CLEA: Closed-Loop Embodied Agent for Enhancing Task Execution in Dynamic Environments
Mingcong Lei, Ge Wang, Yiming Zhao, Zhixin Mai, Qing Zhao, Yao Guo, Zhen Li, Shuguang Cui, Yatong Han, Jinke Ren
2025-03-04
Summary
This paper talks about CLEA (Closed-Loop Embodied Agent), a new AI system that helps robots complete complex tasks more effectively in changing environments by using multiple language models to plan, execute, and adjust actions as needed.
What's the problem?
While large language models are good at breaking down complex tasks into smaller steps, they struggle when it comes to actually carrying out these steps in the real world, especially when things change unexpectedly. This makes it hard for robots to complete long, complicated tasks successfully on the first try.
What's the solution?
The researchers created CLEA, which uses four different AI language models to handle different parts of a task. CLEA can create a plan of smaller tasks based on what it remembers about its surroundings, and it can also check if its actions are likely to work. If something in the environment changes too much, CLEA can remake its plan. They tested CLEA using two different types of robots to find and move objects in a real-world setting.
Why it matters?
This matters because it could make robots much better at handling complex tasks in changing environments, like in factories or homes. In their tests, CLEA was 67.3% more successful at completing tasks than other methods, which could lead to more reliable and adaptable robots for various real-world applications.
Abstract
Large Language Models (LLMs) exhibit remarkable capabilities in the hierarchical decomposition of complex tasks through semantic reasoning. However, their application in embodied systems faces challenges in ensuring reliable execution of subtask sequences and achieving one-shot success in long-term task completion. To address these limitations in dynamic environments, we propose Closed-Loop Embodied Agent (CLEA) -- a novel architecture incorporating four specialized open-source LLMs with functional decoupling for closed-loop task management. The framework features two core innovations: (1) Interactive task planner that dynamically generates executable subtasks based on the environmental memory, and (2) Multimodal execution critic employing an evaluation framework to conduct a probabilistic assessment of action feasibility, triggering hierarchical re-planning mechanisms when environmental perturbations exceed preset thresholds. To validate CLEA's effectiveness, we conduct experiments in a real environment with manipulable objects, using two heterogeneous robots for object search, manipulation, and search-manipulation integration tasks. Across 12 task trials, CLEA outperforms the baseline model, achieving a 67.3% improvement in success rate and a 52.8% increase in task completion rate. These results demonstrate that CLEA significantly enhances the robustness of task planning and execution in dynamic environments.