Abstract
The aim of the study is to investigate driver behaviour with focus on the
variation of different drivers´ ability to steer the vehicle, i.e. path tracking skill. A moving
base driving simulator is used to study drivers´ performance and behaviour when following a
cone track scenario, primarily looking at the driven paths and lateral acceleration levels. The
narrow cone track requires large lateral displacements and relative high attention on
controlling the vehicle, but allows a limited amount of variance in chosen path. To investigate
if there are differences between drivers with varying experience of driving, the recruitment
base is low and high mileage drivers. All drivers are driving the simulator with a vehicle
model of a regular passenger car with both the standard setting as reference and eight other
combinations of steering wheel gear ratio and steering wheel effort for comparison.
The results show that the driving path strategies can be very different from driver to driver,
and that both high mileage and low mileage drivers seem to prefer to hold on to their
strategies despite changes in the vehicle steering wheel ratio and effort. Some strategies result
in twice the lateral acceleration of others, and thus a smaller buffer to the limit set by road
friction. The manoeuvre in the cone track has revealed some of the complexities in
characterising and modelling of drivers by showing that the driver preferred strategy can
differ much between drivers even for a relatively narrow track. This is important to take into
consideration when analyzing path tracking skill on a road or thru a cone track, since the
actual path the driver is trying to follow might differ in many ways from the optimal path, and
thus they strive to fulfil different goals. Even though strategies were similar over the runs, the
variation varied between drivers. The results here support driver classification measuring
three different aspects of path tracking skill:
- choice of path to track - cornering strategy
- deviation from the mean path - repeatability
- change in deviation from the mean path using altered steering parameters -
robustness for changes
The cornering strategy is found to be one good measure for analysing the quality of the
drivers decisions, while the magnitude of standard deviation from the average path seem to be
useful for analysing the repeatability of the drivers, and thereby an indication of the precision
of which the driver controls the vehicle. The ability to handle system changes reflects how
well the driver adapts to the new system.
Many more aspects must be, and have been, studied to improve the knowledge for driver
classification of path tracking skill, but the results here show the potential of analyzing
driving behaviour using moving base simulators with the benefit of those.
Keywords - Skill, Strategy, Path Tracking, Cornering, Moving Base Simulator