【摘要】 为了拓宽钢筋绑扎机器人的研究思路并促进其发展应用，对国内外钢筋绑扎机器人研究现状进行了综述。分析了钢筋网片的绑扎工序并总结得出钢筋绑扎机器人的设计要求，然后详细介绍了钢筋绑扎机器人涉及的绑扎点识别与定位、机器人定位与导航、移动机械臂协调控制等关键技术，最后从多模态传感数据感知融合、自适应智能规划控制、多机协同作业3个方面提出了钢筋绑扎机器人今后的发展趋势。结果表明：通过机器人技术可以提高钢筋绑扎的质量和效率，实现该工序的自动化和智能化，从而促进智能建造的发展；在绑扎点识别与定位方面，主流方法是基于关键点检测的绑扎点识别技术结合基于深度相机或双目相机的绑扎点定位技术；在机器人定位方面，单一传感器的定位技术已无法满足钢筋绑扎机器人在复杂场景的定位精度和稳定性需求，多传感器数据融合定位技术成为研究的热点方向；在机器人导航与控制方面，随着现代智能算法和人工智能技术的发展，机器人路径规划、运动控制和移动机械臂协调控制等技术正朝着智能化的方向发展，然而目前仍存在训练数据需求较大、训练难度大以及将仿真环境里训练的模型部署到真实环境的匹配问题需要解决。更多还原

【Abstract】 In order to broaden the research scope of rebar binding robots and facilitate their development and application, a comprehensive review of the current state of research on such robots both domestically and internationally was conducted. The binding process of rebar mesh was analyzed, and the design requirements of rebar binding robots were summarized. Subsequently, key technologies involved in rebar binding robots, including binding point recognition and positioning, robot positioning and navigation, and coordinated control of mobile manipulators, were detailed. Finally, future trends in the development of rebar binding robots were proposed from three perspectives: multimodal sensor data perception fusion, adaptive intelligent planning control, and multi-machine collaborative operation. The results show that robotic technology can enhance the quality and efficiency of rebar binding, thereby achieving automation and intelligence in the process and promoting the development of intelligent construction. Presently, in terms of binding point recognition and positioning, the mainstream approach combines binding point recognition technology based on key point detection with binding point positioning technology based on depth cameras or binocular cameras. Regarding robot positioning, single-sensor positioning technology has become inadequate to meet the positioning accuracy and stability requirements of rebar binding robots in complex scenes, leading to a focus on multi-sensor data fusion positioning technology as a prominent research direction. In the domain of robot navigation and control, advancements in modern intelligent algorithms and artificial intelligence technologies are driving the development of technologies such as robot path planning, motion control, and coordinated control of mobile manipulators towards intelligence. However, challenges persist, including significant training data requirements, training complexity, and the deployment of models trained in simulated environments to real-world settings.更多还原