Currently developing production ready code for Sensor Fusion and Object Tracking for Level 3/Level 4 vehicles autonomous vehicles.

Previously worked on developing Low Level Radar and Lidar trackers, and Target selection algorithms for ACC and Auto Lane Change functionality.

Developed software in C to calculate the 3D Rigid body transformation between two poses stored as point clouds acquired from a Velodyne sensor using Iterative Closest Point approach.

I worked on finding the 3D Rigid body transformation between two poses represented by point clouds from a velodyne sensor. This 3D transformation can be used for Pose SLAM for Localization and mapping. Our methods were especially useful in conditions were the surface on which the autonomous vehicle was travelling cannot be approximated to a plane and the regular scan-matching approach fails. We developed an algorithm on top of existing ICP (Iterative Closest Points) techniques to find this Rigid body transformation between poses. Our algorithm had significantly less computational time while providing similar results to GICP and NICP, currently used ICP methods.

This is an interesting research problem in the field of Robot localization and mapping to align 3D point clouds and the results of the experiments are pending submission.

Designed and conducted an experiment to verify the relationship between the packing factor with respect to diameter of the filling material and the diameter of the filing cylinder.

Developed a model that integrated the effects of radiation in an existing flame propagation model in packed beds.

Assisted in development of an experimental setup to understand flame propagation behavior in packed beds with novel gasifying agents.

Developed 3D models of the Biomass Reactor and assisted with the procurement of materials for the fabrication of the reactor.