Average Bandwidth Research Articles

Probe-Optimizer: Discovering important nodes for proactive in-band network telemetry to achieve better probe orchestration

By embedding the state data maintained by the programmable data plane into additional customizable probes, proactive in-band network telemetry (INT) can easily achieve flexible, full-coverage and fine-grained network measurement. However, a significant portion of these probes are invalid, failing to capture meaningful network event information, and instead increasing bandwidth occupancy as well as communication overhead between the control plane and the data plane. Furthermore, these invalid probes exacerbate controller overhead, forcing resource-limited CPUs to perform a large amount of meaningless computation and analysis. In this paper, we propose Probe-Optimizer, a novel framework tailored for proactive INT, which can reduce the introduction of invalid probes to comprehensively lower the various telemetry overheads mentioned above. Technically, Probe-Optimizer assigns a unique importance to each node in the telemetry scenario. The importance is significantly related to the probability of network events occurring, which can be used to select important nodes worth monitoring in the topology over a period of time. Then, Probe-Optimizer generates a dedicated set of probe paths for important nodes and another set for the remaining nodes/links, customizing a more appropriate probe frequency for each probe path. Extensive evaluations on both random and FatTree topologies with different scales are conducted. The results show that Probe-Optimizer introduces significantly fewer invalid probes. Benefiting from this, for the topology with a size of more than 200 nodes, compared to the state-of-art proactive INT methods, Probe-Optimizer achieves a higher proportion of probes carrying network events and at least 13%, 42%, and 26% lower communication overhead, CPU usage, and average bandwidth occupancy, respectively.

Quantifying Streamflow Prediction Uncertainty Through Process‐Aware Data‐Driven Models

ABSTRACTThe hydrological model simulation accompanied with uncertainty quantification helps enhance their overall reliability. Since uncertainty quantification including all the sources (input, model structure and parameter) is challenging, it is often limited to only addressing model parametric uncertainty, neglecting other uncertainty sources. This paper focuses on exploiting the potential of state‐of‐the‐art data‐driven models (or DDMs) in quantifying the prediction uncertainty of process‐based hydrological models. This is achieved by integrating the robust predictive ability of the DDMs with the process understanding ability of the hydrological models. The Bayesian‐based data assimilation (DA) technique is used to quantify uncertainty in process‐based hydrological models. This is accomplished by choosing two DDMs, random forest algorithm (RF) and support vector machine (SVM), which are distinctly integrated with two process‐based hydrological models: HBV and HyMOD. Particle filter algorithm (PF) is chosen for uncertainty quantification. All these combinations led to four different process‐aware DDMs: HBV‐PF‐RF, HBV‐PF‐SVM, HyMOD‐PF‐RF and HyMOD‐PF‐SVM. The assessment of these models on the Baitarani, Beas and Sunkoshi river basins exemplified an improvement in the accuracy of the daily streamflow simulations with a reduction in the prediction uncertainty across all the models for all the basins. For example, the nash‐sutcliffe efficiency improved by 54.69% and 10.61% in calibration and validation of the Baitarani river basin, respectively. Equivalently, average bandwidth improved by 79.37% and 71.59%, respectively. This signified the (a) potential of the DDMs in quantifying and reducing the prediction uncertainty of the hydrological model simulations, (b) transferability of the model with an appreciable performance irrespective of the choice of basins having varying topography and climatology and (c) ability to perform significantly irrespective of different process‐based and DDMs being involved, thereby ensuring generalizability. Thus, the framework is expected to assist in effective decision‐making, including various environmental management and disaster preparedness.

Evaluation of the number of events’ influence on model performance and uncertainty in urban data-scarce areas based on behavioral parameter ranking method

Rainfall-runoff data in drainage systems in urban areas is the essential input variable for urban hydrological modeling. Obtaining high quality and sufficient size of urban flood events is always difficult due to the inconvenient underground observation and the untimely capture of the rapid rising and recession stages of urban runoff. It largely constrains the efficiency of model calibration and brings large uncertainty in urban rainfall-runoff simulation. Therefore, the evaluation of the number of events’ influence on model performance and uncertainty is of great significance, which can provide useful information to help the decision makers select sufficient useful observation data for model calibration with relatively fewer events. In this study, we constructed a comprehensive method of behavioral parameter ranking of multi-events (BPROME) based on coupling the Generalized Likelihood Uncertainty Estimation (GLUE) algorithm and the Storm Water Management Model (SWMM). The results in the two small demonstration areas (named Case #A and Case #B) in Shenzhen city of China proved the good performance of the BPROME in uncertainty assessment in urban areas. We found the model uncertainty (average bandwidth (B) and average deviation amplitude (D)) and model performance (containing Ratio (CR) and the maximum value of behavioral samples’ NSE (NSEmax)) became stable at a certain number of events in both two case areas. The B and D decrease and the CR and NSEmax increase as the number of event increases. In particular, the model performance and uncertainty reach their appropriate state concerning the limited observations at a certain range of numbers of events (both three to five in our two case areas). Our results demonstrate the potential influence of the numbers of events for the urban rainfall-runoff modeling calibration which can balance the model efficiency and data collection cost (the number of input rainfall-runoff data). The findings could help decision-makers seek a trade-off between data investment and acceptable model performance.

A Performance Analysis of Security Protocols for Distributed Measurement Systems Based on Internet of Things with Constrained Hardware and Open Source Infrastructures.

The widespread adoption of Internet of Things (IoT) devices in home, industrial, and business environments has made available the deployment of innovative distributed measurement systems (DMS). This paper takes into account constrained hardware and a security-oriented virtual local area network (VLAN) approach that utilizes local message queuing telemetry transport (MQTT) brokers, transport layer security (TLS) tunnels for local sensor data, and secure socket layer (SSL) tunnels to transmit TLS-encrypted data to a cloud-based central broker. On the other hand, the recent literature has shown a correlated exponential increase in cyber attacks, mainly devoted to destroying critical infrastructure and creating hazards or retrieving sensitive data about individuals, industrial or business companies, and many other entities. Much progress has been made to develop security protocols and guarantee quality of service (QoS), but they are prone to reducing the network throughput. From a measurement science perspective, lower throughput can lead to a reduced frequency with which the phenomena can be observed, generating, again, misevaluation. This paper does not give a new approach to protect measurement data but tests the network performance of the typically used ones that can run on constrained hardware. This is a more general scenario typical for IoT-based DMS. The proposal takes into account a security-oriented VLAN approach for hardware-constrained solutions. Since it is a worst-case scenario, this permits the generalization of the achieved results. In particular, in the paper, all OpenSSL cipher suites are considered for compatibility with the Mosquitto server. The most used key metrics are evaluated for each cipher suite and QoS level, such as the total ratio, total runtime, average runtime, message time, average bandwidth, and total bandwidth. Numerical and experimental results confirm the proposal's effectiveness in foreseeing the minimum network throughput concerning the selected QoS and security. Operating systems yield diverse performance metric values based on various configurations. The primary objective is identifying algorithms to ensure suitable data transmission and encryption ratios. Another aim is to explore algorithms that ensure wider compatibility with existing infrastructures supporting MQTT technology, facilitating secure connections for geographically dispersed DMS IoT networks, particularly in challenging environments like suburban or rural areas. Additionally, leveraging open firmware on constrained devices compatible with various MQTT protocols enables the customization of the software components, a crucial necessity for DMS.

Improving canopy transpiration model performance by considering concurrent hot and dry conditions

CONTEXTThe Jarvis-Stewart model has commonly been used to estimate transpiration. However, its assumption regarding independent effects from different influencing factors has seldom been explored. OBJECTIVEThe aim is to test whether the Jarvis-Stewart model can well estimate transpiration under concurrent hot and dry conditions with strong environmental interactions. A secondary objective was to test was to propose a new coupling method without the environmental-independence assumption for transpiration estimation. METHODSThis study tested and compared 24 Jarvis-Stewart models composed of various constraint functions to determine the optimal model structure for an orange orchard under concurrent hot and dry conditions. Meanwhile, a new coupling equation that considers interactive environmental effects on transpiration was proposed, and a total of 48 configurations were examined to determine the most effective configuration. A comprehensive comparison was made between the best Jarvis-Stewart model and the best coupling model based on a Bayesian framework. RESULTS AND CONCLUSIONSResults showed that with fewer parameters, the best new coupling model significantly improved model performance compared with the best Jarvis-Stewart model. Specifically, estimation accuracy was slightly improved, with the average mean relative error decreased from 11.79% to 11.06%, and the average coefficient of determination increased from 0.67 to 0.72. More importantly, estimation uncertainty was significantly decreased, with the average uncertainty band width reduction of 42%. Moreover, the new model generated some water use information that was much more consistent with measured data and did not suffer from over-fitting problems as the Jarvis-Stewart model suffered from. Furthermore, the new model did not require additional data or computational cost. SIGNIFICANCEThe results of our study emphasize the necessity for studying environmental interaction effects on plant water consumption, particularly under the possibility of concurrent hot and dry conditions that are likely to occur under future climate conditions.

An efficient bandwidth-adaptive gradient compression algorithm for distributed training of deep neural networks

In distributed deep learning with data parallelism, communication bottleneck throttles the efficiency of model training. Recent studies adopt versatile gradient compression techniques, with communication sparsification standing out as an effective approach for reducing the number of gradients to be transmitted. However, the deployment of gradient sparsification is adversely influenced by the change of network environment in real systems, and existing methods either neglect bandwidth dynamics during training or experience drastic fluctuation of compression ratios. In this paper, we propose ACE, a novel adaptive gradient compression mechanism with high communication efficiency under bandwidth variation. ACE adapts the sparsification ratio to the average bandwidth in a time window, other than following its dynamics exactly. To accurately compute the compression ratio, we first profile the compression time and model a single iteration time consisting of communication, computation and compression operations. We then present a practical model to fit the needed training rounds till convergence, and formulate an optimization problem to compute the optimal sparsification ratio. We conduct experiments on different DNN models in different network environments and compare various methods in terms of convergence and model quality. The experimental results show that ACE achieves up to 9.39× and 1.28× training speedups over fixed and state-of-the-art adaptive compression methods.

Robust treatment planning for small animal radio-neuromodulation using focused kV x-ray beams.

In preclinical radio-neuromodulation research, small animal experiments are pivotal for unraveling radiobiological mechanism, investigating prescription and planning techniques, and assessing treatment effects and toxicities. However, the target size inside a rat brain is typically in the order of sub-millimeters. The small target inside the visual cortex neural region in rat brain with a diameter of around 1mm was focused in this work to observe the physiological change of this region. Delivering uniform doses to the small target while sparing health tissues is challenging. Focused kV x-ray technique based on modern x-ray polycapillary focusing lens is a promising modality for small animal radio-neuromodulation. The current manual planning method could lead to sub-optimal plans, and the positioning uncertainties due to mechanical accuracy limitations, animal immobilization, and robotic arm motion are not considered. This work aims to design a robust inverse planning method to optimize the intensities of focused kV x-ray beams located in beam trajectories to irradiate small mm-sized targets in rat brains for radio-neuromodulation. Focused kV x-ray beams were generated through polycapillary x-ray focusing lenses on achieving small (≤0.3mm) focus perpendicular to the beam. The beam trajectories were manually designed in 3D space in scanning-while-rotating mode. Geant4 Monte Carlo (MC) simulation generated a dose calculation matrix for each focused kV x-ray beam located in beam trajectories. In the proposed robust inverse planning method, an objective function combining a voxel-wise stochastic programming approach and L1 norm regularization was established to overcome the positioning uncertainties and obtain a high-quality plan. The fast iterative shrinkage thresholding algorithm (FISTA) was utilized to solve the objective function and obtain the optimal intensities. Four cases were employed to validate the feasibility and effectiveness of the proposed method. The manual and non-robust inverse planning methods were also implemented for comparison. The proposed robust inverse planning method achieved superior dose homogeneity and higher robustness against positioning uncertainties. On average, the clinical target volume (CTV) homogeneity index (HI) of robust inverse plan improved to 13.3 from 22.9 in non-robust inverse plan and 53.8 in manual plan if positioning uncertainties were also present. The average bandwidth at D90 was reduced by 6.5Gy in the robust inverse plan, compared to 9.6Gy in non-robust inverse plan and 12.5Gy in manual plan. The average bandwidth at D80 was reduced by 3.4Gy in robust inverse plan, compared to 5.5Gy in non-robust inverse plan and 8.5Gy in manual plan. Moreover, the dose delivery time of manual plan was reduced by an average reduction of 54.7% with robust inverse plan and 29.0% with non-robust inverse plan. Compared to manual and non-robust inverse planning methods, the robust inverse planning method improved the dose homogeneity and delivery efficiency and was resistant to the uncertainties, which are crucial for radio-neuromodulation utilizing focused kV x-rays.

Optimal Scheme Selection Model of Distributed Photovoltaic Consumption in Smart City Based on Gaussian Mixture Model

In an effort to guarantee the stable action and photovoltaic absorption capability of the spread mesh after a macro digit of distributed solar panels are related to the mesh, this paper puts forward the definitive scheme selection model of distributed photovoltaic absorption in smart cities based on Gaussian mixture model. The principal constituent analysis arithmetic is used to process the chronicle data of distributed photovoltaic action, and the processed data are transmitted to the Gaussian compound model. According to this model, the short-term power situation of distributed photovoltaic is predicted, and the external ability and restrict requirements of the definitive programme of distributed photovoltaic consumption are determined according to the prediction fruits. The definitive programme of distributed photovoltaic consumption is selected by the approximate ideal solution sorting method. The measurement fruits show that the model has a good short-term power prediction effect of distributed photovoltaic, and the prediction fruits are in high agreement with the actual fruits, and the coverage rate and interval average bandwidth fruits are above 0.918 and below 0.11 respectively. The photovoltaic consumption of the definitive programme is above 130%.

Load balancing and topology dynamic adjustment strategy for power information system network: a deep reinforcement learning-based approach

As power information systems play an increasingly critical role in modern society, higher requirements are placed on the performance and reliability of their network infrastructure. In order to cope with the growing data traffic and network attack threats in the power information system, we select the power information system data center network as the research object and design an overall system solution based on software defined network, including the application layer, control layer and infrastructure layer. A typical fat tree network topology is simulated and analyzed. We define the load balancing and network topology dynamic adjustment problem as a Markov decision process, and design a data flow path acquisition method based on breadth-first search to construct the action space of each host. Then, a deep reinforcement learning algorithm based on deep Q-network, priority experience replay and target network is introduced to provide solutions for optimizing the performance of power information systems and responding to network attacks. Simulation results show that the proposed method is better than the traditional equal-cost multi-path algorithm in terms of average bandwidth utilization, average jitter and average packet loss, and can reduce the probability of network nodes being attacked by more than 11%.

Optimizing Secure Communication in Distributed Corporate Networks through PPTP and IPSec VPN Protocols

In today’s industry, company’s activities are scattered in several locations and need internet for communication purposes. However, it will be dangerous if third-parties have access to their information. One of the solutions for securing information exchange is by using Virtual Private Network (VPN). The VPN protocols that will be used are Point-to-Point Tunneling Protocol (PPTP) with the help of Internet Protocol Security (IPSec). These protocols are simulated in VirtualBox that has been previously installed with two Mikrotik routers and two Windows 7 operating systems. The PPTP and IPSec protocols are configured in both routers through Winbox software. For the computers, the IP addresses are configured along with iPerf software for bandwidth analysis. In this VPN simulation, both computers can communicate with each other and have no problem in accessing the internet. In PPTP’s network, the average bandwidth is 1.293 Mbits/sec. Meanwhile, in PPTP with IPSec’s network, the average bandwidth is 0.853 Mbits/sec. There is a slight difference considering that in PPTP, the data packets are only encapsulated. On the other hand, the data packets in IPSec protocol are both encapsulated and encrypted.

Investigative Study of the Effect of Damping and Stiffness Nonlinearities on an Electromagnetic Energy Harvester at Low-Frequency Excitations

Ambient vibration energy is widely being harnessed as a source of electrical energy to drive low-power devices. The vibration energy harvester (VEH) of interest employs an electromagnetic transduction mechanism, whereby ambient mechanical vibration is converted to electrical energy. The limitations affecting the performance of VEHs, with an electromagnetic transduction structure, include its operational bandwidth as well as the enclosure-size constraint. In this study, an analysis and design of a nonlinear VEH system is conducted using the Output Frequency Response Function (OFRF) representations of the actual system model. However, the OFRF representations are determined from the Generalised Associated Linear Equation (GALE) decompositions of the system of interest. The effect of both nonlinear damping and stiffness characteristics, to, respectively, extend the average power and operational bandwidth of the VEH device, is demonstrated.

Implemantasi Autentikasi Hotspot Menggunakan Radius Server Mikrotik Router Pada Rumah Sakit Misi Lebak

The Lebak misi hospital area currently mostly using wireless networks where usage is increasing and bandwidth is uneven. Between one user and another and the admin only provides password protection for existing access points. which will cause instability between one user and another, therefore the author uses the Radius server as a bandwidth divider as well as security with user authentication and adding a QR code to the authentication. and the author uses the NDLC (Network Development Life Circle) method. As a reference for the radius server configuration using a proxy router. and the results obtained are the percentage of user bandwidth before using the server radius of 0.99Mbps on download and upload of 1.19Mbps for smartphones on laptops of 6.72Mbps for download and upload of 2.21Mbps then after using the radius of the server the average bandwidth speed using user and password of 0.75Mbps for download and 0.80Mbps upload and the success status is very good, all devices can authenticate. and those who can use a QR code of 0.36Mbps for download and 0.44Mbps for Upload with a poor success status because some devices are not compatible for authentication using a QR code by scanning webqr.com.

Exploration of Cultural Teaching Mechanisms of Ideology and Politics Courses under the Background of “Internet Plus”

Abstract As the cradle of cultivating and conveying talents, colleges and universities should innovate the education of ideological and political courses under the background of “Internet+”, focusing on cultivating students’ cultural self-confidence. This paper first explains blockchain-related technology in Internet technology, including the basic principle and structure of blockchain, the consistency Hash algorithm, the consensus mechanism and the credit mechanism of blockchain. Secondly, it discusses the fit between the blockchain and the cultural teaching mechanism of the Civics course, through which blockchain technology can effectively realize the interactive sharing of the cultural teaching resources of the Civics course. Based on the blockchain technology and MVC network structure, an interactive model of cultural teaching resources for Civic and Political Science courses is constructed, which further improves the way students access cultural teaching resources. Finally, a simulation test is conducted to evaluate the interactivity of the cultural teaching resources in the Civics and Political Science course and the effectiveness of the model. The results show that the average write bandwidth of the consistent Hash algorithm is 756.68MBps, and the average read bandwidth is 215.42MBps, and when the total amount of data exceeds 64GB, the relative standard deviation of the Hash algorithm decreases to about 0.15% and remains stable. The cultural teaching mechanism of the Civics course with the introduction of blockchain technology is based on a digital platform, and when the number of network nodes is 150, the delay of the consensus mechanism is 236.73ms, and the throughput reaches about 98 data interactions per second. The cultural teaching mechanism of the Civics and Politics course in the background of the Internet needs to be based on digitization, fully combining with information technology to realize the common construction and sharing of cultural teaching resources and cultivate new-age talents with higher cultural self-confidence for the development of society.

Exploring frequency-domain nonstationary of seismic ground motions using evolutionary power spectrum and Archimedean Copula functions

In this paper, the frequency-domain nonstationarity of seismic ground motions is investigated using Archimedean Copula functions. To measure the nonstationarity of the ground motions, a group of nonstationarity indexes is introduced based on the evolutionary power spectral density. The values of the proposed nonstationarity indexes of 1034 seismic ground motion records are calculated and their distributions are obtained. The Gumbel, Clayton, and Frank Copula functions among the nonstationarity indexes and intensity measures of the ground motions are estimated from the Kendall rank correlation coefficients. The best-fit Copula functions are determined by minimizing these three estimated Copula functions and the empirical Copula function. The correlation among the nonstationarity indexes and intensity measures from the Copula functions indicates that the average frequency and bandwidth of the seismic ground motion tend to decrease with time increasing, and the ground motion with long duration tends to have remarkable low-frequency components. It is suggested that the nonstationarity of the ground motions may increase the seismic risk of civil engineering structures and could not be ignored in the structural design.

A Feedback-Driven DNN Inference Acceleration System for Edge-Assisted Video Analytics

With the proposal of edge computing, lots of intelligence applications have made significant progress. For enormous video analysis, how to further accelerate the process is still a major challenge. To overcome the challenge, researchers propose various video frame filtering systems to reduce the data transmission. In this work, we propose a Feedback-Driven DNN Inference Acceleration system (FDDIA). FDDIA is committed to further reducing the latency of DNN inference and the transmission according to the feedback information. Specifically, on the device side, FDDIA first uses the detection results of prior frames as feedback to determine the key candidate regions and uses the inter-frame difference to determine the new object regions. Those regions containing large objects are down-sampled to further reduce the frame. On the edge side, FDDIA first integrates all candidate regions into a smaller new image. Then it remaps the detection result for the new image back to the original frame and returns the result to the device. We evaluate FDDIA on different video benchmarks for three object detection tasks. The results show FDDIA improves the average end-to-end latency by 44% and average bandwidth usage by 41% than the existing advanced method while maintaining a high accuracy.

Resonant X-ray emission spectroscopy using self-seeded hard X-ray pulses at PAL-XFEL.

Self-seeded hard X-ray pulses at PAL-XFEL were used to commission a resonant X-ray emission spectroscopy experiment with a von Hamos spectrometer. The self-seeded beam, generated through forward Bragg diffraction of the [202] peak in a 100 µm-thick diamond crystal, exhibited an average bandwidth of 0.54 eV at 11.223 keV. A coordinated scanning scheme of electron bunch energy, diamond crystal angle and silicon monochromator allowed us to map the Ir Lβ2 X-ray emission lines of IrO2 powder across the Ir L3-absorption edge, from 11.212 to 11.242 keV with an energy step of 0.3 eV. This work provides a reference for hard X-ray emission spectroscopy experiments utilizing self-seeded pulses with a narrow bandwidth, eventually applicable for pump-probe studies in solid-state and diluted systems.

Performance evaluation of planar dipole antenna for smart home applications

In a Smart Home wireless module, the antenna generally occupies one-third of the total circuit advances in the Internet of Things (IoTs). Dipole antenna comes in various shapes, but the planar antenna is the most suitable and attractive design for the IoTs application in Smart Homes because it is smaller in size and has a thin profile, making it suitable for mounting on portable equipment. In this research, two antennae, Meandered Dipole and Wideband Planar Antenna, were designed using Advance Design System (ADS) software on a PCB size of 18.8 mm × 43.2 mm, FR4 substrate with 4.6 dielectric constant and 0.8 mm thickness using copper metal as the conducting element. The transmission line has designed width of 1.25 mm, resulting in a typical line impedance of 50 Ω. A comparison of the analysis of simulation and measurement results from feeding port 1 of the Meandered PIFA Antenna shows a better performance of 95.43% efficiency with a gain of 2.54391dBi and a return loss of −24.49 dB at resonant frequency of 2.4 GHz than that from feeding port 2 whose efficiency was only 73.99%. Whereas, the wideband planar antenna shows an average bandwidth of 93.91% from 1.3 to 2.4 GHz. This range covers the GPS applications, the 2.4 GHz ISM band, and the general WCDMA and the LTE 3GPP bands useable for smart home and IoT applications. The proposed antenna provides low power operation, extended range and compact, cost-effective designs to support vendor's customers, and make quality of life smarter and more economical with the use of the smart home device.

Overview of Computer Networks for Data Transmission at Pondok Kopi Islamic Hospital, Jakarta, Indonesia

Network quality monitoring can provide services in the provision of computer network performance in the provision of data transmission according to the level of aspects related to the performance of network infrastructure systems and the number of clients operating for each computer. During the day, at certain hours, many subscribers use the network when the connection is blocked, and the graphics on the screen reach their highest levels. By monitoring computer network installation graphics, Axence netTools software needs to know bandwidth management to overcome network congestion by analyzing the QoS (quality of service) of computer network data transmission. The purpose of this study is to provide quality of service presentation material on LAN infrastructure performance using the TIPHON standard to measure LAN performance and analyze the factors that influence the results of quality of service measurements and obtain packet loss, delay, jitter based on class classification and average bandwidth value, according to TIPHON. The research method used is observation, interviews, literature study, monitoring, and qualitative descriptive research methods. The results showed that the percentage value of quality of service based on the TIPHON standardization was 66.66%. This result was classified as "unsatisfactory" because the average packet loss was 40%. The average percentage of delay is 115ms, and the average percentage of jitter is 7.5ms. The factors that affect the QoS index results are the distance of computer network transmission facilities and the need to monitor computer networks when adding and maximizing bandwidth according to TIPHON standards.

Super-broadband terahertz absorber: an optimized and magnetized graphene-embedded 1D disordered photonic system

In this study, a broadband terahertz absorber was designed and numerically demonstrated. The optical features were computed using 4×4 transfer matrix formalism. The broadband absorption is attained by optimizing the Fermi levels of graphene, the magnetic field, and the thickness of the dielectric layers using the differential evolution algorithm. The results demonstrated that the WMF-optimized scenario offered greater than 90% absorption over a bandwidth of 4.18 THz, and the WoMF&SL scenario provided the shortest bandwidth of 0.89 THz. These findings reveal the significance of the spacer layer to achieve broad absorption. Moreover, the absorption band is tailored to the required spectral range by careful choice of the structural and electrical parameters of the spacer layer; changing the refractive index to 1.2 offers a bandwidth of 4.42 THz and altering the thickness to 12 µm provides a bandwidth of 5.5 THz. The broadband absorption was attained due to the impedance matching provided by the optimized structure over a wide spectral range. Furthermore, the average absorption and bandwidth were enhanced, while fluctuations in the selected absorption band were minimized by engineering the magnetic biasing. The reported super-broadband absorber offers numerous applications in terahertz imaging, terahertz communications, and photodetectors.

AQuA: A New Image Quality Metric for Optimizing Video Analytics Systems

Millions of cameras at the edge are being deployed to power a variety of different deep learning applications. However, the frames captured by these cameras are not always pristine—they can be distorted due to lighting issues, sensor noise, compression etc. Such distortions not only deteriorate visual quality, they impact the accuracy of deep learning applications that process such video streams. In this work, we introduce AQuA, to protect application accuracy against such distorted frames by scoring the level of distortion in the frames. It takes into account the analytical quality of frames, not the visual quality, by learning a novel metric, classifier opinion score , and uses a lightweight, CNN-based, object-independent feature extractor. AQuA accurately scores distortion levels of frames and generalizes to multiple different deep learning applications. When used for filtering poor-quality frames at edge, it reduces high-confidence errors for analytics applications by 17%. Through filtering, and due to its low overhead (14 ms), AQuA can also reduce computation time and average bandwidth usage by 25%. Finally, we discuss numerous new avenues of optimizations of video analytics pipelines enabled by AQuA.