Abstract: The multi-robot cooperative transportation approach had the noteworthy task flexibility and condition adaptability for the manipulation of heavy-duty， large-sized， small-rigidity irregular-shape objects. Firstly， 3 guidance control configurations， mono-camera dual-unit， dual-camera dual-unit， and dual-camera multi-kinematic-pair omni-directional， were investigated for cooperative transportations. The multiple-input and multiple-output motion control models were established. Then， 2 coordinated path tracking control methods were proposed by means of leader-follower strategy which was composed of homogeneous framework based on deviation-conversion prediction control for two robots with fuzzy adjustment of control parameters， and heterogeneous framework based on model prediction control for leader robot and backstepping control for follower robot with reinforcement learning of control parameters. Finally， the multi-robot cooperative transportation prototype systems were developed for the guidance control experiments. The experimental results verify the feasibility and effectiveness of the configurations， models， frameworks， and methods. 

Key words: multi-robot system, cooperative motion, coordinated control, automatic control, path tracking

摘要： 针对大质量、大尺寸、低刚度的异形大部件，采用多机器人的合作搬运方式具有显著的作业灵活性和工况适应性。研究了单目-双驱动、双目-双驱动、双目-多运动副全向等合作搬运导引控制构型，建立了多输入-多输出的运动控制模型；采用领航者-跟随者策略设计了协同路径跟踪控制方法，该方法包括偏差转化预测控制+参数模糊调节的同构架构，以及模型预测控制+反演控制+参数强化学习的异构架构。最后，为进行导引控制实验开发了多机器人合作搬运原型系统，实验结果验证了所提构型、模型、架构与方法的可行性和有效性。

关键词: 多机器人系统, 协同运动, 联动控制, 自动导引, 路径跟踪