Panasonic’s T-Rex Adds Fast Touch Feedback to Robot Manipulation
T-Rex combines visual planning with faster fingertip feedback to improve two-handed robot manipulation in research tests. Photo: T-Rex research team
Brief

Panasonic’s T-Rex Adds Fast Touch Feedback to Robot Manipulation

Panasonic reports that T-Rex combined vision-language planning with faster fingertip corrections, achieving 65% average success across 12 research tasks.

By Laura Bennett • 2 mins read Published: Updated:

Panasonic and its research partners have detailed T-Rex, a robot-manipulation AI that combines vision and language with faster fingertip feedback. In a November 5 release, Panasonic reported a 65% average success rate across 12 contact-rich tasks, compared with 35% for the strongest tested baseline.

T-Rex generates broad movements from visual and language inputs at about five updates per second, while a tactile module adjusts actions at roughly 20 updates per second. This lets the robot respond to slipping objects or excessive pressure without repeating the slower visual-planning process for every correction.

The experiments covered two-handed tasks such as transferring an egg, turning pages, applying toothpaste and using a key. Training included 100 hours of synchronized robot data, with touch measurements from all 10 fingertips, following pretraining on human video. Removing tactile input reduced average success to 42%, Panasonic said.

The project page publishes the dataset and describes remaining weaknesses, including precise alignment and sensor-calibration drift. Those limits matter when comparing laboratory success rates with dependable factory operation. The work is due to be presented at the Conference on Robot Learning in Austin on November 9–12.

The collaboration includes UC Berkeley, NVIDIA, Stanford, Sapienza University of Rome and ItalAI. It complements the wider push toward touch-aware robots, but Panasonic describes manufacturing and logistics applications as future aims rather than announcing a commercial deployment.

Disclaimer: RobotsBeat is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, Robotics, technology, software, and digital innovation sectors. These relationships do not influence RobotsBeat's editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.

News, Science & Tech