Prakash Pawar

Scopus Publications

Improved classification of term and preterm births from electrohysterogram signals using machine learning with model error features
Omkar S. Powar, Arth Kadiya, Prakash Pawar, Mithun Kanchan
Discover Applied Sciences, 2026
Preterm birth (PTB) affects an estimated 15 million infants worldwide each year, raising an enormous challenge in the field of maternal-fetal medicine. This work explores machine learning approaches for distinguishing between term and preterm birth from three-channel electrohysterogram (EHG) measurements of 300 pregnancies, specifically featuring a novel “model error” feature based upon autoregressive modeling of uterine electrical activity. We compared five different machine learning classifiers—K-Nearest Neighbors (K-NN), Support Vector Machine (SVM), Logistic Regression, Naive Bayes, and Gradient Boosting—using a variety of performance indicators such as accuracy, precision, recall, F1-score, and AUC-ROC. Addition of the model error feature, calculating the departures from autoregressive predictions as a measure of unpredictability in uterine contractions, substantially enhanced performance in every case considered. The greatest accuracy (92.50%) and highest AUC-ROC (94.69%) were achieved by Gradient Boosting, in close succession by SVM (92.36% accuracy, 95.85% AUC-ROC). Feature importance analysis verified model error as the strongest predictor. These results, though promising, require external validation in populations with different distributions to establish their potential for clinical use. This work demonstrates the potential for physiologically-informed feature engineering to combine effectively with traditional machine learning methods in PTB prediction, though considerations toward computational efficiency and practical in-patient implementation are worthy of greater scrutiny.
LiDAR-GPS Integrated System for Real-Time Pothole Detection and Visualization Using Google Earth
Vishal B. Pattanashetty, Prakash Pawar
Lecture Notes in Networks and Systems, 2026
"A Comparative Evaluation of YOLOv8 and Detectron2 for Real-World Aerial Segmentation on a Resource-Constrained Dataset"
Deekshith N. R, Dr. Prakash Pawar
Proceedings of 2025 IEEE 22nd India Council International Conference Indicon 2025, 2025
This paper presents a comparative evaluation of two state-of-the-art object detection frameworks, YOLOv8 and Detectron2, for aerial image analysis using unmanned aerial vehicles (UAVs) in resource-constrained environments. A custom UAV-based dataset of high-resolution RGB images was annotated across nine land-use categories, including agricultural fields, buildings, roads, trees, and water bodies, enabling applications in precision agriculture and infrastructure monitoring. The lightweight YOLOv8-nano model was compared with Detectron2 (ResNet-50-FPN) in terms of detection accuracy, inference speed, and memory usage. YOLOv8-nano achieved a mean Average Precision (mAP@0.5) of 91.2, outperforming Detectron2 in inference speed (45 FPS vs. 12 FPS) and memory usage (¡1.5 GB vs. ¿3.8 GB). It also demonstrated superior generalization with higher confidence and fewer false positives, while Detectron2 exhibited increased class confusion under domain shifts. Unlike prior UAV detection studies that rely on high-resource datasets, this work uniquely benchmarks YOLOv8-nano and Detectron2 on a resource-constrained custom UAV dataset, highlighting YOLOv8-nano’s suitability for real-time, edge-deployable UAV applications.
A Novel Algorithm for Aspect Ratio Estimation in SRAM Design to Achieve High SNM, High Speed, and Low Leakage Power
Sanket M. Mantrashetti, Arunkumar P Chavan, Prakash Pawar, H. V. Ravish Aradhya, Omkar S. Powar
IEEE Access, 2025
Performance optimization is a crucial aspect of enhancing the efficiency of electronic systems, and scaling is a primary method for achieving optimal performance while maintaining the integrity of system architecture. This paper introduces a novel algorithm for optimizing transistor sizing in static random-access memory (SRAM) to enhance speed, improve Static Noise Margin (SNM), and reduce leakage power consumption. The SRAM is designed using 45 nm technology and operates at a supply voltage of 1.2 V. To validate the algorithm’s effectiveness, Monte Carlo simulations were conducted under varying process, voltage, and temperature conditions. The results demonstrate read access times of 11.17 ps (HIGH) and 9.97 ps (LOW), and write access times of 12.00 ps (HIGH) and 17.00 ps (LOW). The measured SNM values for the read, write, and hold states were 328.2 mV, 453.7 mV, and 452.3 mV, respectively. The inclusion of precharge and write driver circuits allows for a compact SRAM layout, occupying <inline-formula> <tex-math notation="LaTeX">$9.79~\\mu $ </tex-math></inline-formula> m<sup>2</sup>, with the SRAM cell itself occupying <inline-formula> <tex-math notation="LaTeX">$4.1~\\mu $ </tex-math></inline-formula> m<sup>2</sup>. Furthermore, the proposed SRAM design exhibits low leakage power consumption of 1.64 pW, demonstrating the efficiency and performance benefits of the optimized transistor sizing approach.
Thermal management strategies for lithium-ion batteries in electric vehicles: Fundamentals, recent advances, thermal models, and cooling techniques
Santosh Chavan, Bhumarapu Venkateswarlu, Mohammad Salman, Jie Liu, Prakash Pawar, Sang Woo Joo, Gyu Sang Choi, Sung Chul Kim
International Journal of Heat and Mass Transfer, 2024
Integrating Neural Network Models for Advanced Automation in Analog Amplifier Circuit Design
Arunkumar P Chavan, Shrish S, Sanket M, Abhishek G, Kishan S, Ravish A, Prakash Pawar
Proceedings of Conecct 2024 10th IEEE International Conference on Electronics Computing and Communication Technologies, 2024
Analog Integrated Circuit (IC) design and its automation face challenges due to time-consuming computations and design complexity. Current automation falls short, necessitating a more accurate model and improved dataset collection techniques. Transmitter and receiver in 6G Communication requires amplifiers like the Differential Amplifier (DiffAmp) and Two-Stage Operational Amplifier (OpAmp). This research utilizes Deep Neural Networks to introduce a novel architecture for enhanced prediction of circuit parameters. The significant characteristics of the proposed architecture includes enhanced circuit automation of both DiffAmp and OpAmp using a singular Machine Learning pipeline. A notable contribution is an efficient dataset acquisition technique. The methodologies achieve high accuracy, with 97.3% for DiffAmp and OpAmp.
A Novel TriNet Architecture for Enhanced Analog IC Design Automation
Arunkumar P Chavan, Shrish Shrinath Vaidya, Sanket M. Mantrashetti, Abhishek Gurunath Dastikopp, Kishan S. Murthy, H. V. Ravish Aradhya, Prakash Pawar
IEEE Transactions on Very Large Scale Integration VLSI Systems, 2024
Analog integrated circuit (IC) design and its automation pose significant challenges due to the time-consuming mathematical computations and complexity of circuit design. Though efforts have been made to automate the analog design flow, the current approach falls short in meeting the exact design requirements and plagued by inaccuracies, highlighting the necessity for a more robust approach capable of accurately predicting circuits. In addition, there is a need for an improved dataset collection technique to enhance the overall performance of the automation process. The power management unit (PMU) is a crucial block in any IC that requires the design of low-dropout regulator (LDO) for which amplifiers are fundamental blocks. This research harnesses the capabilities of deep neural networks (DNNs) to automate essential amplifier blocks, such as the differential amplifier (DiffAmp) and two-stage operational amplifier (OpAmp). In addition, it proposes an automation framework for the higher level circuitry of the LDO. This article introduces a novel “TriNet” architecture designed for various parameters of amplifiers, including gain, bandwidth, and power facilitating precise predictions for DiffAmp and OpAmp, and presents a decoder architecture tailored for LDO. A notable aspect is the development of an efficient technique for acquiring larger datasets in a condensed timeframe. The presented methodologies demonstrate high accuracy rates, achieving 97.3% for DiffAmp and OpAmp circuits and 94.3% for LDO design.
An IoT based Intelligent Smart Energy Management System with accurate forecasting and load strategy for renewable generation
Prakash Pawar, Mudige TarunKumar, Panduranga Vittal K.
Measurement Journal of the International Measurement Confederation, 2020
The challenge in demand side energy management lays focus on the efficient utilization of renewable sources without limiting the power consumption. To deal with the above issue, it seeks for design and development of an intelligent system with day-ahead planning and accurate forecasting of energy availability. In this work, an Intelligent Smart Energy Management Systems (ISEMS) is proposed to handle energy demand in a smart grid environment with deep penetration of renewables. The proposed scheme compares several prediction models for accurate forecasting of energy with hourly and day ahead planning. PSO based SVM regression model outperforms over several other prediction models in terms of performance accuracy. Finally, based on the predicted information, the demonstration of ISEMS experimental set-up is carried out and evaluated with different configurations considering user comfort and priority features. Also, integration of the IoT environment is developed for monitoring at the user end.
Design and development of advanced smart energy management system integrated with IoT framework in smart grid environment
Prakash Pawar, Panduranga Vittal K
Journal of Energy Storage, 2019
The day-to-day increased usage of power appliance by consumers is a growing concern in the energy sector, which creates an imbalance in the ratio of demand and supply. Demand-side energy management is an imperative tool to avoid significant deficiency from the supply end and improve energy efficiency. The trend in energy management lays focus on reducing the overall cost of electricity without limiting the consumption counterpart by instead choosing to reduce the power consumption during peak hours. The above issue seeks for design and development of a flexible and portable system to cover a wide variety of consumers for balancing the overall system. The design of smart energy management system is intended to replace the scenario of a complete power outage in a region with partial load shedding in a controlled manner as per the consumer’s preference. Demonstration of experimental work is carried out assuming demand response event and also, considering the maximum demand limit constraint with different cases and changing the order of priority assigned to an appliance. Cost optimization algorithms based on time of usage and user comfort level with sensory information features are embedded within SEMS. Reliable ZigBee communication for home area network is established and also, an IoT environment is developed for data storage and analytics.
Performance analysis of a smart meter node for congestion avoidance and LoS coverage
Prakash Pawar,, Panduranga Vittal K
Aims Energy, 2019
Smart meters are intelligent next-generation energy meters which are used for measuring energy consumption and transmitting information over a network. In a real-time environment, a smart meter network faces congestion and coverage issues. To connect to the network, a network interface card (NIC) is installed for each smart meter. NICs are hardware components that link a host to a network and acts as both the physical and data link layer. Generally, a single node is connected to a NIC. We found that the use of multiple NICs in a node shows significant improvement over the use of a single NIC to overcome network congestion in a real-time environment. The line of sight (LoS) between the transmitter and the receiver is a coverage issue in a smart meter network, which leads to an increase in the packet loss ratio (PLR). In this work, we address network congestion and the coverage issue through multi-channel capability and maximum LoS with single-and multi-hop nodes respectively. The proposed multi-channel network shows two times improvement in the throughput over the regular system; with additional hardware and approximately 1.5 times lower PLR for Node. The experimental results suggest that for a single-hop node, approximately 30 m is the average distance over which LoS communication is possible, and for multi-hop nodes, the distance is 24 m.
Load Scheduling Algorithm Design for Smart Home Energy Management System
Prakash Pawar, Shantanu Sampath, Trisha Ghosh, K P Vittal
2018 IEEE 7th International Conference on Power and Energy Pecon 2018, 2018
Design of smart socket for power optimization in home energy management system
Prakash Pawar, K P Vittal
Rteict 2017 2nd IEEE International Conference on Recent Trends in Electronics Information and Communication Technology Proceedings, 2017