ICRA 2026 has wrapped up, and we’re back from a fantastic week in Vienna! Booth 91 was busy from start to finish, and we wanted to put together a short highlight video to share some of what happened — for everyone who stopped by, and for everyone who couldn’t make it this year.
The Swarm Demo
At the center of our booth was our live autonomous swarm demo — multiple Crazyflies flying autonomously in a controlled indoor environment, with everything tracked, repeatable, and stable across runs. We also could play around with our Lighthouse wand – which was also a great solution for troubleshooting the few misbehaving drones we had during those 3 days.
SwarmGPT, Live and Interactive
One of the highlights of the week was demonstrating SwarmGPT together with the Learning Systems and Robotics Lab (LSY) at the Technical University of Munich. SwarmGPT explores a simple but powerful idea: instead of hand-coding trajectories, you describe the intent — pick a piece of music, prompt a style or expression — and the system handles the planning and safety while the swarm performs it.
This time around, we brought a more interactive version of the demo than our end-of-year collaboration a few months back, and visitors got to try it out for themselves at the booth. Watching people prompt the swarm and then watch their idea come to life in the air was a great reminder of how far natural-language interfaces have come, and how much room there still is to explore in this space.
Research We Saw on the Crazyflie
Beyond our own demos, one of our favorite parts of ICRA is talking with our users and seeing what the community has built. This year was no exception — we spotted Crazyflies appearing in research spanning multi-agent coordination, modular micro-UAVs designed for autonomous mid-air docking, and decentralized swarm control approaches where each drone makes its own decisions based on local information rather than a central planner. Some examples include:
It’s always a bit surreal to see the same small quadcopter we ship from our office end up at the center of such different research questions — from choreography and language-driven control, to docking and modular hardware, to fully decentralized swarms. If you presented work involving the Crazyflie this year, thank you for stopping by and sharing it with us — and if you left a poster behind, it’s already found a home on our office wall.
Thanks for Stopping By
ICRA continues to be one of our favorite events of the year, not just for the demos, but for the conversations. Someone describes a challenge they’re running into in their lab, and a few months later, that conversation has often turned into a feature, a library improvement, or a new piece of hardware. If you stopped by booth 91, told us about your research, or just said hello, thank you. We’re already looking forward to the next one!
If you’d like to dig deeper into any of what’s shown in the video, or want to get started with the Crazyflie yourself, head over to bitcraze.io or reach out at contact@bitcraze.io.
We built a small drone for people who want to understand how things fly. The community took it considerably further than that. The citations keep arriving from directions we didn’t anticipate. Spacecraft dynamics. Tactile human-swarm interaction. Onboard deep learning. Mapping algorithms that fit inside a nano-drone’s compute budget. The platform’s combination of openness, known dynamics, repeatable behavior, and low replacement cost turns out to be useful for a wider set of problems than any single team could have imagined building for.
What follows is by no means a comprehensive survey, but rather a selection of research areas where the Crazyflie has found a home, each illustrated with recent work. We find it genuinely interesting that the same hardware can be useful across this range, and we hope it gives other researchers a sense of what is possible.
1. Decentralized Multi-Agent Coordination and Swarm Control
Multi-agent coordination is probably the research area most closely associated with the Crazyflie, and for good reason. The platform’s light weight, predictable dynamics, and relatively low cost per unit make it practical to run experiments with enough agents to observe emergent swarm behavior, rather than just simulating it. A lab can field a meaningful swarm without the capital outlay that larger platforms would require.
Recent work has pushed this in some interesting directions. Decentralized approaches, where each agent makes decisions based on local information rather than a central planner, are particularly well-served by a platform where individual failures don’t cascade into catastrophic system loss. Research on collision avoidance, formation control, and consensus algorithms benefits from hardware that can absorb the crashes that inevitably happen when you are testing novel coordination strategies.
The ROS 2 ecosystem around the Crazyflie has matured considerably, with frameworks like Crazyswarm2 enabling standardized multi-drone experiments that other labs can replicate. The reproducibility this enables is meaningful: a coordination result demonstrated on Crazyflies in one lab is demonstrable in another (see “CrazyChoir: Flying Swarms of Crazyflie Quadrotors in ROS 2” (arXiv)).
2. Onboard AI and Edge Inference at Nano Scale
What can you fit inside a couple of dozen grams grams and still have compute left over for intelligence? Quite a lot, it turns out, especially when researchers are motivated to find out. The AI Deck, which adds a GAP8 system-on-chip with a camera and Wi-Fi, opened a wave of work on fully onboard perception and inference pipelines on nano-UAVs.
More recently, work using custom expansion decks with the GAP9 processor has enabled onboard SLAM and scan-matching at take-off weights around 46 grams, showing that the platform’s expansion architecture makes it a meaningful target even as compute capabilities grow. See “Ultra-Lightweight Collaborative Mapping for Robot Swarms” (arXiv).
The Crazyflie’s transparent hardware design is important here: researchers can build custom decks, verify the power budget, and integrate new silicon without waiting for a vendor to offer an approved configuration.
3. Spacecraft and Orbital Dynamics Simulation
Researchers at the University of Houston, in collaboration with the US Air Force Research Laboratory, used Crazyflie drones to simulate the relative motion dynamics of spacecraft in formation, specifically the Clohessy-Wiltshire equations that describe how objects move relative to each other in near-circular orbit.
The reasoning is practical: testing spacecraft autonomy on-orbit is expensive and high-risk. Ground-based testbeds using air bearings exist, but are complex and space-intensive. A small fleet of Crazyflies, running scaled versions of orbital trajectories in an indoor lab, offers a much cheaper and more accessible way to validate formation-flying control laws and neural network guidance systems before committing to hardware that will be launched into space.
4. Reinforcement Learning: From Simulation to Real Hardware
Reinforcement learning (RL) for drone control has been a thriving research area for years, but the gap between simulation and physical hardware remains a hard problem to close. The Crazyflie’s well-documented dynamics, consistent manufacturing, and open firmware have made it a preferred target for sim-to-real transfer research, because the sim and the real thing can be brought into close agreement.
Work in this space spans a wide range of problem settings. Multi-agent RL for collision-free navigation, safe RL with control barrier functions, landing on moving targets, and agile trajectory following in cluttered environments have all been demonstrated on Crazyflie hardware. A common thread is that the platform’s low inertia and predictable response make it a fair test: there is nowhere to hide on a platform this light and responsive, and if the policy is sloppy, it falls (see “AttentionSwarm: Reinforcement Learning with Attention Control Barier Function for Crazyflie Drones in Dynamic Environments” (arXiv).
The “LEARN framework”, which claims to run a compact attention-based RL policy on six Crazyflies for multi-robot navigation through 0.2-meter gaps at 2 m/s, is a recent example of how far this line of work has come. The system runs fully onboard, using only time-of-flight sensors, and transfers directly from simulation to real hardware without fine-tuning. See “LEARN: Learning End-to-End Aerial Resource-Constrained Multi-Robot Navigation” (arXiv).
5. Human-Swarm Interaction and Expressive Robotics
An unexpected corner of the research community is the one that adopted the Crazyflie to the human-robot interaction field. It turns out that a swarm of small, quiet, slow-moving drones is a better vehicle for studying how humans interpret and respond to group robot behavior than many ground robot alternatives.
Work in this space ranges from the technical to the almost philosophical. Researchers have studied whether humans perceive swarm motion as intentional and communicative; whether vibrotactile feedback can give operators an intuitive sense of swarm state during physical interaction; how flight formation shapes emotional perception; and how to design impedance-controlled swarms that respond naturally to human touch (see “SwarmTouch: Tactile Interaction of Human with Impedance Controlled Swarm of Nano-Quadrotors” (arXiv).
The Crazyflie’s low injury risk in the event of a collision, its predictable behavior, and its ability to carry sensing and communication payloads make it well-suited to user studies where physical proximity and spontaneous human response are important variables. The fact that the platform is widely available also matters: HRI research benefits from results that can be reproduced in different lab environments with different participant populations.
Looking across these five areas, a pattern emerges. In each case, the research is not about the Crazyflie itself. The platform is a means, not an end.
What the Crazyflie provides is a credible physical substrate that researchers can trust to behave consistently, modify freely, and replace cheaply when something goes wrong. The open source firmware means the dynamics are fully inspectable. The transparent hardware means the platform can be extended with custom decks. The stable software ecosystem means results from one year’s experiments can be compared against another year’s, and against results from other labs using the same platform.
If your research uses the Crazyflie in a direction not represented here, we’d like to hear about it. The research portal at bitcraze.io/portals/research lists some of what we know about, but the community is larger and more inventive than any curated list can capture.
It’s that time of year again! ICRA 2026 (IEEE International Conference on Robotics & Automation) is just around the corner, and this year we’re heading to Vienna. We couldn’t be more excited about this one: Vienna is an incredible city, and we’ve been working on some things we can’t wait to share.
June 1–5, 2026. Come find us!
A reproducible testbed for aerial robotics research
We will be running a live autonomous flight system based on the Crazyflie platform.
The focus is not the flight itself, but what it enables. The system provides a controlled indoor environment where experiments can be repeated, variables isolated, and results compared over time.
This is aligned with how aerial robotics research is actually conducted: iteration speed, reproducibility, and observability matter more than scale in early and mid-stage research. Our platform is designed around those constraints.
Autonomous indoor flight for controlled experimentation
The setup demonstrates autonomous flight under conditions that remain stable across runs.
This allows researchers to evaluate control strategies, perception pipelines, and multi-robot coordination without environmental noise dominating results. It also reduces costs and operational overhead compared to larger platforms, which changes how frequently experiments can be run.
In practice, this makes it feasible to move from idea to validated result faster and with clearer insight into failure modes.
Used in swarm robotics, control, and physical AI research
The Crazyflie platform is used across domains such as swarm robotics, learning-based control, SLAM, and human–robot interaction.
It has been referenced in hundreds of peer-reviewed publications and is often used as a bridge between simulation and larger systems. The value is not in representing the final deployment environment, but in enabling rigorous, comparable experimentation at low cost and risk.
If you are working in these areas, we are interested in how your setup is structured and where constraints appear.
Share your work with us
If you are presenting work that involves the Crazyflie, we would like to see it.
Even better, if you do not need your poster after your session, bring it by the booth! We collect and display these as part of the broader body of work built on the platform. We will make sure it is appreciated properly.
Meet us at ICRA 2026
One of our favourite things about ICRA is getting to meet the community in person, hearing about your research, seeing what you’ve built with the Crazyflie, and exchanging ideas with people who are just as excited about small flying robots as we are. Whether you want to chat about your research, see the demo up close, or just catch up, our booth is the place to be. We love hearing about all the cool projects you’re working on with the Crazyflie, so don’t be shy!
If you are working with the Crazyflie, evaluating platforms, or exploring new research directions, stop by booth 91. You can also reach out at contact@bitcraze.io to schedule time.
Today, we welcome our first blog post by Maurice Zemp. Stay tuned for more of his adventures later!
When I started working on my Matura thesis (a mandatory project in Swiss High School), I wanted to create something that went beyond a purely theoretical project. I was fascinated by the idea of combining cutting-edge technology with a very tangible and exciting challenge: making a small drone fly through a racing course, composed by small gates, completely on its own and as fast as possible. Inspired by the work under Prof. Davide Scaramuzza, I set off for a challenge to find some new alternatives or improvements, while developping my approach from scratch.
What may sound simple at first quickly turns into a highly complex task. A drone needs to perceive its environment, process information in real time, and decide on precise actions within fractions of a second, all without human intervention. Professional drone racing pilots train for years to master this level of control. My goal was to see whether artificial intelligence could achieve something similar, using reinforcement learning as the core technique.
But there was another challenge: I wanted to design a system that wasn’t only powerful, but also affordable and reproducible. Many research institutions use equipment worth tens of thousands of francs for projects like these. I asked myself: Could I build something comparable with a fraction of the budget, and still push the boundaries of what’s possible?
That question became the driving force behind my project, which later brought me all the way to the finals of Schweizer Jugend forscht (short SJf) and saved me a place as the main prize to represent Switzerland at the world’s biggest Youth Science Competition, ISEF 2026. Over the following sections, I’ll share how I built my system step by step, what it was like to present it at the competition, and how it felt when all the effort finally paid off.
Fig. 1: Me with the Crazyflie 2.1 Brushless in the halls of ETH Zurich, where the main event of SJf took place.
Motivation and Objectives
The project I presented at Schweizer Jugend forscht was the result of my Matura thesis, in which I set out to combine my interests in drones, programming, and artificial intelligence. My goal was to develop a complete system for autonomous drone racing, based on Reinforcement Learning (RL), that would not only work in simulation but could also be transferred to real-world conditions.
To achieve this, I focused on three key aspects:
Building a highly efficient simulation environment for training a reinforcement learning agent.
Developing a cost-effective motion-capture system (MoCap) capable of tracking a drone’s position and orientation in real time with high precision.
Integrating both systems in a way that would allow seamless transfer from simulation to real-world experiments with minimal latency.
This combination made the project unique: instead of relying on expensive commercial hardware, I set out to create a solution that would be precise and affordable but still continue the state-of-the-art development in Drone Racing.
Simulation Environment
The simulation was implemented in Python, using Stable Baselines, OpenAI Gymnasium, and NumPy, accelerated with Numba for performance. At its core, the system employed the Proximal Policy Optimization (PPO) algorithm, a state-of-the-art reinforcement learning algorithm known for stability and efficiency.
Unlike general-purpose simulators such as Gazebo, my environment was designed specifically for drone racing. It could process tens of thousands of interactions per second, enabling a training run of a few dozen minutes to correspond to nearly a year of simulated flight time.
Key features included:
A physics-based flight dynamics model accounting for thrust, drag and gyroscopic effects.
A carefully engineered reward function balancing speed, precision and avoiding shortcuts.
A flexible design that allowed different gate sequences and drone parameters to be tested.
A multi-agent environment computing on multiple threads, leading to much shorter training time needed.
Fig. 2: Path of the drone after two hours of training on a given track
With this setup, an RL agent could learn to complete arbitrary racing tracks in near-optimal time1 after only a few hours of training. In simulation, top speeds of up to 100 km/h were achieved, though these exceeded the physical limits of the real drone and were generated with modified drone parameters.
Motion-Capture-System
A second cornerstone of the project was the development of a low-cost motion-capture system. Instead of relying on high-end solutions such as VICON or OptiTrack (which can cost tens of thousands of Swiss francs), I built a custom setup. The drone – a Crazyflie 2.1 Nanocopter (later a Crazyflie 2.1 Brushless) – was fitted with infrared diodes. With four cameras capturing at 120 frames per second, the drone’s position was calculated in real time through triangulation. By using three diodes arranged on the drone, I not only wanted to estimate the position but also the orientation. Unfortunately, due to the cameras being budget and thus not having a high-end resolution, the estimation of the orientation was not feasible and was therefore taken from the onboard IMU.
Fig. 3: Motion-Capture-System Concept
System Integration
For integration, I relied on the ROS2 middleware and the Crazyswarm2 framework, which allowed the simulation and MoCap data to be processed together with minimal latency. This setup ensured that a policy trained in simulation could be executed on the real drone almost seamlessly.
Fig. 4: Sim2Real Integration Concept
Results & Discussion
The results demonstrated the effectiveness of the combined system:
In simulation, the RL agent completed tracks in nearly optimal time1, demonstrating robust generalization across different gate sequences.
The motion-capture system delivered millimeter-level accuracy and reliable tracking in real time despite its simplicity. (see Fig. 5)
In real-world tests, the Crazyflie drone successfully completed tight gate sequences, even at speeds of up to 25 km/h, with a positional deviation of only 5–12 cm compared to the simulated trajectories. (see Fig. 6)
Given that the gates were roughly A3-sized openings (38 × 29 cm), the precision was sufficient to consistently hit every gate, confirming the feasibility of transferring simulated training into real-world racing. (Video)
Fig. 5: Boxplot of the deviation for collected points on a plane in millimeters
Fig. 6: Tracked Real Life Path on a given trackFig. 7: Realisation in my school’s physics laboratory
The system demonstrated that RL-based controllers can generalize effectively from simulation to reality, even with imperfect models. This robustness shows that RL is promising for real-world robotics applications where exact physical modeling is difficult or costly. To further investigate this robustness, I conducted ablation studies in simulation focusing on aerodynamic drag, as well as analyses of the MoCap system’s accuracy in measuring dynamic motion.
As for the aerodynamic drag, I introduced some randomness during evaluation, which deviated largely from what the Algorithm had trained upon. To achieve this, I introduced a parameter k which was multiplied with the correct aerodynamic drag to view how these differences would impact the flight. In Fig. 8 you can see the deviation from the real path over the course of a flight. As obviously more drag results in lower velocities, the two paths had to be fitted using the nearest two possible points. The results are impressive, even with an enormous change to training such as k = 10 the flight still performed somewhat okay, even though the gates probably wouldn’t have been hit anymore.
For the MoCap System, I used a mathematically representable motion, such as a pendulum. Then I performed a parameter optimization on this mathematical model and measured the deviation based on the current speed. There is a clear correlation between speed and deviation (see Fig. 9), which most likely was caused by the missing calibration of the cameras.
Fig. 8: Results of ablation studies on aerodynamic dragFig. 9: Experiment to confirm correlation between deviation and speed
Conclusion and Reflection
This work demonstrates that RL, combined with accessible hardware, enables precise and robust autonomous flight in dynamic indoor environments. The results underscore the potential of low- cost robotics solutions to democratize drone research. Even though it was a large project and there were some hard times, I enjoyed working on it a lot and believe that the result and the memories made with it are even more rewarding than any prize money!
Acknowledgments
To end this guest post, I want to sincerely thank Bitcraze for their amazing work and support during the development of this project. They have genuinely built such an incredible testbed for research in autonomous drones, it’s amazing! Without them, this project wouldn’t have worked out the way it did!
Near-optimal time here references to the theoretical boundary of a drone completing this track, given it’s parameters. To be clear, this is not fitting a polynomial with boundaries on their derivatives onto gate segments, as it was done during earlier approches in the 2010s. It was compared to multiple previous approches by UZH RPG such as Optimal Control (OC). Nevertheless, this comparison could be done more extensively in the future, as time ran short in the end. ↩︎
Setting up a mobile indoor positioning system with the Crazyflie platform—specifically using the Loco Positioning System (LPS)—typically involves walking through a large room with a laser measure, sticky notes, and a bit of patience.
That was exactly my situation at work when setting up UWB anchors for mobile robot demonstrations, which often had to be relocated to different rooms. And it got me thinking—what if there were a more intuitive, flexible way to define and deploy these setups, especially in labs or educational settings where reconfigurability is key?
This question led to the development of XR-PALS, a mixed reality tool that simplifies the entire LPS configuration process. Instead of manually measuring anchor positions and inputting coordinates into cfclient, XR-PALS allows users to create and sync virtual and physical setups almost seamlessly—just by moving virtual anchors in space using a VR headset.
Three-step setup process for the Bitcraze Loco Positioning System, illustrating anchor connection, system configuration, and 3D Crazyflie drone tracking.
What is XR-PALS?
XR-PALS (XR-Powered Assistance for Loco Positioning Systems) is a tool we (Victor Victor and Dominik Grzelak) developed for defining cyber-physical spaces using UWB anchors on tripods. It runs on a passthrough VR headset (we used Meta Quest 3), allowing users to visualize the LPS layout and manipulate anchor placements in a virtual 3D environment until they are virtually and physically aligned.
Once everything is aligned, a single tap on the “Export” button sends the virtual anchor positions to a Rust-based middleware service running on a host machine within the same network as the VR headset.
It’s fast, visual, and eliminates the need for hand-measuring coordinates for our mobile robot demonstrators.
You can watch the system in action in our demo video:
Why We Built It
Usually, the LPS setup involves manual measurements and manual data entry in cfclient. This is time-consuming when environments change frequently—as is often the case in education and research.
With XR-PALS, the goals are to:
Speed up setup – Users in our study completed setups significantly faster.
Reduce error – Anchor placement is visual and interactive. No calculations needed.
How It Works
XR-PALS consists of two main components:
VR application Runs on a VR headset and displays the full anchor layout. Users can drag anchors, view distances, toggle info layers, and adjust alignment using natural hand gestures.
Middleware service A Rust-based REST API that basically bridges the VR app with the LPS anchors. This service receives the coordinates and remotely updates the LPS anchors—enabled by the open-source nature of the LPS firmware: bitcraze/lps-node-firmware. Coordinates are exported as YAML that can be imported by cfclient.
System architecture.
User interactions.
Real-World Evaluation
We compared XR-PALS against a traditional laser measurement tool in a user study involving 18 participants from our Faculty of Computer Science. Each participant had to configure a LPS layout (trapezoidal or rectangular) using one of the two tools.
Results:
XR-PALS users completed the task faster
Fewer measurement errors
Higher usability scores and confidence
Lower cognitive load, especially for the trapezoidal layout
Bar chart comparing task completion times between XR-PALS and laser.Lower values are better.
Measurement error comparison between XR-PALS and Laser. Lower values are better.
The minimal loss in accuracy compared to the time saved is not dramatic for our use cases, compared to the time saved—especially when more participants were able to correctly determine the coordinates overall. In general, users were more confident using VR glasses than a laser measurement device.
Summary and What’s Next
So if you’re working with dynamic robot setups or swarm experiments, XR-PALS might save you a lot of setup time.
The app works natively with the Bitcraze Loco Positioning System and, so to speak, extends the capabilities of the cfclient.
Here’s what we’re currently thinking about:
Supporting larger anchor sets within the app (currently limited to 8 anchors)
Improving user interactions based on feedback from our user study
Displaying anchor diagnostics (e.g., latency, power levels)
Simpler layout reusability (layouts can be saved and shared only via the computer)
Unmanned aerial vehicles (UAV) are invaluable to challenging remote applications, including coastal monitoring, surveillance for safe passage in icy waterways, and search-and-rescue missions. However, after deployment in a remote setting, the functional life of the multirotor is ultimately limited by its battery life. The research we continue to investigate is the ability to cooperatively and autonomously land a multirotor on an uncrewed surface vessel (USV) for recharging. We address this problem in real time with safe control algorithms that we apply on a Crazyflie.
Approach
Our approach enables the Crazyflie to cooperatively coordinate, with a simulated USV, a safe landing in severe wave conditions. It is critical to the autonomy of the system that the agents do not know when or where they’re going to land at the outset, they are cooperating in real time to make these determinations. The novelty of this work is three primary contributions:
Learning a Spatial-Temporal Wave Model as a Gaussian Process
We first learn the local tilt model, representative of the spatial and temporal impact of waves on the tilt angle of a USV, using Gaussian Process (GP) regression. Prior to the execution of the landing, the USV collects Nd noisy observations D = {q, t, φ²}, where q is the position of the USV, t is time, and φ is the tilt angle of the USV. We use GP regression to learn the spatial-temporal tilt model: fw(q, t) = (φ(q, t))² . The predicted tilt and uncertainty, conditioned on the observed data D, at a query point a = [q, t] can be inferred using the posterior distribution.
Distributed Model Predictive Control
Our proposed model predictive control (MPC) architecture combines standard tracking MPCs for the Crazyflie and USV and augments them with additional artificial goal locations. These artificial goals enable the vehicles to coordinate without prior guidance. Each vehicle solves an individual optimization problem for both the artificial goal and an input that tracks it but only communicates the former to the other vehicle. The MPC integrates, into the cost functions for both vehicles, the learned mean and uncertainty quantification of the spatial-temporal wave model from the GP regression. This encourages the agents to converge to calmer waters enabling safer landings in variable wave conditions.
Low-Cost USV Simulation Testbed Platform
To validate the proposed MPC scheme for landing on a USV, we simulate the spatial-temporal motion of a USV in waves of variable intensity using a custom tilting platform. The custom tilting platform has two degrees of freedom (roll, pitch) and is affixed to the deck of a differential-drive unmanned ground vehicle (UGV), the ClearPath Robotics Husky. We selected a differential-drive UGV to replicate the motion of a broad range of USVs, including those with differential drive and conventional rudder steering. Our platform is low-cost, modular, and open source, enabling rapid testing and benchmarking of UAV-USV landing strategies indoors before going onto the water which is high-risk and expensive.
Block diagram of our proposed distributed model predictive control scheme.
Overview of the USV simulation testbed tilting platform for ground vehicles. Left: the platform is shown mounted on the deck of the ClearPath Robotics Husky with the Crazyflie 2.1 hovering above. Center: a closer view of the landing pad configuration and platform components. Right: Section view showing the ball-and-socket joint and the linkage mechanism.
Experimental Setup
Full system architecture.
We evaluate the proposed MPC scheme for UAV-USV cooperative landing in indoor experiments using the Crazyflie 2.1 and our tilting platform. These experiments represent a scaled-down version of real-world harsh wave conditions. Amplitude and frequency are informed by a survey of waves along the coastlines of the Great Lakes of North America. We select a spatially decaying sine wave whose amplitude decreases gradually with increasing x-position. We expect the Crazyflie and the simulated USV to cooperatively select the safest landing spatially and temporally by learning the GP and incorporating it the MPC scheme.
The UAV MPC is run off-board on a Thinkpad X1 Carbon with Intel Core i7-1270P Processor. The MPC runs at a frequency of 50Hz and transmits control inputs to the Crazyflie via long range Crazyradio USB. The Raspberry Pi onboard the Husky runs its MPC at 10Hz. The two vehicles communicate their goals using ROS topics on a local WiFi network and receive their own pose feedback at 240 Hz from a VICON motion capture system.
Experimental Results
Experiment 1. We define the tilt model. Our proposed distributed MPC scheme can locate a low-tilt landing location from all six initial platform positions.
Experiment 2. We learn the wave model using a GP. The MPC scheme can locate a low-tilt landing location from all three initial platform positions.
In the first set of experiments, we assume no uncertainty in the tilt model. We compare four MPC weighting strategies ranging from a purely cooperative strategy in red (where neither vehicle weights wave tilt in the MPC, neglecting spatial-temporal tilt motion), to our proposed strategy in blue (where both vehicles weight wave tilt highly in the MPC). In this set of experiments, our proposed approach (blue) reduces the tilt angle of the platform at landing by between 68-89% and results in a 53% increase in landing success rate over the purely cooperative strategy (red).
In the second set of experiments, we learn the wave tilt model as a GP. We compare the purely cooperative strategy in red from before to our proposed strategy in blue where we weight the posterior mean cost from the GP regression. In this set of experiments, our proposed approach (blue) reduces tilt angle of the platform at landing by 23-32% and results in a 47% increase in the landing success rate over pure cooperation (red).
Future Work
In the work presented above we assume that there is a local region with calm waves that can be reached by both vehicles to then perform a safe landing. However, in practical scenarios, spatial-temporal assumptions are not realistic if an emergency landing is necessary or if time and resources are constrained. Recently, we explored quadratic MPC strategies for landing a Crazyflie on the tilting platform in high frequency high amplitude conditions. In these strategies we include optimization costs that weigh position, attitude, and altitude errors between the multirotor and the platform.
Though all strategies successfully land in low-frequency, low-amplitude conditions, they have low success in the higher-amplitude conditions. Therefore, designing a safe controller that is robust to a wide range of wave amplitudes and frequencies is an ongoing research area.
Poster presented at the workshop ’25 Years of Aerial Robotics: Challenges and Opportunities’ at ICRA 2025.
A couple of weeks ago, we were at ICRA 2025 in Atlanta. This year’s ICRA drew over 7,000 attendees, making it the biggest edition yet. We had a booth at the exhibition where we showed our decentralized swarm demo. The setup included a mix of Crazyflie 2.1+ units with Qi charging decks and Crazyflie 2.1 Brushless platforms with our new charging dock. The entire swarm operated onboard, with two Lighthouse base stations for positioning. More details about the setup can be found in the recent swarm demo blog post.
8 Crazyflies flying simultaneously in our decentralized swarm at ICRA 2025
Some of the brushless drones carried our high-powered LED deck prototype to make the swarm more visible and engaging. One of the drones also had a prototype camera streaming deck, which held up well despite the busy wireless environment.
A Different Perspective
This year we were also invited to participate in a workshop: 25 Years of Aerial Robotics: Challenges and Opportunities, where I (Rik) gave a short presentation about the evolution of positioning in the Crazyflie, from webcam-based AruCo marker tracking to the systems we use today.
Usually, we spend most of our time on the exhibition floor, so being part of a workshop like this was a different experience. It was interesting to hear researchers mention the Crazyflie in their work without needing to explain what it is. That kind of familiarity isn’t something we take for granted, and it was nice to see.
Many thanks to all the participants of the workshop '25 Years Of Aerial Robotics: Challenges And Opportunities' that we had the pleasure to co-organise for #ICRA2025. The speakers, young researchers, chairpersons and, of course, the attendees all made it a unique experience🦾 pic.twitter.com/WNfUUYLtop
The workshop also gave us a chance to talk with both established researchers and newer faces in the field. What stood out most was hearing how people are using the Crazyflie in their work today. It’s very rewarding to see how what we do at the office connects with and supports real research.
Catching Up and Looking Around
One of the most rewarding parts of the conference was the chance to connect directly with people using the platform. We talked to many users, both current and past, and saw new research based on the platform. It was also great to reconnect with Flapper Drones, who build flapping-wing vehicles powered by the Crazyflie Bolt. And it was nice to see HopTo on the exhibition floor for the first time. The company is a spin-off from the Robotics and Intelligent Systems Lab at CityU Hong Kong, which published a Science Robotics paper on the hopcopter concept that used a Crazyflie. We also had the chance to catch up with a maintainer of CrazySim, an open-source simulator in the Crazyflie ecosystem. It’s always valuable to connect with people building on top of the platform, whether through research, hardware, or open-source tools.
Wrapping Up
ICRA 2025 was packed with activity. From demoing the swarm, to the workshop, to hallway conversations, it gave us a lot of valuable feedback and insight. Thanks to everyone who stopped by, joined a talk, or came to say hello.
This week, some of us are on an adventure! Marcus and Tobias will be exploring both the RIG and Embedded World fairs.
RIG showcases the latest innovations in robotics and intelligent systems, while Embedded World is the place to be for cutting-edge embedded technologies. Both events promise amazing demos, insightful talks, and a chance to catch up with some of our collaborators.
Planning to attend either fair? Let’s meet up! We’d love to explore the exhibitions together, chat about cool technologies, or just geek out about the innovations on display. We’ll be wandering through Embedded World on Thursday and hitting RIG on Friday. Send us an email if you’d like to connect – we’re always up for grabbing coffee!
Next May in Atlanta
After our adventures as visitors, we’re thrilled to announce that we’ll be exhibiting at the International Conference on Robotics and Automation (ICRA) 2025! Stop by our booth where we’ll be showcasing our latest demo. We’ll be, as always, available to discuss our newest products, answer your technical questions, and provide insights into how our solutions can transform your robotics applications. We’re also eager to hear your thoughts on what you’d like to see in our upcoming products. Mark your calendars and make sure to find us at Booth #131 – we may even have some presentations in the work, but nothing confirmed yet.
Today in the shop
And, last but not least, the Brushless is now available in a Swarm configuration! Both the Lighthouse Swarm bundle and Loco Swarm bundle have been added to our shop. These new bundles feature all the same components as our standard Swarm packages, but come equipped with the Crazyflie 2.1 Brushless instead of the Crazyflie 2.1+ model.
Drones can perform a wide range of interesting tasks, from crop inspection to search-and-rescue. However, to make drones practically attractive they should be safe and cheap. Drones can be made safer by reducing their size and weight. This causes less damage in a collision with people or the environment. Additionally, being cheap means that the drones can take more risk – as it is less expensive to lose one – or that they can be deployed in larger numbers.
To function autonomously, such a drone should at least have some basic navigation capabilities. External position references such as GPS or UWB beacons can provide these, but such a reference is not always available. GPS is not accurate enough in indoor settings, and beacons require prior access to the area of operation and also add an additional cost.
Without these references, navigation becomes tricky. The typical solution is to have the drone construct a map of its local environment, which it can then use to determine its position and trajectories towards important places. But on tiny drones, the on-board computational resources are often too limited to construct such a map. How, then, can these tiny drones navigate? A subquestion of this – how to follow previously traversed routes – was the topic of my MSc thesis under supervision of Kimberly McGuire and Guido de Croon at TU Delft, and my PhD studies. The solution has recently been published in Science Robotics – “Visual route following for tiny autonomous robots” (TU Delft mirror here).
Route following
In an ideal world, route following can be performed entirely by odometry: the measurement and recording of one’s own movements. If a drone would measure the distance and direction it traveled, it could just perform the same movements in reverse and end up at its starting place. In reality, however, this does not entirely work. While current-day movement sensors such as the Flow deck are certainly accurate, they are not perfect. Every time a measurement is taken, this includes a small error. And in order to traverse longer distances, multiple measurements are summed, which causes the error to grow impractically large. It is this integration of errors that stops drones from using odometry over longer distances.
The trick to traveling longer distances, is to prevent this buildup of errors. To do so, we propose to let the drone perform ‘visual homing’ maneuvers. Visual homing is a control strategy that lets an agent return to a location where it has previously taken a picture, called a ‘snapshot’. In order to find its way back, the agent compares its current view of the environment to the snapshot that it took earlier. The trick here is that the difference between these two images smoothly grows with distance. Conversely, if the agent can find the direction in which this difference decreases, it can follow this direction to converge back to the snapshot’s original location.
The difference between images smoothly increases with their distance.
So, to perform long-distance route following, we now command the drone to take snapshots along the way, in addition to odometry measurements. Then, when retracing the route, the drone will routinely perform visual homing maneuvers to align itself with these snapshots. Because the error after a homing maneuver is bounded, there is now no longer a growing deviation from the intended path! This means that long-range route following is now possible without excessive drift.
Implementation
The above mentioned article describes the strategy in more detail. Rather than repeat what is already written, I would like to give a bit more detail on how the strategy was implemented, as this is probably more relevant for other Crazyflie users.
The main difference between our drone and an out-of-the-box one, is that our drone needs to carry a camera for navigation. Not just any camera, but the method under investigation requires a panoramic camera so that the drone can see in all directions. For this, we bought a Kogeto Dot 360. This is a cheap aftermarket lens for an older iPhone that provides exactly the field-of-view that we need. After a bit of dremeling and taping, it is also suitable for drones.
ARDrone 2 with panoramic camera lens.
The very first visual homing experiments were performed on an ARDrone 2. The drone already had a bottom camera, to which we fitted the lens. Using this setup, the drone could successfully navigate back to the snapshot’s location. However, the ARDrone 2 hardly qualifies as small as it is approximately 50cm wide, weighs 400 grams and carries a Linux computer.
Eachine Trashcan with panoramic camera and Flow deck.
To prove that the navigation method would indeed work on tiny drones, the setup was downsized to a Crazyflie 2.0. While this drone could take off with the camera assembly, it would become unstable very soon as the battery level decreased. The camera was just a bit too heavy. Another attempt was made on an Eachine Trashcan, heavily modified to support both the camera, a flowdeck and custom autopilot firmware. While this drone had more than enough lift, the overall reliability of the platform never became good enough to perform full flight experiments.
After discussing the above issues, I was very kindly offered a prototype of the Crazyflie Brushless to see if it would help with my experiments. And it did! The Crazyflie brushless has more lift than the regular platform and could maintain a stable attitude and height while carrying the camera assembly, all this, with a reasonable flight time. Software-wise it works pretty much the same as the regular Crazyflie, so it was a pleasure to work with. This drone became the one we used for our final experiments, and was even featured on the cover of the Science Robotics issue.
Crazyflie Brushless prototype with panoramic camera.
With the hardware finished, the next step was to implement the software. Unlike the ARDrone 2 which had a full Linux system with reasonable memory and computing power, the Crazyflie only has an STM32 microcontroller that’s also tasked with the flying of the drone (plus an nRF SoC, but that is out of scope here). The camera board developed for this drone features an additional STM32. This microcontroller performed most of the image processing and visual homing tasks at a framerate of a few Hertz. However, the resulting guidance also has to be followed, and this part is more relevant for other Crazyflie users.
To provide custom behavior on the Crazyflie, I used the app layer of the autopilot. The app layer allows users to create custom code for the autopilot, while keeping it mostly decoupled from the underlying firmware. The out-of-tree setup makes it easier to use a version control system for only the custom code, and also means that it is not as tied to a specific firmware version as an in-tree process.
The custom app performs a small number of crucial tasks. Firstly, it is responsible for communication with the camera. Communication with the camera was performed over UART, as this was already implemented in the camera software and this bus was not used for other purposes on the Crazyflie. Over this bus, the autopilot could receive visual guidance for the camera and send basic commands, such as the starting and stopping of image captures. Pprzlink was used as the UART protocol, which was a leftover from the earlier ARDrone 2 and Trashcan prototypes.
The second major task of the app is to make the drone follow the visual guidance. This consisted of two parts. Firstly, the drone should be able to follow visual homing vectors. This was achieved using the Commander Framework, part of the Stabilizer Module. Once the custom app was started, it would enter an infinite loop which ran at a rate of 10 Hertz. After takeoff, the app repeatedly calls commanderSetSetpoint to set absolute position targets, which are found by adding the latest homing vector to the current position estimate. The regular autopilot then takes care of the low-level control that steers the drone to these coordinates.
The core idea of our navigation strategy is that the drone can correct its position estimate after arriving at a snapshot. So secondly, the drone should be able to overwrite its position estimate with the one provided by the route-following algorithm. To simplify the integration with the existing state estimator, this update was implemented as an additional position sensor – similar to an external positioning system. Once the drone had converged to a snapshot, it would enqueue the snapshot’s remembered coordinates as a position measurement with a very small standard deviation, thereby essentially overwriting the position estimate but without needing to modify the estimator. The same trick was also used to correct heading drift.
The final task of the app was to make the drone controllable from a ground station. After some initial experiments, it was determined that fully autonomous flight during the experiments would be the easiest to implement and use. To this end, the drone needed to be able to follow more complex procedures and to communicate with a ground station.
Because the cfclient provides most of the necessary functions, it was used as the basis for the ground station. However, the experiments required extra controls that were of course not part of a generic client. While it was possible to modify the cfclient, an easier solution was offered by the integrated ZMQ server. This server allows external programs to communicate with the stock cfclient over a tcp connection. Among the possibilities, this allows external programs to send control values and parameters to the drone. Since the drone would be flying autonomously and therefore low-frequencies would suffice, the choice was made to let the ground station set parameters provided by the custom app. To simplify usability, a simple GUI was made in python using the CFZmq library and Tkinter. The GUI would request foreground priority such that it would be shown on top of the regular client, making it easy to use both at the same time.
Cfclient with experimental overlay (bottom right).
To perform more complex experiments, each experiment was implemented as a state machine in the custom app. Using the (High-level) Commander Framework and the navigation routines described above, the drone was able to perform entire experiments from take-off to landing.
Using the hardware and software described above, we were able to perform the route-following experiments. The drone was commanded to fly a preprogrammed trajectory using the Flow deck, while recording odometry and snapshot images. Then, the drone was commanded to follow the same route in reverse, by traveling short sections using dead reckoning and then using visual homing to correct the incurred drift.
As shown in the article, the error with respect to the recorded route remained bounded. Therefore, we can now travel long routes without having to worry about drift, even under strict hardware limitations. This is a great improvement to the autonomy of tiny robots.
I hope that this post has given a bit more insight into the implementation behind this study, a part that is not often highlighted but very interesting for everyone working in this field.
The Crazyflie 2.1 Brushless with propeller guards on a prototype charging padThe optimized brushless motor
Finalizing the integration of the Crazyflie 2.1 Brushless into our software ecosystem and expanding its documentation were key steps in preparing for its launch. These efforts ensure compatibility, improve the user experience, and make the platform more accessible to the community. We’re looking forward to a smooth launch and to seeing how the community will utilize the new platform!
This year, we introduced updates to the Crazyflie 2.1 kit, making the 47-17 propellers the new default and including an improved battery. These upgrades enhance flight performance and endurance, culminating in the release of the Crazyflie 2.1+—an optimized iteration of our established platform.
And don’t forget the developer meetings, where we shared some more behind the scenes information and collected invaluable feedback from the community.
We also released a new edition of our research compilation video, showcasing some of the coolest projects from 2023 and 2024 that highlight the versatility and impact of the Crazyflie platform in research.
Team
In the past year, Bitcraze saw significant changes within the team. in February, Rik rejoined the team. Tove started at Bitcraze in April. Mandy, with whom we’ve already worked extensively over the years, joined as our production representative in Shenzen. At the end of the year, we said goodbye to Kimberly, whose contributions will be deeply missed. Additionally, we had Björn with us for a few months, working on his master’s thesis on fault detection, and Joe continued his industrial postdoc at Bitcraze that began in December 2023. Looking ahead, Bitcraze is hiring for two new roles: a Technical Sales Lead and a Technical Success Engineer, to support our ongoing projects and customer collaborations.
Midsummer lunch with the teamChristmas-themed Bitcraze office
As we close the chapter on 2024, we’re proud of the progress we’ve made, the connections we’ve strengthened, and the milestones we’ve reached. With exciting launches, new faces on the team, and continued collaboration with our community, we’re ready to soar to even greater heights in 2025. Thank you for being part of our journey!
