WO2023222961A1

WO2023222961A1 - Method for managing the path of a motor vehicle in its traffic lane

Info

Publication number: WO2023222961A1
Application number: PCT/FR2023/050540
Authority: WO
Inventors: Zoubida LAHLOU; Hamza El Hanbali; Yassine Et-Thaqfy
Original assignee: Stellantis Auto Sas
Priority date: 2022-05-19
Filing date: 2023-04-14
Publication date: 2023-11-23
Also published as: FR3135682A1; FR3135682B1

Abstract

The invention relates to a method for managing the path (TRJ) of an ego vehicle (VE) in its ego lane (VVE), consisting in: detecting a discontinuity (d) in a first marking line (LMG), said discontinuity (d) taking the form of a second marking line (FPO) parallel to the first marking line (LMG) and extending into the ego lane (VVE) in front of the ego vehicle (VE); and, if a vehicle (VCI), called the target vehicle, is detected at the same time opposite the second marking line (FPO) on the same side as the discontinuity (d), then deducing therefrom that the second marking line (FPO) is a false positive resulting from the shadow (OMB) cast by the target vehicle (VCI), and keeping the ego vehicle (VE) in the ego lane (VVE) with the same path (TRJ) as it had before detection of the second marking line (FPO).

Description

DESCRIPTION

Title of the invention: Method for managing the trajectory of a motor vehicle in its lane

The present invention claims priority from French application 2204808 filed on 05/19/2022, the content of which (text, drawings and claims) is herein incorporated by reference. The present invention generally relates to motor vehicles which use driving assistance functions making it possible to automate certain driving functions usually assigned to the driver.

It relates more particularly to a method for managing the trajectory of a motor vehicle in its lane, taking into account false positives in the detection of the lane marking lines.

The driving assistance functions are managed by the vehicle's driving assistance system, also known as the ADAS system (Advanced Driver-Assistance System).

We will deliberately confuse the ADAS system with the ADAS “supervisor” which is the body intended to supervise all ADAS functions within the vehicle.

These functions include in particular adaptive cruise control or “ACC” (English acronym for “Adaptive Cruise Control”) also referred to as “longitudinal regulation” or “longitudinal control”. This function can be supplemented by a function to help keep the vehicle in the center of its lane. This function is also referred to as “LKA” (English acronym for “Lane Keeping Assist”). The positioning of the vehicle in its lane can also be predefined by the vehicle user as instructions. The function of helping to keep the vehicle in its lane is then designated by “LPA” (Anglo-Saxon acronym for “Lane Positioning Assist”).

ADAS systems generally include a plurality of computers also called ECU (Electronic Control Unit). These computers receive and process data delivered by a plurality of sensors on board the vehicle. One of the most commonly used sensors and which is one of the strong components of ADAS systems is a perception sensor such as a front camera, which can be a multifunction camera and which is generally positioned at the front of the vehicle, against the windshield of the vehicle, at the level of inner rear mirror. Such a camera is also sometimes referred to as a CVM camera, or simply CVM (acronym for Multifunction Video Camera).

The digital images captured by the camera are processed by the ADAS system calculators which, in conjunction with other data delivered by the vehicle's other sensors, or on-board databases, autonomously control various components of the vehicle which act on the torque. engine, brakes, steering wheel angle, etc.

It is known from the state of the art of vehicles comprising at least one camera coupled to an image processing device capable of providing modeling data of the environment of the vehicle and in particular the marking lines of a lane. circulation of a vehicle which are used for the functions of keeping the vehicle in its lane (LKA, LPA, etc.). We will subsequently designate by camera the assembly formed by the perception sensor and its associated image processing device comprising in particular image recognition means. State-of-the-art cameras are not always very reliable, particularly when the vehicle is traveling on a poorly maintained lane or when detection of markings is ambiguous. False detections on marking lines, also called "false positives", can cause untimely corrections to the vehicle's trajectory. This reduces the comfort of vehicle passengers and can cause user distrust of semi-autonomous or autonomous driving systems.

These false detections, or “false positives”, can also come from shadows cast on the lane by a large vehicle such as a truck traveling on the lane adjacent to the lane of the vehicle implementing the LPA function, in front of the vehicle. In the remainder of the description, we will designate the vehicle implementing the LPA function by “ego vehicle” and the traffic lane of the ego vehicle, by ego lane. The shadows cast by a truck on the ego lane can be interpreted by the image processing device as being a ground marking line likely to mislead the lane keeping assistance functions, in particular the LPA function, and changing the trajectory of the ego vehicle unintentionally towards a safety barrier or another vehicle (target vehicle) traveling on a lane adjacent to the ego lane. The objective of the present invention is to provide a solution to this problem by making it possible to distinguish between a real marking line and an erroneous marking line, due in particular to a shadow cast by a large vehicle such as a truck.

To this end, the present invention has as its first object a method for managing the trajectory of a motor vehicle, called ego vehicle, in its lane, called ego lane, from modeling data of the marking lines delimiting the ego lane, said data being processed, in a fixed frame of reference relative to the ego vehicle, by trajectory management means on-board in the ego vehicle for maintaining said ego vehicle in the ego lane between the marking lines; said method consisting of detecting a discontinuity in a first marking line, said discontinuity having the shape of a second marking line extending in the ego lane parallel to the first marking line in front of the ego vehicle, and if a vehicle, called the target vehicle, is detected, at the same time, opposite the second marking line on the same side as the discontinuity, then to deduce that said second marking line is a false positive due to the shadow cast by the vehicle target and to maintain the ego vehicle in the ego lane with the same trajectory as that which it had before the detection of said second marking line.

According to a first characteristic, the management method consists, to detect the discontinuity in the first marking line, of calculating the derivative of the lateral distance between the ego vehicle and the first marking line, and if this derivative is greater than a value parameterized determined, to consider that there is a discontinuity in the first marking line.

According to a second characteristic, the method consists, when a target vehicle is detected on the same side as the discontinuity, in calculating a first longitudinal distance between the fixed frame of reference of the ego vehicle and the target vehicle and a second longitudinal distance between the fixed frame of reference of the vehicle ego and the start of the discontinuity and if the difference, in absolute value, between the first and second distances is less than a determined configurable value, then the management method does not take the second line into account. According to another characteristic, the shadow cast by the target vehicle on the ego lane has a general shape of a parallelogram whose large dimension is detected as the second marking line.

The second object of the present invention is a computer program product comprising instructions which, when the program is executed by a computer, lead it to implement the method as described above.

The third object of the present invention is a motor vehicle, called ego vehicle, comprising a camera capturing images of the environment close to the ego vehicle, at the front of the ego vehicle, means for detecting a target vehicle in the environment close to the ego vehicle, at the front of the ego vehicle, means for managing the trajectory of the ego vehicle for maintaining the ego vehicle in its ego lane while respecting a determined positioning instruction of the ego vehicle in its ego lane; said management means implementing the method as described above.

According to a characteristic of said vehicle, said management means are able, in addition, to implement a function of maintaining the ego vehicle in its lane respecting a positioning instruction in the ego lane configurable by the user of the ego vehicle, called function LPA, which when activated, implements the first characteristic of the process as described above.

According to another characteristic of said vehicle, said management means are capable, in addition, of implementing a function of maintaining the ego vehicle in its lane respecting a positioning instruction in the ego lane configurable by the user of the ego vehicle, called LPA function, which when not activated implements the second characteristic of the method as described above.

Other advantages and characteristics may emerge more clearly from the description which follows, given solely by way of non-limiting example and made with reference to the drawings in which:

[Fig. 1] illustrates a road situation in which the trajectory management method according to the invention is implemented;

[Fig. 2] graphically illustrates a first characteristic of the management method according to the invention; And [Fig. 3] graphically illustrates a second characteristic of the management method according to the invention.

In the figures, the same elements are identified by the same references.

The ACC function of the ADAS system implements means of detecting and tracking targets such as a radar called “ACC radar” (or a laser), which is generally positioned at the front of the vehicle behind the front bumper. of the ego vehicle. It is often coupled with the camera to detect and track targets and maintain the ego vehicle in its lane (ego lane).

The signals picked up 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 (and therefore time) information relative to the target.

In addition to detecting and tracking “target” vehicles in the ego lane of the ego vehicle, the ADAS system takes into account vehicles detected in lanes adjacent to that (ego lane) of the ego vehicle and which are likely to change lanes to come and position yourself in the lane followed by the ego vehicle, in front of the ego vehicle. This is particularly the case in so-called unblocking and drawdown maneuvers, on multi-lane road networks, particularly motorways.

Figure 1 illustrates a specific road situation, on a motorway, in which an ego VE vehicle is traveling on its travel lane (ego lane) WE. The ego WE lane is the central lane of a section of motorway framed on its left, by a lane of traffic in the same direction, adjacent VAG and on its right, by a lane of traffic in the same direction, adjacent to the right VAD. The ego VE vehicle is equipped with an ADAS, ADA system. In the situation illustrated in Figure 1, the ego VE vehicle is faced with a risk of false detection (false positive) due to the OMB shadow cast by a VCI truck traveling in the same direction, on the adjacent left lane VAG at a determined distance in front of the ego VE vehicle. The OMB shadow cast by the VCI truck, on the right side of the VCI truck, is superimposed on the LMG marking line separating the VAG and WE traffic lanes, respectively of the VCI truck and the ego VE vehicle, overflowing onto the ego way WE. The OMB shadow of the VCI truck has a general shape of a parallelogram with the long side extending parallel to the LMG marking line. It can thus be interpreted by the ego VE vehicle arriving near the OMB shadow of the VCI truck, as a new FPO marking line (or second marking line considering the LMG marking line as the first marking line) to be taken into account by the ADAS ADA system to correct the TRJ trajectory of the ego VE vehicle in its WE lane. The “false” corrected FTR trajectory is represented in the figure by a dashed arrow surmounted by a cross.

The method according to the invention is based on the detection of a discontinuity in time, of a lane marking line, concomitantly with the detection of the presence of a vehicle, on the same side of the discontinuity, to manage “false positives”.

The method according to the invention considers two scenarios.

A first scenario in which the lane keeping function with positioning instruction configured by the vehicle user (LPA function) is activated and a second scenario in which the LPA function is not activated. In the second scenario (LPA function not activated) line tracking is still operated by the ADAS system using in particular the camera and the ACC radar.

In both cases, it is the detection of the discontinuity over time of a marking line which is taken into account for the management of the trajectory of the vehicle in its lane.

The ego VE vehicle includes a CAM camera and its image processing device capable of providing modeling data for the LMG and LMD marking lines located on either side of the ego VE vehicle. The modeling data allows the LMG and LMD marking lines to be modeled in a fixed XY reference frame relative to the ego VE vehicle. The fixed frame of reference XY comprises an ordinate axis Y substantially parallel to a lateral direction of the vehicle ego VE and an abscissa axis X substantially parallel to a longitudinal direction of the vehicle ego VE, the abscissa axis intersecting at a point of origin O. The longitudinal direction of the ego VE vehicle is for example a straight line passing through the center of the axles of the ego VE vehicle, from the rear towards the front of the ego VE vehicle. The XY reference frame is an orthogonal reference frame. The lateral direction of the ego VE vehicle is perpendicular to the longitudinal direction of the ego VE vehicle, for example, a straight line passing through the right front door and the left front door. The point of origin “O” is located at the CAM camera. The CAM camera is for example a “CVM” type front camera already introduced in the preamble to this description.

The ego VE vehicle further comprises means for detecting a target vehicle VCI in the environment close to the ego vehicle, not shown, using the same camera or another perception sensor such as a radar or a laser coupled RAD or not to the CAM camera. All detection information (marking lines “LMG, LMD”, discontinuity “d”, target vehicle “VCI” are grouped and managed by MGT trajectory management means belonging to the ADAS, ADA system of the ego VE vehicle.

Figure 2 illustrates the first scenario, in which the LPA lane keeping function is activated. The ego VE vehicle then implements the process for managing the change in trajectory of the ego vehicle according to the invention via the MGT management means.

To check the continuity over time of a marking line, the method consists, from the lateral distance between the ego vehicle and the marking lines delimiting the ego lane, in calculating the derivative of the lateral distance separating the ego vehicle marking lines, and if this derivative is greater than a configurable value determined for example after prior calibration, consider that there is a discontinuity in the marking line.

In the example of Figure 2, the method, after calculating the derivative dstLatDeriv of the lateral distance DL of the marking line LMG, determines at a first time t (DLt) that the value of the derivative of the lateral distance dstLatDeriv between the ego VE vehicle and the LMG marking line located to the left of the ego VE vehicle is greater than a predetermined value C1, dstLatDeriv > C1, and at the same first instant t, a VCI truck is detected in front of the ego VE vehicle, traveling on the adjacent left lane VAG of the ego WE lane, and therefore on the same side as the LMG marking line located to the left of the ego VE vehicle. The process then considers that there is a discontinuity “d” and that the new marking line (second marking line) FPO is linked to the OMB shadow cast by the VCI truck and decides not to take this discontinuity into account “ d” which he considers to be a false detection, or “false positive”. The process ignores this discontinuity “d” and therefore does not order a change in trajectory TRJ of the vehicle ego VE in its lane (see ego) WE and resumes monitoring of the first LMG line which was considered, at time t - 1 (DLt-i) preceding the time t of detection of the discontinuity “d”.

Otherwise, in the absence of a target vehicle (truck) VCI, the method takes into account the discontinuity “d” by considering it as a new “real” line and controls a change in trajectory TRJ of the ego VE vehicle depending on of the new marking line.

Figure 3 illustrates the second scenario in which the LPA function for maintaining the ego vehicle VE in its lane (ego lane) WE is inactive.

The method consists of calculating the longitudinal distance A between the ego vehicle VE and the truck VCI detected on the side of the discontinuity “d” determined as for the first case (figure 2) and the longitudinal distance B between the ego vehicle and the start of the discontinuity “d” And if the difference, in absolute value, of the distances A and B is less than a determined configurable value £, |A - fî| < £, then the management process does not take into account the discontinuity “d” which it considers linked to the shadow OMB of the truck VCI, and does not control the change of trajectory TRJ of the ego vehicle VE in its lane WE.

If the difference, in absolute value, of the distances A and B is equal to or greater than £, then the method takes into account the discontinuity “d” by considering it as a new “real” line (second line) and commands a change of TRJ trajectory of the ego VE vehicle according to the new marking line.

The management method implements a computer program product comprising instructions which, when the program is executed by a computer, lead it to implement the method according to the invention.

This program is implemented for example, by one or more processors linked to the ADAS, ADA system, belonging for example to an ADAS supervisor, not shown and already mentioned above, on board the ego VE vehicle. The ADAS supervisor can itself be linked to one or more processors of the IVI (In-Vehicle Infotainment) system, not shown, which is the central organ of the ego VE vehicle dedicated to data processing and communication with the driver. via an HMI (Human Machine Interface), not shown.

Such a program can be updated with the possibility of adding services using an OTA “Over The Air” type update.

Claims

1. Method for managing the trajectory (TRJ) of a motor vehicle (EV), called ego vehicle, in its travel lane (WE), called ego lane, from marking line modeling data (LMG, LMD) delimiting the ego lane (WE), said data being processed, in a fixed reference frame (X, Y) relative to the ego vehicle (VE), by trajectory management means (MGT) on board the ego vehicle (VE ) for maintaining said ego vehicle (VE) in the ego lane (WE) between the marking lines (LMG, LMD), said ego vehicle, comprising a camera (CAM) capturing images of the environment close to the ego vehicle ( VE), at the front of the ego vehicle (VE), means of detection (RAD) of a target vehicle (VCI) in the environment close to the ego vehicle (VE), at the front of the ego vehicle (VE ) ; said method comprising detecting a discontinuity (d) in a first marking line (LMG), said discontinuity (d) having the shape of a second marking line (FPO) extending in the ego lane (WE) parallel to the first marking line (LMG) in front of the ego vehicle (VE), and if a vehicle (VCI), called target vehicle, is detected, at the same time, facing the second marking line (FPO) on the same side as the discontinuity (d) then to deduce that said second marking line (FPO) is a false positive due to the shadow (OMB) cast by the target vehicle (VCI) and to maintain the ego vehicle (VE) in the ego lane (WE) with the same trajectory (TRJ) as that which it had before the detection of said second marking line (FPO).

2. Management method according to the preceding claim, consisting, to detect the discontinuity (d) in the first marking line (LMG), in calculating the derivative (dstLatDeriv) of the lateral distance (DL) between the ego vehicle (VE) and the first marking line (LMG), and if this derivative (dstLatDeriv) is greater than a determined parameterizable value (C1), consider that there is a discontinuity (d) in the first marking line (LMG).

3. Management method according to claims 1 and 2, consisting, when a target vehicle (VCI) is detected on the same side as the discontinuity (d), in calculating a first longitudinal distance (A) between the fixed reference frame (X, Y ) of the ego vehicle (VE) and the target vehicle (VCI) and a second longitudinal distance (B) between the fixed frame of reference (X, Y) of the ego vehicle (VE) and the start of the discontinuity (d) and if the difference , in absolute value, between the first and second distances (A and B) is less than a determined configurable value (s), then the management method does not take into account the second line (FPO).

4. Management method according to one of the preceding claims, in which the shadow (OMB) cast by the target vehicle (VCI) on the ego path (VVE) has a general parallelogram shape whose large dimension is detected as the second line of marking (FPO).

5. Computer program product comprising instructions which, when the program is executed by a computer, lead it to implement the method according to one of claims 1 to 4.

6. Motor vehicle (EV), called ego vehicle, comprising a camera (CAM) capturing images of the environment close to the ego vehicle (VE), at the front of the ego vehicle (VE), detection means (RAD ) of a target vehicle (VCI) in the environment close to the ego vehicle (VE), at the front of the ego vehicle (VE), means (MGT) for managing the trajectory (TRJ) of the ego vehicle (VE ) for maintaining the ego vehicle (VE) in its ego lane (WE) while respecting a determined positioning instruction of the ego vehicle (VE) in its ego lane (WE); said management means (MGT) implementing the method according to one of claims 1 to 4.

7. Motor vehicle (VE) according to the preceding claim, in which said management means (MGT) are capable, in addition, of implementing a function of maintaining the ego vehicle (VE) in its lane (WE) respecting an instruction positioning in the ego lane (WE) configurable by the user of the ego vehicle (VE), called the LPA function, which when activated, implements the method according to claim 2.

8. Motor vehicle (VE) according to claim 6, in which said management means (MGT) are capable, in addition, of implementing a function of maintaining the ego vehicle (VE) in its lane (WE) respecting an instruction positioning in the ego lane (WE) configurable by the user of the ego vehicle (VE), called LPA function, which when not activated implements the method according to claim 3.