FR3075734A1 - METHOD OF DETECTING NEXT FUSION OF THE CIRCULATION PATH OF A VEHICLE WITH AN ADJACENT PATHWAY DURING AN AUTOMATED DRIVING PHASE. - Google Patents

Abstract

The invention relates to a method for detecting a merging of the lane of a vehicle with an adjacent lane, comprising the steps of determining the profiles of the demarcation lines defining said lane of traffic over a visibility distance (200). , extrapolating the profiles between the latter and a predetermined anticipation distance (300), checking the convergence of said lines on said anticipation distance (400), storing the convergence distance (500), and updating, at the end of a predetermined time of said convergence distance (700), of determining simultaneously with said step of updating a second convergence distance by reproducing said determination, extrapolation, verification and storage steps ( 800), and validation of a next merger if the difference between the updated convergence distance and the second convergence distance is smaller than or equal to a predetermined interval (1000).

Description

METHOD FOR DETECTING A NEXT MERGER OF THE VEHICLE TRACK WITH AN ADJACENT TRACK DURING AN AUTOMATED DRIVING PHASE.

Field of the invention

The present invention relates to the field of driving assistance systems for motor vehicles.

It aims in particular at the problem of detecting a next merger of the lane of a vehicle with an adjacent lane during an automated driving phase.

Invention background

Driver assistance systems are now widely used in recent motor vehicles and are experiencing rapid development.

Recently, we have started to see completely autonomous driving assistance systems simultaneously controlling the longitudinal and lateral movement of the vehicle.

This is particularly the case for highway driving assistance systems commonly designated under the acronyms HAD (for “Highway Automated Driving”) or AHDA (for “Automatic Highway Driving Assist”) and capable of completely relieving the driver in situations smooth traffic (speed close to the regulatory limit) on roads with separate carriageways.

In such a driving assistance system, the positioning of the vehicle in its lane is determined relative to the first and second lines of marking on the ground delimiting this lane and detected via a camera and / or external sensors.

In certain situations, the driving assistance system may however have certain difficulties in correctly positioning the vehicle in its lane, in particular when the latter is subdivided into two separate lanes or comes to merge with another adjacent lane.

Indeed, in the first situation (subdivision) the traffic lane widens before there appears at least one intermediate delimitation separating it into two distinct lanes.

In the second situation (merger), one of the two marking lines on the ground defining its old taxiway suddenly disappears while its new merged taxiway is temporarily bounded by two much more widely spaced tagging lines.

Driving assistance systems generally react to this sudden widening of the lane by a marked deviation from the vehicle which can then end up on the path of another vehicle coming from the adjacent lane, which is particularly dangerous. .

In addition, this reaction of the vehicle is a source of stress for the driver who is forced to act and takes control of the steering wheel to put the vehicle back on the right path.

We know from patent application US 2016/0272203 a method for detecting a next subdivision of the vehicle traffic lane into two separate lanes in order to adapt its trajectory.

However, this method does not make it possible to detect cases of merger in front of the vehicle of its lane with another adjacent lane.

Subject and summary of the invention

The present invention aims to improve the situation.

To this end, it proposes a method for detecting a next merger of the lane of a vehicle with an adjacent lane during an automated driving phase, characterized in that it comprises the following steps: - determination, from data captured by optical detection means monitoring the environment located in front of the vehicle, profiles of the two demarcation lines defining the path of travel of said vehicle over the visibility distance of said lines by said detection means; - extrapolation of the profiles of said demarcation lines between said visibility distance and a predetermined anticipation distance; - Verification of the convergence of said demarcation lines over said anticipation distance; - memorization in the affirmative of the convergence distance of said demarcation lines; updating, at the end of a predetermined period, of said convergence distance by subtracting from the previously memorized value the distance traveled by said vehicle during said period; - Determination, simultaneously with said updating step, of a second convergence distance by reproducing said steps of determining the profiles, extrapolating the latter, checking the convergence and memorizing; and - validation of the existence, at said updated convergence distance, of a future merger of the vehicle traffic lane with the adjacent traffic lane if the difference between the updated convergence distance and the second convergence distance is less than or equal to a predetermined interval.

The method according to the invention makes it possible to detect in a particularly reliable manner a forthcoming merger of the vehicle lane with an adjacent lane during an automated driving phase, so as to give the autonomous vehicle the possibility of reacting accordingly by securing its trajectory and / or asking the driver to regain control.

According to preferred characteristics of the method according to the invention, taken alone or in combination: - said predetermined time is between 0.5 and 1.5 seconds; - That said predetermined interval is between 30 and 70 centimeters; - The existence of a next merger of the vehicle lane with the adjacent lane is directly validated when said initially determined convergence distance is less than or equal to the visibility distance of said demarcation lines; - Said method comprises, after validation of the existence of a next merger of the vehicle traffic lane with the adjacent traffic lane, a step of adapting the lateral trajectory of said vehicle so that the latter fits into said adjacent track; - Said method comprises, after validation of the existence of a next merger of the vehicle lane with the adjacent lane, a step of issuing to the driver an alert asking him to resume hands said vehicle; - the extrapolation of the profiles of said demarcation lines is obtained by approximating each one by a third degree polynomial function in a plane affine space; - said detection method is implemented periodically according to a predetermined period of between 20 and 100 milliseconds; - Said optical detection means comprise a video camera; and / or - said optical detection means comprise a LIDAR.

Brief description of the drawings The description of the invention will now be continued with the detailed description of an exemplary embodiment, given below by way of illustration but not limitation, with reference to the appended drawings, in which: - Figure 1 is a functional diagram of an automated driving assistance system; - Figure 2 illustrates a next merger of the vehicle lane with an adjacent lane; and - Figure 3 shows a flow diagram of the method according to the invention for detecting a next merger of the lane of a vehicle with an adjacent lane during an automated driving phase.

detailed description

With reference to FIG. 1, the automated driving assistance system 1 comprises a detection module 10, an actuation module 20, an information and alert module 30, an environmental context assessment module 40 and a control unit 50.

The detection module 10 comprises optical detection means making it possible to assess the environmental context of driving the vehicle and comprising a video camera 11 located behind the windshield.

According to variants not shown, the detection module 10 may include a lidar to replace the video camera 11.

The module 10 also includes a plurality of sensors measuring certain internal driving parameters such as the instantaneous speed of the vehicle and the steering angle of the steering wheel.

The actuation module 20 comprises a plurality of actuators able to act on the steering, acceleration, braking and the gearbox to ensure fully automated driving of the vehicle.

The information and alert module 30 is capable of requesting the driver via the various man-machine interaction (HMI) means of visual, audible, and / or haptic type with which the vehicle is equipped.

The visual interaction means are for example constituted by the touch-screen display of the infotainment device integrated into the central console of the dashboard, by that of the instrument cluster, and / or by that of the head-up display device when the vehicle is fitted.

The sound interaction means are conventionally formed by the loudspeakers of the infotainment system and / or by an independent system alarm.

The haptic interaction means advantageously consist of a plurality of vibrators installed in the driver's seat.

The requests consist for example of the display of an indicator light, the display of an informative or alert text message, the emission of an audible signal, the emission of an informative voice message or alert, activation of the vibrators installed in the seat, or even on the steering wheel.

The driving environmental context evaluation module 40 is able to determine from the data provided by the detection module 10 the type of road taken (motorway, expressway or secondary road), the markings on the ground (color, width and line spacing), the level of fluidity of road traffic and the possible presence of a barrier or a central reservation between the two directions of traffic. The control unit 50 which manages the autonomous driving of the vehicle, comprises at least one computer associated with storage means comprising non-volatile memory of EEPROM or FLASH type and random access memory.

When the context evaluated by the driving environmental context evaluation module 40 is suitable (for example, in the event of heavy traffic or congestion when the vehicle is traveling on a road with separate carriageways), the control unit 50 is able to control the actuation module 20 to ensure fully autonomous driving of the vehicle.

In particular, the control of the lateral movement of the vehicle is controlled by an active assistance module for maintaining the trajectory 51 integrated into this control unit

According to the invention, the non-volatile memory of the control unit 50 stores a process for detecting a next merger of the vehicle traffic lane with another adjacent lane.

We will now describe in detail and with the aid of the flow diagram of FIG. 3, the different stages of this process which is implemented periodically according to a predetermined period preferably between 20 and 100 milliseconds and advantageously equal to 40 milliseconds.

This is initiated when the fully autonomous driving assistance mode is activated (initial step 100), the longitudinal and lateral displacements of the vehicle being then managed entirely autonomously by the control unit 50.

The active trajectory maintenance assistance module 51 then determines, from the data captured by optical detection means 11 monitoring the environment located in front of the vehicle V, the profiles of the two demarcation lines Li, L2 defining the taxiway Vc of this vehicle V (see FIG. 2) over the visibility distance Dv of these lines by the optical detection means 11 (step 200).

It should be noted that the visibility distance Dv of the demarcation lines Li, L2 changes as a function of the driving environmental context (the latter will for example correspond to the range of the optical detection means 11 under optimal conditions but may be considerably less when the weather conditions are bad (rain, snow, fog) or in the event of a traffic jam when these lines are hidden by obstacles or other vehicles located upstream.

The active assistance module for maintaining the trajectory 51 then extrapolates to the next step 300 the profiles of these two demarcation lines Li, L2 between the visibility distance Dv and a predetermined anticipation distance DA of for example between 100 and 200 m.

This extrapolation of the profiles of these two lines of demarcation Li, L2 is obtained by approximating each one by a corresponding third degree polynomial function in a plane affine space provided with a Cartesian coordinate system (X, Y) centered on the vehicle: ii Y = - (initial derivative of the curvature) X3 + - (initial curvature) X2 + (initial course) X + Yo 6 2 where: - X and Y correspond respectively to the longitudinal and lateral components of a point belonging to this line of demarcation ; the initial derivative of the curvature corresponds to the value of the derivative of the curvature at the origin of the reference frame, that is to say at the level of the vehicle at X = 0; - the initial curvature corresponds to the value of the derivative of the curvature at the origin of the benchmark, that is to say at the level of the vehicle at X = 0; - the initial heading (or yaw angle) corresponds to the value of the heading at the origin of the benchmark, that is to say at the level of the vehicle at X = 0; and - Yo corresponding to the initial lateral position of the line of demarcation at the origin of the reference point, that is to say at the level of the vehicle at X = 0.

The process then checks in step 400 if the two demarcation lines Li, L2 converge on the anticipation distance.

If not, the control unit estimates that the taxiway Vc on which the vehicle V is traveling will not merge with the adjacent taxiway VA and the process stops.

If so, the process stores this convergence distance Dc (step 500) and then checks whether the latter is less than or equal to the visibility distance Dv (step 600).

If this condition is fulfilled, the control unit 50 considers that the risk of error is almost zero and validates the existence, at this convergence distance Dc, of a next merger of the taxiway Vc of the vehicle V with the adjacent traffic lane VA (step 1000).

The process then transmits to the actuation module 20 an instruction to adapt the lateral trajectory of the vehicle V so that the latter is inserted in the adjacent lane VA taking into account the vehicles traveling there for braking and acceleration ( step 1100).

In addition to or in replacement of this step 1100, the process can also transmit to the information and alert module 30 an order to transmit to the driver a visual, audible and / or haptic alert asking him to take over vehicle V (step 1200).

We will now return to step 600 in which the control unit 50 determines whether the convergence distance Dc is less than or equal to the visibility distance Dv.

If this condition is not fulfilled (in other words, if the point of convergence is located in the extrapolation zone as in the case of figure of figure 3), then the process estimates that the risk of error is more important and will therefore carry out a check.

To do this, it will perform at the end of a predetermined delay ΔΤ (preferably between 0.5 and 1.5 seconds and preferably equal to 1 second) an update of the convergence distance Dc by subtracting from the stored value in step 500 the distance traveled by the vehicle during said delay ΔΤ (this distance being obtained using the vehicle speed data transmitted by the sensors of the module 10) (step 700);

Simultaneously with said updating step, the control unit 50 will also determine during step 800 a second convergence distance by reproducing the steps of determining the profiles (200), of extrapolating the latter (300) , verification of convergence (400) and storage (500).

The process will then check whether the difference between the updated convergence distance and the second convergence distance is less than or equal to a predetermined interval preferably between 30 and 70 centimeters (step 900).

If this condition is verified, the control unit 50 considers that the risk of error almost zero and validates the existence, at the updated convergence distance, of a next merger of the taxiway Vc of the vehicle V with the adjacent traffic lane VA (step 1000).

Otherwise, the control unit 50 considers that the traffic lane Vc on which the vehicle V is traveling will not merge with the adjacent lane VA and the process stops.

According to variant embodiments, the method does not include step 400 of checking whether the initially determined convergence distance is less than or equal to the visibility distance of the demarcation lines Li, L2, so that steps 700 to 900 are systematically performed and the existence of a next merger of the vehicle lane Vc with the adjacent lane Va is only validated if the difference between the updated convergence distance and the second convergence distance is less or equal to a predetermined interval. In general, it is recalled that the present invention is not limited to the embodiments described and shown, but that it encompasses any variant of execution within the reach of the skilled person.

Claims (10)

1. Method for detecting a next merger of the traffic lane (Vc) of a vehicle (V) with an adjacent lane (VA) during an automated driving phase, characterized in that it comprises the following steps: determination, from data captured by optical detection means (11) monitoring the environment located in front of the vehicle, of the profiles of the two demarcation lines (Li, L2) defining the traffic lane (Vc) of said vehicle ( V) over the visibility distance (Dv) of said lines (Li, L2) by said detection means (H) (200); - extrapolation of the profiles of said demarcation lines (Li, l_2) between said visibility distance (Dv) and a predetermined anticipation distance (DA) (300); - Verification of the convergence of said demarcation lines (Li, L2) over said anticipation distance (DA) (400); - storage in the event of detected convergence of the convergence distance (De) of said demarcation lines (Li, L2) (500); - updating, at the end of a predetermined delay (ΔΤ), of said convergence distance (De) by subtracting from the previously memorized value the distance traveled by said vehicle (V) during said delay (ΔΤ) (700); - determination, simultaneously with said updating step (700), of a second convergence distance by reproducing said steps of determining the profiles (200), of extrapolation of the latter (300), of verification of the convergence (400) and storage (500) (800); and - validation of the existence, at said updated convergence distance, of a next merger of the taxiway (Vc) of the vehicle (V) with the adjacent taxiway (VA) if the difference between the distance of updated convergence and the second convergence distance is less than or equal to a predetermined interval (1000). 2. Detection method according to claim 1, characterized in that said predetermined delay (ΔΤ) is between 0.5 and 1.5 seconds. 3. Detection method according to one of claims 1 or 2, characterized in that said predetermined interval is between 30 and 70 centimeters. 4. Method for detecting a next merger of the traffic lane (Vc) of a vehicle (V) with an adjacent lane (Va) during an automated driving phase, characterized in that it comprises the following steps: determination, from data captured by optical detection means (11) monitoring the environment located in front of the vehicle, of the profiles of the two demarcation lines (Li, L2) defining the traffic lane (Vc) of said vehicle ( V) over the visibility distance (Dv) of said lines (Lx, L2) by said detection means (H) (200); - extrapolation of the profiles of said demarcation lines (Li, l_2) between said visibility distance (Dv) and a predetermined anticipation distance (Da) (300); - Verification of the convergence of said demarcation lines (Li, L2) over said anticipation distance (DA) (400); - memorization in case of detected convergence of the convergence distance (Dc) of said demarcation lines (Li, L2) (500); - validation of the existence, at said convergence distance (De), of a next merger of the taxiway (Vc) of the vehicle (V) with the adjacent taxiway (VA) (1000) if said distance of convergence (De) is less than or equal to the visibility distance (Dv) of said demarcation lines (Li, L2). 5. Detection method according to one of claims 1 to 4, characterized in that it comprises, after validation of the existence of a next merger of the traffic lane (Vc) of the vehicle (V) with the lane adjacent traffic (VA), a step of adapting the lateral trajectory of said vehicle so that the latter fits into said adjacent lane (VA) (1100). 6. Detection method according to one of claims 1 to 5, characterized in that it comprises, after validation of the existence of a next merger of the traffic lane (Vc) of the vehicle (V) with the lane adjacent traffic (VA), a step of transmitting to the driver an alert asking him to take over said vehicle (1100). 7. Detection method according to one of claims 1 to 6, characterized in that the extrapolation of the profiles of said demarcation lines (Li, L2) is obtained by approximating them each by a third degree polynomial function in an affine space plan. 8. Detection method according to one of claims 1 to 7, characterized in that it is implemented periodically according to a predetermined period between 20 and 100 milliseconds. 9. Detection method according to one of claims 1 to 8, characterized in that said optical detection means comprise a video camera (11). 10. Detection method according to one of claims 1 to 9, characterized in that said optical detection means comprise a LIDAR.