Projects
Research systems, papers, and software spanning autonomous robotics, embedded AI, machine learning, and geospatial analysis.
Communication-Integrated Mobile Environmental Monitoring Framework
MCA dissertation: unified mobile robotic platform integrating autonomous YOLO-seg navigation (mask mAP@50 = 0.988), ONNX-optimized to 15.4 FPS / 64.8 ms on Raspberry Pi CPU. PID controller (Kp=13.2, Kd=0.475) achieves 95.37% tracking accuracy and 87.4% RMSE reduction over manual teleoperation across 40 trials.
Manual vs Autonomous Robot Navigation: Experimental Comparison
40-trial study: autonomous YOLO+PID vs human teleoperation. Results: mean RMSE 2.119 cm vs 16.768 cm, 87.4% reduction. Accuracy 95.37% vs 63.32%. Cohen's d = 8.944.
ML-Augmented PID Control for Autonomous Line-Following Robot
Hybrid control: Random Forest trained on 208,983 time steps. Results: 92.1% MAE reduction, 97.9% error energy reduction, 99.6% in-band accuracy. Lyapunov stability formally proven.
Tripathagamini-S / Homography Trajectory Tracking Platform
Web-based teleoperated robot with homography-based trajectory analysis. 4-point interactive calibration, simultaneous path and robot tracking at 30+ FPS. Ground-truth measurement platform.
YOLO Path Segmentation & Real-Time Optimization
Complete model-to-deployment pipeline. 289-image custom dataset. 320x320 ONNX: 15.4 FPS, 64.8 ms (1.57x speedup). 30-min stability: 15.2±0.8 FPS, zero memory leaks.
Adversarial Robustness on CIFAR-10 / FGSM & PGD
Complete adversarial robustness pipeline on CIFAR-10. Implements FGSM and PGD attacks, then PGD adversarial training as defense. Publication-ready visualizations.
Image-to-Sketch Generation using U-Net CNN
U-Net CNN for image-to-sketch transformation with custom MAE+SSIM loss achieving 85% SSIM score. Deployed as Streamlit web app with under 2-second inference.
Sound Detection / Scream & Non-Scream Classification
Binary audio classification system for real-time distress sound detection targeting edge deployment on Raspberry Pi.
Satellite Land Use / Land Cover Classification
ML classifier for multi-class land cover mapping. 92% accuracy across five land classes over 500+ km². Automated GEE workflow.
Water Body Mapping using NDWI & Sentinel-2
Automated water body detection using NDWI on Sentinel-2. Multi-temporal analysis tracking seasonal water extent variations.