Deep Channel Research Articles

Multiscale Convolution-Based Efficient Channel Estimation Techniques for OFDM Systems

With the advancement of wireless communication technology, the significance of efficient and accurate channel estimation methods has grown substantially. Recently, deep learning-based methods are being adopted to estimate channels with higher precision than traditional methods, even in the absence of prior channel statistics. In this paper, we propose two deep learning-based channel estimation models, CAMPNet and MSResNet, which are designed to consider channel characteristics from a multiscale perspective. The convolutional attention and multiscale parallel network (CAMPNet) accentuates critical channel characteristics by utilizing parallel multiscale features and convolutional attention, while the multiscale residual network (MSResNet) integrates information across various scales through cross-connected multiscale convolutional structures. Both models are designed to perform robustly in environments with complex frequency domain information and various Doppler shifts. Experimental results demonstrate that CAMPNet and MSResNet achieve superior performance compared to existing channel estimation methods within various channel models. Notably, the proposed models show exceptional performance in high signal-to-noise ratio (SNR) environments, achieving up to a 48.98% reduction in mean squared error(MSE) compared to existing methods at an SNR of 25dB. In experiments evaluating the generalization capabilities of the proposed models, they show greater stability and robustness compared to existing methods. These results suggest that deep learning-based channel estimation models have the potential to overcome the limitations of existing methods, offering high performance and efficiency in real-world communication environments.

Confined and mined: anthropogenic river modification as a driver of flood risk change

Rivers are often confined by structures and subjected to aggregate mining. In dynamic rivers, these interventions cause changes to riverbed and bank topography that potentially cause changes in hydraulics and flood risk. Repeat, system-scale, high-resolution topographic surveys of the gravel-bed Bislak River, the Philippines, are used to quantify annual morphological change and, using two-dimensional hydraulic modelling, to simulate changes to flood risk. Aggregate mining exports sediment and creates pitted topography, and embankments cause both deeper channels and disconnect the river from its floodplain. The consequently increased channel capacity reduces flood risk, with up to a 5% decrease in inundated areas for 10- to 100-year return periods. Sediment deprivation also increases bed shear stress that can induce scour, infrastructure damage and increased flood impacts. Rising global floodplain populations and increasing demand for aggregate ensure that sustainably managing geomorphologically dynamic rivers to support floodplain development and mitigate flood impacts remains a pertinent challenge.

Sediment Characteristics Study at Sharm Obhur, the Red Sea using Multibeam Backscatter

Sharm Obhur, situated approximately 35 km north of Jeddah City along the Red Sea coast, is a prominent recreational area with diverse benthic habitats. Accurate mapping of seabed sediments is critical for characterizing marine ecosystems and elucidating benthic habitat distribution in this region. This study underscores the importance of assessing sediment distribution patterns in Sharm Obhur, particularly due to the prevalence of recreational activities and navigational traffic. This research investigates sediment dynamics in Sharm Obhur, a coastal creek near Jeddah, utilizing Multibeam Echo Sounder (MBES) backscatter data and grab sample analyses. The study focuses on understanding sediment distribution patterns and their implications for coastal management. MBES technology provides high-resolution data on seafloor composition and roughness, essential for assessing sediment types and bedforms. Through spatial alignment of MBES data with ground truth sediment samples, we analyze sediment characteristics and validate findings. Results indicate predominant sand presence with detectable mud in deeper channels and northern anthropogenically impacted areas. Gravel deposits correlate with coral formations near the creek edges. Tidal influence is evident, with increased sand presence at the creek entrance. The study demonstrates MBES backscatter's efficacy in mapping sediment distributions and seabed characteristics, benefiting marine science and coastal management. This research enhances knowledge of sediment transport processes, facilitating informed decision-making for sustainable coastal development and conservation efforts.

Deep Learning-Based Channel Prediction With Path Extraction

Deep Learning-Based Channel Prediction With Path Extraction

Deep Neural Network-Based Channel Estimation Scheme in Massive MIMO Systems

Channel estimation is one of the most essential technologies of wireless communication, as well as necessary indicators to measure the system’s quality. In recent years, the combination of channel estimation and deep learning techniques has raised wide attention and become one of the most popular research directions of machine learning applications in the wireless communication area. However, in wireless communication systems such as massive MIMO, it’s too complicated to take traditional channel estimation methods with shortcomings like insufficient precision or increased cost. With the combination of deep neural networks (DDN) and the least squares (LS) method, this paper focuses on the deep learning-based channel estimation for massive MIMO systems. Compared with traditional algorithms such as LS or minimum mean-square error method, the loss function of the results is significantly reduced while simulating for the same analog channel. The problem of lack of precision has been improved efficiently and the performance has been improved substantially, which can prove the adequacy of the design of the channel estimator.

Effects of river infrastructure, dredged material placement, and altered hydrogeomorphic processes: The stress ecology of floodplain wetlands and associated fish communities

Recognition of the habitat values of coastal and floodplain wetlands has inspired research and engineering to restore biological functions after widespread species declines. However, the restoration and management of tidal river floodplains requires a more complete understanding of anthropogenic stressors on hydrogeomorphology and ecological processes. River floodplains near the ocean are affected by localized diking and dredging as well as basin-wide stressors such as dams. We evaluated the effects of stressors versus spatial position, both longitudinal and lateral to the mainstem, using physical and biological response variables in river floodplain wetlands. We categorized historical and modern stressors on the hydrogeomorphic regime of the lower Columbia River and estuary, northeast Pacific coast, including basin-scale management and local impacts. Using this categorization, we analyzed 44 attributes using field-collected data from 50 floodplain wetlands. Attributes represent channels, floods, plant communities, and fish communities. Here we show that plant and fish communities are stratified by position along the estuarine–riverine gradient, in contrast to physical habitat characteristics, which are stratified by stressors on hydrogeomorphic regimes and in some cases the lateral distance from the mainstem river on tributaries. Spatial position relative to water-level dynamics and salinity more strongly affect the biota than does stressor history. Stressor effects were greatest on the geomorphology observed in formerly diked, now reconnected wetlands and in wetlands with a history of dredged material placement; historically diked sites had anomalously deep channels with larger cross-sectional areas while sites with dredged material had shallow channels and lower levels of organic carbon in sediment. In wetlands subject only to landscape-scale stressors such as flow alterations by dams, organic carbon levels were higher. These findings provide natural resource managers with opportunities to enhance similarity to natural conditions and better understand future wetland evolution from different baselines of stressor history and river position.

Dissolved oxygen criteria attainment in Chesapeake Bay: Where has it improved since 1985?

Many estuaries, including Chesapeake Bay, have suffered from undesirable conditions of algae blooms, poor water clarity, and low dissolved oxygen (DO). To better understand the status and trends of DO criteria attainment deficit (AD), we conducted a comprehensive assessment using monitoring data in the period of 1985–2022 and focused on the comparison of trends among 13 tidal systems. Our results provide timely and updated information to help address a critical management question, i.e., where has the DO criteria attainment improved since 1985? The overall AD showed statistically significant improvements in both the long term (1985–2022) and the short term (2006–2022), which can be mainly attributed to open water and deep channel designated uses, respectively. Among the 13 tidal systems, ten showed improving trends in the long term, with Mainstem Bay, Pocomoke, Potomac, and Tangier having statistically significant improvements. Over the short term, nine systems showed improving trends, with Chester and Upper Mainstem Bay tributaries having statistically significant improvements. The only statistically significant degrading trend was observed in York. In addition, a report card summary was produced to provide a concise depiction of the trends for the 13 systems and illustrate how their trends are affected by the designated uses, which is the first of its kind in the context of DO criteria assessment. These spatially explicit results can help target locations for further assessment and restoration. Although many of the systems and their applicable designated uses have not improved over time, the detected significant improvements in some systems present new evidence on ecosystem recovery in the context of DO criteria attainment. Overall, this work highlights the combined value of collecting long-term water-quality monitoring data and developing scientific assessment approaches, which is fundamental to understanding the status and trends of complex ecosystems, including Chesapeake Bay and other estuaries.

Global climatological dataset of undersea acoustic parameters derived from the NCEI World Ocean Atlas 2023

Acoustics is most effective in undersea detection, localization, and communication. Establishment of a global climatological dataset of undersea acoustic parameters becomes urgent. In building such a dataset, first we use the Thermodynamic Equation of Seawater-2010 (TEOS-10) to calculate the sound speed (SS) from the gridded temperature and salinity fields of the NOAA/NCEI World Ocean Atlas 2023. Second, we determine the depth of overall minimum from SS profile as the deep sound channel (DSC) axis depth, the depth of overall maximum between the surface and DSC axis as the sonic layer depth (SLD), the depth of the local minimum between SLD and DSC axis as the second sound channel (SSC) depth, and the depth with the SS equalling the maximum SS as the critical depth. Third, we obtain the SS at the surface, SLD, DSC axis, and SSC axis. Fourth, we determine the other acoustic parameters such as In-layer gradient, below layer gradient, DSC strength, SSC strength, depth excess, and surface duct cut-off frequency. The dataset is publicly available.

Low-Complexity Convolutional Neural Network for Channel Estimation

This paper presents a deep learning algorithm for channel estimation in 5G New Radio (NR). The classical approach that uses neural networks for channel estimation requires more than one stage to obtain the full channel matrix. First, the channel has to be constructed by the received reference signal, and then, the precision is improved. In contrast, to reduce the computational cost, the proposed neural network method generates the channel matrix from the information captured from a few subcarriers along the slot. This information is extrapolated by applying the Least Square technique only on the Demodulation Reference Signal (DMRS). The received DMRS placed in the grid can be seen as a 2D low-resolution image and it is processed to generate the full channel matrix. To reduce complexity in the hardware implementation, the convolutional neural network (CNN) structure is selected. This solution is analyzed comparing the Mean Square Error (MSE) and the computational cost with other deep learning-based channel estimation, as well as the traditional channel estimation methods. It is demonstrated that the proposed neural network delivers substantial complexity savings and favorable error performance. It reduces the computational cost by an order of magnitude, and it has a maximum error discrepancy of 0.018 at 5 dB compared to Minimum Mean Square Error (MMSE) channel estimation.

Insights Into Pool Boiling Heat Transfer on Minichannel Surfaces Through Point and Field Measurements

Abstract In this study, saturated pool boiling experiments were conducted on copper minichannel and flat surfaces at atmospheric pressure using water as the working fluid. The heat transfer performance was assessed through point measurements with a heater block, cartridge heaters, and thermocouples, as well as field measurements using a thin-foil heater and infrared thermography. Two copper minichannel surfaces with square cross sections of 1 mm (minichannel-1) and 2 mm (minichannel-2) side lengths were tested and compared to a flat surface. Minichannel-1 and minichannel-2 enhanced the critical heat flux (CHF) by 17% and 45%, respectively, and improved the heat transfer coefficient by 24–40% and 51–75%, respectively, compared to the flat surface. Minichannel-2 exhibited the lowest and most uniform boiling surface temperature, making it the best performer among the three. There was no significant change in departure frequency among the surfaces, and no significant change in departure diameter for the flat surface and minichannel-1. However, minichannel-2 had lower departure diameters due to its deeper channels, which prevented bubble coalescence and maintained low departure diameters. Additionally, minichannel-2 delayed vapor film formation by breaking it with its deeper fins, thereby improving CHF and slightly enhancing bubble dynamics. The enhancement in boiling heat transfer is primarily attributed to the increased surface area provided by the minichannels, with a minor contribution from improved bubble dynamics. However, the dominant factor in enhancing pool boiling heat transfer on minichannel surfaces is the increase in surface area.

Residual channel attention based sample adaptation few-shot learning for hyperspectral image classification.

Few-shot learning (FSL) uses prior knowledge and supervised experience to effectively classify hyperspectral images (HSIs), thereby reducing the cost of large numbers of labeled samples. However, existing few-shot methods ignore the correlation between cross-domain feature channels, and the feature representation ability is insufficient. To address above issue, this paper proposes a novel Residual Channel Attention Based Sample Adaptation Few-Shot Learning for Hyperspectral Image Classification(RCASA-FSL) for hyperspectral image classification (HSIC), which can capture and enhance cross-domain dependencies through multi-layer residual connection and random-based feature recalibration. Specifically, a Deep Residual Feature Channel Attention Mechanism (DRFCAM) is designed to obtain cross-domain dependencies by residual concatenation, and further the residual structure is stacked for mining depth discrimination information. Furthermore, a new Random-based Feature Recalibration Module (RFRM) is proposed to reassign the feature weights via random matrix, which fully explore feature weight relationships to guide the sample adaptation process. Besides, we design a joint loss function with combining the FSL loss and domain adaptive loss for further optimization model. Experiments conducted on several standard hyperspectral datasets demonstrate that the proposed RCASA-FSL is superior to other FSL techniques in both quantitative and qualitative aspects.

A joint self-information and Markovian model-driven deep channel estimation and feedback model for millimeter-wave massive MIMO systems

A joint self-information and Markovian model-driven deep channel estimation and feedback model for millimeter-wave massive MIMO systems

Channel reconstruction through improvised deep learning architecture for high-speed networks

Efficient acquisition of channel state information (CSI) is quite complicated process but immensely essential to exploit probable benefits of massive multiple input multiple output (MIMO) systems. Therefore, a deep learningbased model is utilized to estimate channel feedback in a massive MIMO system. The proposed improvised deep learning-based channel estimation (IDLCE) model enhances channel reconstruction efficiency by using multiple convolutional layers and residual blocks. The proposed IDLCE model utilizes encoder network to compress CSI matrices where decoder network is used to downlink reconstruct CSI matrices. Here, an additional quantization block is incorporated to improve feedback reconstruction accuracy by reducing channel errors. A COST 2,100 model is adopted to analyse performance efficiency for both indoor and outdoor scenarios. Further, deep learning-based model is used to train thousands of parameter and correlation coefficients much faster and to minimize computational complexity. The proposed IDLCE model evaluate performance in terms of normalized mean square error (NMSE), correlation efficiency and reconstruction accuracy and compared against varied state-of-art-channel estimation techniques. Excellent performance results are obtained with large improvement in channel reconstruction accuracy

Water collection through a directional leaf vein pattern by fast laser marker ablation of stainless-steel

Inspired by the natural water harvesting mechanisms of desert beetles, cactus thorns, and leaf veins, we designed a heterogeneous wettability surface with superhydrophilic pattern integrating leaf vein as the directional water transport main channel, attached capillary triangles as auxiliary channel plus a deep rough desorption channel on an overall superhydrophobic surface for an efficient water collection. A superhydrophilic surface was initially fabricated on the stainless steel disc by laser marker ablation allowing 1 μL droplet to spread completely to 0° within 0.12 s, followed by fluorine-containing coating transforming superhydrophilic surface to superhydrophobic one. Directional water transport patterns were then etched on the superhydrophobic surfaces by the secondary laser marker. The surface energy gradient and Laplace pressure induced by the pattern facilitated directional fast transport and efficient desorption of droplets, thus improving water collection efficiency. The enhancement mechanism of the water harvesting behavior for such surfaces was analyzed, with one focus on enhancing collection in hydrophobic regions with capillaries to reduce bouncing off loss and the other on improving balanced cycling of the collection process. At a fog flow rate of 1500 ml/h and 20 cm away from the fog outlet, the directional leaf vein-patterned 19.625 cm2 sized surface demonstrated a fog water collection rate (WCR) of 5.6 Kg·m-2·h-1 and first drop collection at the 49th s, an impressively short time rarely reported. Compared to the superhydrophobic, superhydrophilic samples, and the reference, WCR increased by 180 %, 62 %, and 59 %, respectively, and the first droplet collection time decreased by 73 %, 46 %, and 62 %, respectively. This efficient water collection method has huge potential in arid regions.

Structural Features of Porous Silicon Formed on Heavily Doped Plates of Single-Crystal Silicon with Electron Conductivity

Using scanning electron microscopy, the structures of the surface and internal regions of porous silicon obtained by anodizing heavily doped plates of single-crystal silicon with electron conductivity in a hydrofluoric acid solution at different current densities were studied. It is found that the porous silicon surface has dark gray and light gray pores, which differ in size and surface distribution density. Dark gray pores possess larger sizes, and their density is about 5–10 times less than that of light gray pores. Based on the cross-section imagery, it is shown that light gray pores correspond to underdeveloped channels of small depth, while dark gray pores are the entrance points of deep bottle-shaped channels passing from the surface into the depth of the silicon wafer. The equivalent diameters of light gray pores on the surface of porous silicon are 12–15 nm and are practically independent of the anodic current density. At the same time, the equivalent diameters of dark gray pores and average distances between their centers increase linearly from 15 to 35 nm on the surface and from 35 to 120 nm in the volume of porous silicon when the current density is increased from 30 to 90 mA/cm2. The average thickness of silicon skeleton elements is about 3 nm on the surface and increases to 5–6 nm in the volume. By setting the density of the anode current, it is possible to obtain layers of porous silicon with different structural parameters. The obtained research results have practical significance for the formation of composite materials based on porous silicon, which can be used as a porous matrix for the deposition of metals and semiconductors.

Antarctic ice-shelf meltwater outflows in satellite radar imagery: ground-truthing and basal channel observations

Abstract Ice shelves regulate the flow of the Antarctic ice sheet toward the ocean and its contribution to sea-level rise. Accurately monitoring the basal and surface melting of ice shelves is therefore essential for predicting the ice sheet's response to climatic warming. In this study, we utilize Sentinel-1A synthetic aperture radar satellite imagery combined with shipboard measurements of water temperature and salinity to investigate the presence of surficial meltwater plumes along the Antarctic coastline. Our approach reveals a strong correlation between areas of pronounced low radar backscatter extending from ice shelves and significant decreases in water temperature and salinity, suggesting meltwater-enriched ocean waters. We propose that the low radar backscatter signature of meltwater outflows is caused by stable stratification of the upper water column, driven by density contrasts from buoyant, low-salinity meltwater and surface current shear that reduce Bragg scattering waves. The resulting smooth water surfaces were observed adjacent to the surface expression of deep basal channels, documented in a helicopter survey along part of the Bellingshausen Sea ice edge. We present high-temporal resolution satellite radar as a tool for identifying meltwater release from beneath ice shelves, capable of all-weather, day-and-night imaging.

Facilitating water droplet removal from surfaces using air flow and wettability gradients

A new method is proposed to mitigate ice accretions on wind turbine blades via the creation of a microstructural gradient surface geometry that facilitates spontaneous water droplet motion along the surface. The wettability gradients are formed by laser etching 35 μm wide, 35 μm deep channels into aluminum to form a surface with a gradually increasing solid area fraction. Different design variations are then proposed and systematically evaluated on the merits of their performance. An analytical model is also derived based on a balance of hysteresis and drag forces to predict the critical air speed necessary for droplet movement as a function of the droplet size and surface contact angle. Experimentation has shown good agreement with the model for both the baseline and fixed-pitch channel surfaces and has also demonstrated that, in certain cases, up to 70 % lower critical air speeds are needed to initiate droplet motion on these microstructured surfaces.

Deep learning-driven channel coding: enhancing performance in LDS-CDMA and NOMA systems

This paper presents an innovative approach to channel coding for LDS-CDMA and NOMA systems through the integration of deep learning techniques. By leveraging neural networks within the channel coding process, the proposed model achieves substantial improvements in error correction, leading to a significant reduction in Bit Error Rate (BER) and computational complexity when compared to traditional methods such as Unpunctured Turbo Trellis Coded Modulation (UTTCM). The results demonstrate that the deep learning-based system is particularly effective in challenging communication environments, such as those with low Signal-to-Noise Ratios (SNR) and high levels of multi-user interference, where it outperforms conventional coding techniques. This research underscores the potential of deep learning to enhance the efficiency and adaptability of wireless communication systems, particularly as networks evolve towards 6G and beyond. The flexibility offered by this approach allows for improved handling of dynamic and complex environments, ensuring higher data throughput and reduced latency. Future work may focus on implementing the proposed model in real-time applications, as well as investigating its synergy with emerging technologies like massive MIMO and NOMA, which could lead to even greater performance enhancements in next-generation communication systems.

Improving Non-Line-of-Sight Identification in Cellular Positioning Systems Using a Deep Autoencoding and Generative Adversarial Network Model

Positioning service is a critical technology that bridges the physical world with digital information, significantly enhancing efficiency and convenience in life and work. The evolution of 5G technology has proven that positioning services are integral components of current and future cellular networks. However, positioning accuracy is hindered by non-line-of-sight (NLoS) propagation, which severely affects the measurements of angles and delays. In this study, we introduced a deep autoencoding channel transform-generative adversarial network model that utilizes line-of-sight (LoS) samples as a singular category training set to fully extract the latent features of LoS, ultimately employing a discriminator as an NLoS identifier. We validated the proposed model in 5G indoor and indoor factory (dense clutter, low base station) scenarios by assessing its generalization capability across different scenarios. The results indicate that, compared to the state-of-the-art method, the proposed model markedly diminished the utilization of device resources and achieved a 2.15% higher area under the curve while reducing computing time by 12.6%. This approach holds promise for deployment in future positioning terminals to achieve superior localization precision, catering to commercial and industrial Internet of Things applications.

How human activities reshape carbon burial in estuarine sediment systems: Evidence from the Yangtze Estuary

Estuaries are critical zones for understanding the dynamics of organic carbon (OC) burial, particularly in the context of human intervention, yet detailed information remains limited. This study presents comprehensive field observations aimed at assessing the spatial and temporal variations in sedimentary OC sources, composition, and distribution within the Yangtze Estuary. By employing a three-endmember estimation, we determined the contributions of fluvial, marine, and salt marsh plants to OC in the mouth bar as 52 ± 6 %, 28 ± 12 %, and 20 ± 9 %, respectively. To address the complexities introduced by human activities and dispersal processes, we utilized a PCA-MC four-endmember model, which revealed that redispersal terrigenous organic matter from the outer estuary, categorized as “marine origin”, accounted for 26 ± 10 % re-deposited within the estuary. During the construction of the Deep Channel Navigation Project (DCNP), we observed a shift in terrigenous OC deposition from the North Passage (NP) to the South Passage (SP), along with an increased contribution from salt marsh plants. Post-DCNP completion, sedimentary OC compositions exhibited significant differences between the NP and SP, with the SP enriched in fresh terrigenous OC and the NP showing degraded signals due to dredging activities and source discrimination. Human interventions not only altered the estuarine geomorphology and hydrodynamics but also impacted OC burial and sequestration by modifying sources and compositions. This study highlights the transition of estuarine regions from carbon sinks to potential sources, underscoring the necessity for a comprehensive understanding of carbon cycle dynamics in Anthropocene-influenced estuarine environments.