Manuel Yves Galliker

Development of general-purpose robotic vision-language-action foundation models for robotic manipulation with a focus on real-time execution of action chunking flow policies and a robot agnostic controller runtime for cross-embodiment control.

Developed a torque-based reinforcement learning locomotion policy for the fully tendon-driven humanoid Neo spanning training in Isaac Gym, system identification for sim-to-real transfer, C++ inference, and an upper-body inverse dynamics controller. Integration of MuJoCo simulation for automated autonomy evaluation of RL locomotion and imitation-learned manipulation policies.

Leading and coordinating the successful bring up of our humanoid robot Neo from the first prototype to robust dynamic walking using Whole-Body NMPC and VR teleoperation. Developed a reinforcement learning pipeline in NVIDIA Isaac Sim and set up initial real-time C++ controls software stack.

Developed a real-time motion planning framework for bipedal loco-manipulation using Whole-Body NMPC. Software development, testing and system integration of our custom made torque controlled REVO2 electrical motors using embedded Field Oriented Control (FOC). Growing the team by successfully hiring two controls and one firmware engineer.

Supervising students in their semester thesis at ETH. Technical consulting in the realm of controls, electronics and optics.

Development of optical setup, electronics, controls and calibration algorithms for high precision laser tracking system for automated laser-weeding on a robotic rover platform for smart agriculture. Advancing the business plan of the start-up.

Advanced robotic assessment and therapy of somatosensory hand movement of patients with neurological injuries through software development for UI, data analysis and control for the assistive device ETH MIKE. (C#, Unity3D, SQLite)

Whole-Body Bipedal Locomotion through Nonlinear Model Predictive Control Achieved the first hardware demonstration of online gait generation under consideration of the full system dynamics on a bipedal robot trough developing a whole-body Nonlinear Model Predictive Control approach supervised by the AMBER-Lab and the Robotics Systems Lab. (C++, ROS, Nonlinear Systems and Control Theory, OCS2, Pinocchio, CppAd, Raisim)

Build a software framework to identify the dynamics model of Unmanned Aerial Vehicles (multirotors, fixed-wing, VTOL) from PX4 Autopilot flight logs provided by the default onboard sensor suite as part of my semester thesis at the Autonomous Systems Lab. (Python, C++, Scikit Learn, CVXPY, Aerodynamics, Numerical Optimization, Gazebo, PX4)

Assisting for questions and exercise sessions for students attending the Robot Dynamics master course by the Robotic Systems Lab and the Autonomous Systems Lab. The course provides an overview on how to kinematically and dynamically model and control typical robotic systems such as robot arms, legged robots, rotary wing systems, or fixed wing. (Nonlinear Systems and Control Theory, Matlab)

Enhanced Quality control and reliability in the production of an VTOL drone for high precision aerial mapping through expansion of software automated data collection, analysis and process optimization with a focus on actuators. (KPI assessment, Python, QT, Google Sheets API)

Improved reliability KPIs of drone through leading projects on automated temperature calibration of IMU, barometer and airspeed sensor and automated actuator test bench. (Project Management, Altium, Python, C++, Px4, RPi)

Improved performance and reliability through extensive sensor evaluation, actuator redesign and debugging of the drone and roll-out of new manufacturing processes.(Rapid Prototyping, Matlab, Electric Circuit Analysis & Design, Solidworks)

Coaching and consulting the team of the consecutive iteration of the student project ftero at ETH Zurich

Leading the controls and mechatronics team to develop system modeling, controls, actuation, electronics and sensing for a prototype of an Airborne Wind Energy System. Online presence, media communication and sponsor relations. (Project Management, Control Theory, Aerodynamics, KiCAD, Power Electronics, C++, PX4)

IT support for Windows and Linux Systems concerning drivers, hardware diagnostics and replacement, updates, backups and managing warranty cases. Consulting customers in choosing laptop models according to their needs.

Special Focus on Optimal Control, Planning, System Identification, Machine Learning, Embedded Systems, Numerical Optimization and Computer Vision

Nonlinear Model Predictive Control for Bipedal Locomotion

Special Focus on Mechatronics, Control and Electronics