The MRS UAV System: Pushing the Frontiers of Reproducible Research, Real-world Deployment, and Education with Autonomous Unmanned Aerial Vehicles

TL;DR

The paper introduces The MRS UAV System, a modular, open-source full-stack platform for multirotor research that emphasizes reproducible results through realistic simulation and real-world deployment. It combines a bank of filters for multi-frame state estimation, a heading-based orientation model, and two complementary controllers—an $SE(3)$ geometric force tracker and an MPC-based force feedback controller—alongside a suite of trackers, a robust trajectory generator, and a Gazebo-based simulation environment. The system supports GNSS and GNSS-denied navigation, enables safe outdoor-indoor transitions, and was proven in high-profile competitions like MBZIRC and the DARPA SubT, demonstrating both high performance and scalable, educational value. The authors provide extensive implementation details and open-source code to facilitate replication, extension, and adoption by researchers and students, with real-world deployments informing continuous refinement. Overall, the MRS UAV System offers a practical, well-documented platform that advances reproducible UAV research, multi-robot experimentation, and robotics education, while bridging simulation and reality.

Abstract

We present a multirotor Unmanned Aerial Vehicle control (UAV) and estimation system for supporting replicable research through realistic simulations and real-world experiments. We propose a unique multi-frame localization paradigm for estimating the states of a UAV in various frames of reference using multiple sensors simultaneously. The system enables complex missions in GNSS and GNSS-denied environments, including outdoor-indoor transitions and the execution of redundant estimators for backing up unreliable localization sources. Two feedback control designs are presented: one for precise and aggressive maneuvers, and the other for stable and smooth flight with a noisy state estimate. The proposed control and estimation pipeline are constructed without using the Euler/Tait-Bryan angle representation of orientation in 3D. Instead, we rely on rotation matrices and a novel heading-based convention to represent the one free rotational degree-of-freedom in 3D of a standard multirotor helicopter. We provide an actively maintained and well-documented open-source implementation, including realistic simulation of UAV, sensors, and localization systems. The proposed system is the product of years of applied research on multi-robot systems, aerial swarms, aerial manipulation, motion planning, and remote sensing. All our results have been supported by real-world system deployment that shaped the system into the form presented here. In addition, the system was utilized during the participation of our team from the CTU in Prague in the prestigious MBZIRC 2017 and 2020 robotics competitions, and also in the DARPA SubT challenge. Each time, our team was able to secure top places among the best competitors from all over the world. On each occasion, the challenges has motivated the team to improve the system and to gain a great amount of high-quality experience within tight deadlines.

Figures (20)

• Figure 1: Multirotor UAV platforms equipped for various scenarios carried out by the system presented here.
• Figure 2: The image depicts the world frame $\mathcal{W}$ = $\{\mathbf{\hat{e}}_1$, $\mathbf{\hat{e}}_2$, $\mathbf{\hat{e}}_3\}$ in which the 3D position and the orientation of the UAV body is expressed. The body frame $\mathcal{B}$ = $\{\mathbf{\hat{b}}_1$, $\mathbf{\hat{b}}_2$, $\mathbf{\hat{b}}_3\}$ relates to $\mathcal{W}$ by the translation $\mathbf{r} = \left[x, y, z\right]^{\intercal}$ and by rotation $\mathbf{R}^{\intercal}$. The UAV heading vector $\mathbf{h}$, which is a projection of $\hat{\mathbf{b}}_1$ to the plane $span\left(\mathbf{\hat{e}}_1, \mathbf{\hat{e}}_2\right)$, forms the heading angle $\eta = \mathrm{atan2}\left(\mathbf{\hat{b}}_1^\intercal\mathbf{\hat{e}}_2, \mathbf{\hat{b}}_1^\intercal\mathbf{\hat{e}}_1\right) = \mathrm{atan2}\left(\mathbf{h}_{(2)}, \mathbf{h}_{(1)}\right)$.
• Figure 3: A diagram of the system architecture: Mission & navigation software supplies the position and heading reference ($\mathbf{r}_d$, $\eta_d$) to a reference tracker. Reference tracker creates a smooth and feasible reference $\bm{\chi}$ for the reference feedback controller. The feedback Reference controller produces the desired thrust and angular velocities ($T_d$, $\bm{\omega}_d$) for the Pixhawk embedded flight controller. The State estimator fuses data from Onboard sensors and Odometry & localization methods to create an estimate of the UAV translation and rotation ($\mathbf{x}$, $\mathbf{R}$).
• Figure 4: The bank of filters $\mathcal{K}=\{K_1,K_2,...,K_n\}$. The filters simultaneously estimate $\mathbf{x}_1,\mathbf{x}_2,...,\mathbf{x}_n$. The output hypothesis is chosen by the arbiter.
• Figure 5: An illustration of the implementation diagram of the proposed UAV system. Onboard sensors and actuator modules are depicted as grey blocks. The sensor combination varies depending on the particular UAV task. White blocks represent ROS components responsible for managing sensors or for interacting with the actuators (Mavros). Green blocks stand for feedback controllers (see Sec. \ref{['sec:reference_controllers_impl']}) and red blocks stand for reference trackers (see Sec. \ref{['sec:reference_trackers_impl']}). Purple blocks represent high-level components that provide the controllers and trackers with up-to-date data and maintain the synchronicity of the events. These include controller, tracker and estimator switching, gain and constraint scheduling, and take-off and landing.