Beyond Implicit Force: Evaluating Explicit Force-Torque Proxies in Action Chunking with Transformers

2026-07-16

AI Summary

This work addresses the limited explicit contact awareness of existing vision- and kinematics-based policies in contact-intensive manipulation tasks. While Action Chunking with Transformers (ACT) leverages implicit force cues from teleoperation data, it struggles to generalize to real robotic hardware. To overcome this, we propose an enhanced ACT architecture that shifts its output from master-side commands to slave-side joint states and incorporates an explicit force-torque proxy signal derived from joint torques, replacing unreliable implicit cues. We present the first systematic validation of this torque-based proxy on a physical robot platform. Across four tasks—surface following, insertion, stiffness identification, and force-controlled stopping—the proposed method not only recovers performance lost by discarding implicit signals but also significantly outperforms the original ACT, demonstrating superior contact perception and operational robustness.

Abstract

Contact-rich manipulation requires policies to infer interaction state from signals that are often weakly observable through vision and kinematics alone. Action Chunking with Transformers (ACT) has shown strong performance in fine-grained manipulation, but many deployments collect demonstrations through leader-follower teleoperation, where tracking error between commanded leader motion and executed follower motion implicitly encodes contact, resistance, and constraint violation. This paper examines whether ACT's apparent force-awareness depends on this hidden interaction cue. We introduce an observation-centric ACT variant that predicts future follower joint states instead of leader commands, thereby removing the teleoperation-induced discrepancy signal while preserving the rest of the learning pipeline. We then evaluate whether simple joint-torque proxies, derived from onboard motor current or joint effort, can recover contact-aware behavior without external force/torque sensors. Across four real-world tasks spanning surface following, insertion, stiffness discrimination, and force-based stopping, removing the implicit cue leads to severe failures in force-critical phases. In contrast, torque-augmented policies recover robust contact behavior and improve the base ACT policy. These results demonstrate that, on real hardware, the implicit teleoperation cue is a recoverable source of force-awareness, where torque signals are available, a simple proxy matches, surpasses, or further enhances it.

Problem

• force-torque sensing
• action chunking
• teleoperation
• contact-rich manipulation
• transformer-based policy

Innovation

• Action Chunking with Transformers
• force-torque proxy
• teleoperation discrepancy
• contact-rich manipulation
• joint torque estimation