Enter the search phrase
Author: Sylwia Mac
If you’ve been living in this world long enough, you may remember motor vehicles without seat belts or child restraint systems. We’ve come a long way since, and the vehicles we currently use are becoming safer and safer. This is being achieved with common bottom-up and top-down efforts – on the one hand, manufacturers contribute to the introduction of new in-vehicle safety features with their continuous need for technological development to gain a market advantage, on the other hand, legislation fuels safety improvements in the automotive industry.
Different levels of economic and technical development along with distinct mobility preferences bring diverse challenges to various world regions. While Europe is gradually introducing more and more advanced driver safety features (i.e., driver monitoring systems, speed assist and lateral support systems), a current challenge of the Southeast Asian market is the implementation of ESC (Electronic Stability Control) among two-wheelers which are a highly popular means of transportation there [1]. With the accelerating pace of automotive innovation, legislative developments are needed, and all those regulatory innovations entail the implementation of new technologies. In this cycle, legislative and assessment bodies review state-of-the-art solutions and create regulations that make new technologies both compelling to develop and possible to fulfill.
In today’s blog post, we would like to focus on the latest regulatory specifications regarding the testing of driver distraction. It is, next to driver drowsiness, one of the main causes of road accidents, as these conditions can severely affect driving abilities. In case you missed it, we also encourage you to read one of our previous blog posts that introduces visual-based driver monitoring approaches regarding both driver distraction and drowsiness detection.
As distraction is quite complex as a phenomenon, its measurement and detection can be a challenging task. In its report on advanced driver distraction warning (ADDW) prepared for the European Commission [2], the Transport Research Laboratory points out visual, manual, auditory, and cognitive distractions. Due to technical capabilities, the current scope of interest is on measuring visual and visuo-manual distraction. Focusing on visual distraction alone – gaze patterns may be different while driving through the crowded city and on the highway, so intense looking around will not necessarily indicate distraction, it may as well be a safety-related behavior of monitoring busy surroundings. This is not even mentioning the different physical characteristics of the driver population, which may be yet another challenge for camera-based systems. Given these difficulties, it took quite a long time for regulation with specific guidelines for distraction recognition to be issued. Recently, a draft of the EU’s specific test procedures and technical requirements for the type-approval of motor vehicles in regard to their advanced driver distraction warning systems [3] was released. Along with the European New Car Assessment Programme (Euro NCAP) Assessment Protocol [4], these documents are now leading the testing of drivers’ visual attention monitoring systems.
Effective and well-thought-out testing is the best way to both pass regulatory audits and have a proper background for future system development. As DMS testing is one of our fields of expertise, we comprehensively prepare for conducting such studies – first, we deeply review current regulations and related literature, then after considering the client’s needs, we propose the most suitable testing methodology, including the testing environment selection, equipment specification and synchronization, and scenarios preparation.
Testing environment selection is one of the first things we prepare for a driver monitoring study. As distraction is a safety-critical driver state, we always advise performing testing in a safe setting – a driving simulator or a test track. In our previous post, we compared the main advantages of different testing environments, so that is a good place to start considerations. Some hybrid options are also possible here – as the weather is not always favorable, some testing conditions may be simulated. We can proudly show off our artificial solar system, that can mimic different lighting conditions even on the cloudiest and darkest day (or night 😉). The Driver State Monitoring system, in general, shall be checked at least for operation in day and night conditions, as its performance must not be degraded by lighting between 1 and 100 000 lux.
Source: Robotec.ai
The newest regulations do not recommend any specific testing environment. In case of testing in a simulated environment, we use a high-fidelity simulator built on a real vehicle cabin. For any naturalistic study, we always use an experienced driving instructor as backup, and we recommend equipping the vehicle with dual pedals for the safety of study participants and equipment. Playing for the safer future of transportation, we make the safety of our study participants our priority.
Preparations of a vehicle and its equipment is another significant factor of the successful distraction study or system validation. While the system is still in the development phase, the use of high-precision referencing cameras and motion trackers may be beneficial, and we always remember the importance of examining how these systems may affect each other and ensuring that all measurement devices are properly synchronized. At further stages of preparation for type approval or assessment, the vehicle (in its default configuration) needs to be supplied with a sufficient number of supplementary cameras and a device able to determine its speed to ±1 km/h (as the testing shall be conducted in indicated velocities). To be eligible for scoring points in Euro NCAP’s Driver State Monitoring, the vehicle also needs to be equipped with Seat Belt Reminder (SBR), Autonomous Emergency Braking (AEB), and Lane Support System (LSS).
Source: Robotec.ai
After setting up the equipment, distraction testing scenarios need to be prepared. Euro NCAP presents a complex assessment protocol, that includes different distraction and movement types. In the area of distraction type, it distinguishes:
By movement type:
The protocol also pays attention to driver behaviors that are common behind the wheel but have the potential to affect the performance of the Driver State Monitoring system. Among them, we can find eating, talking, laughing, singing, smoking or vaping, eye scratching or rubbing, and sneezing – all need to be included in the testing protocol and demonstrated in the dossier, in terms of if and how they affect the system’s performance.
In this form, the Euro NCAP assessment seems to be the most advanced, detailed, and demanding document available at the moment. The ADDW draft considers long distraction type only, but that seems a reasonable requirement for the first implementation of distraction recognition systems, that for many manufacturers are still in development or have just recently entered the market.
Last but not least – a proper group of people is needed to perform the testing. While GSR does not specify the testing sample yet, Euro NCAP states numerous requirements related to drivers, both in terms of their demographics and appearance. The perfect sample shall, at a minimum, consist of people of different gender, ages 16 to 80+, AF05 to AM95 stature, with skin color 1 to 6 according to Fitzpatrick Skin Type, and eyelid aperture between 6 and 14 mm. Some of them shall cover occlusions listed in the assessment protocol, like having a face mask, long facial hair, fringe obscuring the eyes, or heavy makeup, as those factors may degrade system performance. We carefully select the composition of the testing sample, ensure recruitment, and assign different noise variables and behaviors for an adequate subsample of the participants. All this makes it possible to effectively verify the operation of the system while satisfying regulatory requirements.
There is still a large field for safety improvement that can be achieved by the common effort of manufacturers, and legislative and assessment authorities. To be ready in advance for upcoming solutions, we are following automotive trends, industry reports, visionary documents, and of course the latest regulations. In the field of Driver Monitoring Systems, which is one of our areas of expertise, we are now focusing on driver drowsiness and distraction recognition and prevention. In the near future, we expect new demand in the form of Child Presence Detection, Driver Stress Recognition, and Driving Under the Influence prevention, just to name a few. We remain at the forefront of upcoming innovation and look forward to all of the exciting challenges ahead.
If you have any thoughts or concerns or would like to discuss potential projects with us, let us know at office@robotec.ai.
Sources:
[1] Towards Zero Foundation 2020 Annual Report, https://static1.squarespace.com/static/602bd660117bd25f81afa72a/t/61fade45ffe6ac40d8290353/1643830867817/TZF-annual-report-2020.pdf
[2] Seidl, et al., 2021, https://ec.europa.eu/docsroom/documents/45901?locale=en
[3] Commission Delegated Regulation (EU) 2023/… of … supplementing Regulation (EU) 2019/2144 of the European Parliament and of the Council by laying down detailed rules concerning the specific test procedures and technical requirements for the type-approval of motor vehicles with regard to their advanced driver distraction warning systems and amending Annex II to that Regulation, https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/13740-Road-safety-advanced-driver-distraction-warning-systems_en
[4] Euro NCAP Assessment Protocol – Safety Assist Safe Driving, Version 10.1.2, February 2023, https://cdn.euroncap.com/media/77138/euro-ncap-assessment-protocol-sa-safe-driving-v1012.pdf