Dheeraj Singh

Masterarbeit in Computer Vision und Reinforcement Learning für Industrie

- Implemented Vision based Deep Learning Models like PoseCNN for identifying multiple robot poses used for mapping. - Utilised NVIDIA Isaac Sim for developing Synthetic Datasets for training Deep Learning Vision Models for pose estimation - Qualitative and quantitative evaluation of different modern Optical Flow Models for robot tracking through monocular camera setup - Experimented with Transformer based Vision model for Monocular Depth Estimation for different datasets

- Design modern infrastructure for thermally optimum data centres with unique implementation methods and by utilizing niche products broadly for sustainable models - Fabricated a mini-robot for autonomous underground cable inspection. - Develop simulation models for efficient server cooling and thermal control utilizing FEA and CFM concepts.

- Developed modern and efficient thermal systems for textile and automotive component production line. - Designed a fault-tolerant mechanism for production line operation by Solidworks modelling and utilizing FEA simulation - Tested multiple materials for the mechanism considering factors of safety and movement fatigues

- Analyzed control parameters for Air Knife system utilized in steel strip galvanization - Obtained the mathematical equation and created a GUI for zinc deposition for customer-specific requirement