Assessment of the cooperation between driver and vehicle automation: A framework

Abstract

As long as the human driver is responsible for part(s) of the driving task during automated driving, the driver and automated driving system are sharing the driving task. Such a shared task is characterized by shared control, in which cooperation between the driver and vehicle automation is essential. However, means to holistically assess the quality of this cooperation are currently lacking. This work addresses how cooperation between driver and vehicle automation can be operationalized and assessed to gain insight into the quality of the shared driving task. Quality indicators and measurement methods are identified across seven dimensions reflective of the quality of cooperation between driver and automation. Based on previous empirical and theoretical studies a total of 34 quality indicators are identified. The methods to measure these quality indicators fall into four categories: 1) Subjective (such as questionnaires); 2) behavioral (such as reaction times, steering response); 3) neurophysiological (such as heart rate and pupil size); and 4) heuristic evaluation. The result is a first step in the development of a framework for the quantitative assessment of cooperation in the shared driving task. Yet, important knowledge gaps remain. For instance, the exact contribution of each quality indicator and their exact interrelationship are currently unclear. Moreover, all quality indicators reflect a requirement that should be met. Further research is needed to define exactly when each requirement is met. Additionally, it should be established to what degree each measurement method can validly and reliably provide insight into their quality indicator. Therefore, to ultimately ensure valid and reliable application of the framework in practice, the framework should continue to be developed and improved upon in future work.