Tools and Methods for Current and Future Controllability Assessment

Driver related evaluation of Advanced Driver Assistance Systems (ADAS) needs to address controllability, effectiveness and user acceptance, which are to some extend interfering with each other. The state of the art in the controllability assessment is currently defined by the Code of Practice of the RESPONSE 3 project which focuses on the driver-system-interaction with single assistance functions like ACC or LKA. However, the controllability evaluation of new assistance functionalities such as ADAS of automation level 2 or automated driving on level 3 (according to SAE definitions) requires a review of the existing methods and tools with regard to necessary adaptations and new developments. For controllability evaluation of future ADAS and systems of higher automation levels the existing methodology needs to be adapted. Aspects to be considered in this context are the increasing amount of information with regards to the automation level. This information needs to be perceived and processed by the driver when interacting with multiple parallel operating assistance functions and complex information and communication systems. The controllability of urban assistance functions and their failures is subject of discussion especially focusing on tools and methods for an urban controllability assessment. To that end, driving simulator experiments, vehicle-in-the-loop and real vehicle studies are conducted analyzing existing controllability methods on their suitability for urban assistance functions. The results show the specific advantages of each applied testing tools and suggest that an overall system evaluation addressing controllability, effectiveness and acceptance combines the advantages of the different testing environments. Next to acceptance and effectiveness the controllability analysis is embedded in the overall evaluation process with focus on the driver and the interaction with the vehicle. The controllability analysis process for higher levels of automation is described. An overview of state of the art controllability evaluation is provided. The problem for future systems is analyzed and possible methods and tools are proposed. The necessary methods and tools are described focusing on next generation ADAS and higher levels of automated driving. The results are limited to the driver interaction with assisted driving. For the assessment of the driver reaction to higher automation levels the use of a high-fidelity driving simulator seems reasonable to achieve a high reproducibility of the driving scenario and a good representation of the driving dynamics.