具身智能工业机器人:体系架构、关键技术与案例研究

Embodied artificial intelligence refers to the integration of artificial intelligence into physical systems,endowing them with the ability to perceive,learn from,and dynamically interact with their environment,so as to achieve their autonomous learning and evolution.To break through the limitations of traditional industrial robots' programmed operations and scene dependence,and promote the upgrade of industrial robots' intelligence,the integration of embodied artificial intelligence and industrial robots was deeply explored,and a new Industrial Robot paradigm—Embodied Artificial Intelligent Industrial Robot(EAI-IR)was discussed.The concept,composition and typical characteristics of EAI-IRs were expounded,and system architecture of EAI-IRs was proposed.Key technologies of EAI-IR were systematically summarized from the perspective of "perception-decision-action-feedback" closed loop.Then,the feasibility and effectiveness of the proposed architecture and technical methods were verified through an assembly case of a planetary reducer.Finally,a summarization of the work was given,and future development trends of EAI-IR were discussed.