@mec.edu.in
Associate Professor and Electrical and Electronics Engineering
Muthayammal Engineering College (Autonomous)
Power Electronics, Renewable Energy Systems and AC/DC Machines
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
V. Sudha, Sathiya Priya Shanmugam, D. Anitha, and R. Raja
IOS Press
An intelligent segmentation and identification of edemas diseases constitutes a most important crucial ophthalmological issues since they provide important information for the diagnosis process in accordance to the disease severity. But diagnosing the different edema diseases using the OCT-images are considered to be daunting challenge among the researchers. The implementation of computational intelligence techniques such as machine learning, deep learning, bio inspired algorithms and image processing techniques may help the doctors for some extent in improving the automatic extraction and diagnosis process consequently improving patients’ life quality. But, these are liable to more errors and less performance, which requires further improvisation in designing the intelligent systems for an effective classification of edema diseases. In this context, this paper proposes the hybrid intelligent framework for the identification, segmentation and classification of three types of edemas such as using the retinal optical coherence tomography (OCT) Images. In this process, Single Feed Forward Training networks (SLFTN) are integrated with Convolutional Layers whose hyperparameters are tuned by using Lion Optimization algorithm. An intensive experimentation is carried out using the Kaggle Retinal OCT Image datasets-2020 with Tensor flow and the proposed framework is trained with the different set of 84,494 images in which performance metrics such as accuracy, sensitivity, specificity, recall and f1score are calculated. Results shows the proposed system has provided satisfactory performance, reaching the average highest accuracy of 99.9% in identifying and classifying the respectively.
R. Raja, Gundala Srinivasa Rao, P. Duraipandy, K. Mathan, and L. Chitra
AIP Publishing
Sakthi C, Yogeshwaran G, Sudha V, and Raja R
IEEE
Chronic kidney disease (CKD) poses an extensive public health task globally, necessitating accurate and early prediction for powerful intervention. In this investigation the predictive competencies of machine learning algorithms utilizing a dataset comprising 25 cautiously selected attributes. CKD's pervasive impact on worldwide health necessitates progressive answers, and this observation stands at the leading edge of pioneering efforts. Interpretability techniques are applied to enhance the transparency of the models, allowing for a deeper understanding of the functions influencing CKD prediction. Validation and evaluation metrics played an important role in guiding the refinement of the model. Precision, recall, and F1 scores are carefully balanced to avoid false positives and negatives. The ADAM (Adaptive moment Estimation) algorithm was deployed to optimize the version's parameters, ensuring fast convergence and advanced predictive overall performance. ADAM is known for being less sensitive to the initial parameter values compared to some other optimization techniques. This adaptability ensures that the algorithm performs optimally throughout a variety of attributes in your dataset.
Adithya D, Rahul R, Sudha V, Vikram N, and Raja R
IEEE
In this project, we leverage cutting-edge deep learning techniques, including CNN, AlexNet, ResNets, and VGG-16, to classify MRI brain images. Our primary goal is early brain tumor detection, addressing the challenge of asymptomatic or vague symptoms. We employ a robust preprocessing pipeline for accurate brain region isolation, incorporating grayscale conversion, Gaussian blurring, thresholding, and artifact removal. Advanced image enhancement techniques like edge-based segmentation and clustering are applied to improve MRI analysis. Our deep learning-based system automates tumor diagnosis and segmentation, easing the burden on radiologists. With more than 2750 MRI images, we use data augmentation to tackle class imbalance, achieving an impressive 98% test accuracy. Beyond CNN, we explore alternative architectures and transfer learning to find the most effective model for MRI classification, revolutionizing early brain tumor detection and aiding medical professionals. This research has the potential to save lives and streamline the diagnostic process significantly.
Harshini D A, Amiritavarshini S, Sudha V, and Raja R
IEEE
This paper presents the design of a smart dispenser system that utilizes Internet of Things (IoT) technology and integrates with a mobile application developed using the Blynk IoT platform. The system aimed to enhance medication adherence and patient safety by providing automated pill dispensing, medication reminders, and real-time monitoring capabilities. The IoT-based pill dispenser consists of a physical device equipped with compartments for storing pills, sensors to detect whether pills are taken or not, and a microcontroller for data processing and communication. The device is connected to the internet, allowing seamless interaction with the Blynk app via Wi-Fi or cellular connectivity. By integrating components like Arduino UNO, nodeMCU, a servomotor, an RTC, an IR sensor, a power supply, and a buzzer, we've designed a robust hardware system for precise medication dispensing. Our system has been rigorously tested and has demonstrated exceptional efficiency, significantly improving medication adherence rates.
R. Raja, L. Jones Nirmal, P. K. Mani, and M. Kavitha
IEEE
Power quality (PQ) concerns have become a critical for electrical engineers as well as electric power companies. PQ disruptions may cause sensitive equipment to malfunction or collapse, which would shut down the process. Devices like Hybrid power filters, active/passive filters and custom power devices (CPDs) are used to address such PQ difficulties. The most effective and efficient CPD is the Dynamic Voltage Restorer (DVR) which aims to reduce voltage sags brought on by problems like short circuits or rapid changes in load. By connecting a DVR in parallel with the load, the suggested approach can introduce a compensating voltage to keep the load voltage within safe parameters. This article describes the enhancement of the PQ in the distribution network by DVR. Simulations in MATLAB/Simulink are used to test if the suggested approach works. The suggested technique has the potential to significantly impact the stability and dependability of power systems as a cost-effective option to improve PQ in distributed systems.
Sekar K, Jayaprakash S, Hemalatha M, and Raja R
IEEE
This article presents the hybrid Renewable Energy Sources (RES) like wind and Fuel Cell (FC) stack used for more efficiency and utilized the current fed inverter. The proposed system has two sources: a current-fed inverter (CFI) and a Proportional Integral (PI) controller controlling the switch. The main objective of the suggested inverter is to provide high voltage gain by two renewable energy sources and produce high voltage for AC loads. Two essential characteristics can achieve this CFI: switching boost inverter and impedance source. Further, the harmonics of the proposed system can be controlled by an LC filter employed nearer to the AC loads. The proposed method is validated by using MATLAB/Simulink software and analyzed through the Simulink waveforms. This THD indicates that the system is executing correctly, and IEEE standards limit THD to 5%. The method with enhanced performance analysis within the range is designed.
Ashok B and Raja R
Inderscience Publishers
V. Sudha, K. Kalyanasundaram, R. C. S. Abishek, and R. Raja
Springer Nature Singapore
R. Raja and B. Ashok
IEEE
In recent times, machine learning (ML) and deep learning (DL) based classification models have been significantly employed in the healthcare sector to determine diseases. Since manual disease diagnosis process is difficult and laborious, automated tools for disease diagnosis have been developed. The ML and DL models can be employed for various disease diagnosis using healthcare data and medical images. Therefore, this article presents an optimal deep representation extreme learning machine (ODR-ELM) technique for medical data classification. The presented ODR-ELM technique mainly intends to identify the occurrence of the disease using the patient medical data. The ODR-ELM technique primarily applies data preprocessing to enhance data quality. Secondly, the DR-ELM classification model is utilized to classify the existence or nonexistence of the disease. To boost the classifier efficacy of the DR-ELM technique, the glowworm swarm optimization (GSO) technique is utilized. The result analysis of the ODR-ELM model take place using three benchmark medical dataset and the results are inspected under several aspects. The simulation results reported the better performance of the ODR-ELM technique interms of different measures.
R. Raja, V. Sudha, Balachandra Pattanaik, and P. Madhumathy
CRC Press
R. Raja and B. Ashok
IEEE
Medical data classification acts as a vital part to enhance the survival rate of the patients and control the health condition to a certain level depending upon the clinical and non-clinical profiles of the patients. To improve the survival rate and decision making process in healthcare sector, necessary disease diagnosis models become essential. This paper presents an efficient Chaotic Grasshopper Optimization Algorithm (CGOA) with Wavelet Kernel Extreme Learning Machine (WKELM), called CGOA-WELM for Medical Data Classification. The CGOA-WKELM model involves preprocessing at the beginning stage to raise the data quality. In addition, WKELM based classification process gets executed which involves the weighted approach to cost function for obtaining the identical effects of weighted ELM. At the same time, in order to effectively set the parameters of the WKELM model, CGOA is applied as a parameter optimizer in such a way to enhance the classification performance. An extensive experimental validation process gets executed to determine the betterment of the CGOA-WKELM model and the results are investigated interms of distinct aspects. The attained experimental values verified the superior performance of the CGOA-WKELM model on all the applied three datasets.
V. Sudha, T. R. Ganesh Babu, N. Vikram, and R. Raja
Computers, Materials and Continua (Tech Science Press)
T. Rajesh, B. Gunapriya, M. Sabarimuthu, S. Karthikkumar, R. Raja, and M. Karthik
Elsevier BV
R. Raja and C. Nagarajan
Bentham Science Publishers Ltd.
C. Navabalachandru, B. Ashok, A. Jagadeesan, and R. Raja
IEEE
Cascaded multilevel inverter is very effective for high power applications. The cascaded multilevel inverter is mostly used compared to other multilevel inverters (MLI). Mainly this paper focuses on cascaded multi level inverter using two unequal dc sources in order to produce a seven level output voltage by using the variable frequency inverted sine pulse width modulation technique (ISPWM). This method reduces the no of dc sources and switching elements. This technique combines the advantages of inverter sine wave variable frequency carriers for a seven level inverter. The performance evaluation of the ISPWM technique for cascaded multilevel inverter is carried out through MATLAB/SIMULINK and the absence of lower order harmonic (THD) is ensured with less switching losses.
R. Raja, R. PrincelynJebaKiruba, A. Jagadeesan, and C. Navabalachandru
IEEE
This paper aims to present a high-step up and non-isolated interleaved dc-dc converter with a common active clamp circuit which makes it highly efficient. The proposed DC-DC boost converter is used as an electronic module for electric vehicle application. Due to the collection of recycled leakage energies collected in a clamp capacitor, it is then recycled to separate load by the clamp boost converter to achieve high efficiency. In addition to this, it avoids mitigation of the output diode's reverse recovery problem and also achieves reduction of voltage stress thereby interleaving the converters. A boost converter is used to clamp the voltage stresses of all the switches in the interleaved converter which is caused by the leakage inductances.