Experimental implementation of state-dependent Riccati equation control on quadrotors

Multirotor unmanned aerial vehicles are well-known and reliable platforms for flight in indoor and outdoor environments. They perform stable flights; standard autopilots have been developed with safety features based on laws and regulations. The safety regulations, which are extremely necessary for outdoor flights, restrict modification of the control structure of the autopilots. Despite the various valuable theory/simulation studies, surprisingly, the experimental implementation of the state-dependent Riccati equation (SDRE) is absent in the literature on flight control, which is the main novelty of this work. Waypoint regulation in an indoor testbed and trajectory tracking of the same waypoints (a square with 6 m edge and 10 cm allowable position error) were practiced. They were compared to show the performance of the system design. The flight experiment was performed on 23 trials to show the reliability of the design and compared with the proportional-integral-derivative (PID), executed onboard without a traditional autopilot. The SDRE and PID were implemented on a customized quadrotor with Raspberry Pi3B+ and Python3 program for onboard implementation. Finding the mean tracking time of the SDRE for the mentioned square 70.86 s, the delay of the PID tracking by 8.98 s confirmed the better performance of the proposed controller over a classical approach. The experimental implementation of nonlinear optimal control is presented for a quadrotor. Flight data and repeatability tests are provided for waypoint control of the flight. The experimental SDRE control implementation is presented. The waypoint control is compared with SDRE trajectory tracking.