在许多工业应用场合，安置传感器测量系统的输出是不可能或非常昂贵的。此外，物体的运动仅能在一些离散点被检测到。在这些系统中，控制目标通常不是跟踪正确的轨迹，而是在传感器的位置获得给定的状态，在这篇论文中，引入RL学习策略来控制湿式离合器，以实现快速的接合。(In many industrial applications, it is impossible or very expensive to place sensor measurement system output. In addition, the motion of the object can be detected only at discrete points. In these systems, the control objective is usually not tracking the right track, but in the position of the sensor for a given state, in this paper, the introduction of RL learning strategies to control the wet clutch, in order to realize the rapid joint.)