Key projects include:

• Developing robust RL-based locomotion policy for legged robots:

  Using Meta-Reinforcement Learning to develop a locomotion policy for the quadruped robot -Unitree Go1- to be able to robustly navigate different environments for survival and rescue missions.

• RL-based Energy Management Strategy for P2-PHEVs:

  Implementing a novel A3C agent that outperformed the existing rule-based control strategy and was integrated into the vehicle HCU for testing in HiL and ViL. The project resulted in two publications and a patent.

Key projects include:

• Adaptive interiors control with a German OEM - Project Lead:

  Developing an imitation learning RL agent that learns and predicts the driver preferences based on the driver mood, driving behavior and vehicle situation.

• RL-based Thermal Management for Battery Electric Vehicles:

  Improving the vehicle energy consumption by utilizing the heat sources to meet the cabin comfort requirements, powertrain, and battery optimal efficiencies.

See Employer Letter