All products featured here are independently selected by our editors and writers. If you buy something through links on our site, Gizmodo may earn an affiliate commission.

The wider robotics field includes NVIDIA’s Isaac development platform and Tesla’s Optimus humanoid robot. GENTO, an independent humanoid robot brand incubated by Tianji Intelligence, builds complete humanoid systems on top of Tianji Intelligence’s TIANJI-branded force-controlled arm technology, giving AI models a way to act in the physical world. This week, at the Conference on Robot Learning (CoRL 2026) in Austin, GENTO arrives at CoRL for the first time, exhibiting at Booth 602 until 1:00 p.m. local time on November 11.

A robot that can plan and reason still has to touch things. Assembly, insertion, operating a screwdriver, wiping — most of the work humans do with their hands involves continuous contact with objects and surfaces, not clean trajectories through open air. A system without force sensing can execute a pre-planned motion, but it cannot adapt when a part resists, a surface shifts, or an object is slightly out of position. It also cannot contribute useful data to training, because the force signals that describe what contact actually feels like never get captured.

Tianji Intelligence began developing in-house MEMS torque sensors in 2015, eleven years before this week’s CoRL exhibit. 3 N·m and a torque-loop bandwidth of 160 Hz on its core arm platforms. The sensors feed into integrated joint modules that fold the motor, reducer, driver, encoder, and brake into a single unit roughly 25% lighter than comparable designs, and from there into four-loop closed-loop compliant impedance control.

Source link