A Review on the Development and Key Technologies of Embodied Intelligent Industrial Robots in Smart Manufacturing

ABSTRACT

Industrial robots are widely deployed in manufacturing systems to execute repetitive and predefined tasks. However, with the rise of Industry 4.0 and the growing demand for flexible, customized, and human‐centered production, conventional industrial robotic systems increasingly face limitations in adaptability, autonomy, and interaction capability. In this context, embodied intelligent industrial robots (EIIRs) are emerging as a promising paradigm for next‐generation smart manufacturing. Unlike conventional advanced industrial robots, EIIRs are characterized not merely by stronger sensing or more powerful models, but by a closed‐loop intelligence paradigm grounded in body‐environment interaction, feedback‐driven adaptation, and task‐oriented action execution. This paper presents a comprehensive review of the development, defining characteristics, key technologies, representative applications, and future trends of EIIRs in smart manufacturing. First, the conceptual foundations and defining characteristics of EIIRs are clarified, with particular emphasis on physical embodiment, perception‐action coupling, feedback‐driven closed‐loop adaptation, and grounded mapping from semantic goals to physical actions. Second, the key technological architecture of EIIRs is analyzed from the perspectives of embodied perception, embodied decision‐making and task planning, and embodied control and execution, together with a comparative discussion of traditional hierarchical architectures and embodied closed‐loop architectures. Third, the practical value of EIIRs is examined through representative industrial scenarios, including robotic welding, robotic grinding, precision assembly, human–robot collaboration, and multi‐robot cooperation. Finally, the major technical challenges, future evolution path, and critical turning points of EIIRs are discussed, including industrial world models, long‐horizon task execution, sim‐to‐real transfer, real‐time deployment, and safety assurance. This review aims to provide a structured reference for understanding how embodied intelligence can drive the transition of industrial robotics from predefined automation toward grounded, adaptive, and context‐aware task execution in smart manufacturing.