Engineering a new generation of material flow

Generating RL based approaches for the design and control of hyper-maneuverable UAVs

• Generating Designs of Multicopters/eVTOLS

• Evaluating Designs in Simulation per Performance Criteria (Maneuverability, Energy Efficiency, etc)

• Implementing Designs in hardware/physical UAVs

• Researching Deep Reinforcement Learning policies for end-to-end vision only control and navigation of autonomous UAVs.

• Exploring Generalization through Simulation to develop a hybrid RL policy that leverages simulated and real-world data for perception and control.

• Engineered Simultaneous Localization and Mapping (SLAM) algorithms for robot perception and ego-trajectory estimation of the Tesla Bot Humanoid Robot.

• Developed a robust perception pipeline that resulted in a 60x improvement compared to the Tesla Autopilot framework (C++, Python).