Technical field
[0001] The present invention relates to an automotive vehicle forward collision avoidance
system in accordance with the preamble of claim 1.
[0002] The present invention also relates to a method in an automotive vehicle forward collision
avoidance system in accordance with the preamble of claim 9.
Background of the invention
[0003] A current trend in the automotive industry is to introduce active safety systems
for avoiding or mitigating collisions. One type of system, with a potentially large
positive impact on accident statistics, is a Forward Collision Avoidance System (FCAS).
An FCAS uses sensors based on technologies such as RADAR (RAdio Detection And Ranging),
LIDAR (Light Detection And Ranging), LASER (Light Amplification by Stimulated Emission
of Radiation) and cameras to monitor the region in front of the host vehicle. In the
FCAS a tracking algorithm is used to estimate the state of the objects ahead and a
decision algorithm uses the estimated states to determine any action, e.g. warning
the driver or autonomous braking.
[0004] Automotive manufacturers are today studying collision avoidance systems providing
warning and auto-braking functionality for an imminent collision with a pedestrian.
Such warning and auto-braking functionality is normally based on the use of sensors,
such as mentioned above, in order to detect the position and motion of pedestrians.
A threat assessor then estimates if the collision avoidance system equipped vehicle
and the pedestrian is on a collision course by predicting the positions a short time
in the future, usually one to three seconds.
[0005] Forward Collision Warning (FCW) is a function that warns the driver in case a collision
with a target object seems likely. The target object may be another vehicle or a pedestrian.
Collision Mitigation by Braking (CMbB) is a function that automatically applies braking
for avoiding or mitigating a collision with a target object.
[0006] Typically, the decision of warning and auto-braking is based on predictions of the
paths of the host vehicle and the target object. The position and motion of the target
object is, as mentioned above, measured with a sensor such as a camera, radar or laser
equipment, or a combination thereof.
[0007] In case the predicted future paths of the host vehicle and target objects intersect
within such a short time that the driver will have to act instantly in order to avoid
a collision, a warning is issued. In case the predicted future paths intersect within
such a short time that avoidance of a collision by either steering or braking is unlikely,
auto-braking is applied in order to avoid or mitigate the consequences of the collision.
[0008] During road following, road users such as vehicles, cars, trucks, buses, motorcycles
and bicyclists usually move in such a way that observation based prediction, i.e.
prediction based on an appropriate motion model and what may be observed and measured
without assuming any change in behaviour, is sufficient for decision-making in FCW
systems and CMbB systems.
[0009] One problem is that if observation based prediction methods are applied, a pedestrian
who is walking towards the host vehicle path is predicted to continue walking into
that path. The same is true for crossing vehicles. Once again, if observation based
prediction methods are applied to a crossing vehicle it will be predicted to move
into the host vehicle path. Both of the above cases are likely to lead to false warnings
and false braking interventions in FCW systems and CMbB systems since the typical
behaviour in such a case is that the crossing vehicle or pedestrian stops before entering
the host vehicle path.
Summary of the invention
[0010] One object of the invention is to provide an improved automotive vehicle forward
collision avoidance system.
[0011] This object is achieved by the method as claimed in claim 1.
[0012] Thanks to the provision of: means for establishing the presence of a target object
in front of a vehicle hosting said system; and means for estimating the position,
velocity and acceleration of the target object; and means for establishing a risk
zone in front of the host vehicle; and means for predicting the future path of the
target object and the host vehicle in order to predict the lateral position of the
target object at a moment when the host vehicle reaches the target object in a longitudinal
direction; and means for executing a collision avoidance task arranged such that:
- if the target object, based on its current position, velocity and acceleration,
is predicted as being able to stop before entering the risk zone through applying
a certain predetermined maximum acceleration in a direction opposite to its velocity,
it is predicted to stop at a position just before entering the risk zone and no collision
avoidance task is executed; and - if the target object, based on its current position,
velocity and acceleration, is predicted as not being able to stop before entering
the risk zone, then the future position of the target object is predicted using the
assumption that the target object will continue to move according to an observation
based motion model and a collision avoidance task is executed, a system is provided
which renders fewer false alarms and interventions and thereby a higher confidence
of real threat when providing for alarms and interventions.
[0013] A further object of the invention is to provide an improved method in an automotive
vehicle forward collision avoidance system.
[0014] This object is achieved by the system as claimed in claim 9.
[0015] Thanks to the provision of the steps of: establishing the presence of a target object
in front of a vehicle hosting said system; and estimating the position, velocity and
acceleration of the target object; and establishing a risk zone in front of the host
vehicle; and predicting the future path of the target object and the host vehicle
in order to find its lateral position at a moment when the host vehicle reaches the
target object in a longitudinal direction; and executing a collision avoidance task
arranged such that: - if the target object, based on its current position, velocity
and acceleration, is predicted as being able to stop before entering the risk zone,
it is predicted to stop at a position just before entering the risk zone and no collision
avoidance task is executed; and - if the target object, based on its current position,
velocity and acceleration, is predicted as not being able to stop before entering
the risk zone, then the future position of the target object is predicted using the
assumption that the target object will continue to move according to an observation
based motion model and a collision avoidance task is executed, a method is provided
which renders fewer false alarms and interventions and thereby a higher confidence
of real threat when providing for alarms and interventions.
[0016] Preferred embodiments are listed in the dependent claims.
Description of drawings
[0017] In the following, the invention will be described in greater detail by way of example
only with reference to attached drawings, in which
Fig. 1 is a schematic illustration of a forward collision avoidance system arranged
in a host vehicle,
Fig. 2 illustrates a typical prior art case where the paths of a host vehicle and
a pedestrian seem to intersect in the near future given an assumption of an observation
based motion model,
Fig. 3 illustrates the risk zone and two hypotheses used for realizing the present
invention.
[0018] Still other objects and features of the present invention will become apparent from
the following detailed description considered in conjunction with the accompanying
drawings. It is to be understood, however, that the drawings are designed solely for
purposes of illustration and not as a definition of the limits of the invention, for
which reference should be made to the appended claims. It should be further understood
that the drawings are not necessarily drawn to scale and that, unless otherwise indicated,
they are merely intended to conceptually illustrate the structures and procedures
described herein.
Description of embodiments
[0019] Figure 1 illustrates schematically a Forward Collision Avoidance System (FCAS) 1,
arranged in a host vehicle 2 in accordance with the present invention. The host vehicle
2 has a breaking system 4 such as an Antilock Brake System (ABS system), e.g. with
brake discs 6 and appertaining calipers 7 associated with each of the front wheels
8 and rear wheels 9 of the host vehicle 2. The host vehicle 2 further usually has
a power steering system 5, which is arranged to control the steering angle of the
front wheels 8. A sensor 3, such as a radar, a lidar, a laser or a camera based sensor
is mounted at the front end of the host vehicle 2 and arranged to monitor the region
in front of the host vehicle 2. The FCAS is operatively connected with the braking
system 4 of the host vehicle 2. Further, the FCAS is arranged to establish the presence
of target objects in front of the host vehicle 2 and to estimate the position, velocity
and acceleration of any target objects established to be present. These estimations
are then used by the FCAS to determine how to avoid or mitigate collision with any
target objects, e.g. by performing collision avoidance tasks such as providing warnings,
e.g. using an acoustic and/or visual alarm actuated through an alarm actuator 12,
or performing autonomous braking.
[0020] The present invention especially relates to prediction of future paths of pedestrians
10. It is also applicable to any other type of target object 10 that typically moves
across a road rather than along it. The solution makes use of the fact that pedestrians
10 are usually aware of the danger of entering the road and thus are not assumed to
continue over the road according to an observation based motion model, such as with
constant velocity, just because they temporarily have a motion towards it.
[0021] The case where the pedestrian 10 stops very late just before entering the host vehicle
2 path is thus, using the prior art assumption of an observation based motion model
(illustrated at A in figure 2), very likely to cause false warnings and brake interventions.
This is a very common situation in urban areas. The same is true for crossing vehicles.
Once again, if observation based prediction methods are applied to a crossing vehicle
it will be predicted to move into the host vehicle 2 path. However, in many situations
the crossing vehicle will stop safely before entering the host vehicle 2 path. Also
this case is likely to lead to false warnings and braking interventions.
[0022] Pedestrians 10 are usually not users of the same road as the host vehicle 2 other
than while crossing it. In case a pedestrian 10 moves towards the road as if it would
cross it, it is likely to stop before crossing it if there is an upcoming vehicle.
Since most pedestrians 10 are aware of the danger of entering a road, they are normally
not likely to enter the road just because they temporarily have a motion towards it.
Thus, using observation based motion prediction of the path of a pedestrian 10 yields
an unacceptably large number of false alarms and interventions, especially in city
environments where pedestrians 10 close to the road move in random directions, usually
without entering the road.
[0023] Altogether, the assumption that a target object 10 will continue to move according
to an observation based motion model is not as easily applicable to pedestrians 10
as it is to vehicles moving along the same road, i.e. non crossing vehicles. Hence,
for pedestrians 10, there is a need for another approach when deciding whether a collision
is likely.
[0024] The solution in accordance with the present invention is based on the use of two
hypotheses. Firstly(illustrated at A in figure 3), that the pedestrian 10 will continue
walking in the same direction and with the same acceleration as it currently is. Secondly
(illustrated at B in figure 3), that the pedestrian 10 will stop before entering a
risk zone in front of the host vehicle 2. When stopping, the pedestrian 10 is assumed
to apply a certain acceleration in opposite direction to its velocity. The assumed
acceleration may be tuned differently depending on the area of application. For collision
warning, the assumed acceleration may reflect the behavior of stopping in an ordinary
way, whereas that of CMbB may rather reflect the acceleration of stopping abruptly.
[0025] The risk zone is an area representing the predicted coverage of the host vehicle
2, i.e. the area between the predicted path of the leftmost and rightmost parts of
the host vehicle 2, i.e. as defined by the host vehicle width w, as illustrated in
figure 3. The prediction with respect to the risk zone takes curvature into account
by using the current yaw rate of the host vehicle 2 (the host vehicle's angular velocity
around its vertical axis), i.e. if the host vehicle 2 is inside a curve, the risk
zone will be curved accordingly. In addition to the host vehicle 2 width w an extra
width d of the risk zone may be applied on each side of the host vehicle 2. The additional
width d may depend on velocity and other factors in order to better reflect the perceived
risk of the situation.
[0026] In the case that the host vehicle 2 is equipped with a lane marker sensor (not shown),
which provides for sufficiently good detection of lane markers 11 (see figure 3),
the area between the left hand side and right hand side lane markers 11 may be used
to define the risk zone, rather than the host vehicle 2 width w and additional width
d. In the case that only lane markers 11 at one side of the host vehicle 2 exist or
that only lane markers at one side of the host vehicle 2 are detected sufficiently
well, these lane markers 11 may be used for defining that side of the risk zone whilst
using the host vehicle 2 width w and additional width d at the other side of the host
vehicle 2.
[0027] It is not uncommon that lane marker sensors provide a good detection also of the
boundary between the road and a sidewalk, i.e. the curb (not shown), which in such
case may be used instead of the line markers for defining the risk zone in accordance
with the present invention. This could be quite useful in urban areas, where line
markers 11 are frequently replaced with curbs.
[0028] In case of using the system 1 for providing a warning the risk zone may be defined
to be wider than the host vehicle 2, e.g. the host vehicle 2 width w and additional
width d or the width as defined by the lane markers 11.
[0029] In the case of using the system 1 for providing for auto-braking (or autonomous braking)
the risk zone may be defined as having the width w of the host vehicle 2.
[0030] It is obvious to the person skilled in the art that the risk zone may be defined
in a number of different ways in order to suit different applications and requirements.
[0031] Using the two hypotheses, the future path of the pedestrian 10 is predicted in order
to find its lateral position y at the moment when the host vehicle 2 reaches the pedestrian
10 in a longitudinal direction x.
[0032] If the second hypotheses (illustrated at B in figure 3), i.e. that the pedestrian
10 will stop before entering a risk zone in front of the host vehicle 2, results in
a predicted position outside the risk zone it is assumed that the pedestrian 10 intends
to stop before entering the risk zone. The second hypothesis is then used when calculating
the future path of the pedestrian 10.
[0033] If the second hypotheses (illustrated at B in figure 3), i.e. that the pedestrian
10 will stop before entering a risk zone in front of the host vehicle 2, yields a
predicted position inside the risk zone, it is assumed that the pedestrian 10 does
not intend to stop (based on the fact that it is too late to stop in a reasonable
manner) before entering the risk zone. The first hypothesis (illustrated at A in figure
3), i.e. that the pedestrian 10 will continue walking in the same direction and with
the same acceleration as it currently is, is then used when calculating the future
path of the pedestrian 10.
[0034] In a similar way as previously described, a third hypothesis may be used. In accordance
with this third hypothesis it is assumed that the pedestrian 10 accelerates in the
other direction, i.e. that the pedestrian 10 will accelerate to escape the risk zone,
rather than stop before entering it.
[0035] Thus, the proposed solution lies in using an improved motion model for pedestrians
10. This improved motion model takes into account that a pedestrian 10 is likely to
stop before entering a host vehicle 2 future path. A predetermined maximum acceleration
for pedestrians 10 is assumed. Thus, using straightforward physical calculations,
it is possible to determine if a pedestrian 10 has the possibility to stop before
entering a risk zone or not. If it is determined that the pedestrian 10 has the possibility
to stop, based on its current position and velocity, it is predicted to stop at a
position just before entering the host vehicle 2 path. In a similar manner, if it
is determined that a crossing vehicle (not shown) has the possibility to stop before
entering the host vehicle 2 path, this crossing vehicle is predicted to stop at a
position just before entering the host vehicle 2 path. On the other hand, if a pedestrian
10, or a crossing vehicle, is estimated as not being able to stop before entering
the host vehicle 2 path, then the future position of this pedestrian 10, or crossing
vehicle, is predicted using conventional motion models. Conventional motion models
typically means that it is assumed that the pedestrian 10, or crossing vehicle, will
continue moving according to an observation based motion model, e.g. with a constant
velocity in the current direction.
[0036] An advantage with the proposed solution, compared to conventional methods, is that
it will lead to fewer false alarms in pedestrian and intersection warning/intervention
safety systems. In the end, it could be the difference between being able to reach
an acceptable level of false alarms or not. Furthermore, when it is predicted that
a pedestrian 10, or crossing vehicle, does not have the possibility to stop before
entering the host vehicle 2 path, the confidence that there is a real threat is higher.
Thus, the positive performance may also be increased, which in this case means earlier
interventions and higher safety benefits by the system 1.
[0037] Thus, in accordance with the present invention is provided an automotive vehicle
forward collision avoidance system 1, having means, such as a radar, a lidar, a laser
or a camera based sensor 3 for establishing the presence of a target object 10 in
front of a vehicle 2 hosting the system 1. Means are provided for estimating the position,
velocity and acceleration of the target object 10. Further means are provided for
establishing a risk zone in front of the host vehicle 2. Also provided are means for
predicting the future path of the target object 10 and the host vehicle 2 in order
to predict the lateral position y of the target object 10 at a moment when the host
vehicle 2 reaches the target object 10 in a longitudinal direction x. Means are also
provided for executing a collision avoidance task, which means are arranged such that:
- if the target object 10, based on its current position, velocity and acceleration,
is predicted as being able to stop before entering the risk zone through applying
a certain predetermined maximum acceleration in a direction opposite to its velocity,
it is predicted to stop at a position just before entering the risk zone, as illustrated
at (B) in figure 3, and no collision avoidance task is executed; and - if the target
object 10, based on its current position, velocity and acceleration, is predicted
as not being able to stop before entering the risk zone, then the future position
of the target object 10 is predicted using the assumption that the target object 10
will continue to move according to an observation based motion model, as illustrated
at (A) in figure 3, and a collision avoidance task is executed.
[0038] In one embodiment, the means for executing a collision avoidance task are arranged
to provide at least one of a warning to a driver of the host vehicle 2 and autonomous
braking of the host vehicle 2.
[0039] In a further embodiment, the means for establishing a risk zone in front of the host
vehicle 2 are arranged to establish the risk zone as an area representing the predicted
coverage of the host vehicle 2 based on the host vehicle 2 width w.
[0040] In a yet further embodiment, the means for establishing a risk zone in front of the
host vehicle 2 are further arranged to establish the risk zone taking curvature into
account by using the current yaw rate of the host vehicle 2 and curving the risk zone
accordingly.
[0041] In a still further embodiment, the means for establishing a risk zone in front of
the host vehicle 2 are further arranged to establish the risk zone through, in addition
to the host vehicle 2 width w, applying an extra width d on each side of the host
vehicle 2, as illustrated in figure 3.
[0042] In another further embodiment, the host vehicle 2 is equipped with a lane marker
sensor (not shown) for detection of lane markers 11 and that either the area between
the detected left hand side and right hand side lane markers 11 is used to define
the risk zone or, in the case that only lane markers 11 at one side of the host vehicle
2 are detected, these detected lane markers 11 at one side of the host vehicle 2 are
used for defining that side of the risk zone whilst the host vehicle 2 width w and
an additional width d are used for defining the other side of the risk zone, as illustrated
in figure 3.
[0043] In a still further embodiment, the means for executing a collision avoidance task
are further arranged such that if the target object 10, based on its current position,
velocity and acceleration, is predicted as being able to accelerate, through applying
a certain predetermined maximum acceleration in the direction of its velocity, and
thereby escape the risk zone before the host vehicle 2 reaches the target object in
a longitudinal direction x no collision avoidance task is executed.
[0044] In yet another further embodiment, the means for estimating the position, velocity
and acceleration of the target object 10 are further arranged such that if the estimated
acceleration of the target object 10 lies within a first range, when attempting to
stop before entering the risk zone through applying a certain predetermined maximum
acceleration in a direction opposite to its velocity and a collision avoidance task
is executed, the collision avoidance task of providing a warning to a driver of the
host vehicle 2 is executed, and if the estimated acceleration of the target object
10 lies within a second range, when attempting to stop before entering the risk zone
through applying a certain predetermined maximum acceleration in a direction opposite
to its velocity and a collision avoidance task is executed, the collision avoidance
task of performing autonomous braking of the host vehicle 2 is executed.
[0045] The present invention also relates to a method in an automotive vehicle forward collision
avoidance system 1 suitable to be utilized in an automotive vehicle forward collision
avoidance system 1 as described above.
[0046] The present invention also relates to an automotive vehicle 2 comprising an automotive
vehicle forward collision avoidance system 1 as described above.
[0047] The invention is not limited to the above-described embodiments, but may be varied
within the scope of the following claims.
[0048] Thus, while there have been shown and described and pointed out fundamental novel
features of the invention as applied to a preferred embodiment thereof, it will be
understood that various omissions and substitutions and changes in the form and details
of the devices illustrated, and in their operation, may be made by those skilled in
the art. For example, it is expressly intended that all combinations of those elements
and/or method steps which perform substantially the same function in substantially
the same way to achieve the same results are within the scope of the invention. Moreover,
it should be recognized that structures and/or elements and/or method steps shown
and/or described in connection with any disclosed form or embodiment of the invention
may be incorporated in any other disclosed or described or suggested form or embodiment
as a general matter of design choice. It is the intention, therefore, to be limited
only as indicated by the scope of the claims appended hereto.
1. An automotive vehicle forward collision avoidance system (1),
characterized in that it comprises:
means for establishing the presence of a target object (10) in front of a vehicle
(2) hosting said system (1);
means for estimating the position, velocity and acceleration of the target object
(10); means for establishing a risk zone in front of the host vehicle (2);
means for predicting the future path of the target object (10) and the host vehicle
(2) in order to predict the lateral position (y) of the target object (10) at a moment
when the host vehicle (2) reaches the target object in a longitudinal direction (x);
means for executing a collision avoidance task arranged such that:
- if the target object (10), based on its current position, velocity and acceleration,
is predicted as being able to stop before entering the risk zone through applying
a certain predetermined maximum acceleration in a direction opposite to its velocity,
it is predicted to stop at a position just before entering the risk zone and no collision
avoidance task is executed; and;
- if the target object (10), based on its current position, velocity and acceleration,
is predicted as not being able to stop before entering the risk zone, then the future
position of the target object (10) is predicted using the assumption that the target
object (10) will continue to move according to an observation based motion model and
a collision avoidance task is executed.
2. A system (1) according to claim 1,
characterized in that the means for executing a collision avoidance task are arranged to provide at least
one of a warning to a driver of the host vehicle (2) and autonomous braking of the
host vehicle (2).
3. A system (1) according to any one of claims 1 or 2,
characterized in that the means for establishing a risk zone in front of the host vehicle (2) are arranged
to establish the risk zone as an area representing the predicted coverage of the host
vehicle (2) based on the host vehicle (2) width (w).
4. A system (1) according to claim 3,
characterized in that the means for establishing a risk zone in front of the host vehicle (2) are further
arranged to establish the risk zone taking curvature into account by using the current
yaw rate of the host vehicle (2) and curving the risk zone accordingly.
5. A system (1) according to any one of claims 3 or 4,
characterized in that the means for establishing a risk zone in front of the host vehicle (2) are further
arranged to establish the risk zone through in addition to the host vehicle (2) width
(w) applying an extra width (d) on each side of the host vehicle (2).
6. A system (1) according to claim 3,
characterized in that the host vehicle (2) is equipped with a lane marker sensor for detection of lane
markers and that either the area between the detected left hand side and right hand
side lane markers (11) is used to define the risk zone or, in the case that only lane
markers (11) at one side of the host vehicle (2) are detected, these detected lane
markers (11) at one side of the host vehicle 2 are used for defining that side of
the risk zone whilst the host vehicle (2) width (w) and an additional width (d) are
used for defining the other side of the risk zone.
7. A system (1) according to claims 1,
characterized in that the means for executing a collision avoidance task further are arranged such that
if the target object (10), based on its current position, velocity and acceleration,
is predicted as being able to accelerate, through applying a certain predetermined
maximum acceleration in the direction of its velocity, and thereby escape the risk
zone before the host vehicle (2) reaches the target object (10) in a longitudinal
direction (x) no collision avoidance task is executed.
8. A system (1) according to claim 1,
characterized in that the means for estimating the position, velocity and acceleration of the target object
(10) further are arranged such that if the estimated acceleration of the target object
(10) lies within a first range when attempting to stop before entering the risk zone
through applying a certain predetermined maximum acceleration in a direction opposite
to its velocity and a collision avoidance task is executed, the collision avoidance
task of providing a warning to a driver of the host vehicle (2) is executed, and if
the estimated acceleration of the target object (10) lies within a second range when
attempting to stop before entering the risk zone through applying a certain predetermined
maximum acceleration in a direction opposite to its velocity and a collision avoidance
task is executed, the collision avoidance task of performing autonomous braking of
the host vehicle (2) is executed.
9. A method in an automotive vehicle forward collision avoidance system,
characterized in that it comprises the steps of:
establishing the presence of a target object (10) in front of a vehicle (2) hosting
said system;
estimating the position, velocity and acceleration of the target object (10);
establishing a risk zone in front of the host vehicle (2);
predicting the future path of the target object (10) and the host vehicle (2) in order
to find its lateral position at a moment when the host vehicle (2) reaches the target
object (10) in a longitudinal direction;
executing a collision avoidance task arranged such that:
- if the target object (10), based on its current position, velocity and acceleration,
is predicted as being able to stop before entering the risk zone, it is predicted
to stop at a position just before entering the risk zone and no collision avoidance
task is executed;
- if the target object (10), based on its current position, velocity and acceleration,
is predicted as not being able to stop before entering the risk zone, then the future
position of the target object (10) is predicted using the assumption that the target
object (10) will continue to move according to an observation based motion model and
a collision avoidance task is executed.
10. An automotive vehicle (2), characterized in that it comprises an automotive vehicle forward collision avoidance system (1) according
to any one of claims 1 to 7.