Design, Modeling and Control for a Tilt-rotor VTOL UAV in the Presence of Actuator Failure

Mohammadreza Mousaei, Junyi Geng, Azarakhsh Keipour, Dongwei Bai, Sebastian Scherer

TL;DR

This paper presents a custom tilt-rotor VTOL (Vertical Take-Off and Landing) UAV designed to survive actuator failures that would ground most aircraft. By leveraging a "dynamic control allocation" scheme, the system can automatically reconfigure its remaining 11 actuators—including tilting rotors and traditional flight surfaces—to maintain stable flight even if a motor dies or a wing flap jams.

What Problem Does This Paper Address?

Unmanned Aerial Vehicles (UAVs) typically fall into two camps: multirotors (great at hovering, bad at distance) and fixed-wings (great at distance, need runways). Tilt-rotor VTOLs offer the best of both worlds, but their complexity makes them prone to mechanical failure.

In a standard quadcopter, losing one motor usually leads to a crash. While some research has explored "dynamic hovering" (spinning in place) for failed multirotors, this isn't ideal for complex missions. This paper addresses the challenge of fault tolerance in hybrid aircraft. The authors ask: Can we use the inherent redundancy of a tilt-rotor design—which has more "control knobs" than it strictly needs—to keep flying normally after a partial hardware failure?

Methods

The researchers developed a custom UAV with 12 actuators: four motors, four tilting mechanisms for those motors, two ailerons, one elevator, and one rudder. Since the aircraft only needs 6 Degrees of Freedom (DOF) to move in 3D space, it has significant actuation redundancy.

Key technical components include:

  • Feasible Wrench Space Analysis: The authors mathematically mapped out the "wrench" (the total force and torque) the vehicle can generate. They visualized how this space shrinks when specific parts fail, providing a theoretical limit on what maneuvers are still possible.
  • Dynamic Control Allocation: Instead of using a static "mixer" (a fixed map of how much each motor spins for a given command), they implemented a real-time optimization algorithm. This "Control Allocator" solves a constrained optimization problem every few milliseconds to find the best way to use the available actuators to achieve the desired flight path.
  • PX4 Integration: The system was built as a lightweight extension to the industry-standard PX4 flight control stack, making it potentially portable to other platforms.

Datasets

The paper does not use a traditional machine learning dataset. Instead, the authors validated their approach through extensive high-fidelity simulations in Gazebo. They have released their source code and simulation environment as an open-source benchmark for the community at theairlab.org/vtol.

Key Insights and Ideas

  • Complexity as a Safety Feature: Usually, more parts mean more things can go wrong. Here, the authors turn that on its head: because the rotors can tilt and there are wing surfaces, the aircraft has multiple ways to achieve the same movement (e.g., pitching up can be done via the elevator OR by tilting the rear rotors).
  • The Importance of "Failure Knowledge": The experiments showed that if the controller knows an actuator has failed, it can recover almost instantly. Without this knowledge, the system continues trying to use the broken part, leading to aggressive rotations and crashes.
  • Cross-Domain Compensation: One of the most interesting "aha" moments is seeing the rotors compensate for wing failures. If the elevator (a flap on the tail) jams during high-speed forward flight, the system can tilt the rotors slightly to provide the necessary pitch moment—a move a traditional airplane simply couldn't make.

Key Results

The system was tested against five major failure scenarios in simulation:

  1. Motor Failure in Hover: The remaining three motors and their tilt servos adjusted to maintain a stable hover.
  2. Tilt Servo Failure in Hover: Even with a rotor locked at a 60° angle, the system reallocated thrust to stay level.
  3. Motor Failure in Cruise: The aircraft maintained forward flight by increasing speed on the remaining motors and adjusting control surfaces.
  4. Elevator Failure in Cruise: With the elevator locked at 6°, the system used rotor tilting to maintain altitude with only a 4-meter deviation.
  5. Aileron Failure in Cruise: The system compensated for a jammed aileron (15°) using differential motor speeds and the remaining healthy aileron.

Limitations and Future Work

  • Failure Detection: The current work assumes the system is immediately "informed" of the failure. In the real world, you need a separate "Failure Detection and Identification" (FDI) module to spot the break first.
  • Structural Stress: The paper notes that while the math allows for drastic rotor tilts at high speeds to compensate for failures, these maneuvers might put extreme physical stress on the aircraft's frame in the real world.
  • Real-World Testing: While the Gazebo simulations are robust, the authors aim to move to physical flight tests to account for unpredictable factors like wind gusts and sensor noise.

Summary updated

Abstract

Enabling vertical take-off and landing while pro-viding the ability to fly long ranges opens the door to a wide range of new real-world aircraft applications while improving many existing tasks. Tiltrotor vertical take-off and landing (VTOL) unmanned aerial vehicles (UAVs) are a better choice than fixed-wing and multirotor aircraft for such applications. Prior works on these aircraft have addressed the aerodynamic performance, design, modeling, and control. However, a less explored area is the study of their potential fault tolerance due to their inherent redundancy, which allows them to tol-erate some degree of actuation failure. This paper introduces tolerance to several types of actuator failures in a tiltrotor VTOL aircraft. We discuss the design and modeling of a custom tiltrotor VTOL UAV, which is a combination of a fixed-wing aircraft and a quadrotor with tilting rotors, where the four propellers can be rotated individually. Then, we analyze the feasible wrench space the vehicle can generate and design the dynamic control allocation so that the system can adapt to actuator failures, benefiting from the configuration redundancy. The proposed approach is lightweight and is implemented as an extension to an already-existing flight control stack. Extensive experiments validate that the system can maintain the controlled flight under different actuator failures. To the best of our knowledge, this work is the first study of the tiltrotor VTOL's fault-tolerance that exploits the configuration redundancy. The source code and simulation can be accessed from https://theairlab.org/vtol.
BibTeX
@inproceedings{mousaei2022design,
  author = {Mohammadreza Mousaei and Junyi Geng and Azarakhsh Keipour and Dongwei Bai and Sebastian Scherer},
  title = {{Design, Modeling and Control for a Tilt-rotor VTOL UAV in the Presence of Actuator Failure}},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
  year = {2022},
  doi = {10.1109/iros47612.2022.9981806},
}

Loading related content...

Figure