摘要
针对移动作业机器人自然语言交互性不足、多模态自主感知能力弱以及自主行为控制复杂等问题,设计了一套基于视觉语言大模型的腿臂机器人自主行为控制实验系统。首先明确了系统的硬件组成,搭建了腿臂机器人实验平台;然后通过自然语言识别与语义解析获取移动作业任务,再利用视觉语言大模型的场景感知与逻辑推理,提出复杂任务分解下智能行为调度策略;同时依托轻量化网络通信实现运动指令下发与机器人状态播报反馈。系列实验验证了该实验系统在综合任务中的自主性与智能性。
To address the challenges of inadequate natural language interaction,weak multi-modal autonomous perception,and complex behavior control in mobile operation robots,we designed an intelligent control system for leg-arm robot based on a large-scale vision-language model.The system clearly delineates the software and hardware configuration.By employing natural language recognition and semantic parsing,the mobile operation tasks can recognize.Based on the ability of scene perception and logical reasoning of large-scale visual models,an intelligent behavior scheduling mechanism for the decomposition of complex tasks is developed.Based on LCM communication,the issuance of motion command and the broadcast feedback of robot status information are realized.A series of experiments with leg-arm robot validated the autonomy and intelligence of leg-arm robot in comprehensive task execution.
作者
陈腾
肖仕钧
荣学文
李贻斌
荣海林
CHEN Teng;XIAO Shijun;RONG Xuewen;LI Yibin;RONG Hailin(School of Control Science and Engineering,Shandong University,Jinan 250061,China)
出处
《实验室研究与探索》
北大核心
2025年第8期67-71,93,共6页
Research and Exploration In Laboratory
基金
山东省自然科学基金项目(ZR2024MF005)
山东大学实验室建设与管理研究项目(sy20243302)
山东大学本科教学改革研究项目(2024Y183)
山东省高等学校青创科技支持计划项目(2023KJ029)。
关键词
腿臂机器人
视觉语言大模型
环境感知
自主行为控制
leg-arm robot
vision-language model
environmental perception
autonomous behavior control
