POTENTIAL BENEFITS OF THE MOTORCYCLE AUTONOMOUS EMERGENCY BRAKING THROUGH DETAILED IN-DEPTH CRASH RECONSTRUCTIONS This is an Author's Original Manuscript of an article submitted for consideration in Traffic Injury Prevention (copyright Taylor & Francis). Traffic Injury Prevention is available online at http://www.tandfonline.com/10.1080/15389588.2013.803280 Giovanni Savino Department of Industrial Engineering, University of Florence, Italy Via di Santa Marta 3, 50139 Florence (Italy)
[email protected] Federico Giovannini
[email protected] Niccolò Baldanzini
[email protected] Marco Pierini
[email protected] Matteo Rizzi Folksam, Sweden S 23, 106 60 Stockholm (Sweden) Department of Applied Mechanics, Chalmers University of Technology (Sweden)
[email protected]
KEYWORDS Autonomous braking, collision mitigation, crash, effectiveness, injury, motorcyclist
ABSTRACT
EC funded project PISa and further detailed in later
Research Question/Objective
studies. The boundary conditions of each simulation
The aim of this study was to assess the feasibility
varied within a range compatible with the uncertainty
and the quantitative potential benefits of the Motorcy-
of the in-depth data and included also a range of
cle Autonomous Emergency Braking (MAEB) in fatal
possible rider behaviours comprehensive of the actu-
rear-end crashes. A further aim was to identify possi-
al one. The benefits of the MAEB were evaluated by
ble criticalities of this safety system in the field of
comparing the simulated impact speed in each con-
PTWs (e.g. any additional risk introduced by the sys-
figuration (no ABS/MAEB, ABS only, MAEB).
tem itself).
Results
Methods and Data Sources
The MAEB proved to be beneficial in a large number
The study selected seven relevant cases from the
of cases. When applicable, the benefits of the system
Swedish national in-depth fatal crash database. All
were in line with the expected values. When not ap-
crashes involved car-following in which a non ABS-
plicable, there was no clear evidence of an increased
equipped motorcycle was the bullet vehicle. Those
risk for the rider due to the system.
crashes were reconstructed in a virtual environment
Discussion and Limitations
with Prescan©, simulating the road scenario, the
MAEB represents an innovative safety device in the
vehicles involved, their pre-crash trajectories, ABS
field of PTWs and the feasibility of such system was
and alternatively MAEB. The MAEB chosen as refer-
investigated with promising results. Nevertheless this
ence for the investigation was developed within the
technology is not mature yet for PTW application.
-1-
The research in the field of passenger cars does not
the AEB can enhance the benefits of other crash
directly apply to PTWs as the activation logic of a
protection systems (i.e. airbags for VRU) by reducing
braking system is more challenging on PTWs. The
collision speeds (Fredriksson and Rosén, 2010),
deployment of an autonomous deceleration would
which is the key point of integrated safety.
affect the vehicle dynamics thus requesting an addi-
The same approach is feasible for motorcycle safety
tional control action of the rider to keep the vehicle
and in 2009 a first prototype of motorcycle AEB (MA-
stable. In addition, the potential effectiveness of the
EB) was developed in the Powered Two Wheeler
MAEB should be investigated on a wider set of crash
Integrated Safety (PISa) project (Grant et al, 2008).
scenarios in order also to avoid false triggering of the autonomous braking.
INTRODUCTION Today motorcycles can represent an important alternative mean of transportation, due to the growing congestion in urban areas and the demand for more time and energy efficient transports. However, previous research has showed that in case of a crash motorcycle riders have much higher injury risks than passenger car occupants (Gabler, 2007). Although Figure 1: Prototype vehicle of the PISa project equipped with
innovative protective equipment for motorcyclists (i.e.
MAEB (in the box: laser scanner mounted in the front fairing).
airbag jackets) has been developed with promising results (e.g. Pellari, 2012), impact speeds are often
This system can detect an obstacle through a laser
too high to prevent severe injuries (Rizzi et al, 2009).
scanner located in the front of the motorcycle, warn
The approach for reducing health-loss due to road
the rider and brake automatically to produce a 0.3 g
crashes is mostly known in the automotive safety
medium deceleration (autonomous braking, AB), if
area as integrated safety: the pre-crash and crash
the rider does not react before the collision becomes
phases (Haddon, 1980) are no longer considered
unavoidable. However, if the rider does brake after
separately but rather interact with each other
this point, maximum braking force is automatically
(Strandroth et al, 2012). With this approach the entire
applied - enhanced braking, EB (Savino et al,
chain of events leading to a crash can be understood
2012a).
and possible countermeasures can be analyzed in
In the present research the MAEB architecture was
order to either prevent crashes or mitigate injuries
reproduced as defined in the PISa project, thus in-
when the crash in no longer avoidable (Strandroth
corporating a standard anti-lock braking (ABS) which
2012).
operated also when the AB was not triggered. The
Following this approach, the development and im-
characteristics of the MAEB are summarized in Table
plementation of Autonomous Emergency Braking (AEB) on passenger cars is well on-going (Eu-
Table 1: Characteristics of the MAEB.
roNCAP, 2012). Most of these systems are designed
Description
Value
to avoid and mitigate car to car rear-end crashes
Detection range
200 m
and/or crashes with pedestrians and cyclists. As-
Detection angle
100°
Detection roll max
sessments based on real-life crashes in USA and Sweden have shown impressive results (HLDI, 2011;
AB reference deceleration
Isaksson-Hellman and Lindman, 2012). In addition
Minimum obstacle width
-2-
8° 3 m/s2, 9 m/s2 (EB) 1m
1.
ness statements from the police and reports from the
Previous studies have retrospectively evaluated the
emergency services. The chain of events leading to
potential of MAEB in real-life crashes through simple
the fatal crash is analyzed and critical events are
analysis of in-depth material (Rizzi, 2011; Savino et
identified; collision speeds are generally derived by
al, 2012b; Strandroth, 2012) showing promising re-
vehicular deformation, while initial driving speed are
sults. Further research is needed to understand more
mostly based on witness accounts, brake skids etc.
deeply the true benefits of MAEB in different crash
As all fatal crashes are included in the sampling crite-
scenarios. Therefore the present study focused on a
rion the material is fully representative for Swedish
specific crash type as an additional step to evaluate
conditions and possibly for northern countries at
MAEB. A sensitivity analysis of the potential benefits
large.
of the MAEB was conducted taking account of factors such as road scenario, dynamical state of the vehi-
Case selection and description
cles involved and rider’s control actions. Therefore
This study focused on rear-end crashes in which the
the use of advanced simulation software was consid-
motorcycles was the bullet vehicle. A typical car-
ered more suitable than a method based on engi-
following scenario can be described as the motorcy-
neering judgments by experts.
cle travelling on a straight path when the lead vehicle (travelling in the same direction) slows down or stops,
Study Objectives
i.e. to turn into a lateral road or because of other
In synthesis, the objectives of the study are the fol-
traffic ahead. Only crashes in which the motorcyclist
lowing:
sustained the deadly injuries in the rear-end collision
further assess the feasibility and the quantitative
itself were included; in other words, crashes involving
potential benefits of the Motorcycle Autonomous
deadly injuries in further secondary collisions (i.e.
Braking (MAEB) with special focus on fatal rear-
other than the initial impact with the lead vehicle)
end crashes;
were excluded from the analysis. In total, 7 fatal
identify possible criticalities of the MAEB in the
crashes involving car-following were identified for the
field of PTWs in fatal rear-end crashes, e.g. any
period 2006-2009 in Sweden. During the same peri-
additional risk for the rider due to the system it-
od, 213 motorcyclists were killed according to Swe-
self.
dish official statistics.
METHODOLOGY
The time of the precipitating event was then defined
Material
as the time at which the major contributory factor took
The present study used in-depth studies of fatal
place, even in addition to other secondary although
crashes collected by the Swedish Transport Admin-
relevant contributory factors (i.e. lead vehicle started
istration (STA). STA has been carrying out in-depth
to heavily brake to avoid a pedestrian or another
studies for each fatal road crash since 1997. Crash
vehicle crossing the intersection with red light). Reac-
investigators at STA systematically inspect the vehi-
tion times, braking distances and other factors (i.e.
cles involved in fatal crashes and record direction of
swerving, driving directions and speeds, etc.) were
impact, vehicular deformation, airbag deployment,
then derived starting from the precipitating event for
tire properties, etc. The crash site is also inspected to
each of the 7 relevant cases mentioned above. All
investigate road characteristics, collision objects,
crashes occurred in daylight, with clear sky and dry
braking skids etc. Further information about injuries
asphalt. All riders did use helmets and were not un-
and use of helmets and protective equipment is pro-
der the influence of alcohol or illegal drugs. No ABS-
vided by forensic examinations, questioning and wit-
equipped motorcycles were involved in these crash-
-3-
es. A brief description of these cases is given below
The lead vehicle (passenger car) stopped at an inter-
and a summary of the cases is provided in Table 2.
section to turn into another rural road and waited for
Case A
some oncoming motorcycles. A motorcycle with two
The lead vehicle (passenger car) slowed down and
riders (males, 71 and 14) following from behind at-
stopped at an intersection to turn left into another
tempted to swerve and crashed in the right rear of
rural road. While waiting for 7 oncoming motorcycles,
the car. The motorcycle driver sustained fatal head
the passenger car was hit from behind by a motorcy-
and thorax injuries.
cle. The rider (male, 66) sustained fatal internal
Case F
bleeding. According to witness accounts the killed
After turning wrong out of a roundabout in an urban
motorcycle rider may have been distracted by the
area, the lead vehicle (passenger car) made a U-turn.
oncoming group of motorcycles.
A motorcycle following from behind hit the left side of
Case B
the car. The rider (male, 21) sustained fatal head
The lead vehicle (light truck) stopped to turn left into
injuries.
a secondary rural road and waited for six oncoming
Case G
motorcycles. A motorcycle travelling along the lane of
The lead vehicle (passenger car) was following an-
the truck hit the rear of the truck. The rider (male, 44)
other car and a bus on a rural road. When the bus
sustained fatal thorax injuries; he might have been
slowed down before a bus stop the two following cars
distracted by the oncoming group of motorcycles.
braked too. The motorcyclist (male, 77) over-braked
Case C
and partly slid into the rear of the lead vehicle, sus-
The lead vehicle (light truck) was about to turn left
taining multiple fatal injuries.
into a private road. A motorcyclist (male, 71) coming from behind may have understood this too late and
Simulations
attempted to swerve while braking. He lost control
Part 1. MAEB vs. Nothing
and sled into the side of the truck sustaining fatal
The seven crashes were reconstructed with a numer-
thorax injuries.
ical tool (PreScan©) in order to perform a sensitivity
Case D
analysis of each case. Such an analysis was intend-
The lead vehicle (passenger car with trailer) was
ed to provide a) a more robust data set for the quanti-
about to turn left into a driveway of a private property.
tative assessment of the MAEB performance com-
A motorcyclist (male, 39) collided from behind in the
pared to the single-shot reconstruction and b) a pre-
trailer's left side. He sustained fatal thorax injuries.
liminary investigation of the importance of various
Case E
factors, which have an influence on the crash outTable 2: description of the analyzed cases (MY: model year; MC: motorcycle) Make, model, MY
Speed limit (km/h)
MC braking before crash
Brake skid (m)
Initial speed (km/h)
Collision speed (km/h)
Impact configuration
Case A
Triumph Thunderbird -03
90
no
-
90
90
Upright
Case B
Suzuki DL1000 -06
70
yes
13
100
65-70
Upright
Case C
Suzuki GSX750 -83
70
yes
17
80
45-50
Case D
Yamaha Teneré 750 -98
50
yes
0
75
60
Case E
Muz Scorpion S -97
70
yes
0
70
50
Case F
Yamaha R6 -99
70
yes
0
105
70-75
Upright
Case G
Suzuki Intruder 1500 -98
70
yes
20
70
45-50
Upright
-4-
Fallen on right hand side before crash Upright, with roll angle Upright, with small roll angle
Table 4: Assumed boundary conditions
Table 3: parameters and associated levels used in the DOE of each crash (percentages are referred to the nominal value of the
Trajectory PTW
Opponent vehicle manoeuvre
Host PTW manoeuvre
Case 1
Straight
Still
None
Case 2
Straight
Still
Hard braking (80f, 60r [bar])
Case 3
Straight
Moving (left turn into lateral street)
Case 4
Impact point out of a slight curve
With trailer, moving (left turn into lateral street)
Case 5
Straight
Still
Case 6
Straight
Performing U-turn
Medium braking (50f 0r [bar])
Case 7
Straight
Decelerating till stop
Medium braking (40f 60r [bar])
variable). Parameter Initial PTW lateral position Initial PTW speed Rider reaction timing
-1/0/1 level -0.5 / 0 / +0.5 [m] 80% / 100% / 120% 0.3 / 0.8 / 1.3 [s]
come. Each case was characterized in detail by the following information:
road scenario: in-scale geometry of the road at the crash site, reconstructed from satellite maps, and type of infrastructure, whenever of relevance;
Hard braking (80f, 80r [bar]) plus right swerving (10° target roll) Poor braking (40f, 40r [bar]) plus left swerving (-10° target roll) Medium braking (30f, 30r [bar]) plus right swerving (10° target roll)
vehicles involved: type and their initial state at precipitating event in terms of position, heading, speed, acceleration, and manoeuvres executed
in nominal conditions (i.e. in the situation of the crash
during the event.
reproduced using the original in-depth information)
The motorcycles used in the simulations were gen-
the rider manoeuvre was initiated 0.8 s after the per-
eral models reasonably similar to the actual ones.
ception of the impending crash. In case of early or
The realistic dynamic behaviour of the motorcycles in
late reaction the rider manoeuvre is anticipat-
the simulations was obtained by using a detailed
ed/delayed of 0.5 s. The variable levels used in the
multi-body vehicle model generated by the simulation
DOE are reported in Table 3. In all cases the output
tool bikeSIM©, and coupled with PreScan© in a
parameter was the impact speed, since it is related to
Matlab Simulink© environment.
the energy dissipated in the impact and to the injuries
A Design of Experiment (DOE) of 3 parameters vary-
of the rider.
ing on 3 levels, using a full factorial scheme, was
A total of 54 simulations (3 variables, 3 values per
applied to each case in order to thoroughly explore
variable, 2 systems) were performed for each case.
variations of the nominal cases, and to quantitatively assess the influence of the parameters on the speed
Part 2. MAEB (w/ and w/o EB) vs. ABS
reduction. A 3-level DOE was chosen due the ex-
In addition to the above mentioned assessment, two
pected non-linear relation of the parameters with the
additional comparisons were performed: MAEB vs.
speed reduction produced by MAEB. The generated
ABS, and pure emergency braking function (i.e. MA-
sets of crash scenarios were simulated both in the
EB with enhanced braking function deactivated) vs.
real conditions (with no assistance system installed
ABS. In these comparisons the full set of crashes
on board of the motorcycle) and with the MAEB.
was considered in nominal conditions and the varia-
Again, the functionalities of the MAEB were de-
bility was introduced only on the rider behaviour.
scribed in Savino et al (2012a).
Apart from the nominal manoeuvre, the rider was
The parameters used in the DOE were the initial
considered able to perform:
PTW lateral position on the road, the initial PTW
speed, and the rider’s reaction time. The quantifica-
panic heavy braking: 80 bar pressure applied to master cylinder both of the front and rear brake;
tion of the latter parameter was done considering that
-5-
speed reduction was respectively 0.2 m/s, 1.0 m/s, Table 5: Summary of the results for the seven cases in the
1.9 m/s and 2.9 m/s.
actual configurations Impact speed reduction (m/s) 1.8
MAEB activation mode pure AB
In case C and D the MAEB did not deploy due to the
swerving attempt. In those cases the ABS functionali-
Case A
Impact speed (m/s) 24.9
Case B
20.3
0.2
EB
Case C
13.2
crash avoided
ABS
Case D
15.2
0.1
ABS
ty was engaged, producing the complete collision
Case E
14.4
1.0
EB
avoidance in case C and an impact speed reduction
Case F
18.0
1.3
EB
Case G
13.7
2.9
EB
inhibition given by the roll angle produced by the
of 0.1 m/s for case D. A synthesis of the results in the actual configurations is provided in Table 5. A diagram showing the impact speed vs. the impact speed
poor braking: 40 bar pressure applied to master
reduction is given in Figure 2. In the diagram, the
cylinder both of the front and rear brake.
theoretical impact speed reduction is plotted in case
The above mentioned pressure values were derived
of fixed obstacle and a) no rider reaction, b) medium
from experimental trials conducted with fully instru-
braking applied at MAEB triggering instant.
mented PTWs, which allowed to link the deceleration with the brake pressures. In case a composite avoid-
Modified configurations
ance manoeuvre resulted from the in-depth database
In most of the crash cases included in the study the
(i.e. swerve plus braking), the swerve action was still
collision with the MAEB off did not take place in the
coupled to the braking behaviours previously de-
modified configurations at Lower Speed and Early
scribed. The latter two braking avoidance manoeu-
Reaction timing (LSER configurations). Those cases
vres are typical of PTW-car crashes (Penumaka
are case B, D, E and F.
2013).
Case A
The assumed boundary nominal conditions are
In all the 26 modified configurations the collision took
summarized in Table 4. A total of 9 simulations (3
place. The MAEB did trigger in pure AB mode, pro-
braking behaviours, 3 systems) were performed for
ducing an impact speed reduction ranging from 1.4
each case.
m/s (at higher speed and with lateral offset on either side) to 1.8 m/s (actual timing, higher speed and
RESULTS
actual lateral position).
A total number of 378 simulations were run for Part 1
Case B
and 84 for Part 2. The results of Part 1 (MAEB vs.
Concerning case B, the collision with MAEB turned
nothing) and Part 2 (MAEB vs. ABS) will be present-
off did not take place in the LSER configurations
impact speed reduction (m/s)
ed in the following paragraphs.
Part 1. MAEB vs. Nothing Actual configurations Starting from the actual configurations of the crash cases, the results of the simulations showed that the MAEB deployed in five out of seven cases: case A, B, E, F and G. In case A where the rider did not at-
Impact speed reduction (pure AB) Impact speed reduction (EB) theoretical speed reduction due to pure AB theoretical speed reduction due to EB (medium braking 6 m/s2)
5 4 3 2
1 0
tempt any braking, the pure AB functionality was
10
activated, producing an impact speed reduction of
15 20 25 impact speed w/o MAEB (m/s)
30
Figure 2. Impact speed reduction (m/s) due to MAEB in the actual configuration. Theoretical speed reduction curves considering no braking and medium braking (6 m/s2).
1.8 m/s. In the other four cases the EB triggered due to the rider’s braking action and the final impact
-6-
only. When the MAEB was on, in those configurations where the collision took place the EB did trigger producing an impact speed reduction ranging from 0.2 m/s (actual configuration) to 2.9 m/s (late reaction, higher speed, lateral offset on the left hand side). Case C In 9 out of 26 modified configurations the PTW collided with the lead vehicle in upright position (in two of those cases with no roll angle). In the remaining con-
a) Case A
figurations the PTW fell down before the impact, similarly to the actual configuration. The AB or EB did not trigger in any of the modified configurations with the MAEB on. The ABS avoided the crash in 17 out of 26 configurations. In those configurations where the collision took place, the impact speed reduction due to the ABS ranged from 0.0 m/s (actual timing, higher speed and actual lateral position) to 1.5 m/s (late reaction, higher speed, actual lateral position).
b) Case B
Case D The collision with the MAEB turned off did not take place in the LSER configurations only. When the MAEB was on, the triggering of the EB took place in three configurations: higher speed and late reaction timing (irrespectively of the lateral position). In those configurations, the impact speed reduction ranged from 1.2 m/s (lateral offset on the right hand side) to 2.3 m/s (actual lateral positioning). The ABS functionality avoided the crash in three
c) Case F
configurations with early reaction of the rider, two of them at the actual speed and one at higher speed. The impact speed reduction of the ABS ranged from 0 up to 2.3 m/s (late reaction, actual speed, lateral offset on the left hand side). Case E With the MAEB turned off the collision did not take place in the LSER configurations only. In the remaining configurations, with the MAEB on the EB triggered in 16 simulations, corresponding to
d) Case G
all the configurations with late reaction and all the Figure 3. Analysis of effects of experiment variables on mean impact speed reduction (m/s) due to MAEB for cases A, B, F and G.
configurations with actual timing except for those at low speed. The impact speed reduction ranged from
-7-
0.4 m/s (current timing, lower speed, actual position)
ality was largely involved. The plots of the remaining
to 2.3 m/s (later reaction, lower speed).
cases are included in Appendix.
The ABS functionality could avoid two collisions and
A diagram showing the impact speed vs. the impact
in the remaining configurations the impact speed
speed reduction for all the configurations including
reduction ranged from 0 m/s to 0.9 m/s (actual timing,
the actual ones, neglecting those where the impact
lower speed, lateral offset on the right hand side).
was completely avoided, is given in Figure 4. In the
Case F
diagram, the theoretical impact speed reduction is
With the MAEB turned off the collision did not take
plotted in case of fixed obstacle and a) no rider reac-
place in the LSER configurations only.
tion, b) medium braking applied at MAEB triggering
In the remaining configurations, with the MAEB on
instant.
the EB triggered in all the configurations and the impact speed reduction ranged from 0.1 m/s (late
Part 2. MAEB (w/ and w/o EB) vs. ABS
reaction, lower speed, offset on the right hand side)
In all the nominal conditions and in all the additional
to 6.3 m/s (early reaction timing, higher speed, offset
configurations obtained with different braking actions
on the right hand side).
(actual, panic and poor braking) the impact speed
Case G
with the MAEB on was equal or lower than with the
With the MAEB off the collision did not take place in
sole ABS on. The diagrams showing the impact
the LSER configurations only.
speed reduction due to the MAEB in comparison with
In all the remaining configurations when the MAEB
the ABS are depicted in Figure 5. A distinction is
was on, the EB triggered producing an impact speed
provided for the pure AB functionality (MAEB without
reduction in the range between 1.2 m/s (late reaction,
EB functionality) and the full MAEB.
higher speed and lateral offset on the left hand side)
The simulations also showed that the sole ABS
and 5.1 m/s (early reaction, actual speed, actual lat-
would have avoided case C crash for any of the three
eral position).
analysed rider braking actions, whereas neither the
Taguchi analysis
ABS nor the MAEB would have significantly affected
For each case, a Taguchi analysis was performed to
the outcomes for case D.
investigate the main effects of the three experiment variables (reaction timing, speed and lateral position).
DISCUSSION
In Figure 3 the plots are given for the main effects of
The MAEB, consisting of ABS, AB and EB function-
cases A, B, F and G in which the AB or EB function-
alities, was relevant in all the 7 crash cases of car following considered in the study and the AB/EB trig-
impact speed reduction (m/s)
5
Case A Case D Case G
4
Case B Case E pure AB
gered in 5 cases. The MAEB is supposed to trigger
Case C Case F EB med. brake
once the collision becomes physically unavoidable either by braking or swerving. Since at high speed
3
the swerve manoeuvre is more efficient than braking 2
(Giovannini 2013), the MAEB activation is delayed
1
until even the swerve action would not avoid the crash, thus limiting the potential impact speed reduc-
0 10
15 20 25 impact speed w/o MAEB (m/s)
tion with respect to lower speed values. The simula-
30
tions indicated an impact speed reduction due to MAEB compatible with the theoretical values in 4 out
Figure 4: Impact speed reduction (m/s) due to MAEB (ABS+AB+EB) in the actual configuration. Theoretical speed reduction curves considering no braking and medium braking (6 m/s2) are indicated.
of 7 cases (see Figure 2). In a fifth case, where the AB/EB was triggered (case B), the benefits were
-8-
minor mainly due to the hard braking action of the
tion. In none of the cases the presence of MAEB and
rider associated to an early reaction (as the crash
its intervention on the vehicle dynamics produced
reconstruction indicated).
evidence of additional threat. No falling events were
In the two cases in which the AB/EB did not trigger
reported with the MAEB turned on.
(case C and D), the system deployment was inhibited by the rider’s action of swerving. In case C the crash
The analysis of the modified configurations, created
was avoided due to the ABS functionality. In another
to assess the robustness of the system to the possi-
case where the rider tried to swerve (case E) the
ble variations of the initial conditions, showed that the
AB/EB first triggered and then turned off due to the
impact speed was never augmented by the MAEB with respect to the standard vehicle. Moreover, no
impact speed reduction (m/s)
3,0
falling events were reported in any of the modified MAEB MAEB (no EB)
2,5 2,0
configurations when the MAEB was on. Concerning the effects of the MAEB in terms of impact speed reduction in the modified configurations,
1,5
positive and negative variations with respect to the
1,0
effects of the MAEB in nominal conditions were re-
0,5
ported depending on the case and on the modified
0,0
A
B
C
D
E
F
variables. The range passed from 0-3 m/s in the ac-
G
tual configurations to 0-4 m/s in the modified configu-
a) Actual behaviour
rations. Globally the values were in the range of impact speed reduction (m/s)
3,0
those estimated with the theoretical model (Figure 4).
2,5
MAEB MAEB (no EB)
2,0
The analysis of the main effects (Figure 3) demonstrate the strong influence of the parameter variation on the MAEB speed reduction and generally confirm
1,5
the non-linear nature of the phenomenon. However
1,0
the data do not allow to draw more general conclu0,5
sions, since few similarities can be identified, thus
0,0
A
B
C
D
E
F
highlighting the relevance of the specific characteris-
G
tic of each crash case. As a consequence an earlier
b) Panic braking
or late timing can either produce a lower or higher impact speed reduction (m/s)
3,0
final impact speed in correlation with the behaviour of
2,5
MAEB MAEB (no EB)
2,0
the opponent vehicle. Similar result is obtained applying different initial speeds and initial lateral positions. The analysis of the interaction plots shows a consid-
1,5
erable mutual influence only between timing and
1,0
speed for the cases B, C, F, and G, while relevant
0,5
interactions among all the parameters are reported
0,0
for the remaining cases. The effects of the interacA
B
C
D
E
F
G
tions are not uniform through the different cases. In
c) Poor braking
conclusion, it is not possible to identify general rules Figure 5: Impact speed reduction due to MAEB in comparison with the sole ABS, considering three braking actions: actual, panic and poor.
although at the same time the MAEB showed to be globally robust with respect to the possible variations in the initial configurations as far as the specific crash
roll angle of the PTW, producing impact speed reduc-
-9-
cases included in the present study concern. As a
by AB (in case of no reaction) and EB (in case the
consequence, further validation studies of the MAEB
rider operated the brakes). Moreover, the positive
could even broaden the variability range of the pa-
effects in terms of speed reduction are expected to
rameters, but they should consider the use of a ran-
be more evident in case of poor braking actions ra-
dom sampling scheme to explore the domain of pos-
ther than in panic hard braking conditions.
sible scenarios generated from nominal scenarios. The present study also illustrates a critical aspect of The results of Part 2 of the study (MAEB vs. ABS)
the MAEB in car following crash scenarios, being the
showed the influence of the AB/EB and pure AB on
possibility of a swerve attempt operated by the rider
the impact speed reduction operated by the MAEB in
prior to collision. In case of swerving, the MAEB is
comparison with the sole ABS functionality. In case
supposed to be inhibited to avoid vehicle destabiliza-
A where the rider did not react before the impact the
tion, according to the decision logic of the system
effects are only due to the pure AB. In the remaining
(Savino et al 2012a). The cases C and D showed
cases the pure AB did not produce a relevant speed
unsuccessful attempts of swerving where the roll
reduction in comparison with the ABS.
angle values above the threshold of 10° inhibited the
In case B in the actual configuration, the full MAEB
MAEB triggering. As a consequence, in those cases
slightly influenced the impact speed and the same
the potential speed reduction operated by the EB
result was obtained in the ‘panic braking’ modified
functionality would not have been exploited, even if
configuration. In fact the rider operated a hard brak-
the roll angle was slightly above the threshold. Fur-
ing. When considering the modified configuration of
ther consideration should be given in the future to
‘poor braking’, the MAEB would result to be more
redesign the inhibition criteria in order to let the cases
effective in terms of impact speed mitigation.
of ineffective swerve attempts potentially benefit of
Case C and D were only affected by the ABS as pre-
the MAEB.
viously mentioned and the MAEB did not deploy in the modified configurations of ‘panic braking’ and
Limitations
‘poor braking’.
While the results of the present study seem promis-
Case E, F and G were influenced mainly by the EB
ing, caution is needed due to a number of limitations.
functionality. For those three cases, the ‘panic brak-
First of all, only fatal crashes from Sweden were in-
ing’ configuration would reduce the effects of the EB
cluded in this study. The analyzed cases were a fully
and particularly in case E. Differently, in modified
representative sample for Sweden, possibly for
configurations of ‘poor braking’, the effects of the
Northern Europe at large, however it could be argued
MAEB would be higher in case E with respect to the
that motorcycling is somewhat different from other
actual configuration.
countries due to the climate in those areas. Therefore the analyzed crashes could differ from crashes in
This study confirms the applicability of the MAEB in
other European areas.
real crash cases in car following scenarios, showing
General caution is also needed when only fatal
possible impact speed reduction in a range of initial
crashes are analyzed as other crashes with lower
conditions. The modified initial conditions could influ-
injury outcomes would probably differ in terms of
ence the final speed reduction operated by the MA-
riding and impact speed ranges. Furthermore, the
EB, although they did not prove to influence the risks
chain of events leading to a fatal motorcycle crash is
associated with the deployment of the MAEB itself. It
probably different from the one with milder injury
was also showed the positive effects of the ABS,
outcomes which means that crash and injury mecha-
although relevant speed reduction is operated both
nisms (braking vs no braking, upright vs prone, etc.)
- 10 -
would probably be different too. Future research
No additional threat was produced by the MAEB in
should address this issue.
terms of vehicle destabilization, rather in none of the cases the PTW fell before the collision when the MA-
A further limitation is due to the material source.
EB was on. The MAEB did not trigger in 2 cases, due
While previous studies have shown that STA’s mate-
to the swerve manoeuvre performed by the rider prior
rial is sufficiently reliable for in-depth analysis
to collision. This represents a critical aspect of the
(Strandroth 2012), it should be noted that most of the
MAEB in car following scenario and in the future it
input values are based on post-crash investigations
should be addressed in order to widen the applicabil-
and not on real-time measurement. For instance,
ity of the MAEB to all the cases of a PTW crashing
collision speeds were not measured with crash pulse
into the lead vehicle.
recorders; vehicle trajectories and driver reactions
The study also simulated modified configurations of
were not recorded either. While input errors would
the respective crash cases to investigate the robust-
clearly affect the simulations and the overall results,
ness of the MAEB applicability to variations in the
an attempt to control for this uncertainty was made by
initial conditions. The simulations highlighted small
analyzing the variation of collision speed depending
variations in the impact speed reduction operated by
on three of the most important parameters (initial
the MAEB in the different configurations, although the
PTW lateral position and speed, rider reaction tim-
peculiarity of each case did not allow for the identifi-
ing). The results showed a reasonable variation of
cation of general rules with respect to the changing
collision speeds, thus suggesting that the simulations
variables.
were robust and consistent.
In the study, simulations conducted with simple ABS, pure AB and full MAEB with three rider braking be-
Finally, at the current state of the knowledge the ex-
haviours showed that the EB functionality could re-
pected values of impact speed reduction due to MA-
duce the impact speed with respect to the ABS (the
EB cannot be correlated with a reduction in the injury
effects were more evident in case of poor braking
outcomes for the rider, yet. In the future, the creation
manoeuvres), while the pure AB functionality affected
of specific motorcycle rider risk curves, similar to
only the case where the rider did not brake prior to
those created for passenger cars, will allow the esti-
collision.
mation of the potential benefits of MAEB in terms of
In conclusion, although the small number of crash
health loss reduction in car following scenarios.
cases and all the given limitations impose cautiousness in the generalization of the results, the new
CONCLUSIONS
material added for evaluating the benefits of the MA-
The study presented the results of virtual reconstruc-
EB in real crash scenarios confirms that the MAEB is
tions of 7 fatal PTW crashes occurred in Sweden and
a potential candidate device to enhance rider safety.
belonging to the car following crash scenario. The simulations were run with and without the MAEB
ACKNOWLEDGMENTS
system to investigate the potential benefits of this
Many thanks to Johan Strandroth at the Swedish
safety device in terms of impact speed reduction. The
Transport Administration for providing access to the
study confirmed the general applicability of the MAEB
in-depth data.
in the car following scenario: the MAEB triggered in 5 out of 7 cases. The impact speed reduction values
REFERENCES
were in the expected range and the maximum value
EuroNCAP, http://www.euroncap.com/ re-
was 2.9 m/s.
sults/aeb/survey.aspx, Accessed 2012-10-02.
- 11 -
Fredriksson R, Rosén E. Integrated pedestrian countermeasures - potential of head injury reduction com-
Rizzi M. The potential of AEB in fatal motorcycle
bining passive and active countermeasures. Pro-
crashes. 55 AAAM Student Research Symposium.
ceedings of the 2010 IRCOBI Conference, Hannover,
Paris, France, 2011.
th
th
th
Germany, 15 –17 September 2010. Rizzi M, Strandroth J, Tingvall C. The effectiveness Gabler HC, The Risk of Fatality in Motorcycle Crash-
of antilock brake systems on motorcycles in reducing
es with Roadside Barriers. Proceedings of the 20th
real-life crashes and injuries. Traffic Injury Preven-
ESV Conference, Lyon, France, paper number 07-
tion, 10(5):479–87, October, 2009.
0474, 2007. Savino G, Pierini M, Baldanzini N. Decision Logic of Giovannini F, Savino G, Pierini M., Analysis of the
an Active Braking device for Powered Two Wheeler
minimum swerving distance for the development of a
application. Proc IMechE Part D: J Automobile Engi-
motorcycle autonomous braking system, Accident
neering. 226(8):1026-1036, August, 2012a.
Analysis and Prevention, forthcoming. Savino G, Pierini M, Rizzi M, Frampton R. Evaluation Grant R, Frampton R, Savino G, Pierini M. PISa –
of an Autonomous Braking System in Real World
Powered two-wheeler Integrated Safety. Project ob-
PTW Crashes. Traffic Injury Prevention 2012b, article
jectives, achievements and remaining activities. Pro-
in press.
ceedings of the 7th International Motorcycle ConferStrandroth J. A Method to Identify Future Potential of
ence, Cologne, Germany.13:319-340, 2008.
Vehicle Safety Technology. Thesis for the degree of Licentiate in Machine and Vehicle systems, Chalmers
Haddon W Jr. Advances in the epidemiology of inju-
University of Technology, Gothenburg, Sweden,
ries as a basis for public policy. Public Health Rep,
2012.
Vol. 95 (5) 411-421, 1980. HLDI, Volvo City Safety Loss Experience – Initial Results. HLDI bulletin, June 2011.
Isaksson-Hellman I, Lindman M. The Effect of a LowSpeed Automatic Brake System - Estimated From th
Real Life Data. Proceedings of the 56 AAAM Conth
th
ference, Seattle (USA), 14 – 17 October, 2012.
Pellari S. Numeric and experimental evaluation of an airbag jacket for motorcyclists. Proceedings of the 2012 International IFZ Motorcycle Conference, Cost
logne (Germany), 1 – 2
nd
October 2012,
Penumaka A, Baldanzini N, Pierini M, Savino G. Indepth investigations of PTW-Car accidents caused by human errors, Journal of Safety Science, forthcoming.
- 12 -
APPENDIX Table 1: Full characteristics of the MAEB. Description
Value
Detection range
200 m
Detection angle
100°
Detection roll max Maximum theoretical deceleration Maximum theoretical roll angle Minimum lateral distance while swerving AB reference deceleration Minimum obstacle width
8° 10 m/s2 45° 0.5 m a) Case C
3 m/s2 (AB), 9 m/s2 (EB) 1m
Acronyms AB
autonomous braking functionality ABS anti lock braking system EB enhanced braking functionality EC European Community MAEB motorcycle autonomous emergency braking PTW powered two wheeler STA Swedish Transport Administration
b) Case D
c) Case E Figure 6. Analysis of the effects of experiment variables on the mean impact speed reduction (m/s) due to MAEB for cases C, D and E.
- 13 -
