Specification-Aware Distribution Shaping for Robotics Foundation Models

2026-03-18

AI Summary

Existing robotic foundation models lack formal guarantees for temporal specifications—such as time-bounded goal reaching, task sequencing, and persistent safety conditions—when executing natural language instructions. This work proposes a specification-aware action distribution optimization framework that, without fine-tuning the pretrained model parameters, enforces hard constraints expressed in Signal Temporal Logic (STL) during inference. At each decision step, the method computes a minimally corrected action distribution by propagating forward dynamics to ensure compliance with STL specifications. To the best of our knowledge, this is the first approach to embed STL hard constraints directly into the inference process of foundation models, enabling zero-training enforcement of complex spatiotemporal requirements. Experiments across multiple simulated environments demonstrate that the framework successfully satisfies diverse STL specifications while preserving the original model’s task performance.

Technology Category

• Natural Language Processing: Safety and Robustness
• Planning, Routing, and Scheduling: Planning with Language Models
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Robotics foundation models have demonstrated strong capabilities in executing natural language instructions across diverse tasks and environments. However, they remain largely data-driven and lack formal guarantees on safety and satisfaction of time-dependent specifications during deployment. In practice, robots often need to comply with operational constraints involving rich spatio-temporal requirements such as time-bounded goal visits, sequential objectives, and persistent safety conditions. In this work, we propose a specification-aware action distribution optimization framework that enforces a broad class of Signal Temporal Logic (STL) constraints during execution of a pretrained robotics foundation model without modifying its parameters. At each decision step, the method computes a minimally modified action distribution that satisfies a hard STL feasibility constraint by reasoning over the remaining horizon using forward dynamics propagation. We validate the proposed framework in simulation using a state-of-the-art robotics foundation model across multiple environments and complex specifications.

Problem

• Robotics foundation models
• Temporal specifications
• Safety guarantees
• Signal Temporal Logic
• Spatio-temporal constraints

Innovation

• Specification-Aware
• Distribution Shaping
• Signal Temporal Logic
• Robotics Foundation Models
• Constraint Enforcement