This report introduces our UniTeam agent - an improved baseline for the "HomeRobot: Open Vocabulary Mobile Manipulation" challenge. The challenge poses problems of navigation in unfamiliar environments, manipulation of novel objects, and recognition of open-vocabulary object classes. This challenge aims to facilitate cross-cutting research in embodied AI using recent advances in machine learning, computer vision, natural language, and robotics. In this work, we conducted an exhaustive evaluation of the provided baseline agent; identified deficiencies in perception, navigation, and manipulation skills; and improved the baseline agent's performance. Notably, enhancements were made in perception - minimizing misclassifications; navigation - preventing infinite loop commitments; picking - addressing failures due to changing object visibility; and placing - ensuring accurate positioning for successful object placement.

翻译：本报告介绍我们的UniTeam智能体——针对“HomeRobot：开放词汇移动操作”挑战的改进基线。该挑战提出了在未知环境中导航、操作新物体以及识别开放词汇物体类别的问题，旨在利用机器学习、计算机视觉、自然语言处理与机器人技术的最新进展，推动具身智能领域的交叉研究。在这项工作中，我们对提供的基线智能体进行了详尽评估，识别了其在感知、导航与操作技能上的缺陷，并改进了基线智能体的性能。具体改进包括：感知方面——减少误分类；导航方面——防止无限循环锁定；抓取方面——解决因物体可见性变化导致的失败；放置方面——确保准确定位以实现成功放置。