A unified end-to-end learning and control framework that is able to learn a (neural) control objective function, dynamics equation, control policy, or/and optimal trajectory in a control system.

Machine Learning + Spacecraft Trajectory Optimisation

Whole-body motion planning for Unitree G1 humanoid in MuJoCo - ZMP preview control, A* footstep planning, MPC balance (49% energy reduction), RL locomotion, and Jacobian IK manipulation

A library for research into legged critters

Gradient-based trajectory optimisation toolbox that speeds up trajectory optimisation by only computing dynamics derivatives at key-points with finite-differencing. Remainder of derivatives are approximated via interpolation.

Geometric reachability-based safety filtering for air traffic control

MATLAB N-Body Orbit Propagator with Space Surveillance (SST) & Collision Avoidance modules. Validated against NASA JPL.

MSc Thesis at Indra on real-time onboard successive convexification (SCvx) with quadratic programming (QP) for powered landing guidance of reusable space launchers