Learning Systems and Robotics Lab (is hiring!)
@learnsyslab
Followers
3K
Following
136
Media
156
Statuses
279
Creating the next generation of robots w/ control, ML & perception! Led by Prof @angelaschoellig at @TU_Muenchen & @UTIAS @TUM_MIRMI @UofTRobotics @VectorInst
TU Munich and U of Toronto
Joined December 2014
RT @SepehrSamavi: Our latest work #SICNav #Diffusion: Safe and Interactive Crowd Navigation with Diffusion Trajectory Predictions has been….
0
2
0
The story about our crowd navigating robot aired this morning on the German public broadcaster ZDF! . ZDF video: w/ subtitles:
We hosted @ZDF @morgenmagazin to film a story about our crowd navigation project running SICNav. It should air on Monday 12 May 2025 between 05:30 and 09:00. SICNav @IeeeTro paper: Find us at #ICRA2025: WeET14.6, Wed. 21 May 17:05, Room 402.
0
0
3
Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.
380
671
3K
We hosted @ZDF @morgenmagazin to film a story about our crowd navigation project running SICNav. It should air on Monday 12 May 2025 between 05:30 and 09:00. SICNav @IeeeTro paper: Find us at #ICRA2025: WeET14.6, Wed. 21 May 17:05, Room 402.
0
2
3
Are you a TUM student with a passion for robotics, machine learning and control? Check out our course offerings for the upcoming summer semester here:
0
1
16
RT @TU_Muenchen: Jack, a wheeled #robot from our university, navigates crowds by predicting people’s paths and adjusting 10 times per secon….
0
6
0
How can a robot traverse crowds of people? Checkout this press release about our crowd navigating robot.
tum.de
Researchers at TUM have built a travelling robot that moves quickly and safely through crowds of people - thanks to interaction between humans and robots.
0
2
12
What happens to your diffusion controller when cameras fail? .Federico's latest paper with NASA's JPL🚀augments diffusion with long-term memory to plan and move even in total darkness using only touch sensors. 📰 📽️
0
1
6
🤖Want to make a safe RL agent? Check out Federico's and @lukas_brunke latest paper on integrating safety filters during training. They show it will boost your performance and sample efficiency!.📰 📽️
0
1
8
Happy Robotic Holidays from the Learning Systems and Robotics Lab! .Episode 3: The Flight before Christmas .Video:
0
3
6
Happy Robotic Holidays from the Learning Systems and Robotics Lab! Episode 2: Santa's Little Helper. Video:
0
0
5
Happy Robotic Holidays from the Learning Systems and Robotics Lab! Episode 1: Delivering Gifts. Video:
0
3
14
RT @angelaschoellig: Last day at #CoRL2024: join us in-person in Munich (ground floor) for an amazing set of speakers at our Master Manipul….
0
5
0
Using our framework, a robot transports a dry sponge at normal speed, but when holding a cup of water, the robot now also considers semantic constraints, for example, not moving above the laptop and avoiding rotating the cup to avoid spillage. (5/5)
0
0
1
For each scene, collision constraints are generated from object point clouds. In addition, semantic constraints are based on point clouds, semantic labels, and LLM-derived safety conditions: spatial (blue text), behavioural (orange text), and pose-based (green text). (4/5)
1
0
1
How does it work? Our semantic safety filter uses a perception module to create a 3D semantic map. LLMs generate constraints on object positions, behaviour, and poses, defining safe actions. Every command is certified before execution, ensuring safe robot motion. (3/5)
1
0
1
Arxiv: Full video: Link: Researchers: @lukas_brunke, Yanni Zhang, @ralfroemer99, @JackNaimer, Nikola Staykov, @zhousiqi9, @angelaschoellig.(2/5).@UofT @TU_Muenchen @VectorInst @UofTRobotics @UTIAS @TUM_MIRMI
1
1
5
How can robots act with common-sense safety in mind? Our new framework combines 3D mapping, LLM reasoning, and safety filters to avoid semantically unsafe actions like carrying a cup of coffee over a laptop. (1/5).#Robotics #AI #RobotSafety #LLM #AIresearch
2
2
12
How can a robot navigate a crowd while jointly predicting human motion & planning robot motion, all while satisfying explicit safety constraints? . Check out our latest work SICNav recently accepted to @ieeeTRO! . preprint, videos, code:
Our latest work Safe and Interactive Crowd Navigation (SICNav) has been accepted to @ieeeTRO! SICNav jointly solves planning and prediction while satisfying safety consts! Thanks @florian_shkurti @angelaschoellig . preprint & code:
0
1
15
🚀 Federico's and @lukas_brunke latest paper is out! They tackle safety in RL by integrating safety filters during training, not just evaluation. Verified on a Crazyflie 2.0, this boosts performance and sample efficiency!.📰 📽️
0
1
6
Big thanks to everyone who visited us at the TUM Open House! It was amazing to meet the next generation of roboticists, with Hello Robot passing out high fives to excited kids!
0
2
9