RESEARCH HIGHLIGHTS
UIUC Smart Bridge - Living Laboratory
Development of digital-twin based monitoring system for UIUC's newly inaugurated Bahl Smart Bridge. Real-time pose-estimation for live pedestrian-traffic monitoring and localization using computer-vision. Checkout the GitHub code here!
Deployable Wireless Machine-vision system for autonomous & synchronized bridge displacement measurement
Developed a deployable and autonomous wireless machine-vision module for on-board real-time displacement estimation of bridges. Real-time pose-estimation for live pedestrian-traffic monitoring and localization using computer-vision.
Monocular Visual Odometry with Deep Learning
Development of a robust learning-based vehicle path tracking based pipeline using relative camera-pose estimation. Utilizing Superpoint CNN network for sparse keypoints generation with RANSAC for updating pose.
Deep Learning based Sensor Data Anomaly Identification
Formulated a multi-channel imagery (time-domain, FFT, Statistics) approach with CNN for anomaly identification. Model achieves high-accuracy even with limited training-data and imbalance in dataset.