CRII: CHS: Human in the Loop: Safety for Semi-autonomous Driving Systems in Emergency Situations

人-网物理系统(CPS)是智能机器和人工操作员的功能很好地集成在一起的系统。从智能制造系统到智能家居再到汽车，这类系统已经变得越来越普遍。现有的智能技术旨在减少人类工作量，提高系统安全性，但它们也带来了新的安全问题。这些担忧在紧急情况下尤其重要，在这种情况下，操作员可能需要在时间压力和高度压力下做出决定。在这些情况下做出的决定往往是最重要的决定，可能会导致严重的安全问题。本次CRII项目以半自动驾驶系统为具体领域，研究了紧急情况下人-CPS系统安全性的影响因素。项目成果将提高我们对半自动驾驶安全关键因素的理解，为未来其他人-CPS系统的设计提供参考，并最终导致一个更高效、更安全的世界。具体地说，在半自动系统中加入人类驾驶员使系统的控制任务变得更加困难，因为控制器必须考虑到来自人类驾驶员的意外输入。大多数现有的关于这类系统安全性的工作都假设了一个先验模型，预计这些模型将完全描述人类的行为。然而，当操作员必须在紧急情况下做出决定时，例如车祸或故障，这样的假设可能会变得无效。本项目的目标是系统地研究人-CPS系统作为一个整体对这种情况的反应，并考察人-CPS系统中驾驶员、车辆子系统以及相互作用的影响因素。拟议的紧急情况下系统安全评估措施包括在突发危险事件中的人的表现，以及在网络攻击导致的车辆故障期间的人的决策。

Human-cyber physical systems (CPS) are systems in which the functions of intelligent machines and human operators are well integrated. Such systems have become increasingly pervasive, from intelligent manufacturing systems to smart homes to automobiles. The existing intelligent technologies aim at reducing human workload and increasing system safety, yet they also bring new safety concerns. These concerns are particularly essential in emergency situations, in which the human operator may be required to make decisions under time pressure and high-level stress. The decisions made in these situations are often the most important ones that can cause severe safety issues. This CRII project investigates influential factors for human-CPS system safety in emergency situations, using the semi-autonomous driving system as a concrete domain. The project outcome will improve our understanding of the safety-critical factors for semi-autonomous driving, inform the future design of other human-CPS systems, and ultimately lead to a more efficient and safe world.Specifically, the inclusion of human drivers in the semi-autonomous systems makes the control tasks of the system more difficult because the controller must allow for unanticipated input from the human driver. Most existing work on the safety of such systems assumes a priori models that are expected to fully characterize human behaviors. However, such assumptions may become invalid when the human operator has to make a decision in emergency situations, such as car crashes or malfunctions. The goal of this project is to investigate systematically how the human-CPS system as a whole reacts to such situations and examine influential factors of the human driver, the vehicle sub-system, and the interaction. The proposed assessment measures of system safety in emergency situations include human performance in response to unexpected hazardous events, and human decision making during vehicle malfunctions caused by cyber attacks.