Learned Incremental Nonlinear Dynamic Inversion for Quadrotors with and without Slung Payloads
For quadrotor control researchers and practitioners, this work offers a practical method to reduce sensor dependency in INDI-based controllers, though the improvement is incremental as it combines existing learning and control techniques.
This paper proposes a learned approach to replace the specialized sensor measurements required by Incremental Nonlinear Dynamic Inversion (INDI) for quadrotor control, using a neural network to predict residual forces. Experimental results show that the neural network can generate smooth approximations of INDI outputs, eliminating the need for rotor RPM sensors while maintaining trajectory tracking performance.
The increasing complexity of multirotor applications demands flight controllers that can accurately account for all forces acting on the vehicle. Conventional controllers model most aerodynamic and dynamic effects but often neglect higher-order forces, as their accurate estimation is computationally expensive. Incremental Nonlinear Dynamic Inversion (INDI) offers an alternative by estimating residual forces from differences in sensor measurements; however, its reliance on specialized and often noisy sensors limits its applicability. Recent work has demonstrated that residual forces can be predicted using learning-based methods. In this paper, we show that a neural network can generate smooth approximations of INDI outputs without requiring specialized rotor RPM sensor inputs. We further propose a hybrid approach that integrates learning-based predictions with INDI and demonstrate both methods for multirotors and multirotors carrying slung payloads. Experimental results on trajectory tracking errors demonstrate that the specialized sensor measurements required by INDI can be eliminated by replacing the residual computation with a neural network.