ABOVE: Mercedes-Benz
drivers perfecting their vehicle
control inputs. This task is
made trickier when the control
software is not predictable
24 CONTROL SOFTWARE
There are acceptable ways to change vehicle behaviours
according to vehicle speed. Purely mechanical vehicles do
this in terms of steering response anyway, where steering
response is reduced as speed increases, in terms of
steering wheel input angle to cornering curvature (1/R).
Drivers of vehicles with such a system learn to expect and
understand the logical nature of the steering behaviour.
Vehicle response gains can be varied smoothly with speed
so that the driver feels how steering responses change
with speed, and can extrapolate what responses they
should expect at, say, higher speeds than they have driven
at before. In a way, the driver effectively increases their
skill level of controlling their car by reverse-engineering
the control software to understand what to expect, but
in reality the driver only needs to learn the input/output
relationships for the controls.
This variability in response is not achieved using
a software language, but rather at a logical level,
suffi cient for the driver to build a mental model of the
vehicle’s response characteristics. I have lost control of
many demonstration vehicles (in controlled test track
conditions) when software has been written in such a way
that the driver has no way of working out what input is
required to get the car to do what he or she wants.
The speed of drivers’ responses is another factor for
consideration in the design of control algorithms. During
one test of a development vehicle I learned that I could
respond to a software failure in 120ms, measured from
Control inputs
Drivers need to be able to understand which control inputs are required in order
to get a vehicle to react how they want it to. Each driver control is a device that
facilitates communication between the driver and vehicle, and each control must
have a defined input/output relationship that identifies the information that is being
transmitted in each direction.
Driver (motion) controls include the steering wheel, brake pedal, accelerator pedal,
parking brake control, and transmission/drive (forwards and reverse). For each
control, the parameters may be displacement and force. If a control parameter is
used for communication between the driver and vehicle, then it cannot logically be
used in the other direction. In the example of the steering wheel, this means that if
the steering wheel angle is defined as the driver’s inputs to the car, then the steering
wheel torque is the only means of informing the driver about what the car is doing.
The principle is similar for the pedals. Brake pedal force informs the car what
amount of braking is required, and the driver can interpret pedal travel to understand
what the car is doing (as happens when ABS becomes active). Currently the
accelerator pedal is generally only used as a one-way communication control, but
logically, some feedback could be provided, and there have been a number of
proposals for haptic accelerator pedal feedback.
Photo: Skoda/Bernhard Huber
VehicleDynamicsInternational.com • May/June 2020
/VehicleDynamicsInternational.com