Since the pneumatic artificial muscle is usually employed to actuate the robot, it is inevitable that it will collide with obstacles and/or other objects such that there will be impacts. The typical impacts include the axial and radial ones. This paper detects these two kinds of impact by a designed differential pressure measurement system, and achieves the goal of identifying them by the autocorrelations of their differential pressure signals in both static and dynamic processes, which are the two main work processes of the pneumatic artificial muscle. In detail, first we design an experiment scheme by connecting the pneumatic artificial muscle to the differential pressure sensor through two pressure tubes; then the axial and radial impact signals are acquired via the impact experiments under different work conditions (impact strength, internal pressure, load, and impact position); finally, the identification of the axial and radial impact signals are achieved based on the autocorrelation technique with a well-selected threshold. The experiments illustrate the effectiveness of the whole method.

中文翻译：

---

基于差压信号自相关的气动人工肌肉在静态和动态过程中的轴向和径向影响识别

由于气动人造肌肉通常用于驱动机器人，因此不可避免地会与障碍物和/或其他物体发生碰撞，从而产生冲击。典型的影响包括轴向和径向影响。本文通过设计的差压测量系统检测这两种冲击，并通过气动人工肌肉的两个主要工作过程静态和动态过程中的差压信号的自相关来达到识别它们的目的。 . 具体来说，首先我们设计了一个实验方案，将气动人造肌肉通过两个压力管连接到差压传感器；然后通过不同工况（冲击强度、内部压力、载荷和冲击位置）；最后，基于自相关技术，选择合适的阈值，实现了轴向和径向冲击信号的识别。实验说明了整个方法的有效性。