Skip to content
latestai.eulatestai.eu
TS

Tactile Step

Research framework adding sole pressure-insole tactile sensing to humanoid locomotion policies, tested on Unitree G1.

UnverifiedRoboticsResearch

Updated EU status checked Last shown

Tactile Step at a glance

What is Tactile Step?
Tactile Step is a research. Research framework adding sole pressure-insole tactile sensing to humanoid locomotion policies, tested on Unitree G1.
Does Tactile Step work in the EU?
Unverified. No explicit evidence found. It is an academic research project with no public product, code or weights listed, so availability is not applicable. Last verified 14 Nov 2026.
How much does Tactile Step cost?
Pricing unknown. Research project; no pricing or release of code/weights stated on the project page.
Where was Tactile Step shown?
In 1 creator video by AI Search. Most recently by AI Search on 4 Oct 2026, at 27:07 in “Gemini 4, GPT 6.1, Plasma conjecture, Napoleon code solved, OpenAI Dots, Claude Sonnet 5.5: AI NEWS”.

Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

TactileStep is a research project (accepted at CoRL 2026) that brings sole pressure sensing into humanoid locomotion control. It targets a gap in humanoid parkour policies: terrain traversal can succeed while landings are harsh, contacts fall on foot edges, or stance is unstable. Humans regulate this through tactile feedback, which most humanoid robots lack.

The framework aligns a tactile simulation with a real pressure insole, so the policy learns from the same contact features available on hardware. During training, tactile and motion cues are used to identify foot-contact phases, and phase-aware rewards encourage safer landings and more stable stance.

Evaluated in simulation and on a real Unitree G1 humanoid across terrains (flat, slopes, stairs, platforms), it reports up to 48.8% lower peak touchdown force, up to 30.1 dB lower peak A-weighted impact noise versus a strong perceptive baseline, and up to 23.8% larger stance contact area. The project page provides videos, method and results; no code, weights or pricing were listed in the material provided. It is an academic work requiring specific insole hardware and a humanoid platform, not a packaged product.

Key features
  • Sole pressure-insole tactile sensing for humanoid locomotion policies
  • Tactile simulation aligned with the real insole for sim-to-real transfer
  • Phase-aware rewards based on recognized foot-contact phases
  • Up to 48.8% lower peak touchdown force and 23.8% larger stance contact area
  • Validated on a real Unitree G1 across stairs, slopes, platforms and flat ground
1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

Keep exploring

Similar AI tools

Closest neighbours by meaning, not just by category.