Mechanical Systems and Signal Processing 25 (2011) 1877–1886

Contents lists available at ScienceDirect

Mechanical Systems and Signal Processing

journal homepage: www.elsevier.com/locate/jnlabr/ymssp

Analysis and simulation of a torque assist automated

manual transmission
E. Galvagno, M. Velardocchia, A. Vigliani 
Dipartimento di Meccanica - Politecnico di Torino, C.so Duca degli Abruzzi, 24, 10129 Torino, Italy

a r t i c l e i n f o abstract

Article history: The paper presents the kinematic and dynamic analysis of a power-shift automated
Received 26 February 2010 manual transmission (AMT) characterised by a wet clutch, called assist clutch (ACL),
Received in revised form replacing the fifth gear synchroniser. This torque assist mechanism becomes a torque
6 September 2010
transfer path during gearshifts, in order to overcome a typical dynamic problem of the
Accepted 22 December 2010
AMTs, that is the driving force interruption.
Available online 12 January 2011
The mean power contributions during gearshifts are computed for different engine
Keywords: and ACL interventions, thus allowing to draw considerations useful for developing the
Assist clutch control algorithms.
Automated manual transmission
The simulation results prove the advantages in terms of gearshift quality and ride
Power-shift transmission
comfort of the analysed transmission.
Torque gap filler
Drivability & 2011 Elsevier Ltd. All rights reserved.

1. Introduction

One of the most challenging issues for the automotive world in recent years has been the improvement of vehicles both
in terms of fuel efficiency and longitudinal behaviour. Since transmissions play a fundamental role for energy saving and
drivability, many researches have focused on enhancing the performance of existing systems and on developing new
technologies [1–6].
Among the several types of transmissions currently available, manual transmissions (MT) show the highest efficiency
value for any type of transmission (96%), while current production automatics (AT) have been improved to provide an
efficiency of about 86% and belt type continuously variable transmissions (CVT) have an overall efficiency of 85%, but their
major advantage consists in allowing the engine to operate most fuel efficiently (see e.g. [7,5,8]).
A recently developed power-shift automated transmission, i.e. the Dual Clutch Transmission, aims at optimising the
advantages of MT and AT (see e.g. Kulkarni et al. [3]), offering high efficiency and excellent shifting quality.
Automated manual transmissions (AMT) are generally constituted by a dry clutch and a multi-speed gearbox, both
equipped with electro-mechanical or electro-hydraulic actuators, which are driven by an Electronic Control Unit (ECU) [9].
In order to overcome the driving torque interruption that leads to undesired vehicle jerks during gear changes [6], different
devices called torque gap fillers (TGF) can be integrated in AMT driveline architectures. A solution has been developed by
Magneti Marelli Powertrain [10]: it mainly consists of an epicyclic gear-set to be added to a conventional AMT allowing to
transfer power from the engine to the secondary shaft during gearshifts.
Recently Hitachi Group [4] has proposed an alternative solution, consisting of a friction clutch mechanism – called
assist clutch (ACL) – that replaces the fifth gear synchroniser on traditional AMTs. The modulation of the assist clutch

 Corresponding author.
E-mail addresses: enrico.galvagno@polito.it (E. Galvagno), mauro.velardocchia@polito.it (M. Velardocchia), alessandro.vigliani@polito.it (A. Vigliani).

0888-3270/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ymssp.2010.12.014
1878 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886

Nomenclature Do angular speed difference

e error of the PID controller
Subscripts
J mass moment of inertia
J%
equivalent mass moment of inertia
c relative to the clutch kP ,kD ,kIPID controller proportional, derivative and
d relative to the driven clutch disk integrative gains
D relative to the driving clutch disk n angular speed (rpm)
e relative to the I.C. engine Pa available power
f relative to the final ratio Pd dissipated power
g relative to the gear R wheel radius
i ¼ I,II, . . . ,V relative to the selected gear ratio t time
in relative to the inertial terms T torque
L relative to the load Vveh vehicle speed
p relative to the primary shaft o angular speed (rad/s)
s relative to the secondary shaft o_ angular acceleration
syn relative to the synchroniser t ¼ oin =oout transmission ratio
w relative to the wheels tf final ratio

allows to shift gears smoothly, without interrupting the driving force [11]; moreover it is a compact and low cost solution
that requires relatively little modification to existing layouts.
This paper deals with this ACL-AMT transmission: the analysis will be focused on the system architecture and on the
dynamic behaviour.
In particular, the authors, after presenting the kinematic and dynamic model of the transmission, investigate and
quantify the power contributions for different engine and ACL interventions arising during various gearshifts in order to
highlight the relative weights between available, dissipated and effectively usable power for the vehicle propulsion; finally
some simulation results are reported.

2. Transmission layout and features

An ACL-AMT (see Fig. 1) consists of a traditional automated manual transmission with a servo-assisted clutch replacing
the fifth gear synchroniser. The system is electronically driven by a central unit that organises the controls of engine, main
clutch, gear selectors and assist clutch in order to optimise the gearshift.
The internal combustion engine (E) is linked through the main clutch (CL) to the primary shaft of the gearbox; at the
other end of the shaft, the ACL allows the connections between the fifth idle gear and the primary shaft (thus replacing the
synchroniser), while the secondary shaft and the differential do not present variations with respect to the classical design.
Such structure allows continuous power transmission between engine and wheels, since during the gearshift phase it is
not necessary to interrupt the coupling between the engine and the primary shaft in order to synchronise and to insert the
next gear. In fact this task is accomplished by the coordinated actions of both ACL and engine that allow to perform the
following phases: discharge of the engaged synchroniser dog-teeth, current gear pair disengagement, primary and
secondary shaft speeds adjustment in agreement with the next required gear ratio, shift of the synchroniser sleeve to
guarantee against the gear disengagement. All these phases take place while the permanent coupling with the secondary
shaft (through the fifth gear) allows a continuous power flow toward the wheels.

Fig. 1. Layout of an AMT with assist clutch (ACL).

 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886 1879

3. Kinematic analysis

The driveline can be described by means of three angular speeds: engine speed oe , primary shaft speed op and
secondary shaft speed os . When a gear is engaged, with the main clutch CL engaged and the assist clutch ACL disengaged
or under regulation, only one of the three velocities is independent, since it holds:

oe ¼ op ¼ os ti , ð1Þ

where ti is the gear ratio of the selected gear.

When the main clutch is slipping, the engine and primary shaft velocities can become different, i.e. oe aop , their values
depending on the dynamic equilibrium of the engine and of the gearbox.
In order to outline the potentiality of the proposed transmission to propel the vehicle through the ACL, let analyse the
angular velocities of its driving (subscript D) and driven (subscript d) discs at fixed gear ratios.
The following equations hold:

Vveh
oACLD ¼ op ¼ t t, ð2Þ
R f i

Vveh
oACLd ¼ os tV ¼ t tV : ð3Þ
R f

It is worth noting that in case the number of gear ratios is greater than five (as in some commercial versions of AMT),
subscript V has to be replaced with the highest available ratio, under the hypothesis to mount the ACL at that gear.
Fig. 2 plots the two speeds of the ACL plates versus the vehicle longitudinal velocity in the range 0–180 km/h for a given
set of gear ratios. The driven discs speed (nACLd ) does not change with the gear ratios and it is superimposed to the speed of
the driving discs (nACLD ) when the fifth gear is selected due to the fact that the ACL is engaged.
The figure allows to appreciate the favourable speed difference DnACL available between the two faces of the multi-disk
clutch for every kinematic condition, thus giving an idea of the possibility to transfer power from the primary to the
secondary shaft during the gearshift:

30
DnACL ¼ op ð1tV =ti Þ ðrpmÞ: ð4Þ
p
Obviously, the available ‘‘kinematic reserve’’ decreases for decreasing gear ratios, up to vanishing when the fifth gear is
engaged (it is worth recalling that the ACL clutch behaves as a synchroniser for the fifth gear). Let us consider for instance a
2nd to 3rd gearshift with the engine running at 3800 rpm (see Fig. 2). The speed difference that has to be eliminated during
the synchronisation process is indicated with DnII-III ; it is clear that only part of the total speed difference DnACL can be
effectively used in order to quickly perform the gearshift process.

5000

4500
1st 2nd 3rd 4th 5th
4000

3500
Δn II−>III
3000 ΔnACL
n [rpm]

2500
nACL
D
2000 nACL
d

1500

1000

500

0
0 20 40 60 80 100 120 140 160 180
Vehicle velocity [km/h]

Fig. 2. ACL velocities at fixed gear.

1880 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886

4. Dynamic analysis

The considered mechanical system shows a different number of degrees of freedom depending on the state of the main
clutch (engaged, disengaged/slipping) and on the state of the AMT (gear engaged, disengaged/under traditional
synchronisation). It requires therefore four sets of dynamic and kinematic equations to describe all of the possible
configurations.
In what follows only the two cases in which the starting clutch is engaged will be analysed, because they are the most
useful to explain the transmission capability.

4.1. One degree of freedom: gear engaged

The system, working as represented in Fig. 3, has a single d.o.f. For the sake of simplicity only the engaged gear pair is
considered, neglecting the resistant effect of the other idle gear pairs inside of the gearbox. The main clutch is engaged,
hence op  oe , while os poe through the gear ratio ti . This situation occurs during the vehicle normal cruise, when the
engine transfers power to the vehicle through the actual gear of the gearbox and eventually, if a clamping force is actuated,
also through the ACL, that is connected to the secondary shaft by means of the fifth gear.
With reference to the free body diagram in Fig. 4, the following dynamic equations can be derived:
Te ¼ TACL þTgi þ ðJe þ Jp Þo
_ e, ð5Þ

ti Tgi þ tV TACL ¼ TL þJL o_ s : ð6Þ

As stated, the system has a single d.o.f. (i.e. os ¼ op =ti ), so it holds:

Te ti ¼ TACL ðti tV Þ þ TL þ½JL þ ðJp þJe Þt2i o

_ s: ð7Þ

It is of interest noting that, since ti Z tV independently of the engaged gear ratio, the presence of the ACL torque has
always a resistant effect on the system dynamics and so a negative influence on the vehicle acceleration. The reason of the
driving torque reduction lies in the fact that part of the torque delivered from the engine reaches the secondary shaft
through the gear ratio tV which is smaller than the one associated with the gear currently engaged ti . Moreover, the
second power path toward the secondary shaft, performed through the ACL actuation, is characterised by a lower
efficiency, due to the dissipations in the clutch, proportional to the speed difference between the discs:
Pd ¼ TACL op ð1tV =ti Þ, ð8Þ

where Pd is the instantaneous power dissipated in the clutch.

Fig. 3. Scheme of the gearbox when the main clutch is engaged and a gear is engaged.

Fig. 4. Free body diagram when the main clutch is engaged and a gear is engaged.
 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886 1881

The transmission efficiency for a fixed value of the ACL torque decreases with the increase of the gear ratio ti , i.e.
passing from higher to lower gear.
In order to better understand the dynamics of the driveline, the torque transferred to the secondary shaft through the
engaged gear pair Tgi (see Fig. 4) can be computed for three specific values of TACL :
TACL ¼ 0-Tgi ¼ Te ðJe þJp Þo
_ e,

_ e -Tgi ¼ 0,
TACL ¼ Te ðJe þ Jp Þo

_ e -Tgi o0:
TACL 4 Te ðJe þJp Þo ð9Þ

The first value corresponds to the condition of disengaged ACL; hence the engine torque, reduced by the inertia torque
at the primary shaft, reaches the secondary shaft only through the gear engaged ti .
The second case corresponds to the ACL threshold torque for which the gearing ti does not transmit torque; therefore,
due to the absence of pressure between the wheels teeth, the gear can be disengaged without exciting torsional vibrations
in the driveline. Moreover, if the additional kinematic constraint op ¼ os tnext is satisfied, the engagement of the next gear
can be performed while the main clutch is engaged without significant shocks for the synchroniser.
Finally, in the third case, if the torque developed by the ACL is even larger, it causes a sign variation of Tgi ; the excess of
torque drawn from the secondary shaft causes a reduction of the vehicle acceleration.

4.2. Two degrees of freedom: gear disengaged

The system represented in Fig. 5 has two d.o.f.: the engine speed oe  op and the secondary shaft speed os . This
situation typically occurs during a gearshift transient, when no gear is engaged. The ACL and the engine can be used
together in order to synchronise the primary shaft, while the ACL gives torque to the secondary shaft filling the torque gap.
The dynamic equations for the primary and secondary shafts are
X Tsyn
Te TACL ¼ i
þ Jp o
%
_ e, ð10Þ
i
ti
X
Tsyni þTACL tV ¼ TL þ JL o
_ s, ð11Þ
i

where Tsyni represents the dynamic friction torque arising from the actuation of the i-th synchroniser sleeve in the
presence of a speed difference between the idle gear and the relative synchroniser ring.
It can be observed that if both the engine torque and the ACL torque are set to zero (i.e. TACL ¼ Te ¼ 0), the two dynamic
equations represent the typical ‘‘synchronisation phase’’ of an AMT, where the synchroniser friction torque causes the
dynamics coupling of the primary and secondary shaft. The only difference, that cannot be neglected, is that here the
primary inertia is increased by the engine contribution due to the fact that the main clutch is engaged; hence, in this
manoeuvre, the actuated synchroniser cannot operate in design conditions.
On the other hand, if the synchroniser torques are set to zero (i.e. Tsyni ¼ 0 8i), because of their small torque capacity,
they can be substituted by the engine and the ACL actions that, working together, can drive the primary shaft at a speed
suitable for the next engagement. Depending on the sign of the sum ðTe TACL Þ, the engine acceleration o _ e can be positive
or negative, thus allowing to speed up or down the primary shaft to obtain the desired synchronisation for downshift and
upshift, respectively. The synchronisers should be used just in the final phase in order to obtain the rigid engagement of
the involved gears (like a frontal teeth coupling). Moreover, it is of interest noting that the ACL also causes a torque
transfer to the wheels; consequently it has a torque gap filling effect even during the synchronisation phase.

Fig. 5. Scheme of the gearbox when the main clutch is engaged and no gear is engaged (for clarity’s sake only one synchroniser is represented).
1882 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886

5. Power contributions during gearshifts

In order to identify the best control strategy for this transmission it is necessary to understand from which sources and
in what proportions it is possible to draw energy for vehicle propulsion during a gearshift. Obviously a clutch cannot
introduce power in the system, but partially dissipates it and partially distributes it to another point of the transmission;
hence the only real sources of energy for torque gap filling are the engine and the kinetic energy stored in the powertrain
inertial components.
The power potentially available for the vehicle propulsion can be computed with the following generic formulation:

X
N X
M
Pa ¼ Te oe  Ji o
_ i oi  Tcj Docj , ð12Þ
i¼1 j¼1

where N is the number of the inertial components of the transmission, while M is the number of clutches. Before reaching
the wheels, the engine power is diminished by the power required for the acceleration of the powertrain inertial
components and by the losses in the clutches.
In particular Eq. (12) can be rewritten for this specific transmission and for the two d.o.f. configuration (with Tsyni ¼ 0 8i)
when the ACL is actuated; the power available at the secondary shaft is
Ps ¼ Te oe Jp o
_ p op TACL ðoD od Þ:
%
ð13Þ

This power is entirely transferred to the secondary shaft by means of the ACL, so it also yields:
Ps ¼ TACL od ¼ TACL os tV : ð14Þ

Starting from these preliminary considerations, in the following sections the intervention of ACL and engine will be
investigated, starting from the case in which only the assist clutch operates, using the stored kinetic energy; then
the combined action of engine and ACL will be presented. The primary shaft inertia, the engine inertia and the engine
torque contributions will be highlighted. In both cases the gear will be considered disengaged, since it is in this condition
that AMTs show a torque gap.

5.1. ACL with null engine torque

Considering Eqs. (10) and (11) describing the two d.o.f. configuration, it is simple to understand that the only way to
propel the vehicle is to actuate the assist clutch. If the engine torque is set to zero, the available energy at the secondary
shaft reduces to the fraction of the kinetic energy variation of the primary shaft not dissipated into the clutch.
Consequently Eq. (13) becomes

Ps ¼ Jp o
_ p op TACL DoACL :
%
ð15Þ

During an upshift manoeuvre, the secondary shaft speed can be supposed constant, due to the large vehicle inertia,
while the primary shaft changes its velocity adapting to the new gear ratio ti þ 1 . Thanks to the former simplification, the
kinetic energy variation that takes place during the gearshift manoeuvre is

DEk ¼ 12 Jp ðo2pi þ 1 o2pi Þ,

%
ð16Þ

where, obviously, at the beginning and at the end of the phase opi ¼ os ti and opi þ 1 ¼ os ti þ 1 .
Depending on the fact that the starting clutch is disengaged or engaged, parameter Jp can assume quite different values:
%

in the first case, Jp coincides with the primary inertia Jp, while when the main clutch is engaged also the engine crankshaft
%

and flywheel inertias must be considered (Jp ¼ Je þ Jp ). It is worth underlining that in practical applications this means that
%

the equivalent inertia can rise up to about 30–40 times the inertia of the primary shaft alone; consequently also the kinetic
energy variation is quite large. As an example, during an upshift from first to second gear with a primary shaft initial speed
of 4000 rpm, the kinetic energy variation is approximatively equal to 0.4 kJ when the clutch is disengaged, while it reaches
12.9 kJ if the clutch is engaged (values obtained for a medium-size sedan).
The next step is to evaluate the fraction of this kinetic energy variation that can be effectively transferred to the wheels
and the duration of the torque gap filling effect that can be guarantee by the inertial contribution. It is necessary to
consider the two d.o.f. dynamic equation (10) with null torque contribution from the synchronisers and from the engine:
jTACL jsgnðop os tV Þ þ Jp o
_ p ¼ 0:
%
ð17Þ

During an upshift, the primary has to slow down (i.e. o _ p o 0) to allow the following gear to be engaged at the end of the
gearshift and, as can be seen in Eq. (17), this task can be executed by the ACL torque as long as the kinematic inequality
op 4 os tV is satisfied.
Let consider a gearshift from gear i to gear i+ 1: the ACL has to vary the primary shaft speed from opi to
opi þ 1 ¼ ðti þ 1 =ti Þopi . Supposing that the torque given by the ACL (TACL ) is constant during the gearshift, then the expression
 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886 1883

of the transient duration Dt can be obtained integrating the differential equation (17):
 
Jp %
Jp ti þ 1 %

ðDti ÞUP ¼  ðDopi ÞUP ¼  1 opi , ð18Þ

TACL TACL ti
where (Dti ÞUP is the time necessary to eliminate the speed difference ðDopi ÞUP as required for passing to the next gear ratio
with constant torque application TACL through the ACL.
Assuming that TACL is the 50% of the maximum engine torque and opi ¼ 4000 rpm, it appears that the two cases
previously discussed lead to very different values of Dt: in the case of disengaged clutch Dt lasts only a few milliseconds,
while when the engine is connected to the primary shaft Dt  0:3 s and hence it is suitable to be used.
It is now possible to compute the average inertial (or synchronisation) power P in due to a constant ACL torque
application during an upshift:
 
DEk 1 t
P in ¼ ¼  TACL opi i þ 1 þ1 : ð19Þ
ðDti ÞUP 2 ti
It must be observed that the inertial power does not depend on the value of Jp and it is also independent of the clutch
%

engagement.
The mean net power available at the secondary shaft, thanks to ACL actuation, is
tV
P s ¼ TACL opi , ð20Þ
ti
while the mean power dissipated in the clutch is
 
1 t tV
P d ¼ TACL opi 1 þ i þ 1 2 : ð21Þ
2 ti ti
Fig. 6 shows the three mean power contributions for the various upshift manoeuvres as a function of the initial primary
speed npi . Passing from 1st-2nd to 4th-5th, it can be observed that

 the power Ps transferred to the secondary shaft increases due to gear ratio ti reduction;
 the inertial power Pin slightly increases because ratio ti þ 1 =ti increases;
 the power loss Pd reduces according to the slip reduction between the clutch discs.

All these power contributions are proportional to the initial primary speed.

1st −> 2nd 2nd −> 3rd

80 80
Pin
60 Pd 60
Power [kW]

Power [kW]

PS
40 40

20 20

0 0
0 2000 4000 6000 0 2000 4000 6000
n p [rpm] n p [rpm]
i i

3rd −> 4th 4th −> 5th

80 80

60 60
Power [kW]

Power [kW]

40 40

20 20

0 0
0 2000 4000 6000 0 2000 4000 6000
n p [rpm] n p [rpm]
i i

Fig. 6. Absolute values of the average powers in the transmission during the synchronisation phase obtainable using ACL alone.
1884 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886

Finally, it is possible to evaluate the torque Tw transmitted to the wheels:

Tw ¼ TACL tV tf , ð22Þ

which is obviously strictly correlated to the vehicle acceleration.

It is worth observing that the torque is independent of the vehicle speed and of the gear to be engaged; however, the
time during which the torque can be delivered (see Eq. (18)) is proportional both to the moment of inertia Jp and to the %

primary speed.
From the previous considerations, it can be stated that if the launch clutch is disengaged, then in practice the ACL is
useless due to the extremely reduced time for torque transmission. On the contrary, if the clutch is engaged then, due to
the additional inertia of engine and flywheel, the torque transmission during an upshift can last for the time necessary for
a fast gearshift even if the engine torque is set to zero. Since during a downshift the primary speed must increase in the
synchronisation phase, the usage of the ACL alone is not useful because its effect is to decelerate it; in this situation the
analysed transmission is not able to fill the torque gap.

5.2. ACL and engine

The system becomes more flexible and easily controllable if also the engine torque is used during the shift transient.
The power contributions P s and P d do not change their expressions (20) and (21), while the inertial power P in becomes
 
1 t
P in ¼ ðTe TACL Þopi i þ 1 þ 1 : ð23Þ
2 ti
This synchronisation power can now change from negative (release of accumulated kinetic energy during upshift) to
positive (storage of kinetic energy during downshift) if the engine torque is greater than the ACL torque. Therefore the
coordinated usage of engine and ACL allows to profitably manage both upshift and downshift.
It is possible to compute the torque difference DT between engine and ACL, both supposed constant in time, needed for
upshift and downshift in order to synchronise in time Dt the primary shaft speed:

Jp
%

DT ¼ os ðti þ 1 ti Þ: ð24Þ

Dt
As an example let suppose that a 1st-2nd upshift or a 2nd-1st downshift is accomplished at 70% of accelerator pedal
position and that the transmission does not modify the engine torque requested from the driver. Fig. 7 plots the engine and
clutch torques, supposed constant, that are required to satisfy a synchronisation transient duration of Dt ¼ 1 s. This map
can be used to compute the reference ACL torque for fixed vehicle speed and engine torque, so that the synchronisation
phase can be completed in time Dt.

300
T
e
|Te−T ACL|
250 T (Upshift 1st−2nd)
ACL
TACL (Downshift 2nd−1st)

200
Torque [Nm]

150

100

50

0
5 10 15 20 25 30 35 40 45
Vehicle Speed [km/h]

Fig. 7. Engine and ACL torques during 1st-2nd and 2nd-1st gearshifts for a transient Dt ¼ 1 s .
 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886 1885

6. Simulation results

The kinematic and dynamic equations are used to model the driveline that is inserted into a complete vehicle model in
order to simulate an acceleration manoeuvre with a series of upshift, from 1st to 4th, as depicted in Fig. 8.
These simulation results are obtained with a simple control of clutch and engine during a gearshift: more specifically,
the ACL follows an open loop trapezoidal torque profile while the engine torque is regulated by means of a standard PID
controller, based on the difference between the actual velocity and the next (desired) speed ðop Þnext of the primary shaft:
e ¼ ðop Þnext op , ð25Þ
Z
de
Treq ¼ TACL þkP eþ kI e dt þ kD , ð26Þ
dt
where Treq is the engine requested torque. When the angular speed error is sufficiently small, the next gear ratio can be
engaged by the shift of the corresponding synchroniser sleeve. The upshift manoeuvres start when engine speed crosses a
constant threshold of 3500 rpm.
In order to show the benefits of the presented ACL-AMT transmission, also a traditional MT is analysed: for the same
vehicle data and driver inputs, Fig. 9 plots engine torque, actual gear and vehicle acceleration of the manual transmission.

6 240
Actual Gear [−]

ACL Torque [Nm]

5 200

4 160

3 120
Acceleration [m/s2] ,

Engine Torque [Nm] ,

2 80

1 40

0 0

−1 −40
0 2 4 6 8 10 12
Time [s]

Fig. 8. Numerical simulation results for a vehicle equipped with ACL: full acceleration test.

6 240
Actual Gear [−]

5 200

4 160
Engine Torque [Nm]

3 120
Acceleration [m/s2] ,

2 80

1 40

0 0

−40
0 2 4 6 8 10 12
Time [s]

Fig. 9. Numerical simulation results for a vehicle with a traditional manual transmission: full acceleration test.
1886 E. Galvagno et al. / Mechanical Systems and Signal Processing 25 (2011) 1877–1886

Comparing the results shown in Figs. 8 and 9, it is evident that the torque gap is reduced and both dynamic
performance and comfort increase. In fact, thanks to the ACL actuation, during the gearshifts the vehicle acceleration does
not fall to negative values as for the MT, but maintains a medium value of about 1 m/s2 for all the synchronisation phases,
when no gear is engaged (actual gear =0 in Fig. 8). For a fixed value of the ACL torque, the contribution in terms of vehicle
acceleration is almost independent of the specific gearshift.

7. Conclusions

From the analysis of the AMT-ACL transmission it is possible to state that the assist clutch proves useful in the following
situations:

 upshift—during each gearshift, the ACL allows to unload the gears currently in mesh and to drive the primary shaft to
the new required speed while transmitting torque to the wheels; the synchroniser is used only during the final phase
and the main clutch is always engaged (apart when starting from standstill);
 downshift during acceleration (kick-down)—in a power downshift, the ACL can help in accelerating the vehicle by filling
the torque gap or at least reducing it, while synchronising the primary shaft together with the engine;
 motoring—the engine braking effect during accelerator pedal release can be improved by using the ACL intervention: in
fact, when the gear is engaged, due to the speed difference sign between its discs, the torque of the assist clutch has the
same sign of the engine friction torque and so the ACL can draw power from the secondary shaft, thus helping the
braking effect of the engine.

Finally, it is worth observing that during a downshift during braking manoeuvre there is no need to deliver driving
torque to the wheels, so it is preferable not to use the ACL and to perform the gear change process as in a conventional
automated manual transmission.

References

[1] S. Kasuya, T. Taniguchi, K. Tsukamoto, M. Hayabuchi, M. Nishida, A. Suzuki, H. Niki, AISIN AW new high torque capacity six-speed automatic
transmission for FWD vehicles, SAE Technical Paper, 2005-01-1020, 2005.
[2] H. Scherer, M. Bek, S. Kilian, ZF new 8-speed transmission 8HP70—base design and hybridization, SAE Technical Paper, 2009-01-0510, 2009.
[3] M. Kulkarni, T. Shim, Y. Zhang, Shift dynamics and control of dual-clutch transmissions, Mechanism and Machine Theory 42 (2007) 168–182.
[4] H. Kuroiwa, N. Ozaki, T. Okada, M. Yamasaki, Next-generation fuel-efficient automated manual transmission, Hitachi Review 53 (4) (2004) 205–209.
[5] A.F.A. Serrarens, S. Shen, F.E. Veldpaus, Control of a flywheel assisted driveline with continuously variable transmission, Journal of Dynamic Systems,
Measurement and Control 125 (3) (2003) 455–461.
[6] A. Sorniotti, E. Galvagno, A. Morgando, M. Velardocchia, F. Amisano, An objective evaluation of the comfort during the gear change process, in: SAE
International Congress, Detroit, 2007-01-1584, 2007.
[7] D.H. Park, T.S. Seo, D.G. Lim, H.B. Cho, Theoretical investigation on automatic transmission efficiency, electronic transmission controls, SAE, 960426,
1996.
[8] M.A. Kluger, D.M. Long, An overview of current automatic, manual and continuously variable transmission efficiencies and their projected future
improvements, SAE Technical Paper, 1999-01-1259, 1999.
[9] G. Lucente, M. Montanari, C. Rossi, Modelling of an automated manual transmission system, Mechatronics 17 (2007) 73–91.
[10] A. Sorniotti, Torque gap filler for automated manual transmissions: principles for the development of the control algorithm, in: SAE International
Congress, Detroit, 2009-01-0952, 2009.
[11] M. Yamasaki, H. Konno, H. Kuroiwa, N. Ozaki, Automated manual transmission with torque assist mechanism for reducing shift shock, SAE Technical
Paper, 2005-01-1783, 2005.