RoboticsAug 17, 2026, 5:59 PM

Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memory

30-second summary

Researchers propose a method to improve long-horizon robot manipulation by using an LLM agent to plan and coordinate subtasks, reducing error compounding in multi-stage tasks.

TickrWire
Key takeaways
  • BATON uses an LLM agent to plan and coordinate subtasks in long-horizon robot manipulation, reducing error compounding.
  • The framework freezes the VLA model and invokes it only for contact-rich segments, improving task reliability.
  • Language-based memory allows the agent to write and adapt instructions for subtask transitions.
  • The method addresses silent constraints between subtasks, a common failure point in multi-stage robotic tasks.
Full story

A new research paper introduces a framework called BATON to address the challenge of long-horizon robot manipulation, where multiple contact-rich skills must be chained together into a single task. While vision-language-action (VLA) models have made progress in mastering individual skills, they often fail in multi-stage tasks due to error compounding and subtask dependencies. The proposed solution freezes the VLA model and places an LLM agent in charge of planning and coordination. This agent operates in free space using analytic primitives, invokes the VLA only for contact-rich segments, and writes adaptation instructions into a language-based memory system. The approach aims to mitigate the limitations of current methods by structuring tasks into manageable subtasks and leveraging the strengths of both LLM agents and VLA models.

The paper highlights two key challenges in long-horizon manipulation: ensuring competence across subtasks and preventing silent constraints between them. By breaking tasks into subtasks and using the LLM agent to manage transitions and adaptations, the method seeks to improve reliability and performance in complex robotic operations. The research demonstrates the potential of combining high-level planning with specialized low-level skills to achieve more robust long-horizon manipulation.

Sponsored
Why this matters
Developers

Provides a new approach to structuring long-horizon robot manipulation tasks, combining LLM planning with specialized VLA models.

Businesses

Could lead to more reliable and scalable robotic systems for industries like manufacturing, logistics, and healthcare.

Investors

Highlights emerging techniques in AI-driven robotics that may attract funding for further development and commercialization.

Everyone

Demonstrates progress in making robots more capable of handling complex, multi-stage tasks in real-world environments.

Glossary
VLA (Vision-Language-Action) models
AI models that combine visual perception, language understanding, and physical action to perform robotic tasks.
Long-horizon manipulation
Robotic tasks that require chaining multiple skills or actions over an extended period, often involving complex interactions.
Sources · 1
Read next
More stories
TickrWire
Security

AI vs AI: Can artificial intelligence contain the fake news epidemic that it has helped unleash? - Genetic Literacy Project

Researchers explore whether AI can detect and mitigate fake news, a problem partly fueled by AI itself.

TickrWire
Security

AI and the New Age of Bioweapons - Foreign Affairs

A Foreign Affairs analysis warns that AI could dramatically lower the barrier to creating bioweapons, accelerating proliferation risks.

TickrWire

Artificial Intelligence: Organizations Across the Americas Urge the IACHR to Address the Environmental and Social Impacts of Rapidly Expanding Data Centers - elciudadano.com

Organizations across the Americas have formally requested the Inter-American Commission on Human Rights (IACHR) to investigate the environmental and social consequences of rapidly expanding data centers, driven by artificial intelligence development.

TickrWire
Security

Suburban man allegedly used AI to create child sexual abuse material: Prosecutors - NBC 5 Chicago

A suburban man is accused of using AI to create child sexual abuse material, according to prosecutors.

Sponsored
TickrWire
Security

Appeals court flags AI-generated fake cases in San Antonio ISD lawsuit - KSAT

A federal appeals court in Texas flagged AI-generated fake cases in a lawsuit involving San Antonio ISD, raising concerns about the reliability of AI in legal filings.

Anthropic’s annualized revenue surges to $65BBusiness

Anthropic’s annualized revenue surges to $65B

Anthropic’s annualized revenue has skyrocketed to $65 billion, adding $18 billion in just two months.

TickrWireAI News Intelligence

We aggregate, verify, summarise and explain the latest artificial intelligence news from open, legal sources.

Daily AI digest

Top AI stories, summarised, in your inbox each morning.

© 2026 TickrWire. Summaries and analysis are AI-generated and may contain errors.