TECHNICAL FIELD OF THE INVENTION
[0001] The invention relates to a method for a host vehicle to assess risk of overtaking
a target vehicle on a pedestrian crossing, and more particularly to assess the risk
when the pedestrian crossing is partially occluded.
BACKGROUND OF THE INVENTION
[0002] Many cities all over the world are making significant effort to reduce number of
fatalities on the roads, including strong focus on pedestrian safety. Overtaking on
a pedestrian crossing is forbidden at least in some countries. Anyway this rule is
being broken sometimes intentionally or by accident due to the driver distraction,
tiredness or bad weather conditions what increases danger for road users. In countries
where there is no such law regulation, the situation for pedestrian is even more critical.
[0003] It is therefore important to propose a new solution to increase safety for pedestrian.
SUMMARY OF THE INVENTION
[0004] The invention relates to a risk evaluation system for a host vehicle that assesses
the risk of overtaking a target vehicle on a pedestrian crossing, said risk being
a risk of hitting a pedestrian on the pedestrian crossing during the overtaking maneuver.
The risk evaluation system is configured to operate when pedestrian crossing may only
be partially seen by the host vehicle as the target vehicle, running toward the pedestrian
crossing, partially obstructs the front field of view of the host vehicle. In such
situation, the host vehicle may estimate a potential risk of the presence of a pedestrian
crossing the road on the obstructed pedestrian crossing, said pedestrian cannot be
seen by the host vehicle.
[0005] In accordance with one embodiment, the risk evaluation system for the host vehicle
for assessing the risk of overtaking the target vehicle on a pedestrian crossing comprises,
an object detector means configured to detect a pedestrian crossing and to detect
the target vehicle in the front field of view of the host vehicle; a controller in
communication with the object detector means, wherein the controller is configured
to determine the pedestrian crossing location and to determine the target vehicle
driving behavior; the controller is further configured to detect a deceleration of
the target vehicle, to evaluate a target vehicle stop distance between the target
vehicle and the pedestrian crossing based on the target vehicle speed if the target
vehicle has been detected in a deceleration mode; to compare the evaluated target
vehicle stop distance with a predetermined target stop distance threshold; to estimate
a risk for the host vehicle of overtaking the target vehicle on the pedestrian crossing
based on the comparison step result of the target vehicle stop distance.
[0006] In another embodiment, a method for a host vehicle to assess risk of overtaking a
target vehicle on a pedestrian crossing comprises the steps of:
detecting a pedestrian crossing on the host vehicle path;
detecting a target vehicle traveling toward the pedestrian crossing;
determining the speed of the target vehicle;
detecting a deceleration of the target vehicle
evaluating a target vehicle stop distance between the target vehicle and the pedestrian
crossing based on the target vehicle speed if the target vehicle has been detected
in a deceleration mode;
comparing the evaluated target vehicle stop distance with a predetermined target stop
distance threshold;
estimating a risk for the host vehicle of overtaking the target vehicle on the pedestrian
crossing based on the comparison step result of the target vehicle stop distance .
[0007] The step of detecting a target vehicle may be a step of detecting the target vehicle
traveling on an immediate adjacent lane of the host vehicle lane. The step of detecting
the speed of the target vehicle may comprise a step of detecting a deceleration motion
of the target vehicle; and wherein the step of evaluating the target vehicle stop
distance may be a step of estimating the target vehicle stop distance between the
target vehicle and the pedestrian crossing by considering a linear deceleration of
the target vehicle.
[0008] The method may further comprises the steps of determining the speed of the host vehicle;
evaluating a host vehicle stop distance between the host vehicle and the pedestrian
crossing based on the host vehicle speed; comparing the estimated host vehicle stop
distance with a predetermined pedestrian crossing stop distance threshold; and adjusting
the estimated risk according to comparison step of the estimated host vehicle stop
distance. More particularly, the step of evaluation a host vehicle stop distance may
comprise a step of estimating the speed of the host vehicle 10 at the pedestrian crossing
point if a maximum force is applied on the brake system of the host vehicle 10.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] Other features, objects and advantages of the invention will become apparent from
reading the detailed description that follows, and the attached drawings, given by
way of example and in which:
- Figure 1 is a traffic situation on a straight road wherein a host vehicle intends
to overtake a target vehicle on a pedestrian crossing;
- Figure 2 is a traffic situation on a curved road wherein the host vehicle intends
to overtake the target vehicle on a pedestrian crossing;
- Figure 3 is an illustration of a risk evaluation system for a host vehicle for assessing
the risk of overtaking a target vehicle on a pedestrian crossing embedded in the host
vehicle in accordance with one embodiment;
- Figure 4 is flow chart of a method for the host vehicle to assess risk of overtaking
a target vehicle on a pedestrian crossing in accordance with one embodiment;
- Figure 5 is a graph illustrating a sequence of the host vehicle during an overtaking
manoeuver of the target vehicle, said host vehicle comprising the risk evaluation
system of figure 3 in accordance with one embodiment.
DETAILED DESCRIPTION OF THE PREFFERED EMBODIMENTS
[0010] Figure 1 and figure 2 are two non-limited examples of usual traffic situations wherein
a host vehicle 10 and a target vehicle 12 drive toward a pedestrian crossing 14 and
wherein a pedestrian 16 can potentially cross a road 18 on the pedestrian crossing
14. The target vehicle 12 and the host vehicle 10 are on two distinct traffic lanes
20, 22. Figure 2 differs from figure 1 only by the road shape, more particularly figure
1 is a situation wherein the road 18 is a straight road 18 and figure 2 is another
particularly situation wherein the road 18 is a curved road 18. According to figure
1 and figure 2, the target vehicle 12 is closer to the pedestrian crossing 14 than
the host vehicle 10 and the host vehicle 10 is not able to detect the potential presence
of the pedestrian 16, said pedestrian 16, if any, being outside the front field of
view 24 of the host vehicle 10. More particularly, a portion 26 of the front field
of view 24 of the host vehicle 10 is occluded by the target vehicle 12 such that the
pedestrian 16 cannot be detected by the host vehicle 10. In other words, if a pedestrian
16 is located at the portion 28 of the pedestrian crossing 14 occluded by the target
vehicle 12 for the host vehicle 10, the pedestrian 16 won't be visible by a driver
of the host vehicle 10 or by any front camera, front or side radar, or lidar of the
host vehicle 10.
[0011] In order to assess the risk of overtaking the target vehicle 12 on the pedestrian
crossing 14, according to figure 3, the host vehicle 10 comprises object detector
means 30 for detecting pedestrians 16 crossing and target vehicles 12. The object
detector means 30 are arranged on the host vehicle 10 such that they are configured
to detect object in the front of the vehicle 10, said object detector means 30 having
a front field of view 24 covering the front direction 29 of the host vehicle 10.
[0012] More particularly, the object detector means 30 comprises road sign detector 32 configured
to detect marks on the road 18 as pedestrian crossing 14 and/or pedestrian traffic
sign. As non-limited example, road sign detector 32 may be a camera 34, a vehicle
to infrastructure communication device 36, or a global positioning system (GPS) 38,
more particularly, the maps information of the GPS. Combination of such road sign
detectors 32 is also a possibility in order to get the most reliable presence information
on the detected pedestrian crossing 14.
[0013] In addition, the object detector means 30 comprises vehicle detector 40 configured
to detect the target vehicle 12. As non-limiting example, the vehicle detector 40
may be a camera 42, a radar 44, or a lidar 46. Combination of such vehicle detectors
40 is also a possibility in order to get the most reliable dynamic information on
the detected target vehicle 12. The host vehicle 10 may have a unique front camera
34, 42 for target vehicle 12 detection and for pedestrian crossing 14 detection.
[0014] The host vehicle 10 further comprises a controller 48 configured to analyze the detected
objects. More particularly, the controller 48 is configured to analyze 50 the detected
pedestrian crossing 14 as the location 52 of said pedestrian crossing 14.
[0015] The controller 48 is further configured to analyze the driving behavior 54 of the
detected target vehicle 12. As non-limiting example, the analysis of the driving behavior
54 of the target vehicle 12 comprises determination of parameters of the target vehicle
12 as location 56, speed 58, and direction 29. Speed 58 estimation of the target vehicle
12 may also comprise estimation of the yaw rate of the target vehicle 12, acceleration
of the target vehicle 12 and deceleration of the target vehicle 12. Estimation of
the location 56 of the target vehicle 12 may also comprise the estimation of the lane
22 of the road 18 on which the target vehicle 10 is driving.
[0016] The controller 48 is also further configured to analyze the driving behavior 54 of
the host vehicle 10 as being in communication with other sensors of (not shown) the
host vehicle 10. As non-limiting example, the analysis of the driving behavior 54
of the host vehicle 10 comprises determination of parameters of the host vehicle 10
as location, speed, and direction. Speed estimation of the host vehicle 10 may also
comprise estimation of the yaw rate of the host vehicle 10, acceleration of the host
vehicle 10 and deceleration of the host vehicle 10. Estimation of the location of
the host vehicle may also comprise the estimation of the lane 20 of the road 18 on
which the host vehicle 10 is driving.
[0017] According to the pedestrian crossing analysis 50 and according to the target vehicle
12 and to the host vehicle 10 driving behavior analysis 54, the controller 48 is configured
to determine a target vehicle 12 to host vehicle 10 relative longitudinal position
according to a host curvilinear coordinate system and also a target vehicle 12 to
a host vehicle 10 relative lateral position 31.
[0018] The controller 48 is also further configured to determine a longitudinal distance
according to the curvilinear coordinate system between the host vehicle 10 and the
pedestrian crossing 14 and between the target vehicle 12 and the pedestrian crossing
14.
[0019] In addition, the controller 48 is configured to determine the road shape according
to the determination of the motion of the host vehicle 10 and of the motion of the
target vehicle 12, said motion of both vehicles 10, 12 being based on a regular determination
of the position of both vehicles 10, 12. The road shape determination may be more
accurately determined, if possible, by using information from the maps of global positioning
system 38 of the host vehicle 10.
[0020] According to the determined parameters and estimated parameters of both target 12
and host 10 vehicle, the controller 48 is configured to assess 60 the risk of overtaking
the target vehicle 12 on the pedestrian crossing 14. The risk is binary coded 62,
as for instance coded as a '0' if no risks are estimated and as a '1' if a risk is
estimated.
[0021] A no risks situation may be a situation wherein the host vehicle 10 determines that
the target vehicle 12 stops before reaching the pedestrian crossing 14 at a sufficient
distance from the pedestrian crossing 14, such that, the host vehicle 10 overtaking
the target vehicle 12 and still driving towards the pedestrian crossing 14, will be
able to detects the previously occluded portion 28 of the pedestrian crossing 14 enough
in advance to check if there is a presence of a pedestrian 16 on the pedestrian crossing
14, and such that the host vehicle 10 will be able to avoid collision with the pedestrian
16, if any, by braking or changing lane 20, 22.
[0022] A risk situation may be a situation wherein the target vehicle 12 is occluding the
field of view 24 of the host vehicle 10 while the host vehicle 10 is overtaking the
target vehicle 12 such that pedestrian 16 on the pedestrian crossing 14 cannot be
detected by the host vehicle 10.
[0023] The binary coding 62 may also present more levels, as for instance a code '00' for
no risks, '01' for low risk, '10' for medium risk, an '11' for severe risk. The granularity
of the risk may be attached to the driving behavior 54 of the target vehicle 12 and
of the host vehicle 10.
[0024] As no limited example, no risk estimation may be the situation previously described.
[0025] The low risk estimation may be a situation wherein the risk evaluation system of
the host vehicle 10 determines that the speed 58 of the host vehicle 10 and the speed
of the target vehicle 12 are slow enough such that collision with the pedestrian 16
may be avoided by a last moment braking action.
[0026] A medium risk estimation may be a situation wherein the host vehicle 10 detects an
emergency braking of the target vehicle 12 such that the target vehicle 12 will stop
at the limit of the pedestrian crossing 14. Said target vehicle 12 behavior detected
by the host vehicle 10 is such that the host vehicle 10 considers a very high probability
of the presence of the pedestrian 16 on the pedestrian crossing 14. Said situation
may be estimated as a 'medium risk' as long as the risk estimation system estimates
that the host vehicle 10 may still have possibility to stop just before reaching the
pedestrian crossing 14, or changing lane 20, 22 to avoid potential collision with
the pedestrian 16.
[0027] A high risk estimation may be a situation wherein the target vehicle 12 behavior
is similar as the medium risk but wherein the speed of the host vehicle 10 and the
distance between the host vehicle 10 and the pedestrian crossing 14 is such that the
risk estimation system estimates that there is no possible usual maneuvers (changing
lane, braking) to avoid passing over the pedestrian crossing during the overtaking
maneuver of the target vehicle 12. In said situation, in case of pedestrian 16 presence,
such collision will only depend on the pedestrian 16 behavior.
[0028] When the risk estimation is not a no risks situation, the controller 48 may be also
configured to estimate a potential location of the pedestrian 16 on the pedestrian
crossing 14 when host vehicle 10 will arrive at the pedestrian crossing 14, such that
the controller 48 may be configured to control 64 the host vehicle 10 to avoid the
collision by bypassing the pedestrian 16.
[0029] Based on the risk estimation of a collision with a pedestrian 16 on the pedestrian
crossing 14, the controller 48 is configured to directly control 64 the host vehicle
10 or may alert the driver of the host vehicle 10 on the best driving behavior to
limit said potential collision. In other words, the controller 48 is configured to
operate 74 the host vehicle 10 in automated mode 76 wherein driver does not control
the host vehicle 10, or to operate 74 the vehicle in manual mode 78 wherein driver
has still control on the host vehicle 10. The controller 48 may be configured such
that the risk estimation system is adapted for manual vehicle and for an automated
vehicle or driverless vehicle.
[0030] Based on the risk estimation of a collision with the pedestrian 16 on the pedestrian
crossing 14, the controller 48 is configured to advice driver or to have control 64
on the host vehicle 10 in order to stop 66 the host vehicle 10, to adjust 68 the speed
of the host vehicle 10 and/or to adjust the direction 70 of the host vehicle 10. Additionally
the controller 40 is configured to get control on the activation of warning devices
72 as horn, or exterior light of the host vehicle 10 to get the attention of the potential
pedestrian 16. In order to get control on the host vehicle 10, as for vehicle comprising
advanced driver assistance systems (ADAS) or for automated vehicle, the controller
48 is thus configured to operate the vehicle controls 80 such as the brakes 82 of
the host vehicle 10, the accelerator 84, the steering 86 of the host vehicle 10, and
the warning devices such as the exterior lights 88 or the horn 90.
[0031] According to figure 4, a non-limited example of a method 100 for the host vehicle
10 to assess risk of overtaking the target vehicle 12 on the pedestrian crossing 14
is shown. The method 100 comprises a step of detecting 110 the presence of the pedestrian
crossing 14 in the front field of view 24 of the host vehicle 10, i.e. on the host
vehicle 10 path. Another step comprises the analysis of the driving behavior 170 of
the target vehicle 12. More particularly, this step 170 consists at first on determining
regularly, i.e. monitoring, 120 the speed of the target vehicle 12 and detecting 130
if the target vehicle 12 is in deceleration mode. If the target vehicle 12 is not
in a deceleration mode, the probability of the presence of the pedestrian 16 is considered
as null. If the target vehicle 12 is in deceleration mode, a further step 140 consists
on that the controller 48 of the host vehicle 10 predicts a null speed location of
the target vehicle 12. Then a further step 150 consists on the analysis of the null
speed location of the target vehicle 12. If the null speed location of the target
vehicle 12 is at the pedestrian crossing 14 or at a very near distance from the pedestrian
crossing 14 before passing on the pedestrian crossing 14, i.e. for instance 1 meter,
then another step 160 of the method 100 consists of on considering a high probability
of the presence of the pedestrian 16 engaged or engaging on the pedestrian crossing
14 such that a high risk of overtaking the target vehicle 12 on the pedestrian crossing
14 is estimated.
[0032] In other words, the controller 48 of the host vehicle 10 consider a high probability
of the presence of the pedestrian 16 engaged or engaging on the pedestrian crossing
14 if the predicted target vehicle stop distance SD_t of the target vehicle 12, i.e.
the distance between the location of the target vehicle 12 at a null speed and the
location of the pedestrian crossing 14, is at a distance from the pedestrian crossing
14 lower than a predetermined target stop distance threshold TSD_th. The predetermined
target stop distance threshold TSD_th is a distance from the pedestrian crossing 14
wherein the controller 48 could estimate that if the target vehicle 12 stops before
said target stop distance threshold TSD_th, the reason of the stop of the target vehicle
12 is not the presence of a pedestrian 16 and that if the target vehicle 12 stops
between said target stop distance threshold TSD_th and the pedestrian crossing 14,
the reason of the stop of the target vehicle 12 is the presence of a pedestrian 16.
[0033] More generally, when the host vehicle 10 considers a high probability of the presence
of a pedestrian 16 engaged or engaging on a pedestrian crossing 14, the host vehicle
10 may estimate a high risk of overtaking the target vehicle 12 on the pedestrian
crossing 14.
[0034] Other parameters may be also evaluated to refine the method 100 for the host vehicle
10 to assess risk of overtaking the target vehicle 12 on the pedestrian crossing 14.
More particularly, the driving behavior of the host vehicle 10 may also be considered.
Therefor the method 100 may comprises a step of determining the speed of the host
vehicle 10 such that an evaluation of the distance, called the host vehicle stop distance
SD_h, between the host vehicle 10 and the pedestrian crossing 14 may be performed
by the controller 48. Said host vehicle stop distance SD_h may be evaluated according
to a maximum braking force of the host vehicle, i.e. in an emergency braking mode
of the host vehicle 10. Then a step of comparing the estimated stop distance SD_h
of the host vehicle 10 with a predetermined safe distance from the pedestrian crossing
at which the controller 48 considers that there is no risks of collision with a pedestrian
16 on the pedestrian crossing 14 is performed. Said safe distance may be used as a
threshold level for the comparison step, said threshold level being called a predetermined
pedestrian crossing stop distance threshold PCSD_th.
[0035] Thus, when the controller 48 of the host vehicle 10 estimates that the host vehicle
10 stop distance could be greater than the predetermined pedestrian crossing stop
distance threshold, it means that the host vehicle 10 is able to stop before reaching
the pedestrian crossing 14, thus, the risk of overtaking the target vehicle 12 such
that host vehicle 10 may be able to stop before potential collision with the pedestrian
16 may be considered as a no risks situation.
[0036] According to figure 5, the target vehicle 12 driving behavior is represented by the
fist curve C1, while the host vehicle 10 driving behavior is represented by the second
curve C2. The situation is comparable as the situation of figure 1 or figure 2. As
the goal of the host vehicle 10 is to overtake the target vehicle 12, the position
of the target vehicle 12 is closer to the pedestrian crossing 14 than the position
of the host vehicle 10.
[0037] The V axis of the graph represents the speed of the target vehicle 12 and of the
host vehicle 10. The X axis represents the location of the target vehicle 12 and of
the host vehicle 10. Location and speed of both vehicles are represented over the
time (T0, T1, T2, T3). The locations X0_h, X1_h, X2_h and X3_h correspond to the location
of the host vehicle 10 respectively at the point of times T0, T1, T2 and T3. The locations
X0_t, X1_t, X2_t and X3_t correspond to the location of the target vehicle 12 respectively
at the point of times T0, T1, T2 and T3.
[0038] The graph is split in three phases P1, P2, P3.
[0039] First phase P1 is the phase wherein the speed of both vehicles 10, 12 is constant.
More particularly, the speed of the host vehicle 10 is higher than the speed of the
target vehicle 12 such that the host vehicle 10 is closer to the target vehicle 12
at the point of time T1 than at point of time T0. During said first phase, even so
the host vehicle detects the pedestrian crossing 14, the controller 48 of the host
vehicle 10 considers a no risks situation as the speed of the target vehicle 12 is
constant.
[0040] The controller may also adjust the risk estimation by comparing the speed of the
target vehicle 12 with a predetermined minimum speed of the target vehicle threshold
MST_th such that, as represented by the graph, risk is considered as a no risks situation
when the speed of the target vehicle 12 is detected as greater than said predetermined
minimum speed of the target vehicle threshold MST_th. A determined speed of the target
vehicle greater than the predetermined minimum speed of the target vehicle threshold
MST_th is interpreted as a driving behavior of the target vehicle 12 such that target
vehicle does not detect any pedestrians in its front field of view.
[0041] On the other hand, when the speed of the target 12 is lower than the predetermined
minimum speed of the target vehicle threshold MST_th, then it would be considered
by the controller 48 of the host vehicle 10 as a potential risk of the presence of
a pedestrian 16 on the pedestrian crossing 14. A determined speed of the target vehicle
lower than the predetermined minimum speed of the target vehicle threshold MST_th
is interpreted as a driving behavior of the target vehicle 12 such that target vehicle
12 does detect a pedestrians in its front field of view.
[0042] The second phase P2 is the phase wherein the target vehicle 12 is in deceleration
mode and wherein the host vehicle 10 is still at a constant speed. During said second
phase P2, in addition to the detection of the pedestrian crossing 14, the host vehicle
10 detects the deceleration of the target vehicle 12 such that a no risks situation
may be reconsidered by the controller 48 of the host vehicle 10. One possibility is
that the controller 48 of the host vehicle 10 estimates the potential target vehicle
stop distance SD_t, by considering a linear deceleration of the target vehicle 12.
Then, the controller 48 may estimate that there is a risk of a potential presence
of a pedestrian 16 on the occluded portion 28 of the pedestrian crossing 14 if the
predicted target vehicle stop distance SD_t is at a distance lower to the pedestrian
crossing 14 than the predetermined target stop distance threshold TSD_th.
[0043] From figure 5, it has to be considered that the predicted target vehicle stop distance
SD_t is at a distance from the pedestrian crossing 14 lower than the predetermined
target stop distance threshold TSD_th such that the controller 48 estimates that there
is a presence of a pedestrian 16 on the occluded portion 28 of the pedestrian crossing
14.
[0044] However, as illustrated at the point of time T2, the host vehicle 10 still runs at
the constant speed as long as the controller 48 estimates that a braking action BA
of the host vehicle 10, especially a maximum braking force, is enough to stop the
host vehicle 10 at a distance SD_h greater than the predetermined pedestrian crossing
stop distance threshold PCSD_th. Thus, during such second phase P2, the manoeuver
of overtaking the target vehicle 12 is a no risks situation.
[0045] The third phase P3 starts at the point of time T3 wherein the risk of overtaking
the target vehicle 12 is considered as not being any more a no risks situation. More
particularly, the third phase P3 starts when the controller 48 of the host vehicle
10 estimates that by applying a maximum force on the braking system of the host vehicle
10, i.e. operating an emergency brake, the host vehicle 10 will stop at a distance
SD_h equal or lower than the predetermined pedestrian crossing stop distance threshold
PCSD_th. In said situation, controller 48 takes control on the host vehicle 10 such
that emergency braking is activated.