Page 113 - Handbook of Biomechatronics

Model-Based Control of Biomechatronic Systems 109

driving experience. The driver-assist systems receive sensory feedback from
the vehicle, and commands such as acceleration, brake, and steering from the
driver. Here, four neuromuscular driver models representing drivers with
different physical strength, age, and gender were developed. These models
were used to design a new model-based EPS controller that adjusts the
steering assistance based on the driver’s physical strength. In the proposed
controller, the EPS characteristic curves (determining the steering assistance)
were precomputed for the predefined driver populations and stored in the
controller. The characteristic curves were optimized such that the drivers
within different populations performing the same steering maneuver have
a similar targeted “steering feel.” The steering feel was defined by a combi-
nation of drivers’ muscular effort and road feel. Finally, the new EPS con-
troller was evaluated in MIL simulations using a high-fidelity integrated
driver-vehicle model. The results showed that the tuned EPS controller
could equally assist drivers with different physical strengths and abilities.

3.1 Introduction
Emerging research has resulted in new models of the interaction dynamics
between the vehicle and its driver, the results of which have given rise to
new driver-assistance technologies—haptic gas pedals, lane keeping, artifi-
cial steering wheel torque feedback (Abbink, 2006), and EPS systems
(Mehrabi and McPhee, 2014a; Farrelly et al., 2007). Steering feel and vehicle
stability are two commonly used criteria in the design of EPS controllers.
Vehicle stability measures are well documented in the vehicle dynamics lit-
erature (Karnopp, 2003), while there is only a limited literature available on
quantifiable steering feel measures. Previous research has found correlations
between steering feel and vehicle handling characteristics; however, these
investigations were limited to a specific driver population (i.e., truck drivers)
(Rothh€amel et al., 2011, 2014). Vehicle manufacturers typically employ
professional drivers to tune steering systems to provide “good” steering feel.
However, this approach has numerous drawbacks. Such experiments can be
expensive, time consuming, and are subject to human error. In addition, the
preferred steering feel is different for vehicles with different handling char-
acteristics (e.g., sport vs luxury cars) (Bertollini and Hogan, 1999), and
simultaneously the optimum steering feel may vary between driver
populations (i.e., drivers with different physical abilities). For example,
young drivers generally have stronger muscles, and thus greater ability to
overcome resistive torques at the wheel, than elderly drivers. Therefore,