[Paper Review] Cascaded Incremental Nonlinear Dynamic Inversion Control for MAV Disturbance Rejection

Ewoud J. J. Smeur, Guido de Croon+1 more|arXiv (Cornell University)|Jan 25, 2017

Adaptive Control of Nonlinear Systems12 references4 citations

TL;DR

This paper proposes a cascaded Incremental Nonlinear Dynamic Inversion (INDI) controller for micro air vehicles (MAVs), combining INDI for attitude and position control to achieve superior wind gust rejection. In windtunnel tests, the MAV exhibited only 21 cm average maximum position deviation when flying through a 10 m/s wind gust—over 7× better than a PID controller.

ABSTRACT

Micro Aerial Vehicles (MAVs) are limited in their operation outdoors near obstacles by their ability to withstand wind gusts. Currently widespread position control methods such as Proportional Integral Derivative control do not perform well under the influence of gusts. Incremental Nonlinear Dynamic Inversion (INDI) is a sensor-based control technique that can control nonlinear systems subject to disturbances. It was developed for the attitude control of manned aircraft or MAVs. In this paper we generalize this method to the outer loop control of MAVs under severe gust loads. Significant improvements over a traditional Proportional Integral Derivative (PID) controller are demonstrated in an experiment where the quadrotor flies in and out of a windtunnel exhaust at 10 m/s. The control method does not rely on frequent position updates, as is demonstrated in an outside experiment using a standard GPS module. Finally, we investigate the effect of using a linearization to calculate thrust vector increments, compared to a nonlinear calculation. The method requires little modeling and is computationally efficient.

Motivation & Objective

• Address the challenge of wind gust disturbances in outdoor and indoor MAV flight, especially near obstacles or in urban environments.
• Overcome limitations of traditional PID controllers in rejecting persistent and aggressive wind disturbances due to gain saturation and slow integral action.
• Develop a sensorless, model-free control approach that uses accelerometer measurements to estimate disturbances and adjust control inputs incrementally.
• Demonstrate the feasibility and robustness of cascaded INDI for both attitude and position control in real-world scenarios, including GPS-based outdoor flight.
• Investigate the impact of linear vs. nonlinear thrust calculation on control accuracy and disturbance rejection performance.

Proposed method

• Implement an inner-loop INDI attitude controller using measured angular accelerations to compute incremental control inputs based on desired vs. actual acceleration.
• Cascade the inner-loop INDI with an outer-loop INDI position controller that uses linear accelerations from the IMU to compute thrust increments for position control.
• Use incremental control updates based on the difference between desired and measured accelerations, avoiding reliance on full system models or wind estimation.
• Incorporate the nonlinear motor-thrust curve (from Figure 8) to compute more accurate thrust vector increments, improving control precision during aggressive maneuvers.
• Apply filtering techniques to reduce noise in accelerometer measurements, ensuring stable control performance.
• Integrate the controller into the open-source Paparazzi autopilot for real-time implementation on a Parrot Bebop quadrotor.

Experimental results

Research questions

• RQ1 Can cascaded INDI control significantly improve disturbance rejection in MAVs during high-velocity wind gusts compared to traditional PID controllers?
• RQ2 To what extent does the use of nonlinear thrust calculation improve vertical acceleration tracking accuracy during aggressive maneuvers?
• RQ3 Can the INDI controller maintain robust performance without frequent GPS updates, enabling reliable outdoor operation?
• RQ4 How does the controller handle actuator saturation and control allocation when control effectiveness changes due to airframe modifications?
• RQ5 Does neglecting the yaw angle’s influence on thrust vector dynamics affect control performance, and can a different Euler angle sequence improve it?

Key findings

• The cascaded INDI controller achieved an average maximum position deviation of only 21 cm when flying through a 10 m/s windtunnel flow, significantly outperforming a comparable PID controller with 151 cm deviation.
• The INDI-based controller demonstrated robustness in outdoor flight using only standard GPS updates, confirming its viability in real-world scenarios.
• The nonlinear thrust calculation reduced the maximum error in vertical acceleration tracking compared to linear approximation, especially during aggressive maneuvers.
• The controller maintained good performance even when the vehicle passed through the zero bank angle point, where thrust requirements change abruptly.
• Online adaptation of control effectiveness enabled the controller to function reliably after airframe modifications, reducing the need for re-tuning.
• Preliminary results suggest that considering axis priorities in control allocation could further improve performance under actuator saturation.