Method and device for detecting a merger of a vehicle into a lane

Description

[0001] The invention is in the field of autonomous vehicle driving assistance systems. In particular, the invention relates to a method and device for detecting insertion into a traffic lane of a vehicle, called a target vehicle, from an autonomous vehicle, called an ego vehicle, for autonomous driving of said ego vehicle as a function of an indicator of the presence of said target vehicle.

[0002] A "vehicle" means any type of vehicle such as a motor vehicle, a moped, a motorcycle, a storage robot in a warehouse, etc. "Autonomous driving" of an "autonomous vehicle" means any process capable of assisting the driving of the vehicle. The process may thus consist of partially or totally steering the vehicle or providing any type of assistance to a natural person driving the vehicle. The process thus covers all autonomous driving, from level 0 to level 5 in the OICA scale, for International Organization of Motor Vehicle Manufacturers.

[0003] The processes capable of assisting the driving of the vehicle are also called ADAS (from the English acronym "Advanced Driver Assistance Systems"), ADAS systems or driver assistance systems. These systems are known to be embedded in a vehicle, called ego vehicle. These systems include sensors capable of perceiving the environment (such as, for example, detecting another vehicle, acquiring information such as the position, speed, acceleration of the other vehicle, detecting and identifying traffic lanes, etc.). These systems also include sensors capable of measuring vehicle parameters and variables (such as, for example, position, speed, acceleration, engine speed, the activation of a turn signal, etc.).Also, these systems include actuators capable of modifying the static and dynamic behavior of the vehicle, and also include means of communication capable of communicating with another vehicle, a server, a passenger of the vehicle.

[0004] In particular, among ADAS systems, an adaptive cruise control is capable of following a vehicle, called the target vehicle, in order to move the ego vehicle according to the same dynamics as the target vehicle while respecting an inter-vehicle time (or inter-vehicle distance; the two are equivalent) and a maximum speed instruction provided, for example, by a driver / passenger of the ego vehicle. A lane keeping device also takes into account the presence of a target vehicle.

[0005] Tracking a target vehicle is very difficult to manage when entering or exiting a bend. Indeed, when a target vehicle shifts slightly laterally relative to the ego vehicle or a trajectory of the ego vehicle, either the vehicle enters a bend or the target vehicle changes lane to, for example, drive in the traffic lane of the ego vehicle. Generally, the trajectory of the ego vehicle is determined by recognition of the traffic lanes and their curvatures from image processing, and recognition of the lane on which the ego vehicle is traveling. In many situations, determining the trajectory of the ego vehicle is temporarily inconsistent with a trajectory of a lane on which the ego vehicle is traveling. Lane recognition by an on-board camera is not possible (road markings masked, erased, and / or not readable due to sunlight, weather, the presence of other objects / vehicles, or roadworks)...). This results, for example, in incorrect deselection of a vehicle or incorrect selection of a vehicle traveling in an adjacent lane.

[0006] Furthermore, the state of the art is known from documents US2018029602A, FR-A-3103438 and EP-A-2676857.

[0007] An object of the present invention is to remedy the aforementioned problem, in particular to select in a robust manner (more availability, functional in a straight line and in a bend) and simple manner an object traveling in front of the ego vehicle. There is no need to identify and model a trajectory of the lanes. To this end, a first aspect of the invention relates to a method for detecting insertion into a traffic lane of a vehicle, called the target vehicle, from an autonomous vehicle, called the ego vehicle, for autonomous driving of said ego vehicle as a function of an indicator of the presence of said target vehicle, said ego vehicle traveling on said traffic lane and comprising sensors capable of providing data from said ego vehicle and from the external environment of said ego vehicle, said method comprising the steps of: Detection, from said sensors, of said target vehicle; Determination, from said sensors, of at least one item of information on the relative positioning of said target vehicle with respect to the ego vehicle; Determination, from said sensors, of a trajectory of said ego vehicle, the trajectory indicating estimated positions of said ego vehicle over a predetermined time horizon; Determination of a minimum distance between said target vehicle and said trajectory, the determination being based on said at least one item of information on the relative positioning; Determination of a first threshold, called the selection threshold, said selection threshold representing a distance and being based on a longitudinal speed of said ego vehicle and on said time horizon; if said minimum distance is less than said selection threshold, increasing the indicator of the presence of said target vehicle;generation of an autonomous driving instruction from the presence indicator.;

[0008] Thus, autonomous driving will be able to adapt the vehicle's steering, just as a driver would, by taking into account future interference, within a given time horizon (equivalent to a duration), between said ego vehicle and another vehicle. It is not necessary to predict or construct the future trajectory of the target vehicle and, in particular, it is not necessary to detect the lines separating the traffic lanes. The threshold being based on the longitudinal speed of the ego vehicle, the detection method is functional in a straight line, when entering or exiting a bend, and when turning. More particularly, when entering or exiting a bend, a vehicle traveling on a lane adjacent to the traffic lane of said ego vehicle will not be considered as present (zero or low probability of presence).

[0009] Advantageously, said selection threshold is equal to Y1(V) = min(Y0; LVo + 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents a minimum function between values, Y0 is a predetermined lateral distance, LVo is a predetermined width of said traffic lane, V is said longitudinal speed, AlatMax(V) is a maximum lateral acceleration as a function of said longitudinal speed, t is said time horizon, and LVC is a predetermined width of said target vehicle.

[0010] Thus, the selection threshold, dependent on the speed and time horizon, is simply set and valid for all longitudinal speeds of the ego vehicle. The development is very simple. It is based on the physics of the vehicle and does not require vehicle tests for different longitudinal speeds and for different cornering radii.

[0011] Advantageously, said method further comprises a step of determining a second threshold, called the deselection threshold, said deselection threshold representing a distance strictly greater than the selection threshold, said deselection threshold being based on a longitudinal speed of said ego vehicle and on said time horizon, if said minimum distance is greater than said deselection threshold, then the presence indicator decreases, otherwise the presence indicator remains unchanged.

[0012] Thus, the determination of the probability of presence will be even more robust, particularly in the event of poor perception of said target vehicle. A target vehicle traveling on the same traffic lane as said ego vehicle will remain considered present (100% or high probability of presence) entering or exiting a bend despite the uncertainties of detection of entry or exit of the bend of the ego vehicle and / or the target vehicle. The difference between the selection threshold and the deselection threshold is called the Unknown zone.

[0013] Advantageously, said deselection threshold is equal to Y2(V) = min(Y0, 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents said minimum function, Y0 is said predetermined lateral distance, V is said longitudinal speed, AlatMax(V) is said maximum lateral acceleration as a function of said longitudinal speed, t is said time horizon, and LVC is said predetermined width of said target vehicle.

[0014] Advantageously, the method is executed periodically, a probability of presence, between 0% and 100%, is determined from a history of determinations of said presence indicator, and said autonomous driving is a function of said probability of presence.

[0015] This way, the detections are consolidated over time. This provides more robustness of detection and allows for fewer variations in the autonomous driving of the ego vehicle, thus making driving smoother and less abrupt.

[0016] A second aspect of the invention relates to a device comprising a memory associated with at least one processor configured to implement the method according to the first aspect of the invention.

[0017] The invention also relates to a vehicle comprising the device.

[0018] The invention also relates to a computer program comprising instructions adapted for executing the steps of the method, according to the first aspect of the invention, when said program is executed by at least one processor.

[0019] Other characteristics and advantages of the invention will emerge from the description of the non-limiting embodiments of the invention below, with reference to the appended figures, in which: [ Fig. 1 ] illustrates a state-of-the-art living situation [ Fig. 2 ] schematically illustrates a device, according to a particular exemplary embodiment of the present invention. [ Fig. 3 ] schematically illustrates a method for detecting insertion into a traffic lane of a vehicle, according to a particular exemplary embodiment of the present invention.

[0020] The invention is described below in its non-limiting application to the case of an autonomous motor vehicle traveling on a road or on a traffic lane. Other applications such as a robot in a storage warehouse or a motorcycle on a country road are also conceivable.

[0021] There figure 1 represents a state-of-the-art real-life situation. An ego vehicle 101 is traveling on a traffic lane 103 delimited by a left side / marking 104 and a right side / marking 105. In many cases (bad weather - rain, snow, low light, etc. -, wear, masking, etc.) the markings 104, 105 will not be detected by a sensor of an ADAS system embedded in the ego vehicle 101. Adjacent to this traffic lane 103, another vehicle 102 is also traveling. For example, the traffic lane 103 may represent a turn entry. The ego vehicle 101 is traveling in a straight line, the sensors of the ego vehicle 101 do not measure any intention to turn and follow a curvature of the traffic lane 103, in particular if the markings 104, 105 are not detected by another system.

[0022] The hatched area 106 represents the trajectory that the ego vehicle 101 will follow over a given time horizon (for example 1.5 seconds). A tracking sensor, of the RADAR or LIDAR type for example, detects the other vehicle 102. The other vehicle is then selected to be, for example, considered as a target vehicle of a driving assistance system. This leads to undesired behavior (heavy braking, etc.) of the ego vehicle 101.

[0023] There figure 2 represents an example of a device 201 included in the vehicle, in a network (“cloud”) or in a server. This device 201 can be used as a centralized device in charge of at least certain steps of the method described below with reference to the figure 2 In one embodiment, it corresponds to an autonomous driving computer.

[0024] In the present invention, the device 201 is included in the vehicle.

[0025] This device 201 can take the form of a box comprising printed circuits, any type of computer or even a mobile telephone (“smartphone”).

[0026] The device 201 comprises a RAM 202 for storing instructions for the implementation by a processor 203 of at least one step of the method as described above. The device also comprises a mass memory 204 for storing data intended to be retained after the implementation of the method.

[0027] The device 201 may further comprise a digital signal processor (DSP) 205. This DSP 205 receives data to format, demodulate and amplify, in a manner known per se, this data.

[0028] The device 201 also comprises an input interface 206 for receiving the data implemented by the method according to the invention and an output interface 207 for transmitting the data implemented by the method according to the invention.

[0029] For example, the input interface 206 can receive the following data: position or geographical location of the vehicle, speed and / or acceleration of the vehicle, set or predetermined positions / speeds / accelerations, engine speed, position and / or travel of the clutch, brake and / or acceleration pedal, detection of other vehicles or objects, position or geographical location of the other vehicles or objects detected, speed and / or acceleration of the other vehicles or objects detected, operating states of sensors, confidence index of data originating from or processed by sensors and / or devices similar to the device 201. For example, the sensors capable of providing data are: GPS associated or not with mapping, tachometers, accelerometers, RADAR, LIDAR, lasers, ultrasound, camera, etc.

[0030] There figure 3schematically illustrates a method for detecting insertion into a traffic lane 103 of a vehicle, called target vehicle, from an autonomous vehicle, called ego vehicle 101, for autonomous driving of said ego vehicle 101 as a function of an indicator of presence of said target vehicle, said ego vehicle 101 traveling on said traffic lane 103 and comprising sensors capable of providing data from said ego vehicle 101 and from the external environment of said ego vehicle 101, according to a particular embodiment of the present invention.

[0031] In one operating mode, the presence indicator is a Boolean to indicate whether a target vehicle is present or not. For example, if the target vehicle is not initially detected as present, the presence indicator is 0, and an increase in said presence indicator results in a value of 1 for the presence indicator. A decrease in the presence indicator results in a value of 0 for the presence indicator.

[0032] In another operating mode, the presence indicator is a number that can increase and decrease. An increase or decrease in the presence indicator is a function of a minimum distance described below. For example, the presence indicator will gradually change from 0 to 1 when the minimum distance is less than a selection threshold described below.

[0033] In another operating mode, a probability of presence, between 0% and 100%, is determined from a history of determinations of said presence indicator. Said autonomous driving is a function of said probability of presence.

[0034] Step 301, Detect, is a step of detecting, from said sensors, said target vehicle. For example, the detection is done from image processing of a camera on board the vehicle ego 101, by processing data from sensors based on electromagnetic, light and / or sound waves.

[0035] Step 302, Det_pr, is a step of determining, from said sensors, at least one piece of information on the relative positioning of said target vehicle with respect to the ego vehicle 101. An object having been detected, it is possible to determine the position of a point of the detected object with respect to a reference frame of the ego vehicle 101, it is a relative position which is determined. For example, in the case of a vehicle detected to the right with respect to the direction of travel of the ego vehicle 101, the closest point, generally the left rear of the target vehicle, will be expressed in a reference frame of the ego vehicle 101, for example in the center at the front of the ego vehicle 101. The position of the detected target vehicle is expressed in a reference frame with respect to the ego vehicle 101 thus facilitating post-processing.

[0036] Step 303, det_traj, is a step of determining, from said sensors, a trajectory of said ego vehicle 101, the trajectory indicating estimated positions of said ego vehicle 101 over a predetermined time horizon. Conventionally, processing of data from these sensors is capable of measuring the longitudinal and transverse dynamics (position, speed, linear or rotational acceleration) of the ego vehicle 101 or of measuring inputs (steering wheel angle, pedal depression, etc.) of a driver or a control member. A temporal processing of this data gives an estimated trajectory of the ego vehicle 101. In a simple case, at a given instant, a measured steering wheel angle of 0° and a measured longitudinal vehicle speed of 10 m / s, gives an estimate of a straight-line advance of the ego vehicle 101 and at the end, for example 1.5 seconds, the vehicle will have advanced 15 meters.

[0037] Typically, the time horizon is in the order of seconds; other values ​​are possible. In an operating mode, the time horizon varies with the vehicle speed or other parameters such as an inter-vehicle time selected by a driver or predefined in the design.

[0038] The trajectory indicates the successive estimated positions of the ego vehicle 101 over several instants. This trajectory is generally expressed parametrically with respect to time in a frame of reference of the ego vehicle 101.

[0039] Step 304, Det_dist, is a step of determining a minimum distance between said target vehicle and said trajectory, the determination being based on said at least one piece of relative positioning information. By expressing in the same frame of reference the coordinates of the target vehicle and the estimated trajectory, the minimum distance between the target vehicle and the estimated trajectory is determined by projection for example.

[0040] Step 305, Det_s1, is a step of determining a first threshold, called the selection threshold, said selection threshold representing a distance and being based on a longitudinal speed of said ego vehicle 101 and on said time horizon.

[0041] The selection threshold is used to characterize the minimum distance, and, for example, to indicate whether the target vehicle is close or not to the ego vehicle 101, to indicate whether the target vehicle is in the trajectory of the ego vehicle 101, etc.

[0042] In one operating mode, the selection threshold varies according to the speed of the vehicle ego 101, also called longitudinal speed. The lower the speed of the vehicle ego 101, the lower the threshold. For example, the selection threshold varies between 0.2 and 1.3 meters. Other values ​​are possible and depend on the reference frame in which the coordinates of the measured positions, speeds and accelerations are expressed.

[0043] Advantageously, said selection threshold is equal to Y1(V) = min(Y0; LVo + 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents a minimum function between values, Y0 is a predetermined lateral distance, LVo is a predetermined width of said traffic lane 103, V is said longitudinal speed, AlatMax(V) is a maximum lateral acceleration as a function of said longitudinal speed, t is said time horizon, and LVC is a predetermined width of said target vehicle.

[0044] The predetermined distance Y0 is of the order of 1 meter, but other values ​​are possible. Y0 can represent the distance between one side of the ego vehicle 101 and an edge of the lane on which the ego vehicle 101 is traveling. In one operating mode, these values ​​are a function of the speed of the ego vehicle 101 and / or of the target vehicle in order to take into account a danger (high speed) or measurement uncertainties varying according to the speed of the ego vehicle 101. For example, in a straight line, Y0 is smaller than the second term / value of the minimum function. Advantageously, the width of the lane LVo is equal to 3.5 meters. Other values ​​are possible. In one operating mode, the width of the lane is a function of measurements from sensors or data embedded or unloaded in the ego vehicle 101.In another operating mode, a value of the target vehicle lane width parameter depends on the ego vehicle reference point considered and / or depends on the position of the ego vehicle 101 relative to the traffic lane 103.

[0045] Advantageously, the width of the target vehicle LVC is equal to 1.6 meters. A width of the ego vehicle 101 is substantially also equal to the width of the target vehicle. Other values ​​are possible. In one operating mode, the width of the target vehicle is a function of measurements from sensors or data embedded or unloaded in the ego vehicle 101. In another operating mode, a value of the width parameter of the target vehicle depends on the point considered on the target vehicle to calculate the minimum distance and / or depends on the position of the ego vehicle 101 relative to the traffic lane 103.

[0046] AlatMax(V) is a maximum lateral acceleration function of said longitudinal speed. Conventionally, AlatMax(V) is a function parameterized according to the speed of the vehicle ego 101. In an operating mode, this function is represented by a table or a map.

[0047] In a preferred operating mode, the AlatMax(V) function represents a maximum lateral acceleration before a loss of comfort. It is a decreasing function with the ego vehicle speed between 6 m / s 2 < and 2 m / s 2 < . For example, the lower the ego vehicle speed, the more we are able to withstand a significant lateral acceleration. The higher the vehicle speed, the less safe we ​​feel when cornering and the lower the maximum lateral acceleration should be.

[0048] Over a fixed time horizon, for example 1.5 seconds, the AlatMax(V) function anticipates the maximum lateral deviation that the vehicle ego 101 will make if it turns and if we wish to remain in a comfortable lateral acceleration.

[0049] The expression "LVo + 0.5*AlatMax(V)*t^2 - 0.5*LVC" represents a lateral deviation or distance in meters from the longitudinal axis passing through the middle of the ego vehicle 101. This deviation takes into account the possible lateral displacement, 0.5*AlatMax(V)*t^2, of the ego vehicle 101 over the duration defined by the time horizon. If the minimum distance is less than this deviation, then there is a risk of collision because the ego vehicle 101 will potentially get too close, during the duration defined by the time horizon, to the target vehicle. The maximum lateral displacement of the ego vehicle 101, based on the ALatMax function, is taken into account over a given time horizon.

[0050] Step 306, Det_s2, is a step of determining (306) a second threshold, called the deselection threshold, said deselection threshold representing a distance strictly greater than the selection threshold, said deselection threshold being based on a longitudinal speed of said ego vehicle 101 and on said time horizon. The deselection threshold is used to characterize a distance of separation, and, for example, to indicate whether the target vehicle is moving away from the ego vehicle 101 or not, to indicate whether the target vehicle remains in the trajectory of the ego vehicle 101, ...

[0051] In an operating mode, the deselection threshold varies according to the speed of the vehicle ego 101. The higher the speed of the vehicle ego 101, the lower the deselection threshold. For example, the deselection threshold varies between 2 and 0.5 meters. Other values ​​are possible and depend on the reference frame in which the coordinates of the measured positions, speeds and accelerations are expressed.

[0052] Advantageously, said deselection threshold is equal to Y2(V) = min(Y0, 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents said minimum function, Y0 is said predetermined lateral distance, V is said longitudinal speed, AlatMax(V) is said maximum lateral acceleration as a function of said longitudinal speed, t is said time horizon, and LVC is said predetermined width of said target vehicle.

[0053] In a preferred mode of operation, ALatMax(V) is the same function, giving a maximum lateral acceleration as a function of the speed of the vehicle ego 101, as that used in determining the selection threshold. In another mode of operation, the same function with a different parameterization is used. In another mode of operation, a different function is used.

[0054] The expression "0.5*AlatMax(V)*t^2 - 0.5*LVC" represents a lateral deviation or distance in meters from the longitudinal axis passing through the middle of the ego vehicle 101. This deviation takes into account the possible lateral displacement, 0.5*AlatMax(V)*t^2, of the ego vehicle 101 over the duration defined by the time horizon. If the minimum distance is greater than this deviation, then there is little risk of collision because the ego vehicle 101 will not be able to get too close, during the duration defined by the time horizon, to the target vehicle. The maximum lateral displacement of the ego vehicle 101, based on the ALatMax function, is taken into account over a given time horizon.

[0055] Step 307, Test1, is a test step to check whether said minimum distance is less than said selection threshold. If so, the ego vehicle 101 and the target vehicle move closer and we move on to step 309.

[0056] Step 308, Test2, is a test step to check whether said minimum distance is greater than said deselection threshold. If so, the ego vehicle 101 and the target vehicle move away and we move on to step 310.

[0057] Step 309, Augm, is a step of increasing the presence indicator of said target vehicle. In one operating mode, this increase is the change from 0 to 1 of the value of the presence indicator. In another operating mode, the increase is a function of a predetermined increment or an increment depending on the minimum distance (the smaller the minimum distance, the greater the increment).

[0058] Step 310, Dim, is a decrease step of the presence indicator. In one operating mode, this decrease or reduction is the passage from 1 to 0 of the value of the presence indicator. In another operating mode, the decrease is a function of a predetermined increment or an increment that is a function of the minimum distance (the greater the minimum distance relative to the deselection threshold, the greater the increment).

[0059] Step 311, Inch, is a step where the presence indicator remains unchanged. We are in the so-called Unknown zone. This zone is dynamic and depends on the speed of the vehicle ego 101. The lower the speed of the vehicle ego 101, the greater the zone, the difference between the deselection threshold and the selection threshold. At very high vehicle speeds, for example greater than 40 m / s, this difference is reduced to 0, Y1(V)=Y2(V)=Y0.

[0060] Step 312, Gen, is a step of generating an autonomous driving instruction from the presence indicator. The presence or absence of a target vehicle likely to interfere on a future trajectory, over the time horizon, of the ego vehicle 101 is essential information for autonomous driving. In the present invention, this presence is made reliable by taking into account uncertainties depending on the speed of the ego vehicle 101. In the event of uncertainty, in the Unknown zone, the presence indicator is not modified.

[0061] Advantageously, the method is executed periodically. The period is of the order of 10 milliseconds. Other period values ​​are possible, such as 10 microseconds or 10 seconds. In an operating mode, the period can even be variable depending on the computing load, the vehicle speed or other conditions. Thus, the detection of a target vehicle is almost permanent as is the generation of the driving instruction, thus allowing regular feedback, and thus being adaptive depending on the traffic context (vehicle speed, presence of target vehicle, road types, etc.).

[0062] Advantageously, a probability of presence, between 0% and 100%, is determined from a history of the determinations of said presence indicator, and said autonomous driving is a function of said probability of presence. Indeed, the lower the presence indicator is than Y1(V), the more certain one is of detecting a vehicle that enters the traffic lane 103 of the ego vehicle 101. The longer the presence indicator remains below Y1(V), the more certain one is of detecting a vehicle that enters the traffic lane 103 of the ego vehicle 101. Conversely, the higher the presence indicator is than Y2(V) or the longer the indicator is higher than Y2(V), the more certain one is of deselecting the target vehicle. By taking the difference into account, a gradient, a variation, of a parameter is determined. For example, the gradient is 0.1 if the indicator is less than Y1(V) and is -0.1 if the indicator is greater than Y2(V).The cumulative sum of this parameter at each determination of the parameter divided in relation to a maximum threshold is representative of a percentage, the percentage being bounded between 0 and 100%. The use of the probability of presence is particularly robust to measurement uncertainties. In an operating mode, statistical tests of threshold crossing are also used.

[0063] In one operating mode, said autonomous driving is an adaptive cruise control function. Adaptive cruise control needs to have reliable detection of the presence of a target vehicle, particularly when entering or exiting a bend. Unreliability leads to very significant variations in the vehicle's speed, which are currently minimized by patches and complex pieces of code that are difficult to configure and do not work generically (either in a straight line or when cornering).

[0064] The present invention is not limited to the embodiments described above as examples; it extends to other variants.

[0065] Thus, an example of an implementation has been described above in which positions, distances, speeds and accelerations are expressed in a reference frame for a given parameter setting. The parameter setting can be modified depending on the reference frame considered. For example, half the width of an object can be equal to 0.

Claims

1. Method for detecting insertion into a taxiway (103) of a vehicle, referred to as a target vehicle, from an autonomous vehicle, referred to as an ego vehicle (101), for autonomous driving of said ego vehicle (101) as a function of an indicator of the presence of said target vehicle, said ego vehicle (101) travelling on said taxiway (103) and comprising sensors capable of providing data of said ego vehicle (101) and of the external environment of said ego vehicle (101), said method comprising the steps of: • Detection (301), from said sensors, of said target vehicle; • Determination (302), from said sensors, of at least one piece of information on the relative positioning of said target vehicle by report to the ego vehicle (101); • Determination (303), from said sensors, of a trajectory of said EGO vehicle (101), the trajectory indicating estimated positions of said EGO vehicle (101) over a predetermined horizon of time; • Determination (304) of a minimum distance between said target vehicle and said trajectory, the determination being based on said at least one item of relative positioning information; • Determination (305) of a first threshold, called the selection threshold, said selection threshold representing a distance and being based on a longitudinal speed of said ego vehicle (101) and on said time horizon; • If (307) said minimum distance is less than said selection threshold, increase (309) of the presence indicator of said target vehicle; • Generation (312) of an autonomous driving instruction from the presence indicator.

2. Method according to claim 1, wherein said selection threshold is equal to Y1(V) = min(Y0; LVo + 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents a minimum function between values, Y0 is a predetermined lateral distance, LVo is a predetermined width of said taxiway (103), V is said longitudinal speed, AlatMax(V) is a maximum lateral acceleration dependent on said longitudinal speed, t is said horizon of time, and LVC is a predetermined width of said target vehicle.

3. Method according to one of the previous claims, wherein said method further comprises a step of determining (306) a second threshold, called the deselection threshold, said deselection threshold representing a distance strictly greater than the selection threshold, said deselection threshold being based on a longitudinal speed of said ego vehicle (101) and on said time horizon, if (308) said minimum distance is greater than said deselection threshold, then the presence indicator decreases (310), otherwise the presence indicator remains unchanged (311).

4. Method according to claim 3, wherein said deselection threshold is equal to Y2(V) = min(Y0, 0.5*AlatMax(V)*t^2 - 0.5*LVC), where min represents said minimum function, Y0 is said predetermined lateral distance, V is said longitudinal speed, AlatMax(V) is said maximum lateral acceleration as a function of said longitudinal speed, t is said time horizon, and LVC is said predetermined width of said target vehicle.

5. Method according to one of the previous claims, in which the method is executed periodically, a presence probability of between 0% and 100% is determined from a log of the determinations of said presence indicator, and said autonomous driving is a function of said presence probability.

6. Method according to one of the previous claims, wherein said autonomous driving is an adaptive speed regulation function.

7. Device (201) comprising a memory (202) associated with at least one processor (203) configured to implement the method according to one of the previous claims.

8. Vehicle comprising the device according to the previous claim.

9. A computer Plan comprising instructions adapted for executing the steps of the method according to one of claims 1 to 6 when said plan is executed by at least one processor (203).