Underwater activities are becoming more and more important, and underwater optical-wireless-communication (OWC) will be needed. In this work, we present and demonstrate an underwater optical-camera-communication (UWOCC) employing long-short-term-memory-neural-network (LSTM-NN). The optical transmitter (Tx) is a PMMA side-glow optical fiber, acting as an "omnidirectional optical antenna" and providing wide field-of-view (FOV) light emission. By utilizing LSTM-NN, the bit-error-rate (BER) of the UWOCC system can be improved and the data-rate is enhanced to 2.7-kbit/s, satisfying the pre-forward-error-correction (FEC) threshold.
In computer vision for aerial images, finding correct point correspondence among images plays an important role in many applications, such as image stitching, image retrieval, visual localization, etc. Most of the research works focus on the matching of local feature before a sampling method is employed, such as RANSAC, to verify initial matching results via repeated fitting of certain global transformation among the images. However, incorrect matches may still exist. Thus, a novel sampling scheme, Pentagon-Match (PMatch), is proposed in this work to verify the correctness of initially matched keypoints using pentagons randomly sampled from them. By ensuring shape and location of these pentagons are view-invariant with various evaluations of cross-ratio (CR), incorrect matches of keypoint can be identified easily with homography estimated from correctly matched pentagons. Experimental results show that highly accurate estimation of homography can be obtained efficiently for planar scenes of the HPatches dataset, based on keypoint matching results provided by LoFTR.
Foundation models represented by ChatGPT, have initiated an outstanding revolution across various domains. With the pre-trained foundation model in specific fields, numerous downstream tasks exhibit state-of-the-art performances. This paper extends this paradigm to automatic modulation classification (AMC), employing a masked autoencoder vision transformer (ViT-MAE) as a foundation model to advance AMC. The experimental results show that our signal constellation diagram-based foundation model outperforms traditional deep learning methodologies, underscoring the vast potential of foundation models in AMC and wireless communication systems.
To save power during surveillance, birds or bats can perch and rest on branches. To date, few flapping-wing drones can perch to a branch despite being capable of flying agilely in multiple flight modes. Due to their small size and limited payload, the flight endurance and surveillance times of flapping drones were short. For example, our 63-gram prototype of the hoverable flapping drone can only hover in the air for less than 5 minutes using an 11-gram lithium polymer battery. Extra loading with a camera and grippers will further reduce the endurance by 40%. The passive perching requires no flight power. This motivated our development of a 6-gram swing hook for branch perching just like the hangar hook with a lifting strength of up to 2 kilograms. With a hook onboard, the drone can perch upside down from a branch, significantly extending the surveillance duration. In addition, we outlined the principle of how pilot vision helps to aim the hook engagement to a branch by proper leveling and approaching speed. In the future, a firstperson view could further help the flight maneuver to perch indoors.
Integrated space-air-ground network (SAGIN) has become a key approach to achieve seamless global coverage and efficient information transmission. However, the requirements for task latency and the limited energy supply of unmanned aerial vehicle UAV) pose significant challenges when it acts as edge nodes to provide computation offloading services. To address the computation offloading problem in SAGIN, this paper proposes an optimization solution based on the Deep Deterministic Policy Gradient (DDPG) algorithm. We introduce behavior noise and employ state normalization to preprocess the observed state. Our aim is to minimize the energy consumption and task processing latency of UAV by jointly optimizing the UAV's flight trajectory, transmission power, task offloading ratio, and offloading destinations selection. Experimental results validate the effectiveness and feasibility of the proposed method, showing that the DDPG algorithm has significant advantages over some other intelligent algorithms in reducing energy consumption and processing latency.
Vehicular edge computing (VEC) is a crucial paradigm in B5G networks, aiming to offload computation-intensive and latency-sensitive tasks for vehicular users (VUs). However, not all tasks are time-critical; some are soft real-time (SRT), i.e., tasks with soft deadlines, and some are hard real-time (HRT), i.e., tasks with hard deadlines. Most of the prior studies primarily focus on either SRTs or HRTs and the realistic network with the culmination of both SRTs and HRTs with more priority to HRTs is needed to be explored more. In this paper, we investigate a non-orthogonal multiple access (NOMA) assisted VEC network and propose an offloading and resource allocation approach that considers the criticalness of task deadlines. To improve the prioritized service miss function, we propose joint computational offloading and resource allocation schemes under allowable power and computational resource constraints. This results in a mixed integer non-linear programming problem that is solved using continuous relaxation, successive convex approximation, and alternating optimization. Simulations are conducted to illustrate the efficacy of the proposed scheme.
Massive access plays an important role in establishing future 6G networks. In an asynchronous massive multiple-access channel (MAC) where user activities and channel delays are unknown, it is critical to detect and estimate them in order to ensure reliable data transmission. Even in asynchronous unsourced MAC, a promising data transmission scheme for massive MAC, this information is also crucial. In this paper, we propose a tropical-arithmetic method for a joint delay and user activity detection, which can support a massive number of users with only a few orthogonal wireless resources. A necessary condition for the maximum path delay to ensure correct detection is also identified. Simulation results verify that our designs have zero detection errors even when the number of users scales quadratically with the number of orthogonal resources.
Polarization-based, direct-detection modulation for-mats, such as Stokes vector modulation (SVM) for single-mode links and mode vector modulation (MVM) for multimode links, have garnered significant attention due to their potential to improve the spectral efficiency and energy consumption of short-haul communications. In this invited paper, we review the latest studies on SVM and MVM, including the optimized geometric constellation shaping and bit-to-symbol mapping, the accurate modeling of MVM propagation over few-mode fibers (FMFs), and the the experimental implementation of various SVM/MVM constellations.
The Discontinuous Reception (DRX) plays a vital role in user equipment (UE) power saving with a tradeoff of additional latency. Heterogeneous UEs may have specific DRX parameters that best suit their communication and hardware requirements. As 3GPP improved the signaling process of UE assistance information, UE can now send requests for their preferred DRX configuration to the base station (BS), while the BS still holds the right of actual configuration. In light of this opportunity, we propose Configuration as a Service (ConfigaaS), where Mobile Network Operators (MNOs) commit to configuring UEs according to their preferred DRX parameters indicated in the UE assistance information under a pricing framework. To gauge the profit potential for MNOs, we explore the value function of UEs in relation to different DRX configurations. From the perspective of an MNO, we develope a framework for maximizing profits through rate-setting, under different degrees of prior knowledge about UEs and two pricing strategies. Our simulations demonstrate the achievable profits with uniform and differentiated pricing compared to the full-information benchmark. Moreover, the analysis delves into the conditions and the extent that the profit differs between these two pricing schemes.
To address the ill-posed nature of image restoration tasks, recent research efforts have been focused on integrating conditional generative models, such as conditional variational autoencoders (CVAE). However, how to condition the autoencoder to maximize the conditional evidence lower bound remains an open issue, particularly for the restoration tasks. Inspired by the rapid advancements in CVAE-based video compression, we make the first attempt to adapt a conditional video compressor for image restoration. In doing so, we have the low-quality image to be enhanced, which plays the same role as the reference frame for conditional video coding. Our scheme applies scalar quantization in training the autoencoder, circumventing the difficulties of training a large-size codebook as with prior works that adopt vector-quantized VAE (VQ-VAE). Moreover, it trains end-to-end a fully conditioned autoencoder, including a conditional encoder, a conditional decoder, and a conditional prior network, to maximize the conditional evidence lower bound. Extensive experiments confirm the superiority of our scheme on denoising and deblurring tasks.
This study proposes an autonomous vision-based UAV system for low-altitude road following, overcoming drawbacks of remote operation and GPS reliance. The system autonomously tracks road centerlines and avoids obstacles. Validated through experiments, it has 0.55-meter centerline error and 99.33% of the time within the range on a 1-kilometer, 7-meter wide road, demonstrating feasibility and robustness. It is capable of long-distance road following without heavy GPS reliance and effective obstacle avoidance, making it suitable for various lowaltitude road following missions.
Deepfake detection becomes necessary because Deep-fakes allow anyone's image to be co-opted and lead to the severe trust issues. Despite the popularity of deepfake videos, very few temporal-based solutions rely can be found. In this paper, we consider temporal information in deepfake detection. In particular, we consider a temporal-attention module, in addition to a spatial-CNN for spatial features. By taking advantage of the temporal consistency, our method significantly improves generalization ability. Our method outperforms the prior work in the cross-dataset setting and demonstrate the temporal-attention module's importance.
This study investigates the effectiveness of multi-user multiple-input and multiple-output (MU-MIMO) Tomlinson-Harashima precoding (THP) in low Earth orbit (LEO) satellite communication (SATCOM) systems. The objective is to explore the potential of complementing terrestrial networks with LEO SATCOMs for coverage enhancement in sixth-generation mobile communication systems (6G). In particular, we focused on line-of-sight environments unique to LEO SATCOMs and adopted THP as a spatial-correlation-tolerant MU-MIMO precoding scheme. To evaluate the performance of THP, we utilized a channel model based on 3GPP technical report 38.811, which represents the communication between the LEO satellite equipped with a uniform planner array antenna and terrestrial user terminals. The system capacity of THP was derived from a mod-Lambda-channel-based analysis. To provide a comprehensive comparison, we conducted computer simulations to contrast the performance of THP and that of the traditional linear precoding (LP). The numerical results obtained from this study shed light on the applicability of THP in LEO SATCOM systems.
Affective computing is an active area of research, but how to efficiently transmit the sensed data to the server for affective computing is less investigated. In this paper, we propose affective communication for affective computing, with the wireless link from the affection-sensing devices to the affective-computing server being semantic communication. The semantic communication problem asks how precisely the transmitted symbols convey the desired meaning, and thus the semantic communication for affective computing is the most efficient if the meaning of the sensed affection data is conveyed for affective computing with minimum wireless resources used. Deep neural networks (DNNs) are adopted as the semantic-channel encoder and the semantic-channel decoder for end-to-end joint design. Simulations using the FER2013 facial expression recognition dataset shows the effectiveness of the proposed DNN-based semantic communication codec in affective communication for affective computing. Furthermore, federated learning for affective communication is investigated for privacy concerns.
Metasurfaces have attracted much attention because they offer versatile optical functionalities with ultracompact footprints. This session presents several emerging topics on metasurfaces, covering the design, fabrication, and applications.
In practical machine learning applications, image classification often involves discerning the category to which an object belongs. In this paper, we present a novel approach called Binary Episode Classifier (BEpiC) within the context of meta-learning. BEpiC aims to optimize episode generation by strategically selecting training samples. The methodology involves training the initial class with a comprehensive set of similar images while the contrasting class is exposed to a diverse array of dissimilar images. We create highly dissimilar image sets by randomly forming multiple image clusters. Identifying the cluster centers and selecting the representative image closest to each center facilitates the determination of the among clusters. The distance of these images from the other class is then calculated. The cluster whose representative image exhibits the greatest distance is chosen as the definitive class. Our proposed method showcases significant efficacy for binary meta-learning classification across diverse classes. While our experimentation primarily focused on Model-Agnostic Meta-Learning (MAML), the adaptability of the episode generation strategy extends to a spectrum of meta-learning classifiers. Empirical findings substantiate that the proposed method attains remarkable accuracy in one-shot classification scenarios and moderately higher accuracy in few-shot classification tasks.
This paper introduces a novel compressed sensing channel estimation method, termed Attenuated-RMMP, tailored for non-orthogonal waveforms-based orthogonal time frequency space (OTFS) systems in high fractional Doppler shift communications. Leveraging a reduced cyclic prefix OTFS (RCP-OTFS) framework with rectangular pulse shaping, the proposed approach constructs a sensing matrix and applies regularized multipath matching pursuit (RMMP) for channel estimation. To address computational complexity, an attenuated regularization step is introduced in the RMMP process, reducing runtime while maintaining BER performance. Additionally, a Doppler pre-compensation mechanism is proposed to handle high Doppler shifts effectively. Numerical simulations validate the effectiveness of the proposed scheme in terms of bit error rate (BER) and computational complexity.
By incorporating apodization and double-layer grating structures, a broadband silicon photonics Mode-Division-Multiplexing (MDM) grating coupler is successfully designed with a broad bandwidth of 62nm above 40% coupling efficiency for all three LP 01 , LP 11a , and LP 11b modes of the few-mode fiber.
Currently, unmanned aerial vehicles (UAVs) have a wide range of applications. One such application is the detection of marine pollution, for which numerous methods have been proposed. While validation on public datasets has shown promising results, our practical experience has revealed that the color of the sea can vary across different regions, leading to inaccurate detection results. This paper presents a marine pollution detection model that is highly stable and capable of accurately detecting pollution in various environments without requiring additional training data. Our evaluation results show that the proposed model improves the IoU by about 13% and 14%, and the F1-score by about 10% and 23.9% for ocean and marine pollution, respectively, compared to the baseline model on unseen domains, proving its effectiveness.
Unlike traditional narrowband communication antennas that operate within specific frequency bands, UWB antennas utilize frequencies across a wide range, typically spanning from 3.1 GHz to 10.6 GHz. However, most of the existing antennas are characterized by low gain, and there is a growing need to extend the existing UWB frequency range. To achieve this, suitable antennas are required, especially those with high gain and extended bandwidth. On this note, a compact coplanar waveguide (CPW)-fed ultra-wideband (UWB) elliptical patch antenna supported by a reflectarray metasurface (MS) for gain augmentation is reported in this article. The antenna and MS have planar dimensions of 40 mmx40 mm and 60 mmx60 mm respectively. The CPW ground features a semi Omega-shaped slot and a 12-gon elliptical slot for enhanced surface current distribution and bandwidth, respectively. Besides, a simple equivalent circuit representation for the UC is proposed. Our design attains a peak gain of 10.8 dBi, a total efficiency above 90%, and a 149% -10 dB fractional bandwidth (FBW) (2.6-17.75 GHz). The proposed structure is ideal for UWB energy harvesting, sensing, and communication within the C-band (4-8 GHz), X-band (8-12 GHz), and Ku-band (12-18 GHz) due to its large FBW, high total efficiency, and quasi-end-fire radiation performance.
