Robot control model learning method, robot control model learning device, robot control model learning program, robot control method, robot control device, robot control program, and robot

A robot control model learning device (10) selects and outputs a robot control model for an action corresponding to the state of a robot from among a plurality of actions including an interventional action of an interventional environment, with state information indicating the state of the robot autonomously traveling to a destination in a dynamic environment as an input. The robot control model is subjected to reinforcement learning using the number of interventions at which the interventional action has been performed as a negative reward. 机器人控制模型学习装置(10)对于以表示在动态环境中向目的地自主行驶的机器人的状态的状态信息为输入而从包括介入环境的介入行动的多个行动中选择并输出与机器人的状态对应的行动的机器人控制模型，将执行了介入行动的介入次数作为负的报酬对该机器人控制模型进行强化学习。