This was a graduate robotics team project I led. The goal was to learn the control theory behind propulsive rocket landings — think Falcon 9 — without the cost of real rocket hardware. A single-wing monocopter that spins about its vertical axis was chosen as a minimum-hardware testbed.
The control challenge: inputs must be phase-shifted relative to the blade's rotation angle to produce the intended real-world response. We designed a PD controller with Kalman-filtered IMU data for azimuth estimation and applied the appropriate phase compensation to the control law.
In practice the hardware was extremely difficult to stabilize — the vehicle would destroy itself before any useful data could be collected. The project shifted to a full simulation environment where the PD control law was implemented and validated properly.
Gallery
