为降低任务规划的复杂度并提高机器人对动态环境的适应能力,提出一种基于人机协同的任务规划方法,将任务规划全过程划分为任务级和指令级.在任务级规划中,用户意图输入、任务描述和人形机器人的任务解析与智能推理相结合,为指令级规划提供依据.在指令级规划中,基于任务的三元分解模型建立模块化的规划结构,以直观的图形化方式组合任务分解出的三元素,构成各种不同的任务,实现对任意任务的实时规划,同时机器人系统通过指令解析和预判辅助规划过程,形成既能满足人的任务需求又符合规划合理性的指令序列.最后用实例验证所提方法的有效性与可行性. In order to reduce the complexity of task planning and improve the adaptability of robot to the dynamic environment, a task planning method based on human-machine collaboration is proposed, which divides the whole task planning into task-level and instruction-level.In the task-level planning, Input, task description and mission analysis of humanoid robot combined with intelligent reasoning to provide the basis for instruction level planning.In instruction level planning, task-based ternary decomposition model to establish a modular planning structure, intuitive graphical combination The three elements of the task decomposition form various tasks to realize the real-time planning of any task. At the same time, the robot system resolves and assists the planning process through the instructions to form the instruction that meets the needs of people and meets the requirements of planning rationality Finally, an example is used to verify the validity and feasibility of the proposed method.