arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.00455cs.ROcs.SYeess.SY

基于测试驱动的代理框架用于可靠机器人控制器

Test-Driven Agentic Framework for Reliable Robot Controller

  • Department of Electrical and Computer Engineering(电气与计算机工程系)
  • University of California, Riverside(加州大学河滨分校)
  • Department of Mathematics(数学系)
  • Department of Computer Science(计算机科学系)
  • University of Colorado, Boulder(科罗拉多大学波德分校)

机构由 AI 辅助整理,请以论文原文为准。

Shivanshu Tripathi, Reza Akbarian Bafghi, Maziar Raissi

更新

中文总结 AI 辅助

本文提出一种基于测试驱动的代理框架,用于生成可靠的机器人导航控制器,通过迭代优化提升控制器的可靠性和鲁棒性。

中文摘要 AI 辅助

在本工作中,我们提出了一种基于测试驱动的代理框架,用于合成可部署的低级机器人控制器,用于导航任务。给定一个带有超声波传感器机器人图像的2D地图,或一个3D机器人仿真环境,我们的框架通过结构化测试套件的诊断反馈迭代优化生成的控制器代码,以实现任务成功。我们提出了一种双层修复策略来优化生成的代码,该策略在提示级别优化和直接代码编辑之间交替进行。我们在2D导航任务和Webots仿真器中的3D导航中评估了该方法。实验结果表明,基于测试的合成在单次控制器生成方面显著提高了控制器的可靠性和鲁棒性,尤其是在初始提示不充分的情况下。源代码和演示视频可在:https://shivanshutripath.github.io/robotic_controller.github.io 获取。

英文摘要

In this work, we present a test-driven, agentic framework for synthesizing a deployable low-level robot controller for navigation tasks. Given a 2D map with an image of an ultrasonic sensor-based robot, or a 3D robotic simulation environment, our framework iteratively refines the generated controller code using diagnostic feedback from structured test suites to achieve task success. We propose a dual-tier repair strategy to refine the generated code that alternates between prompt-level refinement and direct code editing. We evaluate the approach across 2D navigation tasks and 3D navigation in the Webots simulator. Experimental results show that test-driven synthesis substantially improves controller reliability and robustness over one-shot controller generation, especially when the initial prompt is underspecified. The source code and demonstration videos are available at: https://shivanshutripath.github.io/robotic_controller.github.io.

↑