Real-time robot actuation is one of the main challenges to overcome in human-robot interaction. Most visual sensors are either too slow or their data are too complex to provide meaningful information and low latency input to a robotic system. Data output of an event camera is high-frequency and extremely lightweight, with only 8 bytes per event. To evaluate the hypothesis of using event cameras as data source for a real-time robotic system, the position of a waving hand is acquired from the event data and transmitted to a collaborative robot as a movement command. A total time delay of 110 ms was measured between the original movement and the robot movement, where much of the delay is caused by the robot dynamics.

翻译：实时机器人驱动是人机交互中需要克服的主要挑战之一。大多数视觉传感器要么速度过慢，要么其数据过于复杂，难以向机器人系统提供有意义的信息和低延迟输入。事件相机的数据输出频率高且极为轻量，每个事件仅需8字节。为验证将事件相机作为实时机器人系统数据源的假设，本研究从事件数据中获取挥手动作的位置信息，并将其作为运动指令传输至协作机器人。测得原始运动与机器人运动之间的总延迟为110毫秒，其中大部分延迟由机器人动力学特性造成。