DATAMETA LabDataMeta Lab

Dr. Arvind R. Singh

Associate Professor
Sichuan Technology and Business University

About

Arvind R. Singh (Senior Member, IEEE) received the master’s degree in electrical power system from the Department of Electrical Engineering, College of Engineering, Pune, India, in 2009, and the Ph.D. degree in electrical engineering from Visvesvaraya National Institute of Technology Nagpur (VNIT), Nagpur, India, in 2014. He is currently working as Associate Professor in School of Intelligent Building and Manufacturing, Sichuan Technology and Business University, Chengdu, Sichuan, China. He worked as an Associate Professor with the School of Physics and Electronic Engineering, Hanjiang Normal University, Shiyan, China. He was a Lecturer (Contract) with the University of Pretoria, Pretoria, South Africa. From 2019 to 2021, he completed two postdoctoral research fellowships with Shandong University, Jinan, P.R. China, and from 2017 to 2018, and 2023 with the University of Pretoria, Pretoria, South Africa. Prof. Singh was the recipient of the Excellence Technical Education Quality Improvement Programme (TEQIP) master’s scholarship by the Ministry of Human Resource and Development (MHRD), Government of India, for his master’s degree. He was also the recipient of the Indian Government MHRD TEQIP Scholarship for his doctoral studies. He acquired special first category merit-based funding from the Shandong Provincial Government under the Provincial Postdoctoral Advance Program in 2019 for his work on the reliable protection scheme for microgrids. He is currently a Senior Member of the Industry Applications Society and Power Energy Society, USA.

Experience

Associate Professor

Hanjiang Normal University

Jul 2023 - Jun 2026
Shiyan Hubei, P. R. China

Associate professor, School of Physics and Electronic Engineering, Department of Electrical Engineering, Hanjiang Normal University, Hubei Shiyan 442000, China: Teaching and Research.

Lecturer

University of Pretoria

Jan 2022 - Aug 2022
South Africa

• Teaching: Teaching EEV 732 Power Distribution Engineering course for graduate and Honors students. • Arranged expert meets for students on a periodic basis to create an interest about the subject. Encouraged students to build committees among themselves

Postdoctoral Research Fellow

Shandong University

Jan 2019 - Dec 2021
Jinan, P. R. China

• Research achievements: Development of novel microgrid protection algorithm on first principle of synthesis and execute China postdoctoral science foundation grant research project. Final report submission and post project publications. • Teaching: Teaching Post Graduate English foundation course for research methodology and ethical manuscript writing.

Postdoctoral Research Fellow

University of Pretoria

Oct 2017 - Dec 2018
South Africa

• Research achievements: Research on smart grid course development fully funded by SANEDI and renewable energy sources (microgrid). Publication of research article in joint collaboration with various stakeholders. • Teaching: Teaching and tutorials various courses in the area of power system analysis, electrical engineering fundamentals and renewable energy.

Assistant Professor

K. L. University

Sep 2022 - Jun 2023
India

• Course curriculum development, research guidance and teaching courses in the area of Power and Energy Applications. • Faculty advisor to undergraduate or graduate students which includes helping students with their thesis/dissertation, their research project or any other duties. • Actively participate in international joint research and execution of such joint research projects.

Associate Professor

Sichuan Technology and Business University

Aug 2026 - Present
Chengdu, Sichuan, P. R. China

Associate professor, School of Intelligent Building and Manufacturing, Sichuan Technology and Business University, No. 65 Xueyuan Street, Tuanjie Street, Pidu District, Chengdu, Sichuan Province Postal Code: 611745: Teaching and Research.

Education

No education added yet.

Skills

No skills

Other Publications

A blockchain-enabled multi-agent deep reinforcement learning framework for real-time demand response in renewable energy grids

Energy Strategy Reviews, Sep 2025

The increasing integration of renewable energy into smart grids introduces challenges of demand-supply imbalance, peak load stress, and cyber-physical vulnerabilities. Existing demand response (DR) frameworks often lack scalability, privacy-preserving data sharing, and secure transaction mechanisms, which limit user participation and grid resilience. To address these challenges, this study proposes GridSyncNet, a blockchain-enabled multi-agent deep reinforcement learning framework for real-time demand response. The framework integrates federated learning to enhance decentralized forecasting accuracy, blockchain consensus to ensure transparent and tamper-proof energy trading, and actor–critic based DRL agents to dynamically optimize load scheduling and energy dispatch across prosumers. Extensive simulations demonstrate that GridSyncNet outperforms benchmark models such as OD-CNN, D-FCAS, and USTCF. Specifically, it achieves a 98.2 % demand response efficiency, 30.6 % reduction in carbon emissions, and 97.4 % forecasting accuracy. Comparative analysis with multi-agent DRL (MADRL) approaches further confirms that GridSyncNet provides superior scalability, privacy, and security in decentralized environments. The proposed framework contributes to the design of secure, resilient, and sustainable energy management systems, offering practical insights for accelerating the transition toward net-zero energy communities. By combining blockchain, federated learning, and multi-agent reinforcement learning, GridSyncNet establishes a comprehensive pathway for trustworthy and adaptive smart grid operations.

Real-time traffic flow optimization using large language models and reinforcement learning for smart urban mobility

Applied Soft Computing, Sep 2025

The exponential growth of metropolitan populations causes transportation network congestion, which increases fuel usage, travel time, and environmental damage. Traditional traffic management systems (TMS) seldom handle these issues in real time. Recently developed Large Language Models (LLMs), especially those using Reinforcement Learning (RL), may enhance urban transportation systems. Traffic management technology's real-time flexibility and shifting congestion patterns provide improved potential. Traditional approaches cannot estimate traffic flow or adapt to urban settings. A strong AI-driven method is needed to improve urban mobility and traffic flow. This paper introduces the LLM-RL Traffic Optimization Framework (LLM-RL-TOF). LLMs analyze real-time traffic data and give predictive insights in this context. Due to these new insights, the RL algorithm can improve traffic flow in real time and reduce congestion via dynamic traffic management. IoT sensors and urban traffic cameras capture real-time traffic data, including traffic volume and incidents. This data helps the LLM estimate bottlenecks, accidents, and traffic congestion. An RL agent uses LLM outputs to adjust traffic signal timing and suggest alternate routes. With real-time alternatives, traffic flow and urban mobility may be optimized. The junction throughput rate rose 17.5 %, the queue length accumulation index fell 22.3 %, and the average vehicle delay fell 18.6 %. The decrease in average vehicle delay enabled all these gains.

DRL-Driven Optimal User Association and Load Balancing in Hybrid RF/LiFi Based IoT Systems

IEEE Transactions on Network and Service Management, Sep 2025

The proliferation of Internet of Things (IoT) devices has raised considerable difficulties in the identification of users, optimization of power, and load balancing in hybrid RF/LiFi networks. As interconnection among devices increases, ensuring optimal performance while managing network resources efficiently becomes quite complex. This complexity arises due to accommodating the possibly diverse users’ needs, fluctuating channel conditions, and varying interference levels, all necessitating sophisticated management solutions to provide seamless connectivity and dependable communication. To tackle these issues, a deep joint hybrid system (DJHS) technique is presented, which employs proximal policy optimization (PPO), a cutting-edge deep reinforcement learning (DRL) algorithm. DJHS aims to effectively handle the intricate problems surrounding user association and load balancing while optimizing power usage in dynamic contexts. DJHS continuously updates its approach based on real-time network data through adaptive learning methods, allowing it to make intelligent decisions that improve overall system performance regarding data throughput and power optimization. Simulation results demonstrate that DJHS outperforms existing approaches such as sac, a2c, td3, and trpo regarding crucial metrics, including data rate and power transmission. Notably, DJHS’s ability to adjust to variations in signal-to-interference-plus-noise ratio (SINR) allows for enhanced resource allocation and network stability. This flexibility ensures that users receive optimal service even in changing conditions, enhancing the overall user experience.

Enhancing Indoor IoT Edge Intelligence With Deep Reinforcement Learning in Hybrid WiFi/LiFi Networks

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , Aug 2025

The increasing demand for high-speed wireless connectivity in indoor environments has driven the development of hybrid wireless (Wi-Fi) and light fidelity (LiFi) networks. These systems combine the ubiquitous coverage of Wi-Fi with the high data rates of LiFi, providing an effective solution for dynamic indoor applications such as smart homes, offices, and industrial automation. As a critical component of edge intelligence, hybrid networks enable real-time data processing at the network edge, addressing the connectivity demands of indoor IoT systems. However, deploying these networks in dynamic indoor settings introduces challenges, including managing downlink transmission power, resource allocation, energy consumption, and access point selection amidst varying user mobility and interference. This study proposes a machine learning framework for enhancing edge intelligence, leveraging deep reinforcement learning with a proximal policy optimization algorithm (DRL-PPO) to optimize hybrid Wi-Fi/LiFi network performance in indoor environments. The proposed framework improves channel allocation and load balancing at the edge, ensuring energy efficiency while achieving high data throughput. We also introduce a comprehensive system model to manage dynamic traffic in indoor scenarios and develop resource allocation algorithms for efficient bandwidth distribution. Extensive simulations show that our approach, DRL-PPO, achieves up to an average of 15% higher throughput and 20% lower transmission power compared to baseline methods, demonstrating its potential for real-world deployment in indoor IoT scenarios.

Quantum-Inspired Metaheuristic Algorithms for Trust-Based Privacy Agreements and Secure Access in the Internet of Vehicles

IEEE Transactions on Intelligent Transportation Systems, Aug 2025

Internet of Vehicles (IoV) is a new type of network that enables communication among vehicles, infrastructure, and users, providing better traffic systems. When exchanging information, the IoV systems face challenges such as data privacy, access control, and trust management issues. The security mechanisms are utilized to address privacy and trust issues. However, existing solutions still face adaptability, scalability, and efficiency issues when handling the dynamic requirements of IoV ecosystems. This study addressed security issues by designing the Quantum-Inspired Metaheuristic Framework (QIMF), which integrates a quantum optimization algorithm and adaptive privacy agreements to enhance security between IoV entities. During the analysis, the trust score is evaluated for all entities, which helps to minimize unauthorized access and manage trustworthiness using the quantum-inspired approach while exchanging information. The private agreement is assessed via the computed trust score, which improves the policy’s construction. According to the policy, secure access control is achieved by integrating the quantum superposition and tunnelling features to ensure scalable and secure transactions. The discussed system is implemented using the Python tool, and the CICIoV2024 dataset is used to evaluate the discussed system’s efficiency. The QIMF achieves 99.2% trust accuracy with a minimum false acceptance and rejection rate. The integrated approach ensures scalable and secure solutions for improving vehicular networks.

Enhanced CPCV algorithm for improving power quality in electric vehicle battery charging

Scientific Reports, Jul 2025

There are several difficulties to overcome when integrating electric vehicles (EVs) into power distribution networks, especially when it comes to preserving power quality (PQ) because of the harmonic distortion produced throughout battery charging. These issues are not sufficiently addressed by conventional charging algorithms like Constant Current Constant Voltage (CCCV), which frequently leads to higher Total Harmonic Distortion (THD), decreased system efficiency, and generally insufficient performance PQ. The Constant Power Constant Voltage (CPCV) charging algorithm, which is a revolutionary approach to addressing these issues, dynamically modifies the charging power according to the battery’s state of charge (SoC). Compared to conventional techniques, this creative technology efficiently controls harmonic emissions and enhances power quality. Three distinct EV models—the Tesla Model 3, the BYD ATTO 3, and the Kia EV3 Long Range—were used in simulations to assess the algorithm’s performance. The findings show that the CPCV algorithm works noticeably better than the CCCV method with respect to of lowering harmonic distortion; for the 3rd, 5th, 7th harmonics, THD values were lowered to as low as 0.41%. Furthermore, comparing to the greater loss seen in CCCV (3.85 kWh to 5.89 kWh), the CPCV algorithm also demonstrated significant decreases in energy losses, ranging from 2.72 kWh to 3.51 kWh. Furthermore, the CPCV method boosted efficiency by guaranteeing a power factor that was almost constant throughout all charging conditions. These results demonstrate the CPCV algorithm’s efficacy in improving power quality, maximizing energy use, and facilitating the smooth integration of EVs into contemporary distribution systems. In order to meet the increasing demand for EV charging infrastructure and guarantee an additional dependable and sustainable power system, CPCV offers an attractive option by reducing harmonic distortion and increasing general charging efficiency.

A Comparative Analysis of A* and Ant Colony Algorithms for EV Routing

Artificial Intelligence for Integrated Smart Energy Systems in Electric Vehicles, Jul 2025

The shift towards sustainable transportation systems underscores the need to enhance the routing of electric vehicles (EVs). This work delves into the Electrical Vehicle Routing Problem (EVRP), by considering the unique requirements of electric vehicles, such as battery range limitations and the availability of charging stations. A significant addition to this work is the inclusion of different driving modes such as Eco, Ride, Sport, and Warp which influence energy consumption rates and travel times. Also, it aims to optimize routes not only for efficiency but also by factoring in these variable driving modes to adapt to drivers' preferences or situational needs. The A* algorithm is employed for its effectiveness in quickly determining the shortest and least costly path by estimating future costs. This algorithm's performance is compared with that of the Ant Colony Optimization (ACO) algorithm, which mimics the way ants find efficient paths. Our comparative analysis seeks to justify the obtained result along with swiftness handling the complex of EVRP, including varied vehicle modes and accurate time estimations. Further, the customized version of the routing problem for both EV bikes and EV cars with the consideration of speed modes, State of Charge (SOC), discharge level calculation, energy used, and time of travel is optimized. Also, EV route-based feasibility and charging metric analysis are performed.

Automated Open Circuit Fault Detection in Cascaded H-Bridge Multilevel Inverters Using Artificial Intelligence Technique for EV Application

Artificial Intelligence for Integrated Smart Energy Systems in Electric Vehicles, Jul 2025

Cascaded h-bridge multilevel inverters (CHBMLI) represent a promising solution for high-power applications due to their ability to produce high-quality output voltage with reduced harmonic distortion. This reduced THD makes the Cascaded MLIs useful for EV applications. CHBMLIs are more suitable for PV-based EV Charging applications due to their structure. However, like any complex electrical system, CHBMLIs are susceptible to faults, which can lead to system failures if not promptly detected and addressed. In this research, we propose an automated fault detection method for identifying open circuit (OC) and source faults in CHBMLIs using artificial neural networks (ANN). The ANN is trained using a dataset comprising voltage and current measurements under normal operating conditions and various fault scenarios. Once trained, the ANN is capable of accurately detecting OC faults in real-time, thus enabling proactive maintenance and preventing potential system downtime. Simulation results using MATLAB/Simulink demonstrate the effectiveness and reliability of the proposed fault detection approach, highlighting its potential for enhancing the reliability and performance of CHBMLI systems in practical applications.

A blockchain consortium-based framework to enhance interoperability, standardization, and secure demand response management in smart grid applications

Results in Engineering, Jul 2025

The rapid advancement of smart grids, propelled by the integration of distributed energy resources (DERs) and renewable energy technologies, has exposed key challenges such as interoperability, standardization, and data security. These issues impede the efficient operation and scalability of smart grid applications, particularly in enabling decentralized, real-time energy trading and demand response management. Effective management of DERs and demand response systems relies heavily on seamless information exchange, data-driven decision-making, and robust digital communication frameworks to maintain system stability and operational efficiency. To address these challenges, this study presents the Blockchain Consortium-Based Demand Energy Trading System (BC-DETS), a blockchain-powered framework designed to enhance interoperability, security, and standardized energy trading mechanisms within smart grid ecosystems. Comprehensive simulations have validated the effectiveness of BC-DETS using Key Performance Indicators (KPIs) such as transaction latency, demand response participation rate, operational cost reductions, and overall grid efficiency under diverse scenarios. Findings reveal a 35 % boost in grid efficiency through optimized energy distribution and minimized energy losses, alongside a 15 % decrease in operational costs due to reduced transaction overhead and improved energy allocation. Moreover, demand response participation rates increased by 40 %, facilitated by secure and transparent blockchain-enabled real-time energy transactions. These numerical findings underscore blockchain’s transformative potential in enhancing the scalability, security, and inclusivity of smart grids, establishing a foundational platform for future advancements in blockchain standardization and sustainable energy management solutions.

Air pollution prediction with advanced preprocessing and deep ensemble learning

Atmospheric Pollution Research, Jun 2025

The research presented investigates the impact of model selection and preprocessing techniques on air pollution prediction performance, particularly pertaining to achieving Sustainable Development Goals (SDGs). Accurate training of predictive models necessitates effective handling of missing or null values in environmental datasets. To address this challenge, Probabilistic Principal Component Analysis (PPCA) and the Extra Tree Regressor for data imputation are employed, followed by scaling using Robust Scaler, Min-Max Scaler, and Standard Scaler. A thorough comparison of these preprocessing methods revealed that PPCA is the most suitable choice for imputing missing values in air quality datasets, while the Robust Scaler provided the most reliable and accurate scaling. Additionally, Stochastic Gradient Descent (SGD) is integrated as an optimization technique to enhance model performance. The Weighted Average ensemble method, combining PPCA imputation and Robust Scaler, demonstrated superior predictive capabilities. This research highlights the potential for further improvements through additional ensemble techniques and model optimization strategies, opening avenues for future research focused on improving prediction precision and advancing the achievement of Sustainable Development Goals linked to environmental sustainability.

Air pollution prediction with advanced preprocessing and deep ensemble learning

Energy 4.0 Trends, Challenges, and Applications, Apr 2025

Renewable energy, often green energy, will play a critical role in energy generation in the next years. The existing system’s SOC percent of the battery at load is only 65%. However, our proposed system’s SOC percent of the battery at load is 85%. The battery’s SOC is an important metric that refects its performance. Accurate SOC estimation not only protects the battery by preventing overcharging or discharge and increasing its life, but it also enables the application to use reasonable control mechanisms to meet the goal of energy saving.

A deep learning and IoT-driven framework for real-time adaptive resource allocation and grid optimization in smart energy systems

Scientific Reports, Jun 2025

The rapid evolution of smart grids, driven by rising global energy demand and renewable energy integration, calls for intelligent, adaptive, and energy-efficient resource allocation strategies. Traditional energy management methods, based on static models or heuristic algorithms, often fail to handle real-time grid dynamics, leading to suboptimal energy distribution, high operational costs, and significant energy wastage. To overcome these challenges, this paper presents ORA-DL (Optimized Resource Allocation using Deep Learning) an advanced framework that integrates deep learning, Internet of Things (IoT)-based sensing, and real-time adaptive control to optimize smart grid energy management. ORA-DL employs deep neural networks, reinforcement learning, and multi-agent decision-making to accurately predict energy demand, allocate resources efficiently, and enhance grid stability. The framework leverages both historical and real-time data for proactive power flow management, while IoT-enabled sensors ensure continuous monitoring and low-latency response through edge and cloud computing infrastructure. Experimental results validate the effectiveness of ORA-DL, achieving 93.38% energy demand prediction accuracy, improving grid stability to 96.25%, and reducing energy wastage to 12.96%. Furthermore, ORA-DL enhances resource distribution efficiency by 15.22% and reduces operational costs by 22.96%, significantly outperforming conventional techniques. These performance gains are driven by real-time analytics, predictive modelling, and adaptive resource modulation. By combining AI-driven decision-making, IoT sensing, and adaptive learning, ORA-DL establishes a scalable, resilient, and sustainable energy management solution. The framework also provides a foundation for future advancements, including integration with edge computing, cybersecurity measures, and reinforcement learning enhancements, marking a significant step forward in smart grid optimization.

Collaborator Notes

No notes yet.

Sign in to write a note.

Sign in to message, connect with or follow Dr. Arvind R. Singh.

Create free account