The TUM team demonstrates exceptional technical maturity by seamlessly integrating multi-sensor fusion with high-precision mechanical systems. It is a textbook example of academic excellence that prioritizes robust, mission-ready engineering over experimental novelty.
深度探索
先修知识
- 暂无数据。
后续步骤
- 暂无数据。
深度探索
WARR x HORYZN | European Rover Challenge 2026 | Qualification Video本站添加:
Hi, I'm Hirona from Horizon.
>> And I am Szilard from WARG, and together we are competing for the European Rover Challenge. Our joint team consists of two student initiatives from the Technical University of Munich.
>> Hi, my name is Leonard and I'm the software lead of project creator.
>> I'm Bastian and I'm the mechanical lead.
>> I am Szilard, the project lead, and together we build our rover, Felicia.
>> As mentioned earlier, the development of the rover is taken care by WARG Space Robotics. The team is structured to three divisions, to the mechanical, the electrical, and the software team. Each of these teams focuses on solving its respective set of challenges. The development of the drone is taken care by Horizon. They dedicated one part of their team to build the drone, Phobos.
And why do we want to participate in the ERC?
>> Because participating in the ERC for me means challenging myself by applying the skills and knowledge I've learned in university and by building the rover.
>> The ERC creates a competitive environment that motivates me and we also like to see cool project of other teams.
>> And we want to stress test the rover and we can't wait to do it on a real Mars yard.
>> [music] >> Our rover, Felicia, was designed with a focus on reliability and maneuverability.
We use four independently steerable wheels, which ensures a very capable navigation system and allows for maneuvers such as spot turning and crabbing.
This, in combination with our purposely designed wheel profiles, allow for excellent navigation in challenging terrains. Further improving our navigation abilities, Alicia utilizes a rocker suspension. The rockers are differentially linked, which ensures sufficient ground contact at all times.
All in all, our drive system has been proven to allow precise and reliable navigation.
>> In the electronics team, we are doing both hardware and firmware development.
For the hardware, we are developing our own custom-made PCBs, and for the firmware, we are writing our own code.
The power coming from the battery is being distributed by our power board.
For the electrical side of the drive system, we are using off-the-shelf driver boards from Trinamic. To ensure safety, we implemented a PCB that protects against reverse polarity, over voltage, and over current. To also protect against back EMF of the motors, a 2.3 mF capacitor bank was added to the circuit board.
>> The rover uses four sensors for odometry measurement. One RGB-D camera facing forward, one facing backward, a lidar facing the ground, and an IMU. All four odometry sources are fused using an extended Kalman filter to provide the rover with a final combined odometry estimation. Additionally, the lidar is scanning the terrain in front of the rover and incrementally building a cost map, where cell cost increases with terrain steepness. This gives the rugged and highly inclined areas high cost, which will prevent the rover from trying to traverse them, instead making the rover search a path around them. We are simulating the Mars yard with the Aruco tag landmarks placed according to the coordinates that you provide to us. When the rover sees at least three landmarks, it triangulates its location to correct odometry errors using point and plane solver algorithms from the Open Computer Vision Library.
When the rover receives a goal, it computes a path to it using the A* algorithm.
It then proceeds to autonomously navigate that path, paying attention to the local cost map computed from the lidar scan to assess whether the terrain in front of the rover is traversable.
>> The rover can show its ability to drive [music] forward, backwards, and because each wheel has a steering motor, it is possible to do a spot turn.
It also can drive uphill.
>> [music] [music] [music] [music] [music] >> And in case of an emergency, all power is cut as the emergency button is pressed.
>> We use a Quanser quadcopter made of glass fiber reinforced plastic. The platform includes an optical flow for indoor navigation, a down-facing camera for tag detection, and a one-board computer and Wi-Fi link for remote access.
>> Our drone can move upwards, forwards, backwards, to the left, to the right.
It can also rotate and land manually.
A safety system was implemented in such a way that pushing a button on the controller initiates auto landing.
>> Unfortunately, we ran into some delays concerning our robotic manipulator.
Therefore, the best we can show you is this excerpt from our multi-body dynamic simulation.
We use a 6° of freedom arm with a total reach of approximately 1.2 m.
All links are made of trusses using carbon fiber tubes and aluminum gusset plates. Where possible, we use slip rings to allow for infinite rotation, especially in axis 1, 4, and 6.
What we can show you is our version of a robotic gripper.
It functions based off a rack and pinion gearing system and jaws made of foaming TPU, which allows them to deform around our target objects and improves our lifting capabilities.
In this clip, you can also see our current storage system for both rock and soil samples.
It uses a single point load cell for accurate weight measurements.
Storage containers are mounted on linear rails, which allows for easy and quick removal of the samples.
Our gripper jaws are interchangeable and will be adapted for the task at hand.
Our deep sampling system remains unchanged from last year's ERC due to insufficient manpower and time constraints.
The drill system uses two lead screws for a coupled vertical motion of both the drill itself and its housing relative to the rover.
For the drilling itself, we use a specialized auger, which allows for taking deep samples at specified depths.
>> Thank you very much for watching our video. We all hope to see you soon in Krakow. And now, let's do some stress testing.
>> Woo!
>> [music] [music] >> Woo!
相关推荐
Beyond Robotics | European Rover Challenge 2026
beyondrobotics
189 views•2026-06-01
Beatbot Sora70: JetPulse Technology and AI obstacle avoidance and navigation!
DroidModderX
26K views•2026-06-02
Tesla FSD 14.3.3 Hits Phoenix Streets - FIRST LOOK
anthonystesla
114 views•2026-05-29
NVIDIA Just Unveiled a 75 DOF Humanoid Robot
DPCcars
74K views•2026-06-01
NVIDIA's NEW Humanoid Robot Isaac Groot Revealed!
DPCcars
3K views•2026-06-01
Reachy Mini: the $300 open source robot you can actually hack — Andres Marafioti, Hugging Face
aiDotEngineer
662 views•2026-05-29
GTU ROVER | ERC 2026 Final Video
GTURover
401 views•2026-05-31
Tesla's $25 Trillion Dollar Problem | Will Optimus Win?
RyanShawtech
253 views•2026-05-30
热门趋势
Why Batman Lets The Joker Live 🤨
zackdfilms
9222K views•2026-05-30
They're Complete Trash
penguinz0
558K views•2026-06-04
The Murder of Deputy Caleb Conley
MidwestSafety
810K views•2026-06-04
I Bought FAKE HopeScope Merch (and paid a subscriber to give it a makeover) | Hopeful Hauls
HangWithHopescope
158K views•2026-06-04











