SupportingTrust in Autonomous Driving

Proceedings of the 22nd International Conference on Intelligent User Interfaces Pub Date : 2017-03-07 DOI:10.1145/3025171.3025198

Renate Häuslschmid, Max von Bülow, Bastian Pfleging, A. Butz

引用次数: 96

Abstract

Autonomous cars will likely hit the market soon, but trust into such a technology is one of the big discussion points in the public debate. Drivers who have always been in complete control of their car are expected to willingly hand over control and blindly trust a technology that could kill them. We argue that trust in autonomous driving can be increased by means of a driver interface that visualizes the car's interpretation of the current situation and its corresponding actions. To verify this, we compared different visualizations in a user study, overlaid to a driving scene: (1) a chauffeur avatar, (2) a world in miniature, and (3) a display of the car's indicators as the baseline. The world in miniature visualization increased trust the most. The human-like chauffeur avatar can also increase trust, however, we did not find a significant difference between the chauffeur and the baseline.

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

支持对自动驾驶的信任

自动驾驶汽车可能很快就会进入市场，但对这种技术的信任是公众辩论的主要讨论点之一。那些一直完全控制着自己汽车的司机被期望心甘情愿地交出控制权，盲目信任一项可能会杀死他们的技术。我们认为，可以通过驾驶员界面来增加对自动驾驶的信任，该界面可以可视化汽车对当前情况的解释及其相应的行动。为了验证这一点，我们在用户研究中比较了不同的可视化效果，叠加到驾驶场景中:(1)司机的化身，(2)微型世界，(3)汽车指示器的显示作为基线。微缩可视化的世界最能增加信任。人形司机头像也可以增加信任，然而，我们没有发现司机与基线之间有显著差异。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

Proceedings of the 22nd International Conference on Intelligent User Interfaces

自引率

0.00%

发文量

期刊最新文献

I-SED: An Interactive Sound Event Detector Identifying Frequent User Tasks from Application Logs Label-and-Learn: Visualizing the Likelihood of Machine Learning Classifier's Success During Data Labeling Negative Relevance Feedback for Exploratory Search with Visual Interactive Intent Modeling Guidelines for Tree-based Collaborative Goal Setting