Towards a Holistic Safety Monitoring in Intelligent Vehicle Control
Tim Köhler, Martin Schröer
In Proceedings of the 10th International Conference on Informatics in Control, Automation and Robotics, (ICINCO-2013), 29.7.-31.7.2013, Reykjavik, SciTePress Digital Library, volume 1, pages 583-588, Jul/2013. ISBN: 978-989-8565-70-9.
Today, the state of the art in vehicle safety follows an explicit design flow. Specific sensors measure a particular dimension (e.g. distance to other vehicles) and “safety” is defined as a specific range of allowed values (e.g. minimal distance). The disadvantage of such an approach is that safety issues which were unconsidered at design time are not detectable. Furthermore, a detection of issues that are only indirectly measurable is difficult to realize. In this paper, a holistic safety monitoring approach is presented that makes use of all
available sensor data and tries to find an implicit definition of “safety”. By such an inverse approach vehicle safety issues which are hard to be directly measurable might be detectable, too. For instance, an identification of driver-initiated critical situations (e.g. caused by distraction) could be possible if taking multiple sensor modalities into account and having an implicitly defined “safe” state. Furthermore, the article describes the selection of potential test platforms and shows already collected test data of a mobile robot platform. Presented in this work-in-progress paper is the concept of definition, implementation, and detection of implicit vehicle safety.
Intelligent Vehicle Control, Car Safety, Electric Vehicle, Model-Based Prediction, Fault Detection, Human- Machine Interface