Scientific research, robot learning, perception, planning, control, and new artificial intelligence methods for robotics.
A new imitation learning method for robotic disassembly, based on language and visual context, improved task performance by 35 percentage points compared to the baseline....
A new method, called PACDM, enables the efficient solution of kinematic problems in closed-loop mechanisms through modular reduction. Studies were conducted on a seven-degree-of-freedom manipulator…
The B-Human team from Germany won the world championship in robotic soccer and took two titles at the World Humanoid Robot Games tournament in Beijing. Their...
In July 2026, the German Institute of Robotics (RIG) launched the first permanent Industrial Advisory Board - a new structure supporting the integration of scientific research with real-world industry needs....
Researchers have identified a key challenge in reinforcement learning for legged robots- some tasks, such as transitioning from walking to standing on their hands,...
The new CT-SAFR framework, presented at the IEEE CAI conference in May 2026, can detect flawed reasoning in AI models used by autonomous robots.
On September 2, 2026, the European Parliament in Brussels became the setting for one of the most important robotics demonstrations in Europe - fifteen advanced...
A 2026 study showed that a voice mediator can largely replace humans in triadic robotic collaboration, achieving comparable results in two of...
A new method called MotiMem-Omega can reduce energy consumption in the memory interface of an autonomous vehicle by up to 36%, reducing the bit density by a value of 1 per…
The University of Michigan is making a number of robotics projects and datasets publicly available, which support academic research and technology development. These include tools...
At the University of Michigan, studying robotics is not just about theory - it’s about practice, research, and competitions. The program is one of the first of its kind at a top-tier university…
New FLARE framework enables manipulation robots to autonomously recover from errors in challenging tasks that require physical contact. The research shows that the approach improves effectiveness...
Researchers have introduced RAEM - a new framework for autonomous exploration of multi-story environments by quadrupeds. Simulation and real-world tests have demonstrated its ability to continuously and stably explore...
No matching language
Napisz krótko, co nie działa. Informacje o stronie i urządzeniu dodamy automatycznie.
