Autonomous driving (AD) is currently one of the most eagerly-pursued goals of the automotive industry. In order to realize this goal, numerous challenges in complex scenarios, such as daily urban traffic, must be overcome.
For at least the next ten years, it can be assumed that the driver will still need to take control in situations which the vehicle is unable to manage alone. Hence, when it comes to the development of conditionally automated (level 3) driving systems the observation of driver behavior, in order to assess their ability to assume control of the vehicle when necessary, is a critical milestone.
Such driver monitoring systems (DMS) are already in use in some vehicles in active advanced driver assistance systems (ADAS). DMS are also being implemented in other areas such as passive safety, driver-vehicle interaction and the monitoring of driver health data.
The demand for DMS is expected to increase dramatically over the coming years. This means that OEMs and Tier 1 & 2 suppliers specializing in image processing for these types of systems will need to demonstrate the performance of their DMS in order to position themselves optimally in the marketplace.
Our turn-key solution for the validation of driver monitoring systems (DMS) includes a full range of services, products and competences designed to deliver optimal results for every project scenario. From specialist consultation and support in every project phase to ready-to-use ground truth systems and high-quality training data sets, we have got all the necessary steps right up to final validation covered.
A driver monitoring system (DMS) needs to be reliable, efficient and precise. Our mission is to guarantee these qualities via our high-quality online GT systems for head and gaze tracking and eyelid closure detection. The DMS generates signals relating to head position, direction of gaze, degree of eyelid closure, as well as providing information about the identity of the driver, facial expressions (e.g. neutral, smiling, sad, angry) and heart rate.
With the help of all these signals, driver behaviors such as head shaking or nodding and blinking (amplitude, duration etc.) can be observed, and, with that, driver alertness can be monitored and fatigue detected.
Our GT system for head tracking consists of a motion capturing system which measures the position of the head using infrared-reflective markers. It is also possible to temporally and spatially synchronize the GT system with the DMS. This means that the position of the head as measured by the GT system is translated into the image coordinate system of the DMS camera. Maximum accuracy is achieved with the help of image processing algorithms and highly-precise calibration targets. The overall accuracy of the system is 6 mm in all directions.
The GT system for the tracking of head position can also be used as a GT system for instructed gazes. In this system, LEDs are mounted in the driver's field of view. With the help of a 3D measurement system, the positions of the LEDs are then calibrated with the coordinate system of the GT system. While stationary or driving, the driver is prompted to look at an LED, which is either lit or blinking for traceability, and press a button to confirm it has been seen. Using this method, the position of the head, as well as the positions of the LEDs, are determined so that a gaze direction vector can be calculated. With this system, an accuracy of 2 degrees can be achieved.
The GT head tracking system can be used to measure the position of facial landmarks, such as the corners of the eyes, which can then be converted into image coordinates. In this way, it is possible to obtain an online function for the annotation of the corners of the driver's eyes.
With the help of our GT systems, we provide ground truth data for head positioning and through our annotation process we provide ground truth data for eyelid closure values as well as size and position of the pupil.
For the development of an accurate driver monitoring system (DMS) which can, for example, detect whether the driver has their eyes closed or is no longer paying attention to the road, raw data must be collected and enriched to obtain GT data regarding the degree of eyelid closure.
These are used in the validation of the DMS. Depending on the level of accuracy required, we can offer three different approaches for annotating eyelid closure:
Everything you need for the test and validation of your DMS in a 6-step process:
We provide everything you need for the test and validation of your DMS, including:
We have customers worldwide- vehicle manufacturers as well as Tier 1 & 2 suppliers- who want to test and validate their own driver monitoring systems, or who simply require data for the training of high-precision algorithms. Our broad expertise and high-quality service in the areas of testing and validation, ground truth system development, data enrichment, system integration, vehicle prototyping and sensor calibration enable us to develop the most reliable driver monitoring systems.