Embedded Files
OTHER FIELD PUBLICATIONS
OTHER FIELD PUBLICATIONS
1. Enhancing healthcare with AI: Sustainable AI and IoT-Powered ecosystem for patient aid and interpretability analysis using SHAP
1. Enhancing healthcare with AI: Sustainable AI and IoT-Powered ecosystem for patient aid and interpretability analysis using SHAP
Contributors : Biplov Paneru , Bishwash Paneru , Sanjog Chhetri Sapkota , Ramhari Poudyal
Highlights :
IoT-based system monitors health parameters: SpO2, body temperature, pulse rate, room temperature, and humidity.
SVC model achieved testing accuracy of 98.83% and training accuracy of 98.71%, with cross-validation accuracy of 99%.
SHAP analysis identified SpO2 level as the most significant feature for health monitoring.
Real-time supervision enabled through the Blynk IoT system and cloud-based data integration.
Robust decision-making model designed for efficient and precise IoT healthcare applications.
Enhances patient care by predicting risks and ensuring timely intervention.
Status :
Published in Measurement: Sensors ( Elsevier)
Link :
2. An Exploration of Effects of Dark Mode on University Students: A Human Computer Interface Analysis
2. An Exploration of Effects of Dark Mode on University Students: A Human Computer Interface Analysis
Contributors : Awan Shrestha, Sabil Shrestha, Biplov Paneru, Bishwash Paneru, Sansrit Paudel, Ashish Adhikari, Sanjog Chhetri Sapkota
Highlights :
Explored the effects of dark mode on students using e-Learning sites, focusing on behavior and user experience.
Addressed issues like blue light exposure during late hours, which disrupts circadian rhythm and causes eye strain and headaches.
Utilized HCI techniques such as surveys, interviews, and evaluation methods to analyze students' interactions with e-Learning platforms.
Developed an e-Learning site with a dark mode theme and gathered student feedback.
Survey findings: 79.7% preferred dark mode on phones, and 61.7% supported adding dark mode to e-Learning websites.
Concluded that dark mode reduces blue light emission and improves usability for students, especially during late-hour studies.
Status :
Published in Arxiv Preprint
Link :
2. Advancing sustainable mobility: Dynamic predictive modeling of charging cycles in electric vehicles using machine learning techniques and predictive application development
2. Advancing sustainable mobility: Dynamic predictive modeling of charging cycles in electric vehicles using machine learning techniques and predictive application development
Contributors : Biplov Paneru , Durga Prasad Mainali , Bishwash Paneru , Sanjog Chhetri Sapkota
Highlights :
Trained machine learning models (Gradient Boosting, Random Forest, Decision Trees, and Linear Regression) to predict charging cycles in EV battery systems.
Performance Metrics:
Gradient Boosting achieved the highest R² score of 0.87, showcasing superior prediction accuracy.
Random Forest model followed with an R² score of 0.83, effectively capturing data nuances.
Highlighted the importance of selecting optimal models to improve prediction accuracy for battery charging cycles.
Developed an EV Battery Charging Cycle Predictor App, enabling accurate predictions to support maintenance scheduling and energy management decisions.
Emphasized the role of advanced ML methods in enhancing EV battery efficiencies and driving electric mobility technologies.
Suggested future work to include real-world data and integrate the application with general energy systems for broader impact.
Status :
Published in Systems and Soft Computing, Elsevier
Link :
4. EEG Right & Left Voluntary Hand Movement-based Virtual Brain-Computer Interfacing Keyboard with Machine Learning and a Hybrid Bi-Directional LSTM-GRU Model
4. EEG Right & Left Voluntary Hand Movement-based Virtual Brain-Computer Interfacing Keyboard with Machine Learning and a Hybrid Bi-Directional LSTM-GRU Model
Contributors : Biplov Paneru, Bishwash Paneru, Sanjog Chhetri Sapkota
Highlights :
Focused on EEG-based Brain-Machine Interface (BMI) to detect and simulate keystrokes for individuals with motor impairments.
Classified EEG data into three categories: resting state (0), 'd' key press (1), and 'l' key press (2).
Developed and trained models, where SVC achieved the highest classification accuracy of 90.42%. Other models include MLP (89%), Catboost (87.39%), Logistic Regression (90.81%), KNN (72.59%), Gaussian Naive Bayes (79.21%), and a Bi-Directional LSTM-GRU hybrid (89%).
Utilized ERP windows for feature engineering and signal segmentation for accurate event classification.
Created a tkinter-based GUI for real-time keypress prediction and simulation using the trained MLP model.
Provided an innovative approach for BCI keyboard simulation to enhance accessibility for users with motor impairments.
Status :
Published in Arxiv Preprint
Link :
Contributors: Biplov Paneru , Ankit Adhikari , Bishwash Paneru , Krishna Bikram Shah , Sanjog Chhetri Sapkota , Ramhari Poudyal , Khem Narayan Poudyal
Highlights :
This study presents a simulation of an IMU sensor using Matplotlib to estimate satellite pose. It focuses on analyzing roll, pitch, and yaw values obtained from the BNO-055 IMU sensor.
Determine a satellite's position. This mainly examines whether a low-cost RF-433Mhz transceiver module will be incorporated into the ground station setup for efficient communication between the satellite body and the ground station.
Highlights using a four-turn helical antenna designed via the 4nec2 simulation program, achieving a VSWR of 1.432 and a directivity of 18.581 for RF-based communication.
Roll, pitch, and yaw values are observed in specific ranges and effectively visualized using Matplotlib, providing insights into the CubeSat's body pose during satellite-ground station communication.
Connecting the IMU data with a Matplotlib program using PySerial enhances the real-time visualization of satellite movements for ground station monitoring.
Status:
Published in Measurement: Sensors ( Elsevier)
Link :
Page updated
Google Sites
Report abuse