METHOD FOR SELECTING A MOVABLE OBJECT FOR IMPLEMENTING AN ADAPTIVE VEHICLE CONTROL FUNCTION FOR A MOTOR VEHICLE

Description

[0001] The present invention relates generally to driver assistance systems also known as ADAS (Anglo-Saxon acronym for "Advanced Driver-Assistance System") equipping certain motor vehicles and is more particularly interested in a method of selecting a moving object for the implementation of an adaptive speed regulation function, also called ACC function, for a motor vehicle, called carrier vehicle.

[0002] The management of vehicles, called moving objects, evolving in the environment of the carrier vehicle, is implemented by the ACC function which is one of the driving assistance functions offered by the ADAS system.

[0003] The ACC (Adaptive Cruise Control) function, also known as adaptive speed regulation, adds extra assistance to the cruise control function found in most modern vehicles. ACC adds distance control to the speed control. To achieve this, ACC uses telemetry (or a telemetry system) such as radar or laser to detect and track moving targets in the same lane as the vehicle, in front of it.

[0004] Following the acquisition and processing of signals from the telemetry system, the ACC function measures the distance and approach speed of a vehicle preceding the "carrier" vehicle and defines an Inter-Vehicle Time (IVT), which can be configured before activating the ACC function. The "carrier" vehicle is the one equipped with the ACC function. The carrier vehicle is sometimes also referred to as the "ego vehicle" or "host vehicle." The vehicle preceding the carrier vehicle is also called the "target vehicle" or simply the "target."

[0005] The ACC function, on the other hand, automatically adjusts the speed of the carrier vehicle to maintain a safe distance that takes into account the speed of both the carrier vehicle and the target vehicle. This safe distance helps prevent collisions between the two vehicles. The ACC function is controlled by the ADAS system, which also manages the vehicle's cruise control, also known as ACC.

[0006] The ACC function of the ADAS system uses target detection and tracking methods such as an "ACC radar" (or a laser), which is generally positioned at the front of the vehicle behind the front bumper. It is often coupled with another type of camera, a front and / or multifunction camera, which is located at the rearview mirror at the top of the windshield.

[0007] The signals captured by the ACC radar (or laser) are processed by software means specialized in the recognition of radar (or laser) signals and are translated into distance and speed information (and therefore time).

[0008] In addition to detecting and tracking "target" vehicles in the carrier vehicle's lane, it is also necessary to consider vehicles or "moving objects," simply referred to as "objects," detected in lanes adjacent to the carrier vehicle's lane. These objects may change lanes and position themselves in front of the carrier vehicle. This is particularly relevant during maneuvers known as lane change and lane change maneuvers on multi-lane road networks, especially highways. To address this, the ACC function can identify a lane change (the entry of an object into the carrier vehicle's lane, making it a target), and a lane change (the exit of the object from the carrier vehicle's lane).

[0009] Therefore, the ACC function must detect and select, or conversely deselect, objects that may either enter or leave the lane of the carrier vehicle, neither too early nor too late.

[0010] Among the major flaws encountered in the selection of detected objects are the late selections or deselections of objects that make a lane change: selection for objects that come from adjacent lanes and enter the lane of the carrier vehicle and deselection of objects that were considered as "targets" and that leave the lane of the carrier vehicle towards the adjacent lanes on the left or right.

[0011] Selecting a detected object as a target after it has entered the lane of the carrier vehicle is considered a late selection, which results in strong braking of the carrier vehicle and / or jerky speed regulation.

[0012] And deselecting a target after it has left the carrier vehicle's lane is also considered a late deselection because it results in a loss of time in selecting the next target, with the same drawbacks as for a late selection.

[0013] In addition, the prior art is known from documents DE102009007885A and US2017305422A.

[0014] The present invention aims to remedy these drawbacks by proposing a solution that allows for the anticipation of "cut-ins" and "cut-outs" for the selection of potential targets and the deselection of targets, and thus makes the ACC function proactive with respect to these events.

[0015] To this end, the present invention has as its first object a method for selecting a movable object for implementing an adaptive speed control function, called the ACC function, of a vehicle, called the carrier vehicle, consisting of:to detect objects traveling in the same lane as the carrier vehicle and / or in one or both lanes adjacent to the carrier vehicle's lane, in the same direction of travel as the carrier vehicle; for an object detected in the carrier vehicle's lane, to calculate the probability P(LLC) that the object makes a lane change from the carrier vehicle's lane to the left-hand adjacent lane and the probability P(RLC) that the object makes a lane change from the carrier vehicle's lane to the right-hand adjacent lane; for an object detected in the left-hand adjacent lane of the carrier vehicle's lane, to calculate the probability P(RLC) that said object makes a lane change to the right; for an object detected in the right-hand adjacent lane of the carrier vehicle's lane, to calculate the probability P(LLC) that said object makes a lane change to the left;and to select an object in the left adjacent lane and / or in the right adjacent lane when the result of the probability equation for lane change to the right lane P(RLC) and / or for lane change respectively to the left lane P(LLC) is equal to "1" and the said object(s) have been detected in the left adjacent lane and / or in the right adjacent lane; to deselect an object located in the carrier vehicle's lane when the result of the probability equation for lane change to the left P(LCC) or the result of the probability equation for lane change to the right P(RLC) is equal to "1" and the said object has been detected in the carrier vehicle's lane.

[0016] Depending on one characteristic, the process consists of: in a first preliminary step, to select first and second sets of determined binary variables suitable for defining physical characteristics of a moving object and information relating to the configuration of a lane, in situations of preparation for a lane change respectively on the left or on the right; each of the variables being able to take the state "0" or "1" depending on a determined threshold; in a second preliminary step, to establish causal and dependency relationships between the variables of each first and second set; in a third preliminary step, from the causal and dependency relationships between the variables of the first and second sets, to establish first and second probability equations P(LLC) and P(RLC) that a detected moving object will initiate a lane change on the left, respectively on the right;the selection or deselection of a detected moving object being conditioned by the result of the first and second probability equations P(LLC) and P(RLC). ;

[0017] According to another characteristic, the first set of variables includes the variable EVG defining the existence of a lane adjacent to the left of an object, the variable VLG defining the left lateral velocity of the object, the variable CV defining the lane curvature, the variable ALG defining the left yaw angle; these variables allow the calculation of the probability P(LLC) that a detected moving object, present in the lane or in the lane adjacent to the right of the carrier vehicle's lane, will initiate a lane change on the left, using the formula: P ( LLC | VLG,CV,ALG,EVG ) = P ( EVG ) * P ( LLC | VLG,CV,ALG), and in which the second set of variables includes the variable EVD defining the existence of a lane adjacent to the right of an object, the variable VLD defining the right lateral velocity of an object, the variable CV defining the lane curvature, and the variable ALD defining the right yaw angle; these variables allow the calculation of the probability P(RLC) that a detected moving object, present in the lane or in the lane adjacent to the left of the carrier vehicle's lane, will initiate a lane change on the right, from the formula: P ( RLC | VLD, CV, ALD, EVD ) = P ( EVD ) * P ( RLC | VLD, CV, ALD ) .

[0018] According to another characteristic, the process involves selecting an object if P ( RLC | VLD, CV, ALD, EVD ) = 1, and if the detected object is in the adjacent lane, to the left of the carrier vehicle's lane.

[0019] According to another characteristic, the process involves selecting an object if P ( LLC | VLG, CV, ALG, EVG ) = 1, and if the detected object is in the adjacent lane, to the right of the carrier vehicle's lane.

[0020] According to another characteristic, the process involves deselecting an object if, P ( RLC | VLD, CV, ALD, EVD ) = 1 or P ( LLC | VLG, CV, ALG, EVG ) = 1, and if the detected object is in the lane of the carrier vehicle.

[0021] The second object of the invention is a computer program product comprising instructions which, when the program is executed by a computer, lead the computer to implement the steps of the process as described above.

[0022] The invention has as its third object, a motor vehicle implementing the process as described above.

[0023] Other advantages and features will become clearer from the following description, given solely as a non-limiting example and with reference to the drawings in which: [ Figs. 1 ] illustrates a road situation in which the method according to the invention is implemented; [ Figs. 2 ] illustrates a block diagram of a vehicle implementing the method according to the invention; [ Figs. 3 ] illustrates the ALG variable taken into account by the process according to the invention; [ Figs. 4 ] illustrates the ALD variable taken into account by the process according to the invention; [ Figs. 5 ] illustrates the DLG variable taken into account by the process according to the invention; [ Figs. 6 ] illustrates the DLD variable taken into account by the process according to the invention; [ Figs. 7] illustrates a block diagram of a first architecture of causal relationships and dependencies between variables of a first set of variables, taken into account by the process according to the invention; [ Figs. 8 ] illustrates a block diagram of a second architecture of causal relationships and dependencies between variables of a second set of variables, taken into account by the method according to the invention; and [ Figs. 9 ] illustrates a flowchart of the steps of the process according to the invention.

[0024] There figure 1This illustrates a specific road situation in which four vehicles, VP, VC, OG, and OD, are traveling in three adjacent lanes: a central lane (VVP) and two adjacent lanes, respectively, to the left (VOG) and right (VOD) of the central lane (VVP). All four vehicles (VP, VC, OG, and OD) are traveling in the same direction. The adjacent lanes (VOG and VOD) are separated from the central lane (VVP) by two broken lines, LG and LD. The first vehicle, the carrier vehicle (VP), is traveling in the central lane (VVP). It is preceded in its lane (VVP) by a second vehicle, the target vehicle (VC). A third vehicle, the moving object (OG), is traveling in the adjacent left lane (VOG), and a fourth vehicle, the moving object (OD), is traveling in the adjacent right lane (VOD).The four vehicles VP, OG, OD and VC are in the VIS field of view of the carrier vehicle VP via a camera and / or radar (or laser) mounted in the carrier vehicle VP and introduced below.

[0025] There figure 2 illustrates schematically in the form of a block diagram, a passenger vehicle comprising an ADAS 10 system, implementing the process according to the invention.

[0026] The ADAS 10 system is coupled with at least one ACC radar (or laser) 11. It is configured to implement the ACC function for the detection and tracking of moving objects in cooperation with a front (or multifunction) camera 12 and other sensors 13 of the carrier vehicle VP including: pedal sensors (brake, accelerator, clutch), a steering angle sensor (steering wheel rotation angle), a gear ratio sensor, a turn signal indicator, a yaw angle or heading sensor, a longitudinal and lateral speed sensor, etc.

[0027] The ADAS 10 system acts on control elements of the dynamic behavior of the VP vehicle, including at least one longitudinal control element 14 (cruise control) and one lateral control element 15.

[0028] The ADAS 10 system further includes a processing unit 16 arranged to calculate, based on information provided by the ACC radar 11, the camera 12, and other sensors 13 of the carrier vehicle, the probability of an object changing lanes in the lane or adjacent lanes of the carrier vehicle. This calculation is based on causal relationships and dependencies determined between various chosen variables, respectively defining the physical characteristics of surrounding objects (longitudinal and lateral positions, longitudinal and lateral speeds, etc.) and information relating to the lane configuration (curves, lane existence, line polynomials, etc.). This processing unit 16 implements, in particular, one or more algorithms 17 that implement the models.

[0029] The binary variables used by the method according to the invention for calculating probabilities are introduced below. These variables are chosen based on the relevance of the quantities they characterize and their relationships in situations involving preparation for, or intention to change lanes.

[0030] Depending on whether a specific threshold is reached or not by the physical characteristics or information to which the variables relate, they take on a state of "1" or "0".

[0031] The variable EVD(Existence Right Lane) defines the existence of a lane adjacent to the right of the object. If the quality of the lane adjacent to the right of the object is greater than a threshold S1 (0.8 for example), and if the existence of the lane adjacent to the right of the object is greater than a threshold S2 (0.8 for example), then we consider that there is indeed a lane adjacent to the right of the object, and therefore EVD = 1, otherwise EVD = 0.

[0032] Existence is an indicator that reflects the probability of a track existing; in other words, a confidence indicator for the track's existence. It is generally between 0 and 1. A threshold S2 greater than 0.8 indicates above-average confidence in the track's existence. Track quality is an indicator based on track width, the adequacy of track markings, and the span of track markings. This indicator is also between 0 and 1. A threshold S1 greater than 0.8 indicates a track of above-average quality.

[0033] The variable EVG(Left Lane Existence) defines the existence of a lane adjacent to the left of the object. If the quality of the lane adjacent to the left of the object is greater than a threshold S3 (0.8 for example), and if the existence of the lane adjacent to the left of the object is greater than a threshold S4 (0.8 for example), then it is estimated that there is indeed a lane adjacent to the left of the object, and therefore EVG = 1, otherwise EVG = 0.

[0034] The variable CV (Track Curvature) defines the possibility for the object to make a track change given the track curvature.

[0035] For this, we define a curvature threshold S5 of the track below which we consider a track change possible, and therefore CV = 1, and above which a track change is unlikely, and therefore CV = 0.

[0036] The variable VLD(Right Lateral Velocity) defines the lateral dynamics of the object to the right. If the object has a right lateral velocity greater than a determined threshold S6 (threshold other than 0), then VLD = 1, otherwise VLD = 0.

[0037] The variable VLG (Left Lateral Velocity) defines the lateral dynamics of the object, towards the left. If the object has a left lateral velocity greater than a determined threshold S7 (threshold other than 0) then VLG = 1, otherwise VLD = 0.

[0038] The variable ALG (Left Yaw Angle) defines the yaw angle α of the object VO between the vehicle center and the tangent TGT to the LG line that the object VO, traveling on its VVO lane, must cross to reach the destination (adjacent) lane on the left VAG ( Figure 3 ).

[0039] If the angle α is greater than a determined threshold S8, this means that the driver is turning towards a lane change, and therefore ALG = 1, otherwise ALG = 0.

[0040] The variable ALD (Right Yaw Angle) the yaw angle α of the object VO between the vehicle center and the tangent TGT to the line LD that the object VO, traveling on its lane VVO, must cross to reach the destination (adjacent) lane on the right VAD ( Figure 4 ).

[0041] If the angle α is greater than the threshold determined S9, this means that the driver is turning towards a lane change, and therefore ALD = 1, otherwise ALD = 0.

[0042] The variable DLG (Distance to Left Line) defines the distance of the object VO from the left line LG separating the lane VVO of the object VO, from the adjacent left lane VAG ( Figure 5 ).

[0043] If the distance is less than a threshold S10, then DLG = 1, otherwise DLG = 0.

[0044] The variable DLD(Distance to Straight Line) defines the distance of the object VO from the straight line LD separating the lane VVO of the object VO, from the adjacent right lane VAD ( Figure 6 ).

[0045] If the distance is less than a threshold S11, then DLD = 1, otherwise DLD = 0.

[0046] The process exploits architectures (or models) of causal relationships and dependencies between these variables to calculate the probability that an object will make a lane change to the left lane P(LLC) or a lane change to the right P(RLC).

[0047] There figure 7 illustrates a first architecture of causal relations and dependencies between a first set of chosen variables determined allowing to calculate the probability P(LLC) that an object makes a change of lane towards the left lane "LLC", Anglo-Saxon acronym for "Left Lane Change".

[0048] This first set of variables includes the VLG, CV, ALG and EVG variables defined above.

[0049] To calculate the probability that an object makes a lane change to the left lane LLC, the process calculates the conditional probability that the probability P(LLC) is equal to "1" by considering the variables VLG, CV, ALG and EVG introduced above.

[0050] This conditional probability is expressed by the formula below: P LLC VLG CV ALG EVG = P EVG ∗ P LLC VLG CV ALG

[0051] For an object detected via the ACC function: If: P LLC VLG CV ALG EVG = 1

[0052] So, if the detected object is in an adjacent lane (suggesting a "cut-in") then the process selects the object in question.

[0053] If the detected object is in the path of the carrier vehicle (foreshadowing a "cut-out") then the process deselects the object.

[0054] There figure 8illustrates a second architecture of causal relations and dependencies between a second set of chosen variables determined to calculate the probability P(RLC) that an object makes a lane change to the right lane "RLC", an Anglo-Saxon acronym for "Right Lane Change".

[0055] These variables are VLD, CV, ALD and EVD.

[0056] To calculate the probability that an object makes a lane change to the right lane, the process calculates the conditional probability that the probability P(RLC) is equal to "1" by considering the variables VLD, CV, ALD and EVD introduced above.

[0057] This conditional probability is expressed by the formula below: P RLC VLD CV ALD EVD = P EVD ∗ P RLC VLD CV ALD

[0058] For an object detected via the ACC function: If: P RLC VLD CV ALD EVD = 1

[0059] So, if the detected object is in an adjacent lane (suggesting a "cut-in"), the process selects the object in question.

[0060] If the detected object is identified in the path of the carrier vehicle (foreshadowing a "cut-out") then the process deselects the object.

[0061] The process according to the invention, illustrated by the flowchart of the figure 9 , consists in a first step 100 of detecting objects OG, VC, OD circulating on the same lane VVP as that of the carrier vehicle VP and / or on one or both of the lanes adjacent VOG, VOD to the lane VVP of the carrier vehicle VP, in the same direction of travel as the carrier vehicle VP.

[0062] In a step 200, for a VC object (target) detected in the VVP lane of the carrier vehicle VP, the method calculates the probability P(LLC) that the VC target makes a lane change, from the VVP lane of the carrier vehicle VP to the adjacent left lane VOG and the probability P(RLC) that the VC target makes a lane change, from the VVP lane of the carrier vehicle VP to the adjacent right lane VOD.

[0063] In step 300, for the OG object detected in the left adjacent lane VOG, the process calculates the probability P(RLC) that this OG object will change lanes to the right adjacent lane. In step 400, for the OD object detected in the right adjacent lane VOD, the process calculates the probability P(LLC) that this OD object will change lanes to the left adjacent lane. In step 500, the result of each of these probability equations P(LLC) and P(RLC) is used to select an OG or OD object as the target, or to deselect a VC target object, depending on the lane VOG, VOD, or VVP in which the OG, VC, or OD object was detected.

[0064] Thus, if the object OG, OD is located in the adjacent left lane VOG, or right lane VOD respectively, and the result of its probability equation for changing lanes to the right P(RLC), or to the left P(LLC) respectively, is equal to "1", then this object OG, OD is selected by the ACC function. The target VC, located in the carrier vehicle's lane VVP, is deselected when the result of the probability equation for the target changing lanes to the left lane P(LLC) or to the right lane P(RLC) is equal to "1".

[0065] The process implements a computer program product containing instructions for executing the steps of the process described above.

[0066] This program is implemented by one or more processors of the vehicle's ADAS system (ADAS supervisor) in connection with one or more processors of the IVI (In-Vehicle Infotainment) system, which is the central component of the vehicle dedicated to data processing and communication with the driver.

[0067] Such a program can be updated with the possibility of adding functions using an OTA (Over The Air) type update.

Claims

1. Method for selecting a moving object for implementing an adaptive speed control function, called the ACC function, of a vehicle, called the carrier vehicle (VP), consisting of: - to detect (100) objects (OG, VC, OD) travelling on the same lane (VVP) as that of the carrier vehicle (VP) and / or on one or both of the lanes adjacent (VOG, VOD) to the lane (VVP) of the carrier vehicle (VP), in the same direction of travel as the carrier vehicle (VP); - for an object (VC) detected in the lane (VVP) of the carrier vehicle (VP), to calculate (200) the probability P(LLC) that the object (VC) makes a lane change, from the lane (VVP) of the carrier vehicle (VP) to the adjacent left lane (VOG) and the probability P(RLC) that the object (VC) makes a lane change, from the lane (VVP) of the carrier vehicle (VP) to the adjacent right lane (VOD); - for an object (OG) detected in the left adjacent lane (VOG) of the lane (VVP) of the carrier vehicle (VP), to calculate (300) the probability P(RLC) that said object (OG) makes a lane change to the right; - for an object (OD) detected in the right-hand adjacent lane (VOD) of the carrier vehicle's (VP) lane (VVP), calculate (400) the probability P(LLC) that said object (OD) will change lanes to the left; and - to select an object (OG, OD) on the left adjacent lane (VOG) and / or on the right adjacent lane (VOD) when the result of the probability equation for lane change to the right lane P(RLC) and / or for lane change respectively to the left lane P(LLC), is equal to "1" and that the said object(s) (OG, OD) have been detected on the left adjacent lane (VOG) and / or on the right adjacent lane (VOD); - to deselect an object (VC) located on the lane (VVP) of the carrier vehicle (VP) when the result of the probability equation for a leftward lane change P(LCC) or the result of the probability equation for a rightward lane change P(RLC) is equal to "1" and said object has been detected on the lane (VVP) of the carrier vehicle (VP).

2. A method according to claim 1 consisting of: - in a preliminary first step, to select first and second sets of specific binary variables suitable for defining the physical characteristics of a moving object and information relating to the configuration of a track, in preparation situations to a change of lane respectively on the left or on the right; each of the variables being able to take the state "0" or "1" depending on a determined threshold (S); - in a second preliminary step, to establish causal and dependency relationships between the variables of each first and second set; - in a third preliminary step, from the causal relationships and dependencies between the variables of the first and second sets , to establish the first and second probability equations P(LLC) and P(RLC) that a detected moving object engages in a change of lane on the left, respectively on the right; the selection or deselection of a detected moving object being conditioned by the result of the first and second probability equations P(LLC) and P(RLC).

3. A method according to the preceding claim, wherein the first set of variables comprises the variable EVG defining the existence of a lane adjacent to the left of an object, the variable VLG defining the left lateral velocity of the object, the variable CV defining the curvature of the lane, the variable ALG defining the left yaw angle; these variables allowing the calculation of the probability P(LLC) that a detected moving object, present in the lane (VVP) or in the lane adjacent to the right (VOD) of the lane (VVP) of the carrier vehicle (VP) will initiate a left lane change (LLC), from the formula: P(LLC | VLG,CV,ALG,EVG) = P(EVG) * P(LLC|VLG,CV,ALG), and wherein the second set of variables comprises the variable EVD defining the existence of a lane adjacent to the right of an object, the variable VLD defining the right lateral velocity of an object, the variable CV defining the curvature of the lane and the variable ALD defining the right yaw angle; these variables allow the probability P(RLC) to be calculated that a detected moving object, present in the lane (VVP) or in the lane adjacent to the left (VOG) of the lane (VVP) of the carrier vehicle (VP) will make a lane change to the right (RLC), from the formula: P(RLC|VLD,CV,ALD,EVD) = P(EVD) * P(RLC|VLD,CV,ALD).

4. A method according to the preceding claim, consisting of selecting an object (OG) if P(RLC|VLD,CV,ALD,EVD) = 1, and if the detected object is in the adjacent lane, to the left (VOG) of the lane (VVP) of the carrier vehicle (VP).

5. Method according to claim 3, consisting of selecting an object (OD) if P(LLC|VLG,CV,ALG,EVG) = 1, and if the detected object is in the adjacent lane, to the right (VOD) of the lane (VVP) of the carrier vehicle (VP).

6. Method according to claim 3, consisting of deselected an object (VC) if, P(RLC|VLD,CV,ALD,EVD) = 1or P(LLC|VLG,CV,ALG,EVG) = 1, and if the detected object (VC) is in the lane of the carrier vehicle (VP).

7. Product computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the process according to any one of claims 1 to 6.

8. Motor vehicle (VP) implementing the method according to one of claims 1 to 6.