Licentiatavhandling (sammanfattning) Göteborg : Chalmers tekniska högskola, 2020

Automotive Original Equipment Manufacturers (OEMs) compete worldwide to stand out with new trends and technologies. Automated Driver Assistance Systems (ADAS) are an example of advanced solutions where a lot of effort is put into the development and utilization of vehicle data. ADAS systems range from different types of information/warning systems to adaptive functions designed to assist the driver in the driving tasks and ensure more efficient and comfortable driving. These types of systems have become standard at many OEMs, including Tesla, Cadillac, BMW, Mercedes, Volvo Cars, and others. Volvo Cars is well-known for the development of such ADAS functions as ACC (Adaptive Cruise Control) and PA (Pilot Assist). These functions offer lateral and/or longitudinal support, but leave the driver in full control and with responsibility for the driving task.

The ADAS systems are not fully automated. These systems have a number of limitations related to the context where they can operate. Previous studies have demonstrated that the drivers’ understanding and adoption of these systems is not definite and may vary from full technology acceptance to complete ignorance. Therefore, in-depth understanding and interpretation of driver behavior and needs regarding the use of ADAS can significantly help developers to reflect on and improve the systems to meet the users’ expectations.

Recently, the availability of data coming from the in-vehicle sensors network has increased significantly. The amount of received data potentially enables the in-depth quantitative driver behavior evaluation in a time-efficient and reliable way. Moreover, the ability of vehicle sensors and actuator data to synchronize the driver and system performance and assess the driving conditions in the moment of driver-system interaction can contribute to the comprehensive context-aware ADAS evaluation.

Developing methods for objective assessment of driver behavior is a task with a high level of complexity. This process requires (i) investigation of the driver behavior assessment area where vehicle data can be useful; (ii) identification of the influencing factors for evaluating ADAS functions; (iii) definition of the relevant data for the data-driven driver behavior evaluation; (iv) investigation of the ways to improve the feasibility of vehicle data.