Skip to content

IROS 2026 FILIC

Jump to bottom
hwoo.han edited this page Sep 7, 2026 · 2 revisions

IROS 2026 β€” FILIC: Dual-Loop Force-Guided Imitation Learning with Impedance Torque Control

Venue: IROS 2026 (Pittsburgh) Β· paper #1885 Β· Tsinghua University Β· HKUST(GZ) Β· DISCOVER Robotics (Ge, Jia, Li, … Zhou). Paper: arXiv 2509.17053 Β· code. Representative of: force-aware learning from demonstration β€” a dual-loop framework that makes a position-centric IL policy force-informed and force-executed, even on arms without F/T sensors. Companions: Tactile VLA Β· Real-Time Execution Β· Dexterous Manipulation Β· IROS 2026 survey.

FILIC's dual-loop structure β€” outer loop: two-view ResNet features + an external force estimator (from joint torque Ο„) feed, via cross-attention, a Transformer encoder-decoder that outputs a 25 Hz action pose sequence; inner loop: an impedance controller (2 kHz, virtual spring-damper) + gravity compensation (250 Hz) execute compliantly through IK (architecture figure from Ge et al., arXiv 2509.17053, Β© the authors)

1. Problem

Many contact-rich tasks need precise force regulation, but most imitation-learning policies are position-centric and force-unaware, and adding F/T sensors to collaborative arms is costly + extra hardware.

2. Method

FILIC = Force-guided Imitation Learning + Impedance torque Control, in a dual-loop structure:

  • A Transformer-based IL policy paired with an impedance controller β†’ compliant, force-informed, force-executed manipulation.
  • For arms without F/T sensors: a cost-effective end-effector force estimator from joint-torque measurements via analytical Jacobian inversion, compensated with model-predicted torques from a digital twin.

3. Results

  • FILIC significantly outperforms vision-only and joint-torque-based methods, achieving safer, more compliant, adaptable contact-rich manipulation. Code released.

4. Why it matters

FILIC is IROS 2026's force-from-demonstration representative and a pragmatic answer to a recurring gap (data-pyramid Β§4: force/tactile is the signal lost first): it makes IL force-aware without new sensors by estimating EE force from joint torque + a digital-twin correction. The dual-loop (learned intent + impedance execution) is the transferable idea β€” the control-theoretic counterpart to the multi-rate tactile loops in MrGrasp/T-Rex: instead of adding a tactile sensor stream, it recovers force analytically and executes compliantly. This lowers the hardware bar for contact-rich LfD.

Limitations (reviewer): joint-torque force estimation is coarser than a real F/T sensor (bounded by digital-twin fidelity); impedance control assumes a compliant/torque-controllable arm; contact-rich-scoped.

5. Links

← Back to IROS 2026 survey Β· Home

Navigation

πŸ“– Reviews

🏷 Model lineages

🧠 ML foundations

πŸ—“ Conferences

(each page indexes its per-paper pages)

πŸ“Œ Foundational

Clone this wiki locally