Raj Kumar.J.S.
@karunya.edu
Assistant Professor in the Department of Computer Science and Engineering
Karunya Institute of Technology and Sciences
RESEARCH INTERESTS
Device Modelling
High power VLSI Design
HEMT
Nano electronics
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Darwin Roach D, Pearly Princess. J, V. Vijula, J. Raj Kumar, and Anit Mary Shibu
IEEE
This research study presents the analysis of learning disability detection and classification using two classification algorithms. Learning disabilities are neurodevelopmental disorders that affect the ability to acquire and process information. Detecting and classifying these disabilities accurately is crucial for providing appropriate interventions and support to affected individuals. This research analyzes two categorization algorithms: Random Forest (RF) and Support Vector Machine (SVM). Although RF is an ensemble learning method that incorporates several decision trees to produce predictions, SVM is a supervised machine learning algorithm that employs hyperplanes to split data points into distinct groups. This analysis compares the two algorithms' performance in terms of F1-score, recall, accuracy, and precision. A dataset consisting of features related to cognitive abilities, academic performance, and behavioral characteristics of individuals with and without learning disabilities is used for the analysis. The results show that both SVM and RF achieve high accuracy in learning disability detection and classification. However, SVM performs better than RF because it has greater precision and recall. This indicates that SVM is better at correctly identifying individuals with learning disabilities and minimizing false positives and false negatives. These algorithms can assist in early identification and intervention, enabling targeted support for individuals with learning disabilities.
D Elizaroshan and J.S. Raj Kumar
IEEE
This study presents an improved methodology of wildfire detection in the forest using PyTorch, a wise machine learning framework, and the Flask application for real-time monitoring. Focusing on accuracy and speed in detection, the methodology mainly refers to data acquisition, and preprocessing of satellite image datasets in its essence, meanwhile featuring the training of robust models through data augmentation. The experimental results using the PyTorch model, which depends on DenseNet201, have proven its ability to determine patterns and reflect highly dynamic contrasts between vertical and horizontal pans; this is especially true when the principal element of the considered satellite image is diffuse and indeterminate. The framework has been integrated into a user-friendly Flask web app, and it gives outputs of immediate wildfire detection upon image upload. The approach is also shown to deliver better metrics thus a more efficient solution as proven by the experimental results. The study informs environmental technology about introducing an easy-to-apply, applicable, and adaptive tool that may have potential associations with environmental agencies for environmentally sustainable approaches to wildfire management.
Yakubraj M, Rajkumar J S, Sophas Samuel S, Matthew Palmer G, and Ancy Jenifer J
IEEE
In this article, the problem of using an intelligent system to ensure the use of personal protective equipment (PPE) in industrial safety is discussed. The real-time solution provides fast feedback and easy-to-use monitoring to safety managers and users alike by combining sensors and machine learning algorithms. The analysis shows how it can improve industrial safety regulations and assesses its efficacy. Strategically placed sensors gather copious amounts of data about PPE compliance, and the system integrates wearable, vision-based, and environmental technologies. Novel techniques such as Computer vision algorithm is used object detection and recognition. SSD single-shot multi-box detectors and (YOLO) enable efficient object detection and quick PPE item identification. Furthermore, it has been observed that Region-based Convolution Neural Network(R-CNNs) increase workplace safety standards by improving the system's adaptability to variations in the usage of safety gear over time.
J. S. Raj Kumar, D. Nirmal, H. Victor Du John, and K. Binola Jebalin
AIP Publishing
J.S. Raj Kumar, H. Victor Du John, Binola K Jebalin I.V, J. Ajayan, Angelin Delighta A, and D. Nirmal
Elsevier BV
S. Angen Franklin, Binola K Jebalin I. V, Subhash Chander, Raj Kumar, J. Ajayan, and D. Nirmal
The Electrochemical Society
In this research work, design of the Extended Field Plate Length (E-FPL) T-gate with Fe doped AlGaN buffer structure on the graded Aluminum Gallium nitride (AlGaN)/ Gallium nitride (GaN) high electron mobility transistor (HEMT) is proposed. The gate length of 60 nm with an ExFPL up to 50 nm towards the drain shows remarkable improvement in breakdown voltage. Meanwhile, the drain current (IDS) and transconductance (GM) is further improved by the Fe doped AlGaN Buffer design. In radio frequency (RF) small signal analysis this device exhibits a peak current-gain cutoff frequency fT of 148 GHz. This device has improved transconductance of 24% with high frequency has compared with conventional GaN HEMT device. It is highly compatible with military applications such as Radio Frequency (RF) upstream transmitters, ship and aircraft communication transmitters and High-frequency Radars (HFRs).
Navya V S, H. Victor Du John, J.S. Raj Kumar, Angen Franklin, and D. Nirmal
IEEE
To validate the suitability for radiofrequency (RF) applications, the characteristics of the proposed Ga<inf>2</inf>O<inf>3</inf> based Negative Capacitance Field Effect Transistor (NCFET) with a stacked ferroelectric HfO<inf>2</inf> and Al<inf>2</inf>O<inf>3</inf> gate dielectrics is simulated and studied using TCAD simulation platform. The key Parameters used to investigate the proposed device’s RF performance include intrinsic capacitances (C<inf>gs</inf> & C<inf>gd</inf>), transconductance (g<inf>m</inf>) and the results attained for the proposed device is analyzed. The findings of this work recommends that the proposed Ga<inf>2</inf>O<inf>3</inf> based stacked NCFET design will be advantageous for successful device integration because it will offer benefits like a steep slope and reduced supply voltage operation for RF applications.
J. S. Raj Kumar, D. Nirmal, J. Ajayan, and Shubham Tayal
Springer Science and Business Media LLC
J. S. Raj Kumar, D. Nirmal, Manish Kumar Hooda, Surinder Singh, J. Ajayan, and L. Arivazhagan
Springer Science and Business Media LLC
J.S. Raj Kumar, D. Nirmal, H. Victor Du John, S. Angen Franklin, and G. Samuel
IEEE
The impacts of a HEMT with a 60 nm T-gated mini field plate device are examined in this study research work, along with the DC/RF properties of the device. The implementation of T- Gate field plate along with the Graded barrier layer configuration in the device provides an extreme impact in the drain current (Ids) of the device. A graded channel AlGaN/GaN HEMTs with the implementation of mini field plate (MFP) exhibits with Ids =1.6 A/mm. In the proposed de vice the InAlGaN layer is simulated which enhances breakdown voltage of 96% while maintaining RF performances of the simulated device. The simulated device exhibits breakdown and fT of 98V and 185GHz respectively.
L. Arivazhagan, D. Nirmal, Subhash Chander, J. Ajayan, D. Godfrey, J. S. Rajkumar, and S. Bhagya Lakshmi
Springer Science and Business Media LLC
P Pavan Kumar Reddy, S Bhagya Lakshmi, L Arivazhgan, J S Raj Kumar, and D Nirmal
IOP Publishing
Abstract For the first time, AlGaN/GaN HEMT is demonstrated for bio-sensing application using transconductance analysis. A lot of HEMT based biosensors were developed experimentally but very few reported on sensing applications. These devices are ideal for sensing or tracking biomolecule because of the spontaneous and piezoelectric polarization properties. AlGaN/GaN HEMT with nanogap cavity is used to detect different biomolecule like streptavidin, protein and uricase. The sensitivity of AlGaN/GaN HEMT is investigated through drain current (ID) and transconductance(gm) and is analyzed using Technology Computer Aided Design (TCAD) tool. The result shows a noticeable change in drain current on introducing different biomolecules below the gate cavity region. Higher sensitivity was obtained for with Transconductance analysis than with drain current analysis.
Arivazhagan L, Anwar Hasan Mohammed Jarndal, Subhash Chander, Godfrey D, Raj Kumar J S, S Bhagyalakshmi, Pavan Kumar Reddy, and D. Nirmal
IEEE
Impact of substrate thickness on self-heating effect for AlGaN/GaN HEMT is analyzed. Self-heating effect is analyzed using Technology Computer Aided Design (TCAD) simulation. GaN-HEMT with gate width of 1mm and gate length of 0.7 µm is considered for the analysis. In the simulation, substrate thickness is varied from 100 µm to 200 µm to study its impact on self-heating effect and drain current. In addition, the ambient temperature is varied from 300K to 700K and its impact is analyzed. Impact of gate-gate spacing on thermal performance is also analyzed. Furthermore, Kink effect on drain current at VGS=-2V is observed.
L. Arivazhagan, D. Nirmal, J. Ajayan, D. Godfrey, J. S. Rakkumar, and S. Bhagya Lakshmi
AIP Publishing
Model of self-heating for AlGaN/GaN High Electron Mobility Transistor (HEMT) is proposed. In the model, degradation of thermal conductivity effect is included. Model data of the device is obtained using MATLAB. Physical simulation of GaN-HEMT is carried out using Technology Computer Aided Design (TCAD). The model data is compared with the simulation result to validate the model. In the simulation, temperature and defect dependent thermal conductivity is used. The drain current and temperature are analyzed using pulsed I-V simulation. It is found that the proposed model data shows excellent fit with simulation characteristics.Model of self-heating for AlGaN/GaN High Electron Mobility Transistor (HEMT) is proposed. In the model, degradation of thermal conductivity effect is included. Model data of the device is obtained using MATLAB. Physical simulation of GaN-HEMT is carried out using Technology Computer Aided Design (TCAD). The model data is compared with the simulation result to validate the model. In the simulation, temperature and defect dependent thermal conductivity is used. The drain current and temperature are analyzed using pulsed I-V simulation. It is found that the proposed model data shows excellent fit with simulation characteristics.
L. Arivazhagan, D. Nirmal, J. Ajayan, D. Godfrey, J. S. Rajkumar, and S. Bhagya Lakshmi
AIP Publishing
AlGaN/GaN HEMT with AlN passivation is proposed and investigated. Effectiveness of AlN passivation is analyzed and compared with SiN, SiO2, Al2O3, and HfO2 passivation materials. The performance of these passivation materials are analyzed using Technological Computer Aided Design (TCAD) simulation. The potential, polarization charge, energy band diagram of the GaN HEMT are analyzed. GaN device with AlN passivation exhibits higher drain current than other passivation materials. Hence, AlN passivation is an excellent passivation material for GaN-HEMT in higher drive current application.
L. Arivazhagan, D. Nirmal, D. Godfrey, J. Ajayan, P. Prajoon, A.S. Augustine Fletcher, A. Amir Anton Jone, and J.S. Raj Kumar
Elsevier BV
J. S. Raj Kumar, D. Nirmal, L. Arivazhagan, and Pratik P Pandit
IEEE
A novel AlGaN/GaN metal-oxide-semiconductor High-electron mobility transistor (MOS-HEMT) using Hf02 as a gate dielectric, grown on a silicon substrate was successfully designed and simulated. In this, we have shown the comparison between HEMT and MOS-HEMT, which eventually proves MOS-HEMT is better than HEMT in certain cases. Different Significant study has been made on gate oxides such as SiO2, Gd2O3, Al2O3, etc but Hf02 was proved to be efficient in reduction of leakage current on GaN-based platforms [3], [4]. Here based on the device performance, Ids of MOS-HEMT is 0.25 A at VGS =6 V is comparatively higher than HEMT which is 0.2 A. The resulted MOS-HEMT also shows improved characteristics of drain current 0.38 A at Vgs= 0 and Vds= 5, hip-her than HEMT.
Pratik P. Pandit, L. Arivazhagan, P. Prajoon, J.S. Rajkumar, J. Ajayan, and D. Nirmal
IEEE
In this work, we analyzed the DC performance of asymmetric AlGaN/GaN High Electron Mobility Transistors (HEMTs) on SiC substarte using Silvaco-TCAD software. The highlights of the proposed HEMT are intrinsic GaN channel, AlN nucleation layer, AIGaN barrier layer and asymmetric gate technology and GaN cap layer. The $\\mathrm {L}_{\\mathrm {g}} = 50$ nm proposed HEMT on SiC substrate exhibits a $\\mathrm {g}_{\\mathrm {m}_{-}\\max }$ of 170 mS/mm and $\\mathrm {I}_{\\mathrm {D}\\mathrm {S}_{-}\\max }$ of 800 mA/mm and breakdown voltage of 550 V. The proposed HEMT on SiC substrate exhibits a threshold voltage of -5V which indicates its D-Mode operation of the device. This excellent DC and breakdown characteristics of the proposed HEMT makes them an excellent candidate for future high power and high frequency applications.