Rohith Kalathur

Working on AI assisted manipulator robotics, along with computer vision.

. Programming of algorithms for vision based grasping for robotic manipulation using adaptive trajectory search in latent space. . Used Reinforcement learning algorithm(SAC) for learning manipulation along with autoencoder based pose detector for vision.

Programming of Deep learning based algorithms for RGB based Novel Object Detection. Used Domain Randomization to generate data set along with programming of Region Proposal Networks and Mask -RCNN for Few shot learning.

Built object Detection and tracking module for an autonomous ship with a ship making company Schottel. Developed an object detection and tracking module using Faster-RCNN. Programming of integration of Ultrasound, LiDAR and IMU sensors for sensor fusion on ROS Kinetic. Successfully tested the whole platform on a real ship.

Responsible for the competency development in the system level modelling and simulation of complete Electronic Power Steering System in AMESim. Obtained competency in Multiport Object Oriented modeling of mechanical and electrical systems on system level simulation software called AMESim (Siemens software analogous to MapleSim).

Embedded Control, Automation and industrial communication, Advanced control, Digital control, Mechatronic systems, Software Engineering, Modelling and Simulation, Sensorics, Actorics, Machine Dynamics and Systems Dynamics, Robotics, Deep Learning, Computer Vision