The Influence of Study Design on Results in Hmi Testing for Active Safety

Active safety systems show great potential in preventing a large number of accidents. However, unless the system is completely autonomous, its actual effect will depend on how well it interacts with the driver. Therefore, Human-Machine-Interface (HMI) testing for active safety systems has become central in their development. For reasons of reproducibility and safety, HMI testing is usually carried out in a driving simulator or test track environment. These environments are different from real life driving. Unless the study design accurately reflects the conditions under which the system will be used, results will have low validity. Hence, study design becomes very important. The influence of study design was shown in two HMI-studies of Forward Collision Warning (FCW) modalities carried out by Volvo Cars and Ford Motor Company in VIRTTEX, Ford's motion-based driving simulator. In each study subjects were exposed to a surprise FCW event, with most subjects receiving a FCW. Results show that distracted drivers’ reactions to the warning correlated to their degree of previous exposure to warnings as well as the type of warning. Drivers who had received other warnings in the vehicle prior to the surprise FCW event responded as intended to all warning types. Drivers who neither trained with nor were informed about any vehicle warnings prior to the surprise FCW event responded partially as intended to the warnings, with an interesting exception for verbal warnings. The results show that to achieve high validity in HMI evaluations, the study design can benefit from exposing drivers to warnings in a way that reflects their normal awareness of warnings in real life driving. It also suggests that developers could tailor HMI design to frequency of use, as well as benefit from keeping drivers adequately aware of the warning types a vehicle can provide.