Chapter | Four

Kinematic Steering
In Chapters 4 and 5, attention is paid to vehicle behavior. This chapter dis-
cusses when tire forces are neglected, which is usually referred to as
Ackermann steering. Chapter 5 covers handling and stability behavior.

4.1 AXIS SYSTEMS AND NOTATIONS

To study the response of a vehicle in order to control inputs or disturbances,
it is necessary to specify one or more coordinate systems to measure the posi-
tion of the vehicle. The SAE method (SAE: Society of Automotive Engineer)
[61] will be followed here. There are two main systems to measure the posi-
tion of the vehicle. The first is an earth-fixed system, denoted by XYZ. The
second system is a vehicle-fixed system xyz (lower case), as indicated in
Figure 4.1. The origin of the xyz system is usually taken at the vehicle’s cen-
ter of mass. The orientation of the vehicle’s axis system xyz with respect to
XYZ is given by a sequence of three angular motions:

ϕ: roll rotation angle about the vehicle’s x-axis

θ: pitch rotation angle about the vehicle’s y-axis
ψ: yaw rotation angle about the vehicle’s z-axis

A reference situation is considered in the case of earth-fixed and vehicle-

fixed coordinate systems coinciding. The x-axis is taken from the vehicle’s
central plane, which is pointing forward and horizontal in the reference situa-
tion. The y-axis points to the driver’s right-hand side and is horizontal in the
reference situation. The z-axis points downward. The velocity of the vehicle
is taken as the velocity of the center of mass, as measured in the XYZ system.
Its components in the local xyz system are referred to as:

u: longitudinal velocity along the x-axis

v: side velocity along the y-axis
w: normal velocity along the z-axis

111
Essentials of Vehicle Dynamics.
r 2015 Joop P. Pauwelussen. Published by Elsevier Ltd. All rights reserved.
112 Kinematic Steering

Clearly, these velocity components, in general, will not be parallel to the

ground plane. Therefore, we define the following velocity components:

forward velocity: horizontal velocity component\y-axis

lateral velocity: horizontal velocity component\x-axis

The angular velocities relative to the local xyz system are denoted as:

p: roll angular speed

q: pitch angular speed
r: yaw angular speed (yaw rate)

A top-down view of a car following a path with speed V is shown in

Figure 4.2. Three angles are distinguished here that describe the projected

Pitch
v, y

u, x
Roll
Yaw
w, z

FIGURE 4.1 Vehicle local axis xyz.

FIGURE 4.2 Projected view.

 4.2 Ackermann Steering 113

orientation of the vehicle, with respect to the earth-fixed coordinate

system:

• The heading angle ψ between the projected x-axis and the global X-axis.
• The side slip angle β between the forward velocity (path tangent) and the
projected x-axis (taken positive clockwise).
• The course angle η between the forward velocity and the global X-axis.

Consequently,
ðt
η 5 ψ 1 β 5 ψðt 5 0Þ 1 rðτÞUdτ 1 β ð4:1Þ
0

4.2 ACKERMANN STEERING

In general, steering of a vehicle at a finite velocity leads to side forces at the
wheels, which will counteract the lateral force, acting on the vehicle. These
side forces correspond to tire slip angles, according to certain lateral tire
characteristics. In this chapter, we discuss a situation where the vehicle veloc-
ity is very small, such that no significant lateral forces act on the vehicle.
Consequently, no side forces are required at the wheels and, ideally, maneu-
vering of the vehicle can be done by pure rolling of the wheels if we neglect
tire turn slip. Any steering mechanism that satisfies these conditions is
referred to as Ackermann steering (Figure 4.3). The pole, or center, of the
vehicle rotation is on the line connecting the rear wheels.

FIGURE 4.3 Low-speed cornering.

114 Kinematic Steering

From this figure, one easily obtains, for pure rolling of wheels
t t
cotðδL Þ 2 cotðδR Þ 5  ð4:2Þ
a1b L
for track width t and wheelbase L. Unfortunately, no practical steering mech-
anism satisfies Eq. (4.2); some examples are illustrated next, discussed earlier
by Genta and Morello in Refs. [10] and [11]. The Ackermann share of a
steering mechanism is defined as (see Ref. [16], Chapter 1):

δR 2 δL
Ackermann share: 3 100½% ð4:3Þ
δR 2 δL;AM

with inner wheel steering angle δR, outer wheel steering angle δL, and exact
Ackermann outer steering wheel angle δL,AM, which satisfies Eq. (4.2).
Pure Ackermann steering means that the inner and outer wheel steering
angles are not identical. The reader may easily verify that Eq. (4.2) results in
a steering angle difference, as shown in Figure 4.4, for t 5 1.5 [m] and
L 5 2.76 [m]. One observes a small difference for wheel steering angle, up
to 5 , which is the range for normal handling situations. Clearly, applying
the same steering angle for both the inner and outer wheels is an
acceptable approximation for handling analyses (discussed in Chapter 5). For
parking conditions with a large steering angle, one may expect a difference
between inner and outer wheel steering angle, up to 10 .
Let us consider the quadrilateral steering mechanism, shown in Figure 4.5.
We follow the analysis in Ref. [10]. The mechanism consists of four bars, con-
nected at four rotational joints. The length of the lower bar, f2, is found from

f2 5 f 2 2UdUsin γ

The lateral horizontal distance H between the lower two joints follows
from the lengths of the remaining bars and the steering angles:

H 5 f 2 dUsinðγ 2 δL Þ 2 dUsinðγ 1 δR Þ

12

10

8
δR – δL [°]

0
0 5 10 15 20 25 30 35 40
δR [°]

FIGURE 4.4 Difference between inner and outer steering angle, pure Ackermann
steering.
 4.2 Ackermann Steering 115

Further, H follows from the orientation of the lower bar after steering:

H 2 5 f22 2 ½dUcosðγ2δL Þ2dUcosðγ1δR Þ2

These two results are sufficient to determine δL from δR for this steering
mechanism, as we have done. We determined the Ackermann share according
to Eq. (4.3) for different inner wheel steering angle and angle γ. Figure 4.6
shows the results for f 5 1.3 [m] and d 5 0.2 [m] (right plot). We also deter-
mined the difference between the outer wheel steering angle and the optimal
Ackermann value according to Eq. (4.2), as shown in the left plot of Figure 4.6.
One observes errors in the outer wheel steering angle, compared to pure
Ackermann steering, on the order of 2 to 3 [ ], when the angle γ is varied
between 16 and 24 [ ]. The value γ = 20 [ ] appears to be a reasonably good
choice. The relative deviation in terms of Ackermann share varies from 60%
to 130%, with γ = 22 [ ] apparently giving a share closest to 100% over the

FIGURE 4.5 Example steering mechanism.

2 130

120
δR – δL,Ackermann [°]

1
Ackermann share [%]

110

0 100

γ = 16° 90 γ = 16°
–1 γ = 18° γ = 18°
80
γ = 20° γ = 20°
70 γ = 22°
–2 γ = 22°
60 γ = 24°
γ = 24°
–3 50
0 10 20 30 40 0 10 20 30 40
δR [°] δR [°]

FIGURE 4.6 Deviation of outer wheel steering angle to Ackermann steering (left plot)
and Ackermann share (right plot).
116 Kinematic Steering

FIGURE 4.7 Rack and pinion system.

full range for the inner wheel steering angle. Note that, for small steering
angle and therefore a small absolute deviation from Ackermann steering, the
Ackermann share can still be quite different from 100%.
Next, we will consider the rack and pinion system, shown schematically
in Figure 4.7. The system is assumed to be in one horizontal plane, defined
by the parameters a, h, d, f, and angle ζ.
The rack is positioned at a distance a from the axle. There are six revo-
lute joints, with fixed kingpins on the axle. We will investigate the kinematic
properties for variations for these parameters.
The steering angles δ1 and δ2 can be determined for different positions of
the rack, as indicated in the lower part of Figure 4.7. The end positions of the
rack are known, the positions of the kingpins on the axle remain unchanged, and
the positions of the other joints are found from the fact that the bar lengths
remain unchanged. We completed this analysis for the following reference data:

a 5 0.25 [m]
h 5 0.50 [m]
d 5 0.32 [m]
f 5 0.65 [m]
ζ 5 5 [ ]

Results are shown in Figures 4.8 and 4.9. One observes:

• There is an optimal value for d when the other parameters are unchanged.
The error, with respect to the Ackermann steering, can be reduced to less
than 0.5 [ ].
• With the a-value reduced (rack is positioned closer to the axle), the error
is reduced as well.
• The same sensitivity is observed when the angle ζ is reduced to zero.
Indeed, a quick survey confirms that this angle is chosen small in general,
which likely is a consequence of packaging restrictions.
• Finally, the value for rack length h appears to be rather optimal for our
choice of the other design parameters.
 4.2 Ackermann Steering 117

5 2

4 a = 0.20 [m]

δL – δL,Ackermann [°]
a = 0.25 [m]
δL – δL,Ackermann [°] 3 1.5
a = 0.30 [m]
2

1
1
0
–1 d = 0.28 [m]
d = 0.30 [m] 0.5
–2 d = 0.32 [m]
d = 0.34 [m]
–3
d = 0.36 [m]
–4 0
0 10 20 30 40 0 5 10 15 20 25 30
δR [°] δR [°]

FIGURE 4.8 Deviation of outer wheel steering angle to Ackermann steering, for vary-
ing parameters d and a.

2.5 5
ζ = 0 [°]
4
δL – δL,Ackermann [°]

δL – δL,Ackermann [°]
2 ζ = 5 [°]
ζ = 10 [°] 3
1.5
2

1 1

0
0.5 h = 0.40 [m]
–1 h = 0.50 [m]
h = 0.60 [m]
0 –2
0 5 10 15 20 25 30 0 10 20 30 40
δR [°] δR [°]

FIGURE 4.9 Deviation of outer wheel steering angle to Ackermann steering, for
varying parameters ζ and h.

A design for an improved rack and pinion steering system has not been
attempted here. The preceding analysis demonstrates the sensitivity of the
steering system configuration with respect to certain parameters (it provides a
qualitative analysis), with reference to pure Ackermann steering. We con-
clude that one will not reach the ideal Ackermann performance, but one
could get satisfactorily close to it.
Steering is done with the intention to maneuver the vehicle, which means
the driver is aiming for a certain trajectory curvature for a given steering
angle δ, i.e., for a trajectory curvature gain. We choose the steering angle δ
as the average of the left and right steering angle, as follows:
Rrear
cot δ 5 12Uðcot δL 1 cot δR Þ 5
L
which is near the direct average of the steering angles. For kinematic steer-
ing, the following gains are distinguished (see also Figure 4.3):

1=R 1 1
Trajectory curvature gain: 5 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ð4:4aÞ
δ δU b 1 L Ucot δ
2 2 2 L
 
β 1 b b b
Body slip angle gain: 5 Uarctan pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   ð4:4bÞ
δ δ R2 2 b2 RUδ L
118 Kinematic Steering

where the curve radius R is assumed to be large, compared to the wheelbase

L. From Eq. (4.4a), the important expression for the Ackermann steering
angle δAM follows as:

L
δ 5 δAM 5 ð4:5Þ
R

This expression means that for negligible velocity, and therefore negligible
lateral acceleration, the axle steering angle of a vehicle is equal to the ratio of
wheelbase and path curve radius. In Chapter 5, we will discuss the relationship
between steering angle and lateral acceleration under steady-state conditions.
Expression (4.5) provides the first point on that curve, i.e., for ay 5 0.
We close this section with some remarks concerning the need for
Ackermann steering. As mentioned previously, the situation of pure Ackermann
steering is never reached, but can be closely approximated. In addition:

• The condition of negligible vehicle velocity is usually not satisfied.

• There is always side slip because of toe-in (usually present).
• Aligning effects lead to roll-induced steering.
• Suspension compliance and steering compliance lead to additional steer-
ing and therefore side slip.

Conversely, too large a deviation from pure Ackermann steering could

result in significant tire wear, which in itself will affect the feedback of the
steering performance and road conditions to the driver. Turn slip plays an
important role in this, and this feedback preferably should not be influenced
by side slip response or aligning torque effects.

4.3 THE ARTICULATED VEHICLE

The analysis of low-speed vehicle maneuverability is of interest in deter-
mining the amount of space required by the vehicle. In general, the
designer’s intention is to reduce the requirements for maneuvering space,
i.e., to limit this space as much as possible. For a single passenger car,
this analysis is straightforward. For articulated vehicles, such as a
carcaravan or trucktrailer combination, a limited maneuverability space
is not obvious. In general, the different articulations (car, caravan, trailer,
etc.) follow different curve radii. As a result, the difference between the
minimum inner radius and the maximum outer radius for the vehicle com-
bination during maneuvering can be significantly larger than the width of
the vehicle combination. This difference is called the swept path, which
should preferably be as small as possible. The optimal situation is when
the larger trailer axle is following exactly the same curve as the first vehi-
cle’s axle. Consider a cartrailer combination, as schematically shown in
Figure 4.10. The trailer is assumed to have one axle that may be steered.
 4.3 The Articulated Vehicle 119

FIGURE 4.10 Cartrailer combination with steerable trailer axle.

The question is how to steer this axle so that the curve radii at vehicle’s
front axle and the trailer’s back axle are identical to create a swept path as
minimal as possible.
Let us begin with the situation with the trailer axle as not steered, i.e.,
δ2 5 0, and the radius Rrear perpendicular to the trailer. Assume that all curve
radii are large compared to the dimensions of the vehicle. In that case, one
may write
   
L2 f L2 1 f
γ 5 π 2 arccos 2 arccos 
R1 R1 R1
L1 L1
R1  R 5 
tanðδ1 Þ δ1

A combination of these relationships gives the trailer angle gain:

γ L2 1 f
5 ð4:6Þ
δ1 L1

In other words, a large articulation angle γ is obtained under low-speed

conditions, if the coupling overhang f is large or there is a large distance
between kingpin and trailer axle.
120 Kinematic Steering

The radii Rfront and Rrear can be expressed as follows:

L1
Rfront 5
sin δ1

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L21
Rrear 5 f 2 1 2 L22
tan2 δ1

Eliminating δ1 leads to

R2rear 5 R2front 2 L21 2 L22 1 f 2

As illustrated, we have determined the relative off-tracking

Rfront 2 Rrear
Rfront

in percent, for different values of L2 and f, for Rfront 5 10 [m], and a fixed car
wheelbase of 2.76 [m] (Figure 4.11). One observes a dominant effect from
the position of the trailer axle, with respect to the kingpin between car and
trailer. The parameter f has only a minor effect.
If δ2 6¼ 0, the radius Rrear is no longer perpendicular to the trailer. We use
the cosine rule in the triangle RrearR1L2 and find, under the same assump-
tion of large curve radii, that

γ L2 1 f δ2
5 2 ð4:7Þ
δ1 L1 δ1

18

16
Relative off-tracking [%]

14

12

10
L2 = 2 [m]
8
L2 = 3 [m]
6 L2 = 4 [m]
4 L2 = 5 [m]

2
0 0.5 1 1.5 2
Coupling overhang f [m]

FIGURE 4.11 Off-tracking for different vehicle parameters.

 4.3 The Articulated Vehicle 121

We choose δ2 such that Rrear 5 R. The cosine rule then results in the
following relationship:

R2 1 f 2 5 R21 5 R2rear 1 L22 2 2URrear UL2 Usin δ2  R2 1 L22 2 2URUL2 Uδ2

Hence,

L22 2 f 2
R5
2UL2 Uδ2

However, we have also

L1
R5
δ1
Consequently,

δ2 L2 2 f 2 δ2 L2 2 f 2
5 2 and 5 2 ð4:8Þ
δ1 2UL1 UL2 γ ðL2 1f Þ2

where we used Eq. (4.7). Expression (4.8) describes how the trailer axle
steering angle should be linked to the kingpin angle γ to minimize the swept
path at low speed. We plotted this ratio in Figure 4.12 for various values of
L2 and f; this ratio was plotted for the same values for Rfront and wheelbase
L1 as in Figure 4.11. Clearly, small values of f (such as for compact cars)
require the largest trailer axle steering, which is consistent with the largest
off-tracking in Figure 4.11. Less steering is required for larger values of f
and smaller values of L2. Note that, for a conventional trailer, the kingpin
angle γ increases with L2 and f, cf. Equation (4.6), and therefore, a smaller
gain is required to minimize the off-tracking.

0.8

0.6
δ2 / γ

L2 = 2 [m]
0.4
L2 = 3 [m]
L2 = 4 [m]
0.2
L2 = 5 [m]
0
0 0.5 1 1.5 2
Coupling overhang f [m]

FIGURE 4.12 Optimal trailer axle steering gain with respect to kingpin angle.