Forte: Low-cost 3D-printed robotic arm for service robotics; Inter-LLM algorithm boosts household robot multi-object task success
Key Takeaways
- Forte: Low-cost, fully 3D-printed robotic arm with sub-millimeter repeatability introduced.
- Inter-LLM algorithm boosts household robot multi-object task success by 30%.
- GR-3 VLA model and ByteMini robot advance adaptable service robotics.
- FollowUpBot: LLM-driven robot improves postoperative patient monitoring adaptively and privately.
- Adaptive visuo-tactile fusion achieves 93% success in dexterous robot tasks.
Top Stories
Forte: Low-cost, 3D-printed 6-DoF robotic arm with industrial-grade precision
On July 21, 2025, Forte, a fully 3D-printable robotic arm with 0.63 kg payload and sub-millimeter repeatability, was introduced to advance affordable service robotics. It minimizes backlash and control complexity without costly electronics, enhancing practical robot deployment.
Inter-LLM algorithm boosts household robot multi-object task success by 30%
On July 21, 2025, researchers presented Inter-LLM, an algorithm improving household robot multi-object collection by 30% in success and efficiency. It enhances robot intelligence for better human command fulfillment in service robotics.
GR-3 VLA model and ByteMini robot advance adaptable service robotics
On July 21-22, 2025, GR-3, a vision-language-action model, and ByteMini, a bi-manual mobile robot, were introduced to improve adaptability to new objects and tasks in service robotics.
FollowUpBot: LLM-driven robot improves postoperative patient monitoring
On July 21, 2025, FollowUpBot was introduced as an LLM-powered edge robot providing adaptive, private postoperative care with high interaction coverage and report accuracy.
Adaptive visuo-tactile fusion achieves 93% success in dexterous robot tasks
On July 21, 2025, a force-guided attention fusion method was published, enabling robots to adaptively combine visual and tactile data, achieving 93% success in fine manipulation tasks.
Real-World Robots
Gaze-informed LLM framework enhances human-robot interaction adaptability
Researchers developed a gaze- and speech-informed LLM framework improving adaptability and engagement in bi-directional human-robot interaction for service robots.
Human-in-the-loop teleoperation enhances safety in EV battery recycling
On July 20, 2025, a teleoperated system was proposed to safely disassemble electric vehicle batteries, reducing hazards and enabling future automation in service robotics recycling.
Mobile robots achieve 80% kiwifruit detachment and autonomous pollination
On July 21, 2025, research detailed mobile robots for kiwifruit harvesting with over 80% detachment and autonomous pollination, advancing agricultural service robotics.
Tech Advances
Adaptive robot task planning improves human-robot collaboration efficiency
On July 20, 2025, a task planning framework was introduced to optimize task allocation and scheduling, enhancing team performance and human preferences in collaborative service robots.
Data-driven MPC improves control accuracy of flexible cable-driven robotic arms
On July 21, 2025, a data-driven model predictive control method was presented, achieving 2.070 mm positioning accuracy and reducing solution time by 80% for cable-driven robotic arms.
LOAT framework enables zero-shot object navigation in mobile robots
On July 20, 2025, the LOAT framework was published, integrating LLMs to improve object-goal navigation generalization and zero-shot deployment in service mobile robots.
EMP framework enables stable humanoid upper-body motion imitation
On July 21, 2025, a reinforcement learning framework with Executable Motion Prior was introduced to enable humanoid robots to imitate human upper-body motions while maintaining stability.