research

Legged robots are expected to run, jump, and carry loads in the real world, but their performance is bounded long before the controller runs — by the actuators, gearboxes, and linkages chosen at design time. Torque density, reflected inertia, gear ratio, and thermal limits decide what behaviours are physically reachable, and these choices are usually fixed by hand, early, and in isolation from the control problem they constrain.

My research treats hardware design as an optimization problem in its own right, and couples it to the locomotion behaviours it has to support. I work on computational design and optimization of planetary-gearbox actuators, co-design frameworks that tune morphology and actuation against a target task, and the robot platforms needed to validate that any of it survives contact with reality. The emphasis throughout is on practical, reproducible systems: automated CAD, open tooling, and hardware experiments rather than simulation alone.

The work is organized around three themes.