The paper introduces a specification‑aware action distribution optimization framework that enforces Signal Temporal Logic (STL) constraints during the execution of pretrained robotics foundation models without altering their parameters. At each decision step, the method minimally adjusts the action distribution to satisfy hard STL feasibility constraints by reasoning over the remaining horizon with forward dynamics propagation. The approach is validated in simulation using state‑of‑the‑art robotics foundation models across multiple environments and complex specifications.
arXiv:2608. 13678v1 Announce Type: cross Abstract: A central goal of robot learning is to enable robots to execute rich instructions specified at runtime.
arXiv:2502. 19135v2 Announce Type: replace Abstract: We present PLANTOR, a framework for generating and executing multi-robot task plans from natural-language task descriptions through LLM-assisted knowledge-base construction.
arXiv:2606. 23754v1 Announce Type: cross Abstract: Deploying foundation models for robot control raises a central challenge: the expressive power that enables rich, multimodal perception also makes these models opaque and difficult to analyze formally, rendering them intractable for existing verification tools.
arXiv:2606. 00104v1 Announce Type: cross Abstract: Foundation models are increasingly used to drive autonomous systems, yet existing approaches either keep the model in a tight control loop, raising latency and hallucination risk, or compile natural language into opaque end-to-end policies that are hard to explain, constraint and require domain-specific datasets and fine-tuning.
arXiv:2406. 03367v2 Announce Type: replace Abstract: Large Language Models (LLMs) possess extensive foundational knowledge and moderate reasoning abilities, making them suitable for general task planning in open-world scenarios.
The paper proposes a unified, architecture‑agnostic probabilistic framework for embodied foundation models using stochastic kernels. It defines these models through conditional output distributions and action‑realization kernels that together produce task‑conditioned robot policies. The framework links pretraining and post‑training phases to policy deployment, and is instantiated for OpenVLA and DreamZero, covering both autoregressive action‑token prediction and joint video‑action generation.
arXiv:2606. 18519v1 Announce Type: cross Abstract: Though robotic systems are now being commercialized and deployed in various industries, many of these systems are highly specialized and often require an advanced skill set to operate and ensure they perform as instructed.
The paper introduces ExecVLA, a framework for vision‑language‑action models that separates goal‑oriented behavior from execution‑specific constraints using a bi‑level action representation and supervised reasoning tokens. It adds explicit objectives to keep goal representations stable while preserving execution constraints, and builds datasets with fine‑grained reasoning annotations in simulation and on a Realman‑75 robot. Experiments on LIBERO and the real robot demonstrate better goal completion and notably more reliable adherence to instruction‑specified execution constraints.
The paper introduces a hierarchical feedback framework that guides large language models (LLMs) to produce executable plans for cyber‑physical systems (CPSs) from natural language commands. It prompts LLMs to generate both a structured task plan and signal temporal logic (STL) constraints, then uses a sound algorithm to detect contradictions and provide feedback for plan revision. At execution time, diagnostic information from failed control synthesis—such as violated constraints and conflicting component assignments—is fed back to the LLM to enable task replanning, achieving an 87.9% success rate on seven heterogeneous CPS scenarios.
The paper introduces iAm.md, a Markdown-based framework that enables robot skill self-assessment through agentic introspection. By mapping open‑vocabulary semantic detections and object segmentations to persistent records, the system anchors autonomous behavior generation in deployment evidence, mitigating grounding failures. Experiments on a simulated TIAGo robot demonstrate that this representation supports both skill self‑assessment and executable task generalization for navigation‑and‑manipulation tasks.
arXiv:2609.37519v1 Announce Type: cross Abstract: Video-based policy learning is particularly promising, as it illustrates target behaviors without requiring action annotations or embodiment-matched...
The paper introduces an oracle-guided differentiable safeguard for flow-matching robot policies that ensures safety constraints are met without explicit constraint models. By using binary feasibility queries during both forward and backward passes, the method corrects unsafe updates and incorporates local boundary information into learning signals. Experiments on obstacle-aware robotic manipulation show that the approach achieves 100% constraint satisfaction while maintaining task performance.