In 2015, the Federal Communications Commission approved the usage of the 3.55 GHz - 3.70 GHz (i.e., band~48) spectrum opportunistically for commercial purposes. Since then, the 150 MHz bandwidth, also known as the Citizens Broadband Radio Service (CBRS), has attracted the attention of the leading telecommunication industries and the research communities. The CBRS spectrum has a hierarchical access architecture where the original owners of the CBRS spectrum, i.e., the incumbents, have the highest and uninterrupted transmission authority. As the incumbent transmission is sparse, the unused spectrum is opportunistically shared between the Primary Access License (PAL) users and the General Authorized Access (GAA) users. Both the user groups get approval to communicate on the CBRS spectrum from the Spectrum Access System (SAS), which acts as a cloud-based centralized administrator of the CBRS and is responsible for ensuring non-compromised incumbent operation on the CBRS. Note that the GAA users do not purchase the spectrum access through auction as PAL does and has the slightest preference to communicate on the CBRS. In this paper, we study the communication reliability of a GAA CBRS Device (GAA CBSD) when the approval from the SAS administrator to communicate on the CBRS spectrum is considered. Our experimental observations show that the deployed GAA CBSD remains uncommunicative for a sizable amount of the observed time duration as the SAS administrator does not allow the GAA CBSD to communicate on the CBRS to protect the incumbent/PAL transmissions.

The Orders/second and latencies of Order Processing with software obtained from a Stock Exchange

In this paper we present an account of state-of the-art in Acoustic Scene Classification (ASC), the task of environmental scenario classification through the sounds they produce. Our work aims to classify 50 different outdoor and indoor scenario using environmental sounds. We use a dataset ESC-50 from the IEEE challenge on Detection and Classification of Acoustic Scenes and Events (DCASE). In this we propose to use 2000 different environmental audio recordings. In this method the raw audio data is converted into Mel-spectrogram and other characteristics like Tonnetz, Chroma and MFCC. The generated Mel-spectrogram is fed as an input to neural network for training. Our model follows structure of neural network in the form of convolution and pooling. With a focus on real time environmental classification and to overcome the problem of low generalization in the model, the paper introduced augmentation to achieve modified noise based audio by adding gaussian white noise. Active researches are going on, in the audio domain and we have seen a lot of progress in the past years.

The unprecedented growth in the number of connected devices has given rise to the Internet-of-Things (IoT) and led to an increasing demand for additional computational and communication resources. Within this context, unmanned aerial vehicles (UAVs) have shown to provide extended coverage, flexibility, and reachability. Motivated by this, in this paper, we develop a UAV-assisted data dissemination framework for IoT networks.  To this end, we formulate a joint optimization problem that aims to minimize the total energy expenditure, i.e., the sum of the energy consumed by the UAV and all the spatially-distributed IoT devices. We propose a deep reinforcement learning approach to solve the joint device classification, device association, and path planning optimization problem. In particular, we aim to 1) train a double deep Q-network agent in order to classify devices into two classes, and then using this classification we 2) develop an association algorithm based on the nearest-neighbor heuristic for device association, and 3) develop a path planning algorithm based on the Lin-Kernighan heuristic. Simulation results show that the proposed approach efficiently reduces the energy consumption as compared with the benchmark approaches, i.e., brute force approach and baseline approach. Furthermore, obtained results show that our approach provides a near optimum solution with a fraction of the time required compared to the brute force approach.

The emergence of Internet-of-Things (IoT) has led to the development of new energy efficient and long-range wireless technologies, which are necessary for the successful realization of IoT applications. Within this context, long range (LoRa) has emerged as one of the prominent low power wide area network (LPWAN) technologies that is envisioned to accommodate the future IoT requirements. However, current LoRa frameworks suffer from limited network capacity, particularly in dense deployments. Consequently, power domain-superposition modulation (PD-SPM) is integrated with LoRa to alleviate the aforementioned problem. In this paper, the error rate performance of a LoRa-enabled PD-SPM system is investigated over Rayleigh fading channels. Specifically, symbol error rate (SER) expressions are derived to characterize the performance of an arbitrary number of users under extreme fading conditions. Furthermore, Monte Carlo simulations are presented to validate the analytical framework.

This paper targets a dynamic statistical threshold detection algorithm of fiber Bragg grating (FBG) spectral peaks at the presence of changing Signal-to-Noise Ratio (SNR) in an optical fiber of a sensing application. The proposed post-demodulation SNR-based detection implements sliding window technique. Its detection threshold is adapted by the targeted probability of false alarms and by background noise statistics. The proposed detection algorithm is independent of FBG spectral peaks shapes and simple enough computationally to implement. It has been demonstrated and validated using sensor network with a deployed group of FBG-based sensors, by implementing simplified sliding window technique. When the adjacent FBG spectral peaks overlap partially, it provides a high degree of certainty in rejecting false FBG detection. The algorithm can be used for „spectral windowing“ or precise measurement of FBG spectral peaks parameters, especially in densely populated FBG sensor networks. This can lead to a significantly higher spectral bandwidth utilization in FBG sensing applications.

Today, the majority of the web’s content and user data is controlled by a few large tech companies. There is a growing movement to devolve this control evenly across the entire internet, representing the transition to Web3. In order for this movement to be successful, technologies and protocols must be developed to enable web users to use the web securely without trusting any other user. That is, today’s web is structured so that users must trust these companies, so trustless alternatives haven’t already been developed. Broadly, this movement emphasizes developing peer-to-peer networks, blockchains, and distributed storage systems. These systems make use of cryptographic primitives to guarantee security.

This study introduces a new multi-antenna array synthesizer based on Slepian functions. The synthesizer concentrates beamforming (BF) gain within a spatial region (i.e., an angular sector), optimizing Shannon capacity of the targeted region, which is suitable for codebook-based analog BF. Starting with the mean capacity formula incorporating the effect of BF, Jensen’s inequality was used to set upper and lower bounds of the mean capacity. Then, a novel method was introduced by combining the two bounds into a new approximation of the mean capacity that outperform both bounds. Finally, the approximation was formulated to a solvable Slepian optimization problem that yielded the weights of the synthesizer. The properties of the synthesizer were listed, including a discussion on how it behaves by changing the width of the targeted region. The steering method was derived, and simulation results were presented. 

Purpose: To compare the diagnostic accuracy and explainability of a new Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and Resnet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and to identify the salient areas of the photographs most important for each model’s decision-making process. Study Design: Evaluation of a diagnostic technology Subjects, Participants, and/or Controls: 66,715 photographs from 1,636 OHTS participants and an additional five external datasets of 16137 photographs of healthy and glaucoma eyes. Methods, Intervention, or Testing: DeiT models were trained to detect five ground truth OHTS POAG classifications: OHTS Endpoint Committee POAG determinations due to disc changes (Model 1), visual field changes (Model 2), or either disc or visual field changes (Model 3) and reading center determinations based on disc (Model 4) and visual fields (Model 5). The best-performing DeiT models were compared to ResNet-50 on OHTS and five external datasets. Main Outcome Measures: Diagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map generation strategies. Results: Compared to our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all five-ground truth POAG labels; AUROC ranged from 0.82 (Model 5) to 0.91 (Model 1). However, the AUROC of DeiT was consistently higher than ResNet-50 on the five external datasets. For example, AUROC for the main OHTS endpoint (Model 3) was between 0.08 and 0.20 higher in the DeiT compared to ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting the use of important clinical features for classification, while the same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc, Conclusions: Vision transformer has the potential to improve the generalizability and explainability of deep learning models for the detection of eye disease and possibly other medical conditions that rely on imaging modalities for clinical diagnosis and management.

Dynamic traffic flow forecasting remains an open issue to this day. As other spatio-temporal problems, traffic prediction deals with both temporal and spatial nonlinear relationships, with the particularity that nearby points in the Euclidean space might be allocated in different roads, adding another layer of complexity. Traffic prediction has witnessed a revolution with the appearance of deep learning, with graph neural networks being prominently responsible for a steep increase in forecasting accuracy. In this paper, we consider the use of an automatic attention mechanism in order to improve the prediction capabilities of a traffic graph convolutional network. This model is based on the composition of gated recurrent units and graph convolution networks to model space and time simultaneously. To overcome the spatial modelling limitations of the original model, our proposal replaces the graph convolutional layer with a graph attention mechanism. Our aim is to model spatial relations in an automatic, more dynamic way. In order to prove the validity and usefulness of our proposal, we have performed a thorough experimentation over two known traffic datasets used in previous research, plus a new, complex one which we have curated and published. Our results portray a clear and statistically significant advantage with the inclusion of spatial attention, surpassing the performance of a wide set of state-of-the-art models on every tested scenario.

With regard to the separating isosurface of two classes of objects in two-dimensional space, the transition from rectilinear to curvilinear is documented. Marching Squares and OpenCV were used to calculate the data that were used for illustrations in this paper. It is found that the curvilinearity is adjustable via variable grid resolution, as well as via convolution such as Gaussian blur. This paper's reference section is rather small, and the hunt for the literature is left to those who have access to the journals.

Abstract—Hardware security is considered a major design and manufacturing target area with a broad range of research and development topics such as protection of intellectual property (IP), metering of hardware, detection of hardware Trojans, and a lot of other topics. This paper discusses Trojan realization in integrated circuits (ICs), as well as the possible security measures, also exploring the usage of the 3-D integration in hardware security where additional hardware can be mounted after fabrication to foster secure execution just for those systems which need it.

The generation of a laser filament through clouds produces a shockwave which displaces the small water droplets from the air to create a quasi-transparent channel through which, information can be transmitted. We present a robust method that utilizes the channel for free-space optical communication using structured light beams coupled with a femtosecond filament. Structured light beams based on their spatial profiles are perfectly suited to propagate around the filament. We have also developed a 4-bit communication channel using a segmented annular beam. Our system demonstrates resilience to noise and is unaffected by the filament. This method can improve the scalability, robustness, and capacity of the free optical channel.

In this paper, an inductorless common source (CS) - common gate (CG) complementary push-pull balun LNA is proposed for sub-GHz low power applications. A dual stage noise cancelation technique is proposed to reduce the noise figure (NF) of the circuit. To maximize the linearity performance, a tunable nMOS based source follower feedback in shunt to the CS stage is employed. Tunable feedback maximizes the IIP2 in the IIP2 mode while it maximizes the IIP3 in the IIP3 mode. The circuit operates in two power modes - low power mode (LPM) (1.2 V supply) and high linearity mode (HLM) (1.8 V supply). In the LPM , IIP2 performance is improved while IIP3 performance is limited by the third order output conductance non-linearity. In the HLM, both IIP2 and IIP3 performance is maximized. The circuit, implemented in 0.18µm RFCMOS process, achieves a maximum gain of 15.5 dB with minimum NF of 3.2 dB while maximum 3-dB bandwidth ranges from 65 MHz to 840 MHz. In the LPM, the maximum IIP2 and the maximum IIP3 are 48 dBm and 1.3 dBm respectively while in the HLM, the maximum IIP2 and the maximum IIP3 are 43 dBm and 8.3 dBm respectively. In LPM, it dissipates a maximum total power of 3.75 mW while in the HLM, it dissipates a maximum total power of 6.4 mW. To the authors’ best knowledge, the proposed balun LNA achieves the best linearity performance reported till date.

Antenna array mutual impedances are measured from embedded element patterns. To do this practically, additional relationships are developed for Thevenin-driven embedded element patterns where the driven element has a different source impedance from the non-driven elements.

This paper proposes an approach to the integration of the Business Process Model and Notation (BPMN) and Economic Order Quantity (EOQ) models for supply chain business process simulation. This approach considers the synergy of Business Process Management (BPM) and Supply Chain Management (SCM) concepts both strive to reduce costs and satisfy customers. Therefore, we consider the extension of BPMN-based models with EOQ-driven parameters to simulate supply chain business processes and support decision-making in the context of supply chain management.

A document by Qijia Huang . Click on the document to view its contents.

Clustering is a fundamental tool of scientific analysis, ubiquitous in disciplines from biology and chemistry to astronomy and pattern recognition. We propose a novel clustering algorithm based on the natural idea that a cluster and its nearest neighbor with higher mass should be merged into one cluster, unless they both have relatively large masses and the distance between them is also relatively large. The find of mass and distance peaks reveals the mergers that don’t conform to the rule and should be removed. The algorithm is parameter-free and harnesses this idea to recognize any cluster and find the proper number of clusters and noise autonomously. Experiments on numerous synthetic and real-world data sets show the enormous versatility of the proposed algorithm that remarkably outperforms the best compared algorithm. Additionally, we also compare it with latest state-of-the-art deep clustering algorithms on several challenging image data sets. The proposed algorithm without any deep representation achieves better or close performance than deep clustering algorithms on image clustering.Our codes are available: https://github.com/JieYangBruce/TorqueClustering .