Page 19 - ATZ WORLDWIDE

AUTHORS CONCEPT FROM SUBJECTIVE PERCEPTION
TO PHYSICAL PARAMETERS
In order to achieve objective ratings on
how well driver assistance systems are To objectify the documented subjective
performing with respect to driving com- perceptions, it is necessary to define
fort and perceived safety the well-estab- physically measurable variables that cor-
lished AVL-Drive was enhanced with relate with these subjective impressions.
Dr. Jürgen Holzinger environmental detection sensors. Using In the case of driver assistance systems
is Project Manager ADAS at the information of these sensors (radar, lane we are dealing with, the physical param-
Business Division for Assessment detection camera) the driving mode is eters are very manifold by definition.
and Development of Driving detected automatically, if a particular Typically an ACC influences the vehicle’s
Attributes – Vehicle
at AVL List GmbH ADAS function is activated. For each of longitudinal dynamics, whereas the LKA
in Graz (Austria). these detected operation modes physical and the LCA affect the lateral dynamics.
parameters are being calculated. With
these parameters an objective rating is LATERAL CONTROL
applied which is independent of the AND LANE KEEPING
driver. [1, 2, 3]
Human errors of drivability experts, Before developing assessment criteria
that inevitably sneak in therefore are and functions for automated lateral con-
Dipl.-Ing. Erik Bogner eliminated. Due to the fact that ADAS trol functions, extensive studies focused
is Skill Team Leader at the like ACC or LKA are quite novel with on subjective impressions on drivers and
Business Division for Assessment low market penetration it was decided to passengers were carried out, as men-
and Development of Driving
Attributes – Vehicle use test drives both with end-customers tioned before.
at AVL List GmbH and drivability experts for subjective rat- An analysis of the measured data from
in Graz (Austria). ings. For the alignment of the ratings the test drives revealed that achieving a
only the feedback of the experts was mathematical description of the subjec-
used. tive perception of lateral control is an
The AVL-Drive ratings are based on intricate task. The perception of choice-of-
the ISO rating scale. Ratings above 7.0 driving-line and lane control is not only
GAIN IN COMFORT AND SAFETY meets most customer’s expectations and characterised by a multitude of physical
will not cause customer complaints. variables, but also on the way the envi-
Advanced Driver Assistance Systems Lower ratings will cause increased cus- ronment and the course of the route are
(ADAS) increase the safety on the road tomer complaints. Some customers will perceived. The perceived safety of an LKA
by alerting drivers to critical situations, not use the ADAS because of its unsatis- system’s choice-of-course depends on the
or even by taking action autonomously fying behaviour. difference between the current and the
to avoid collisions or to reduce the sever- expected choice of course.
ity of accidents. Additionally they MEASURING METHODS The lane control rating corresponds
enhance the driving comfort with fea- with the control accuracy of the yaw angle
tures like automatic parking, lane keep- Either the build-in sensors of the vehi- in the lane – as well as the chosen driving
ing and distance keeping assistance. cle-under-test or additional radar and line. Not every trajectory within the lane is
Drivers and passengers will only per- lane detection sensors can be used for perceived as pleasant and safe for the
ceive this as a gain in comfort and safety the objective assessment of ADAS func- driver and the co-driver in every driving
if this support is provided in a safe, relia- tions, depending on the fact whether the situation. Only driving lines that corre-
ble and pleasant manner. measurement signals are in reach or not. spond with large radiuses and smooth
Objective assessment criteria aiming The subjective assessments were gath- changes of direction are perceived as calm,
at driveability perception, comfort and ered by male as well as female persons well-controlled and safe. The main criteria
perceived safety for driver and co-driver of all age groups and with different lev- for the driving line assessment that could
therefor are crucial for developers in els of driving experience. With the anal- be identified out of the feedback of normal
order to provide meaningful feedback ysis of the subjective ratings, in combi- drivers and driving experts are defined as
of potential shortcomings of the appli- nation with the questionnaires, the nec- mentioned in TABLE 1.
cations in question, allowing developers essary criteria to assess the functions Due to limited range of sight of actual
to eliminate possible weaknesses ACC, LKA and latter LCA could be camera systems mostly the centre-lane-tra-
effectively. extracted. jectory as target trajectory in the oncoming
From the broad variety of ADAS func- For the majority of the test candidates street segment is used. AVL-Drive is able
tions nowadays, AVL is dealing with the the most important criteria was the to detect when the LKA is switched on in
basic driving functions like distance decelerating/braking behaviour of an order to assess the typical lateral parame-
keeping (Adaptive Cruise Control, ACC), ACC. This criterion reflects comfort as ters. The runtime measurement for a LKA
Lane Keeping (LKA) and Lane Changing well as perceived safety. Interestingly ride is portioned into 2-to-5 s time win-
(LCA) control which are coming across some criteria had a very diverse impres- dows, each parameter than stands for one
independently of the particular SAE level. sion on the driver and the co-driver. [4] lateral assessment criteria.
ATZ worldwide 09|2017 17