Periodic Reporting for period 3 - MEDIATOR (MEdiating between Driver and Intelligent Automated Transport systems on Our Roads)
Berichtszeitraum: 2022-05-01 bis 2023-04-30
The MEDIATOR project aimed to develop an in-vehicle system, the Mediator system, that intelligently assesses the strengths and weaknesses of both the driver and the automation and mediates between them, while also taking into account the driving context. It assists the timely take-over between driver and automation and vice versa, based on who is fittest to drive. This Mediator system optimises the safety potential of vehicle automation during the transition to full (level 5) automation. It would reduce risks, such as those caused by driver fatigue or inattention, or on the automation side by imperfect automated driving technology. MEDIATOR has facilitated market exploitation by actively involving the automotive industry during the development process.
The Mediator system, as developed in the project, consists of four main modules: a driver monitoring system, an automation monitoring system, a human-machine interface and a decision making module.
In the first phase of the project a state-of-the art knowledge overview was created focussing on all knowledge required to develop the Mediator system (D1.1). In addition, knowledge gaps were identified and prioritised forming the basis for the research agenda in the first phase.
Driver monitoring
With state-of-the-art knowledge and research results from within the project the functional requirements for the Mediator driver monitoring system were determined (D1.2). The Mediator driver monitoring system was integrated in different prototypes (D2.11) that were evaluated in the project. MEDIATOR focussed on monitoring fatigue, distraction, and discomfort while taking driving context into account.
Based on state-of-the-art knowledge from literature with knowhow from the industry and evaluation results from the project, guidelines for monitoring degraded driver performance were developed (D4.3). The guidelines focussed on fatigue, distraction, and discomfort, led by functionality constraints, technological possibilities, safety relevance and feasibility.
Automation monitoring
The automation state monitoring system assesses the automation “fitness”; the ability to drive in a certain automation level at the current point in time and in the near future. In the first phase of the project the functional requirements for the automation state monitoring system were defined (D1.3)
The automation state monitoring system integrated in the prototypes (D2.11) was running in real-time. The output of the automation state monitoring, together with the output of the driver monitoring module, forms the key input for the decision making module. In addition, the output of automation state monitoring system allowed the HMI showing time budgets to the driver; the time remaining till the next change in automation level (availability).
User centred HMI design
The Mediator HMI design took a user centred and holistic approach. The HMI design is framed by five principal Design Guidelines (D1.5) dealing with the following main challenges for a human interface in the transition towards automated driving:
• Transfers of control
• Transparency & information load
• Keeping the driver in the loop
• OEM design space
• Negotiating conflicts
The Mediator HMI design was integrated into several prototypes (D2.10 and D2.11). Based on the evaluation results, general HMI guidelines were defined (D4.2).
Decision logic
The Mediator decision logic is really the brain of the system. With the initial functional requirements (D1.4) al prototype was developed by means of a computer simulation (D3.2). The decision logic was integrated in a vehicle prototype (D2.9 D2.11).
The central roles of the Mediator decision logic are to:
• Recommend switches between available automation levels between automation and human driver
• Trigger corrective actions
• Enforce emergency actions
• Provide supportive information to the HMI
Prototypes and evaluation trials
The Mediator system was integrated into six different prototypes: two driving simulators, one virtual reality, two vehicle prototypes and a computer simulation (D2.9 2.10 2.11) each focussing on different aspects of the system.
These prototypes were used for extensive evaluation with over 200 participants resulting in a collection of 360 datasets. All evaluation results are available (D3.2 D3.3 D3.4) and integrated in a final overview of the evaluation results (D3.5).
The evaluation studies revealed good usability scores and high acceptance rates. Most importantly, several safety benefits were observed such as a reduction of distracted driving and an increase in use of the automated systems.
Impact and roadmaps
Based on the evaluation results, MEDIATOR estimated the potential safety benefits and related societal benefits of the system developed (D4.1). Using the experiences gained in the project, protocols for low-cost laboratory testing (D4.4) and recommendations on legal and regulatory aspects (D4.5) were defined.
In order to facilitate the uptake of the MEDIATOR results by the transport industry, a roadmap for exploitation for road transport (D5.9) and a roadmap for exploitation for aviation, maritime, rail (D5.10) were developed with the involvement of external stakeholders. In addition a total of five workshops with external stakeholders and a well visited and high rated final event were organized to disseminate the results.
Mediator could have prevented 17,738 and 13,012 extra-urban rear-end collisions in Germany in 2019 and 2020, respectively, with an assumption of 50% effectiveness rate. The financial benefit ranges from 0.003% to 0.26% of the gross domestic product (GDP) depending on the country. The societal benefits in terms of avoided road injuries also range from around 0.01% to 0.2% of the GDP.
The Mediator system could reduce driver distraction up to 40%. According to the European Road Safety Observatory, it is generally estimated that distraction plays a role in 5 - 25% of crashes in Europe. The findings indicate that the Mediator system can contribute significantly to reducing crash risks associated with distraction. Different (social) media channels we used to raise awareness of the results achieved by the project.
Based on lessons learned and web survey results, a total of 18 recommendations are formulated for potential policy changes that could improve the current regulatory framework. These recommendations ensure a better safety transport culture and propose improvements to policy makers for the next generation of high-level vehicle automation.