Design and development of autonomous vehicles capable of moving safely on roads by sensing the environment has motivated researchers to focus on design of pedestrian detection systems. Similarly, Convolution Neural Networks (CNN) is considered as one of the preferred image classification algorithms. Most of the papers reported in literature employ standard object detector modules available online for pedestrian detection. Design of a real-time pedestrian detection system using CNN for autonomous vehicles is proposed and the system is designed from scratch without using any standard module available. The performance evaluation of proposed system is carried out using INRIA dataset, PETA–CUHK dataset and realtime video input. The CNN parameters were also tuned to achieve best possible recognition accuracy. Recognition accuracies ranging between 96.73 – 100% is obtained on using the system, based on dataset employed. The results obtained are also compared with the results reported in literature and the system designed can be considered on par with those reported in literature. The proposed real-time pedestrian detection system can also be employed as driver assistance system for non-autonomous vehicles.

Design and Evaluation of a Real-time Pedestrian Detection System for Autonomous Vehicles

This study optimized the latest YOLOv5 framework, including its subset models, with training on different datasets that differed in image contrast and cloudiness to assess model performances based on quantitative metrics and image processing speed. The hyperparameter in the feature-extraction phase was configured based on the learning rate and momentum and further improved based on the adaptive moment estimation (ADAM) optimizer and the function reducing-learning-rate-on-plateau to optimize the model’s training scheme. The optimized YOLOv5s achieved a better performance, with a mean average precision of 98.6% and a high inference speed of 106 frames per second. The ADAM optimizer with a detailed learning rate (0.0001) and momentum (0.99) fine-tuning yielded a sufficient convergence rate (0.69% at 55th epoch) to assist YOLOv5s in attaining a more precise detection for underwater objects.

https://ieeexplore.ieee.org/ielx7/6287639/6514899/09773108.pdf

Optimizing the Hyperparameter Tuning of YOLOv5 for Underwater Detection

This study optimized the latest YOLOv5 framework, including its subset models, with training on different datasets that differed in image contrast and cloudiness to assess model performances based on quantitative metrics and image processing speed. The hyperparameter in the feature-extraction phase was configured based on the learning rate and momentum and further improved based on the adaptive moment estimation (ADAM) optimizer and the function reducing-learning-rate-on-plateau to optimize the model’s training scheme. The optimized YOLOv5s achieved a better performance, with a mean average precision of 98.6% and a high inference speed of 106 frames per second. The ADAM optimizer with a detailed learning rate (0.0001) and momentum (0.99) fine-tuning yielded a sufficient convergence rate (0.69% at 55th epoch) to assist YOLOv5s in attaining a more precise detection for underwater objects. 

The novel SARS-CoV-2 coronavirus caused a global pandemic in 2020 with millions of diagnosed cases and a staggering number of deaths. As a preventive measure, many governments issued social...

https://ascelibrary.org/doi/pdf/10.1061/%28ASCE%29NH.1527-6996.0000492

Deep Learning for Visual Analytics of the Spread of COVID-19 Infection in Crowded Urban Environments

Smart grid integrates computer, communication, and sensing technologies into existing power grid networks to achieve significant informatization-related advantages. It will provide communication between neighbors, localized management, bidirectional power transfer, and effective demand response. Smart grids (SG) replace conventional grids by integrating various operational measures, including smart automation appliances, smart meters, and renewable energy sources. Regarding energy management and resolving energy issues, SG is one of the most cutting-edge and potentially game-changing innovations. Even still, its complexity suggests that decentralization may provide significant gains. Because of its increasing digitization and interconnectedness, it is also vulnerable to cyber threats. Blockchain, in this sense, is a potential SG paradigm solution that provides several great benefits. Even though blockchains have been widely discussed to decentralize and strengthen smart grid cybersecurity, they have not yet been researched in depth from an application and architectural standpoint. Blockchain-enabled SG applications are the subject of an in-depth research investigation. Electric vehicles (EVs), home automation, energy management systems, etc., are only a few of the many examples that have prompted the proposal of blockchain designs for their respective use cases. Information communication network security is of paramount importance. However, this evolving system raises cybersecurity issues. This paper aims to guide researchers in the right manner so they may build blockchain-based, secure, distributed SG applications in the future. This article also summarizes cybersecurity threats pertaining to smart grids. It starts with a description of a blockchain followed by the blockchain infrastructure, challenges, and solutions for different smart grid applications. A look back at the tried-and-true methods of securing a power grid is offered, and then it discusses the newer and more complex cybersecurity threats to the smart grid. In addition, models of common cyberattacks are presented, and the methods of defense against them are examined.

/pdf/incorporation-of-blockchain-technology-for-different-smart-7ywhh45l.pdf

Incorporation of Blockchain Technology for Different Smart Grid Applications: Architecture, Prospects, and Challenges

Recent developments in pedestrian detection techniques shows that former algorithms cannot satisfy accurate and speedy detection for practical applications. Adding to the list of modern algorithms in deep learning, these methods are capable to fulfill the requirement of modern applications for pedestrian detection. In this paper, a deep convolutional neural network (D-CNN) based single class You Only Look Once (YOLOv3) state-of-the-art approach is proposed to overcome the problem of pedestrian detection in the contemporary application namely advanced driving assistance system (ADAS), and video surveillance system in terms of false detection (FD) and miss rate (MR). The proposed model is trained on INRIA datasets, which are universally applicable for pedestrian detection. Furthermore, it is effectively demonstrated in different scenarios.

An Improved D-CNN Based on YOLOv3 for Pedestrian Detection

Autonomous hydrogen-based solar-powered energy system for rural electrification in Balochistan, Pakistan: An energy-economic feasibility analysis

More electric aircraft challenges: A study on 270 V/90 V interleaved bidirectional DC–DC converter

In modular power converters, the dc-link capacitor has the highest failure rate due to thermal stress, reducing system performance and reliability. Current harmonics are one of the prominent causes of thermal stress on the dc-link capacitor that may get instigated by the resonance between the dc-link capacitor and stray inductance of the busbar. The dc-link capacitor current harmonics mitigation through busbar design optimization has not been reported. The challenge lies in analyzing the profound resonance characteristics of high order dc-link busbar impedance network as no insightful information can be obtained from impedance analyzers. This paper features a new strategy for high power modular converters that splits the high order dc-link busbar impedance network into corresponding lower-order sub-units. Then, the required relationship between resonating components and corresponding resonance frequencies is attained by exploiting the structural symmetry of the standard modular power unit as well as the inheritance relationship between sub-circuit units and the combined network. Finally, leveraging the derived resonant frequencies expressions, busbar structure is optimized, and up to 23% reduction in capacitor RMS current is achieved. Simulation and experimental results are provided to validate the efficacy of the proposed analysis method. The in-depth theoretical study featured in this paper can dictate not only the busbar design optimization but also a key to improve capacitor lifetime and reliability of the system.

Unit Partition Resonance Analysis Strategy for Impedance Network in Modular Power Converters

Nowadays, power converters with reduced cost, compact size and high efficiency are evolving to overcome the emergent challenges of renewable energy integrations. In this context, there is an increased demand for well-designed power converters in renewable energy applications to reduce energy utilization and handle a variety of loads. This paper proposes a center-tapped bridge cascaded series-resonant LC dual active bridge (DAB) converter for DC-DC conversion. The low part count of the proposed converter enables a high-power density design with reduced cost. The proposed converter offers reduced conduction losses as the reverse current is eliminated by adopting current blocking characteristics. Reverse current blocking also enables zero voltage switching (ZVS) and zero current switching (ZCS) over a wide operating range. Therefore, using a simple fixed frequency modulation (FFM) scheme offers a wide operating range compared to a conventional DAB converter. A thorough comparison of the proposed converter and a conventional DAB converter is provided based on conduction losses and switching losses to illustrate the performance improvement. Lastly, the effectiveness of the proposed converter is validated through simulation and experimental results.

/pdf/practical-evaluation-of-loss-reduction-in-isolated-series-2ag2dd2j.pdf

Mustafa Tahir

Papers

An Improved D-CNN Based on YOLOv3 for Pedestrian Detection

Autonomous hydrogen-based solar-powered energy system for rural electrification in Balochistan, Pakistan: An energy-economic feasibility analysis

More electric aircraft challenges: A study on 270 V/90 V interleaved bidirectional DC–DC converter

Unit Partition Resonance Analysis Strategy for Impedance Network in Modular Power Converters

Practical Evaluation of Loss Reduction in Isolated Series Resonant Converter with Fixed Frequency Modulation