Robot agent skill learning and generalization method based on case reasoning

The invention relates to the field of robot agent skill learning, in particular to a case reasoning-based robot agent skill learning and generalization method, which comprises the following steps of: constructing a case library, acquiring a target case which is most similar to each state vector of a robot agent in a current task scene in the case library, and obtaining an initial action strategy in the current task scene, obtaining a target action strategy, and obtaining a final action strategy. According to the method, the problem of low interpretability of the action strategy caused by using the deep neural network to fit the action strategy is solved, and the task efficiency of the robot intelligent agent is improved. 本发明涉及机器人智能体技能学习领域，具体涉及一种基于案例推理的机器人智能体技能学习与泛化方法，包括：构建案例库，获取案例库中与当前任务场景下的机器人智能体的每个状态向量最相似的目标案例，获取当前任务场景下的初始动作策略，获取目标动作策略，获取最终动作策略。本方法克服了使用深度神经网络拟合动作策略带来的动作策略可解释性低的问题，提高了机器人智能体的任务效率。