CAPER: Constrained and Procedural Reasoning for Robotic Scientific Experiments

TL;DR

CAPER is a framework that enhances robotic scientific experiments by explicitly enforcing procedural constraints and structured reasoning, leading to improved robustness, controllability, and data efficiency in complex manipulation tasks.

Contribution

It introduces a responsibility-separated planning and control pipeline that enforces procedural correctness and robustness in robotic scientific experiments.

Findings

Improves success rate and procedural correctness in experiments.

Enhances robustness and data efficiency in low-data regimes.

Demonstrates effectiveness on scientific workflow and manipulation benchmarks.

Abstract

Robotic assistance in scientific laboratories requires procedurally correct long-horizon manipulation, reliable execution under limited supervision, and robustness in low-demonstration regimes. Such conditions greatly challenge end-to-end vision-language-action (VLA) models, whose assumptions of recoverable errors and data-driven policy learning often break down in protocol-sensitive experiments. We propose CAPER, a framework for Constrained And ProcEdural Reasoning for robotic scientific experiments, which explicitly restricts where learning and reasoning occur in the planning and control pipeline. Rather than strengthening end-to-end policies, CAPER enforces a responsibility-separated structure: task-level reasoning generates procedurally valid action sequences under explicit constraints, mid-level multimodal grounding realizes subtasks without delegating spatial decision-making to large language models, and low-level control adapts to physical uncertainty via reinforcement learning with minimal demonstrations. By encoding procedural commitments through interpretable intermediate representations, CAPER prevents execution-time violations of experimental logic, improving controllability, robustness, and data efficiency. Experiments on a scientific workflow benchmark and a public long-horizon manipulation dataset demonstrate consistent improvements in success rate and procedural correctness, particularly in low-data and long-horizon settings.