Calibrating trust through knowledge: Introducing the concept of informed safety for automation in vehicles
Calibrating trust through knowledge: Introducing the concept of informed safety for automation in vehicles
复制标题
DOI：
10.1016/j.trc.2018.07.001
复制
发表时间：
2018-11-01
影响因子：
8.3
通讯作者：
Jennings, Paul
中科院分区：
文献类型：
--
作者：
Khastgir, Siddartha;Birrell, Stewart;Jennings, Paul
车辆自动化的发展越来越受到人们的关注，因为它具有许多潜在的好处，如安全、提高交通效率、减少排放等。影响公众接受自动车辆技术的关键因素之一是他们的信任程度。建立信任是一个动态的过程，需要将其调整到安全部署的正确水平，以确保适当使用这类系统。影响信任的因素之一是向司机提供的关于系统的真实能力和局限性的知识。在对56名参与者进行驾驶模拟器研究后，作者发现，随着对自动化系统真实能力和局限性的了解的引入，与没有提供关于该系统的知识时相比，对自动化系统的信任增加了。参与者体验了两种不同类型的自动化系统：低能力自动化系统和高能力自动化系统。有趣的是，随着知识的引入，低性能和高性能自动化系统的平均信任级别是相似的。在实验结果的基础上，作者引入了知情安全的概念，即告知驾驶员自动化系统的安全限度，使他们能够将他们对系统的信任校准到适当的水平。
There has been an increasing focus on the development of automation in vehicles due its many potential benefits like safety, improved traffic efficiency, reduced emissions etc. One of the key factors influencing public acceptance of automated vehicle technologies is their level of trust. Development of trust is a dynamic process and needs to be calibrated to the correct levels for safe deployment to ensure appropriate use of such systems. One of the factors influencing trust is the knowledge provided to the driver about the system's true capabilities and limitations. After a 56 participants driving simulator study, the authors found that with the introduction of knowledge about the true capabilities and limitations of the automated system, trust in the automated system increased as compared to when no knowledge was provided about the system. Participants experienced two different types of automated systems: low capability automated system and high capability automated system. Interestingly, with the introduction of knowledge, the average trust levels for both low and high capability automated systems were similar. Based on the experimental results, the authors introduce the concept of informed safety, i.e., informing the drivers about the safety limits of the automated system to enable them to calibrate their trust in the system to an appropriate level.