Paper Overview
Field: Computer Vision (CV) Authors: Sassan Mokhtar, Lars Doorenbos, Marius Bock, Sina Mokhtar, Katharina Beckh, Dominik Bach, Juergen Gall Published: 2026-11-10 arXiv: 2611.14470
Abstract
Preventing human error in physical assembly tasks requires assistance systems that can intervene before mistakes are made. This is especially challenging in settings where users are free to choose the construction order and correctness depends on the evolving assembly rather than a fixed next step. We present TanCRAFT, a four-stage real-time predictive assistance system that tracks the scene state, reasons about the current partial assembly, forecasts the near-future motion of manipulated pieces, and intervenes when the predicted placement is incompatible with the target. We instantiate our framework and confirm its effectiveness in a controlled tangram assembly task through offline benchmarks and a within-subject user study. Our findings show that the predictive feedback significantly reduces errors compared with unguided and reactive sessions, while achieving the lowest average perceived workload and the shortest average completion time. Overall, these results demonstrate the value of anticipatory feedback for error prevention and mark a step toward cost-aware assistance in automated and human-robot shared workcells.
Key Points
- Problem: Assist users in free-order assembly tasks, where correctness depends on the evolving partial assembly rather than a fixed next step.
- Approach: Four-stage pipeline — scene state tracking, reasoning over the partial assembly, near-future motion forecasting of manipulated pieces, and intervention on predicted incompatible placements.
- Evaluation: Controlled tangram assembly task, offline benchmarks plus a within-subject user study.
- Findings: Predictive feedback significantly reduces errors versus unguided and reactive sessions, with the lowest perceived workload and fastest completion times.
*Auto-collected on 2026-11-12*