TRACT: Temporally Routed Action Chunks with Chronological Phase Authority for Contact-Rich Manipulation

2026-07-31

AI Summary

This work addresses temporal misalignment in action chunking caused by uniform labeling across task-phase boundaries, as well as response latency and execution stalls during contact-rich manipulation. To resolve these issues, the authors propose a temporally routed action chunking mechanism that decouples each action chunk into the current phase and a single future boundary. Task-local graphs enforce phase-wise temporal authority, while cumulative boundary distributions route future queries to their corresponding phase trajectories. Additionally, a causal response-deficit integrator compensates for suppressed motor responses. The approach preserves phase semantic consistency while substantially improving robustness and success rates in contact tasks, achieving 10/10 full-task success and a 99.00% median wiping completion rate across six real robots—without phase ambiguity or execution stalls—significantly outperforming baseline methods.

Abstract

Action chunking shortens the effective decision horizon of robot imitation learning by predicting multiple future actions, while conventional phase conditioning describes the current control instant. When a predicted horizon crosses a procedural boundary, assigning the current phase to the entire chunk creates a structural temporal mismatch. We present TRACT, which factorizes phase-structured action chunking into an accepted current phase and a single CURRENT-to-NEXT boundary inside the future horizon. A task-local graph constrains chronological phase authority, and a cumulative boundary distribution monotonically routes future queries through phase-specific query and action paths. For contact execution, a causal response-deficit integrator compares policy intent with ACK-eligible subsequent motion, accumulates arm compensation when directional response is suppressed, and decays after confirmed recovery. Across six real-robot variants with ten trials each, full TRACT achieves 10/10 full-sequence success, 99.00 [88.75, 100.00]% median [min, max] wipe completion, zero observed phase ambiguity, and zero stalls. Under the current complete method package and evaluation setting, the routed representation obtains better observed task results than the flat package (6/10 vs. 3/10 success; 77.08% vs. 8.03% median wipe completion). Chronological authority reduces observed phase ambiguity from 8/10 to 0/10, and response integration reduces stalls from 4/10 to 0/10. The package comparison does not isolate routing from other generator-package differences.

Problem

• action chunking
• phase conditioning
• temporal mismatch
• contact-rich manipulation
• procedural boundary

Innovation

• action chunking
• phase authority
• temporal routing
• contact-rich manipulation
• response-deficit integration