Implementing human factors within the design process of advanced driver assistance systems (ADAS)
PhD candidate: B. M. van Waterschoot, MSc (email@example.com)
Promotor: Prof. dr. ir. F.J.A.M. van Houten (firstname.lastname@example.org)
Co-promotor: Dr.ir. M.C. van der Voort (email@example.com)
Period: April 2008 – March 2012
This research project aims at developing a design approach for ADAS applications in which human factors (including stakeholder feedback and objective performance measures) are accounted for. Because the driving task is part of a complex (traffic) system, with a large number of interacting components, ADAS design is confronted with choices for which the influence on the system, and the driving performance in particular, cannot be predicted. Therefore, providing relevant feedback during the design process, about the consequences of specific choices,would increase the efficiency and safety of driver assistance systems.
Keywords: human factors engineering, advanced driver assistance systems (ADAS), interaction design, complex systems
Background and problem definition
The emerging trend in driver support systems is aimed at reducing requirements placed on the driver. By equipping vehicles with sensors, navigation and motion planning the driving task is shared between human actors and the supporting assistance systems. Ultimately, by adding and improving cognition and control techniques, this could lead to autonomous vehicles in which the driving task is controlled by the vehicle and the responsibility is shifted towards the vehicle and its manufacturer.
Although legal issues and high infrastructural demands will prevent the introduction of such autonomous vehicles in the near future, research has already provided (semi-) automated concept cars in which no (or minimal) intervention of human actors is required. Meanwhile, different assistance systems are already supporting the driver by means of sensory information (e.g. visibility aids or lane departure warnings), correction (e.g. anti-lock braking system or traction control) or even control (e.g. automatic parking). A serious implication of the growing amount of these assistance systems in modern day cars, is the unknown effect different types and quantities of information can have on the driving performance. Different studies provided evidence of reduced users’ workload while supported by assistance systems (e.g. Stanton and Young, 2005). However, these studies did not take into account what effects a combination of support systems (and hence, with different configurations and amount of information) can have on the drivers’ performance. Information which is of prior importance in order to produce safe and efficient cars in the future.
The present research is aimed at providing ADAS designers with insight in what influence different types and configurations of information, provided by the assistance systems, can have on the driving performance. Ultimately, our goal is to improve the efficiency and safety of driver assistance systems. This research aims to develop a design approach that allows designers to obtain relevant feedback during the design process, regarding the consequences of their choices. For this, we intend to set up an adaptive design environment in which both stakeholders’ preferences and driving performance can be measured online, in order to bring about designs in which human factors are well accounted for.
Stanton, N.A., Young, M.S. (2005). Driver behaviour with adaptive cruise control. Ergonomics, 48(10), 1294-1313.
Scientific and societal relevance
In general, designers are confronted with stakeholders’ preferences and technical constraints. Once the iterative design process reaches its final post-production phase, evaluation and conclusions are used to either improve future products or re-design the one at hand. This means that -although preferences and constraints are being met- the overall performance remains uncertain until the product has been used in its ‘natural’ environment with its given scenarios.
Providing a design toolkit that is able to access both stakeholder preferences and performance measures during the design process, would therefore provide insights for car manufacturers and the scientific community.
Expected practical and scientific results
Traditional research approaches, which can be characterized as rigid (testing a given configuration by means of prototype or stimulus-response research in a controlled laboratory), typically involve two main shortcomings. Results of testing a given ADAS application will only be beneficial for future design and does not provide relevant insights about different configurations. In addition, isolated stimulus-response effects cannot be generalized to different contexts or environments.
An adaptive design approach that provides online performance measures and stakeholders’ preferences within (theoretically) any given context would therefore provide a more ‘naturalistic’ research environment. This environment would provide insight about driving performance under different stimulus-response configurations, and would support the design process by giving both objective and subjective feedback about the design choices made.
This research is part of the knowledge centre AIDA, a collaboration between TNO and the University of Twente.