternately to achieve a high-quality, stable solution. Simulation results demonstrate that the proposed system significantly reduces task processing latency, and ensures both user sum rate and radar sensing performance, offering a promising approach to enhance ISCC systems under resource-constrained environments.

This study presents an advanced radar-based multitarget human motion recognition (M-HMR) approach using multiple-input-multiple-output (MIMO) frequency-modulated-continuous-wave (FMCW) millimeter-wave (mmWave) radar. First, leveraging the scattering point cloud obtained from radar, a multitarget localization algorithm is proposed, which iteratively filters the scattering points to improve the accuracy of target localization. A signal separation algorithm is then developed to obtain motion signals for each target, wherein an adaptive beamforming algorithm with dynamic beamwidth and empowered by linearly constrained minimum variance (LCMV) is proposed to alleviate the interference among different targets. Based on the results of signal separation, the M-HMR problem is converted into single-target HMR, and a convolutional neural network (CNN) is used to achieve the classification of different motions. Tested on real datasets collected using the mmWave radar system, MMWCAS-RF-EVM, from Texas Instruments, the proposed approach demonstrates high accuracy in isolating target signals and classifying motions, showcasing its potential for complex HMR applications.

To better understand whale vocalizations, it is essential to simultaneously extract the components of whistles and clicks. However, prevailing methods in whale signal processing have focused on analyzing individual whistles or clicks. In this letter, we propose a novel method for separating whistles and clicks, even when these components overlap. This method combines a time-frequency (TF)-varying Gaussian window with the synchro-compensating Chirplet transform to achieve variable anisotropy, concentrating the TF energy along the ridges of the whistle and click components in the TF representation (TFR) and computing their directional angles. By using the enhanced TFR and these directional angles, a bidirectional ridge splitter is developed to separate the components of overlapping whistles and clicks. The effectiveness of this method is validated through comparative experiments using both simulated and real whale signals.

This paper considers multiuser millimeter wave (mmWave) MIMO communication systems where both the base station (BS) and the user equipments (UEs) use analog beamforming (BF). We propose a beam acquisition approach to find the best directions for the narrow beams, required to achieve BS-UE links above a minimum signal-to-noise ratio (SNR) threshold. The multiple UEs first select their beams simultaneously, based on signals transmitted from the BS on wide training beams. The BS then switches to a beam acquisition mode in order to select the best beam direction for each UE, based on training signals the UEs send on the uplink. We propose a novel parallel beam acquisition scheme for this uplink phase. We also propose a user rejection policy, in which weak UEs that do not satisfy a target SNR threshold, equivalent to a certain bit error rate, are rejected. To do so, we derive the posterior probability that a BS beam has a resultant SNR below the target SNR threshold. We also propose a generalized likelihood ratio test (GLRT) to reduce the computational burden. We derive closed-form error probability expressions for the proposed GLRT approach, and examine the trade-off between the beam acquisition time and accuracy in selecting the beams. We show that the beam acquisition time of our proposed approach is as low as 3% of existing approaches.

Transmission lines are an important part of the power system, and the normal operation of the transmission line is a key step to ensure the stable operation of the power system. The failure of transmission line can cause interruption of power supply and cause serious economic losses. This work uses Convolutional Neural Networks (CNN) to detect and locate the transmission line faults. In the model, a time window is used to perform sliding sampling on the fault data to generate the data set, which not only speeds up the training process, but also retains the temporal correlation of the data. Two CNNs are used to achieve fault detection and location, respectively. In addition, Gaussian white noise is added to the test set to evaluate the anti-interference ability of the model. The results show that the model has good performance and high stability for transmission line fault detection and location.

Deep learning methods have exhibited remarkable perforamance for fault diagnosis in power grid transmission lines. However, specific power systems often demand dedicated classifiers. Alterations to the structure of grid can negatively impact the accuracy of the previously trained models. Retraining a new model will cost a lot of time and sacrifice more economic resources. In such cases, transfer learning is able to transfer the knowledge acquired from a previous task to a new target task. Consequently, this notably decreases the training costs. In this study, a convolutional neural network (CNN) with transfer learning is implemented for fault diagnosis in AC/DC hybrid power systems with various structures. Through a series of simulations, feasibility and efficiency of this method for fault diagnosis are proved. This method also shows strong robustness in fault detection with varying fault inception angle, fault resistance, fault location, and system frequency shuffling. Moreover, the improved CNN model in this paper exhibits powerful generalization ability on test data. The aforementioned advantages illustrate the potentiality of this method for fault diagnosis in real power grid.

In this paper, we investigate a millimeter wave communication-centric integrated sensing and communication (ISAC) system with hybrid analog and digital (HAD) architecture with the application to vehicle to vehicle (V2V) network. Each vehicle node needs to perform two types of sensing: the panoramic close-range and forward long-range. We adopt multibeam technology for simultaneous forward long-range sensing and communication. Compared to traditional sensing-first-communicating-later (SFCL) scenario, simulation results show that the considered ISAC with HAD scenario can obtain higher communication throughputs while maintaining an acceptable sensing performance.

Covert communication prevents legitimate transmission from being detected by a warden while maintaining certain covert rate at the intended user. Prior works have considered the design of covert communication over conventional low-frequency bands, but few works so far have explored the higher-frequency millimeter-wave (mmWave) spectrum. The directional nature of mmWave communication makes it attractive for covert transmission. However, how to establish such directional link in a covert manner in the first place remains as a significant challenge. In this paper, we consider a covert mmWave communication system, where legitimate parties Alice and Bob adopt beam training approach for directional link establishment. Accounting for the training overhead, we develop a new design framework that jointly optimizes beam training duration, training power and data transmission power to maximize the effective throughput of Alice-Bob link while ensuring the covertness constraint at warden Willie is met. We further propose a dual-decomposition successive convex approximation algorithm to solve the problem efficiently. Numerical studies demonstrate interesting tradeoff among the key design parameters considered and also the necessity of joint design of beam training and data transmission for covert mmWave communication.

Millimetre wave (mmWave) beam tracking is a challenging task because tracking algorithms are required to provide consistent high accuracy with low probability of loss of track and minimal overhead. To meet these requirements, we propose in this article a new cost-effective analog beam tracking framework namely Adaptive Tracking with Stochastic Control (ATSC). Under this framework, beam direction updates are made using a novel mechanism based on measurements taken from only two beam directions perturbed from the current data beam. To achieve high tracking accuracy and reliability, we provide a systematic approach to jointly optimise the algorithm parameters. The complete framework includes a method for adapting the tracking rate together with a criterion for realignment (perceived loss of track). ATSC adapts the amount of tracking overhead that matches well to the mobility level, without incurring frequent loss of track, as verified by an extensive set of experiments under both representative statistical channel models as well as realistic urban scenarios simulated by ray-tracing software. In particular, numerical results show that ATSC can track dominant channel directions with high accuracy for vehicles moving at 72 km/hour in complicated urban scenarios, with an overhead of less than 1%.

In this paper, we propose an adaptive beam search algorithm for the initial alignment of millimetre-Wave beams. The proposed algorithm works by gradually deactivating beams that are unlikely the best beam from a pre-synthesised codebook to save overhead, based on a Bayesian probability criterion with a uniform improper prior. The beam deactivations can be implemented with low-complexity operations that require computing a low-degree polynomial or a search through a look-up table. The proposed algorithm does not require prior knowledge of channel statistics or signal to noise ratios (SNRs) to optimise the amount of searching time, and uses a suitable amount of time to achieve satisfactory beam search accuracy in different SNRs and fading scenarios. Numerical results confirm that the proposed algorithm can adapt to a wide range of channels with a fixed algorithm parameter, and can achieve better balance between beam search overhead and accuracy than non-adaptive approaches with fixed overhead.

Millimeter wave (mmWave) multiple-input multiple-output (MIMO) communication systems with a large number of antennas are power hungry when using conventional high-resolution analog-to-digital/digital-to-analog converters (A/Ds). To reduce the power consumption of mmWave MIMO systems, existing studies have considered hybrid structures with a reduced number of high-resolution or low-resolution A/Ds at either the transmitter or the receiver side. In this paper, we propose and investigate a multi-user hybrid architecture with low-resolution A/Ds equipped at both the transmitter and the receivers. To mitigate the impact of utilizing low-resolution A/Ds at the transceivers, we propose a novel data transmission scheme, which exploits a weighted phased-array to synthesize the beamforming matrix in the analog domain so as to mitigate inter-user interference. Under the scheme proposed, we derive the achievable rate and the energy efficiency to establish guidelines on the optimal resolution choice of A/Ds for hybrid mmWave systems. For a typical total transmit power at the BS, e.g., 30 dBm, the proposed scheme with 5~6-bit A/Ds can significantly improve the energy efficiency by as much as 100% over that of the conventional hybrid MIMO architecture with high-resolution A/Ds (10-bit A/Ds), without significant degradation in data rate performance.

Millimeter Wave (mmWave) communications rely on highly directional beams to combat severe propagation loss. In this paper, an adaptive beam search algorithm based on spatial scanning, called Iterative Deactivation and Beam Shifting (IDBS), is proposed for mmWave beam alignment. IDBS does not require advance information such as the Signal-to-Noise Ratio (SNR) and channel statistics, and matches the training overhead to the unknown SNR to achieve satisfactory performance. The algorithm works by gradually deactivating beams using a Bayesian probability criterion based on a uniform improper prior, where beam deactivation can be implemented with low-complexity operations that require computing a low-degree polynomial or a search through a look-up table. Numerical results confirm that IDBS adapts to different propagation scenarios such as line-of-sight and non-line-of-sight and to different SNRs. It can achieve better tradeoffs between training overhead and beam alignment accuracy than existing non-adaptive algorithms that have fixed training overheads.

Swift and accurate alignment of transmitter (Tx) and receiver (Rx) beams is one of the fundamental design challenges to support directional transmission in millimeter-wave cellular communications. In this paper, we propose a new Optimized Two-Stage Search (OTSS) algorithm for Tx-Rx beam alignment via beam training. In contrast to one-shot exhaustive search, OTSS judiciously divides the training energy budget into two stages. In the first stage, OTSS explores and trains all candidate Tx-Rx beam pairs and then discards a set of less favorable pairs learned from the measured received signal. In the second stage, OTSS takes an extra measurement for each of the remaining pairs and combines with the previous measurement to determine the best one. For OTSS, we derive fundamental bounds on its misalignment probability under a single-path channel model with ideal codebooks and establish a guideline on its optimized parameter choices. Numerical results have confirmed the advantage of OTSS over the state-of-the-art baselines.

Swift and accurate alignment of transmitter (Tx) and receiver (Rx) beams is a fundamental design challenge to enable the reliable outdoor millimeter-wave communications. In this paper, we propose a new optimized two-stage search (OTSS) algorithm for Tx–Rx beam alignment via spatial scanning. In contrast to one-shot exhaustive search, the OTSS judiciously divides the training energy budget into two stages. In the first stage, OTSS explores and trains all candidate beam pairs and, then, eliminates a set of less favorable pairs learned from the received signal profile. In the second stage, OTSS takes an extra measurement for the each of the survived pairs and combines with the previous measurement to determine the best one. For the OTSS, we derive an upper bound on its misalignment probability, under a single-path channel model with training codebooks having an ideal beam pattern. We also characterize the decay rate function of the upper bound with respect to the training budget and further derive the optimal design parameters of OTSS that maximize the decay rate. OTSS is proved to asymptotically outperform the state-of-the-art beam alignment algorithms and is numerically shown to achieve better performance with limited training budget and practically synthesized beams.

In millimeter wave cellular communication, fast and reliable beam alignment via beam training is crucial to harvest sufficient beamforming gain for the subsequent data transmission. In this paper, we establish fundamental limits in beam-alignment performance under both the exhaustive search and the hierarchical search that adopts multi-resolution beamforming codebooks, accounting for time-domain training overhead. Specifically, we derive lower and upper bounds on the probability of misalignment for an arbitrary level in the hierarchical search, based on a single-path channel model. Using the method of large deviations, we characterize the decay rate functions of both bounds and show that the bounds coincide as the training sequence length goes large. We go on to characterize the asymptotic misalignment probability of both the hierarchical and exhaustive search, and show that the latter asymptotically outperforms the former, subject to the same training overhead and codebook resolution. We show via numerical results that this relative performance behavior holds in the non-asymptotic regime. Moreover, the exhaustive search is shown to achieve significantly higher worst case spectrum efficiency than the hierarchical search, when the pre-beamforming signal-to-noise ratio (SNR) is relatively low. This paper hence implies that the exhaustive search is more effective for users situated further from base stations, as they tend to have low SNR.

In this paper, we develop an analytical framework for the initial access (also known as base station (BS) discovery) in a millimeter-wave communication system and propose an effective strategy for transmitting the reference signals (RSs) used for BS discovery. Specifically, by formulating the problem of BS discovery at user equipments (UEs) as hypothesis tests, we derive a detector based on the generalized likelihood ratio test and characterize the statistical behavior of the detector. The theoretical results obtained allow analysis of the impact of key system parameters on the performance of BS discovery, and show that RS transmission with narrow beams may not be helpful in improving the overall BS discovery performance due to the cost of spatial scanning. Using the method of large deviations, we identify the desirable beam pattern that minimizes the average miss-discovery probability of UEs within a targeted detectable region. We then propose to transmit the RS with sequential scanning, using a pre-designed codebook with narrow and/or wide beams to approximate the desirable patterns. The proposed design allows flexible choices of the codebook sizes and the associated beam widths to better approximate the desirable patterns. Numerical results demonstrate the effectiveness of the proposed method.

Coordinated scheduling/beamforming (CS/CB) is a cost-effective coordinated multipoint (CoMP) transmission paradigm that has been incorporated in the recent long-term evolution cellular standard. In this paper, we study CS/CB with the aim of developing low-complexity multicell coordinated user scheduling policies. We focus on a class of multicell interfering broadcast networks in which base stations have only local data and local channel state information, but each has sufficient antennas to serve multiple users using zero-forcing beamforming. The coordination problem is formulated as finding scheduling decisions across the cells such that the network sum rate is maximized. Starting from the two-cell model, we uncover the structure for a good scheduling decision, which in turn leads to the definition of two distributed scheduling policies of differing complexity and intercell coordination. Asymptotic theoretical bounds on the average sum rate are derived to predict the performance of the policies proposed. We extend to some example networks containing more than two cells and develop network-wide coordination policies. Numerical results confirm the effectiveness of the proposed policies and shed light on practical coordinated system design.

Heterogeneous wireless networks (HetNets) provide a means to increase network capacity by introducing small cells and adopting a layered architecture. HetNets allocate resources flexibly through time sharing and cell range expansion/contraction allowing a wide range of possible schedulers. In this paper, we define the capacity of a HetNet down link in terms of the maximum number of downloads per second, which can be achieved for a given offered traffic density. Given this definition we show that the capacity is determined via the solution to a continuous linear program (LP). If the solution is smaller than 1 then there is a scheduler such that the number of mobiles in the network has ergodic properties with finite mean waiting time. If the solution is greater than 1 then no such scheduler exists. These results have clear implications for network planning. The above results continue to hold if a more general class of schedulers is considered.

We investigate a joint user association and resource allocation problem in a three-tier heterogeneous network. Orthogonal resource allocation among different tiers is assumed. The problem is formulated as minimizing the total resources required to satisfy given user traffic demands. We first examine the structure of the optimal solution for this convex optimization problem and show that the optimal user association and hence resource allocation is determined by a bias value on the data rate offered by each base station. We then develop distributed algorithms based on the dual ascent method in determining the optimal rate bias for each BS. Numerical experiments demonstrate that the developed algorithms converge fast and produce close-to-optimal solutions. It is also shown that under cross-tier orthogonal resource allocation, the three-tier deployment provides significant performance improvement over a two-tier deployment, using the same set of base stations in each case.

We investigate a coordinated scheduling problem in a two-cell network where in each cell, two users are scheduled for simultaneous communication. Zero-forcing (ZF) beamforming is employed at each base station to suppress both intra- and inter-cell interference. The coordinated scheduling/beamforming problem is formulated as finding proper scheduling decisions and hence beamformers across the network such that a weighted sum-throughput is maximized. We propose three distributed scheduling policies that only require local data and local channel state information at each cell, and consume much less computation and communication overhead than the global optimization approach via exhaustive search. The proposed policies illustrate the complexity-performance tradeoff for the coordinated system. Nevertheless, numerical results show that at all levels of complexity, the proposed policies perform close to the global optimization approach with ZF beamforming and outperform the scheme with matched filtering beamforming even with global coordination.

We consider a coordinated multiuser scheduling problem for a multicell mutually interfering broadcast network. In particular, we focus on a two-cell cluster, where both base stations have only local data and local channel state information, but each has sufficient number of antennas to serve multiple homogeneous users under a full zero-forcing beamforming transmission. The scheduling problem is formulated as finding proper scheduled users and hence beamformers across the cells such that the sum rate is maximized. We uncover the structure for a good scheduling decision, which in turn motivates three distributed coordinated scheduling policies of different levels of complexity. For the simplest policy, we derive a lower bound on the expected achievable sum rate. It is shown in the large user population limit, the simplest policy suffices to preserve the best possible multiplexing gain and multiuser diversity gain for the model studied, but it does induce a pairing loss on the sum rate due to the limited coordination between cells.

This paper proposes a novel and mechanically decoupled leg and wheel hybrid transformable robot called HyTRo-I that combines two mobility concepts. For example, while wheeled vehicles shares higher speed than legged and tracked machines on a flat ground, they have relatively lower degree of flexibility than the other two on irregular terrain. The HyTRo-I robot evolves three motion modes: wheeled rolling, quadrupedal walking and leg-wheel hybrid mode. Despite the over-whelming complexity of obstacles, only several typical obstacles are selected for the study, which are stairs, large protrusions and ditches. Firstly, the transition locomotion mechanism between wheeled rolling mode and quadrupedal walking mode should be studied in detail. In the course of reciprocal transition locomotion, the static and reversible transformation gait not only guarantees the shifting stability and a small number of transition steps, but also the relatively balanced torque of joints. Secondly, after HyTRo-I converting to effective locomotion mode, the adaptive gait control strategies are proposed to traverse three types of obstacles. Finally, a serial of experiments were performed to verify the validity of the proposed transformation gait and adaptive step-up gaits.

This work studies a novel power-efficient precoder design problem for a linear cellular array with base station (BS) cooperation: data symbols intended for mobile stations (MSs) are drawn from discrete finite alphabets, precoding is performed among BSs to produce appropriate signals transmitted over the channel, symbol-by-symbol detection is performed at each MS, and a minimum Symbol Error Probability (SEP) for detection is introduced as the Quality-of-Service (QoS) metric at each MS. With regular constellations such as 16-QAM deployed as system data inputs, the SEP constraints are formulated and characterized by a set of convex relaxations on the received signals. A convex power optimization problem is then formulated subject to the SEP constraints. By the primal-dual decomposition approach, a distributed algorithm is developed to solve the problem in which only local communication among BSs is required. Our scheme is shown to significantly outperform linear zero-forcing precoder in terms of transmit power consumption.

In this letter, we propose a low-complexity channel-estimate based adaptive linear equalizer. The equalizer exploits coordinate descent iterations for computation of equalizer coefficients. The proposed technique has as low complexity as operations per sample, where and are the equalizer and channel estimator length, respectively, and is the number of iterations such that and . Moreover, with dichotomous coordinate descent iterations, the computation of equalizer coefficients is multiplication-free and division-free, which makes the equalizer attractive for hardware design. Simulation shows that the proposed adaptive equalizer performs close to the minimum mean-square-error equalizer with perfect knowledge of the channel.