Driver assistance system with virtual goal for adaptive cruise control

Description

[0001] The invention relates to a driving assistance method with a virtual target for ACC regulation. It finds an advantageous application in the form of a driving assistance method for an ego vehicle in a motor vehicle equipped with such a driving assistance system.

[0002] It also relates to a driver assistance system for an ego motor vehicle.

[0003] It also relates to a motor vehicle comprising a powertrain, braking means and such an assistance system.

[0004] Driver assistance technologies are becoming increasingly widespread and are no longer limited to high-end vehicles.

[0005] These technologies make it possible to simplify the driving of motor vehicles and / or to improve the reliability of driver behavior.

[0006] Automated speed management systems are commonly installed on current vehicles, generally operating on the basis of regulating the distance between the equipped vehicle, also called the ego vehicle, and the one in front of it in its lane of travel, called the target vehicle or simply target.

[0007] It is well known to equip a motor vehicle with an adaptive speed control device, in particular referred to as the ACC system, from the English "Adaptive Cruise Control".

[0008] Such an adaptive speed control system includes a means of detecting the vehicle's environment such as radar and can therefore detect other vehicles or objects on the roadway, in particular it can detect a vehicle ahead on the roadway.

[0009] These systems are designed, for example, to control the vehicle so that its speed is equal to a command given by the driver, except in the presence of an event on the road requiring the vehicle to slow down (following a vehicle which has a speed different from the command given by the driver, traffic jam, traffic light...), in which case the speed of the vehicle is controlled accordingly.

[0010] Next, the speed of a vehicle equipped with such an adaptive cruise control system can be adjusted to maintain a virtually constant safe distance from the vehicle in front. This is achieved by interacting with the engine control system and / or the braking system to accelerate or decelerate the vehicle.

[0011] State-of-the-art adaptive cruise control systems are sometimes unreliable. Late speed corrections can cause abrupt acceleration or deceleration, leading to discomfort and a lack of safety. Such situations can occur, in particular, when the target vehicle changes lanes or when another vehicle merges into the target vehicle's lane.

[0012] An embodiment of a vehicle equipped with a controller implementing a conventional adaptive speed regulation method is described below with reference to the figure 1 .

[0013] The motor vehicle 10 is a motor vehicle of any type, including passenger cars and commercial vehicles. In this document, the vehicle incorporating the means for implementing the invention is referred to as the "ego" vehicle. This designation serves only to distinguish it from other surrounding vehicles and does not in itself impose any technical limitations on the motor vehicle 10.

[0014] The motor vehicle 10 or ego vehicle 10 includes a conventional adaptive speed control DISP device implementing a conventional adaptive speed control process, notably in the form of a controller.

[0015] The classic DISP adaptive speed control system of a motor vehicle can be part of a more comprehensive driver assistance system.

[0016] In the rest of this document, the target vehicle is defined as a vehicle located in the surrounding traffic of the ego 10 vehicle, whose kinematic attributes, including position, speed and acceleration, are taken into account in the calculation of the longitudinal target speed of the ego vehicle.

[0017] A target vehicle can be a motor vehicle of any type, including a passenger vehicle, a commercial vehicle, or a motorcycle.

[0018] The DISP adaptive cruise control system of a motor vehicle requires the perception of vehicles in surrounding traffic, notably through sensors such as perception and localization sensors. Indeed, determining the kinematic attributes Att_Cible—namely, lane presence, position, speed, and relative acceleration of surrounding vehicles—using these sensors is necessary for accurate localization. Data from the perception sensors is processed, in particular, through fusion, to identify and characterize surrounding objects along with their kinematic attributes. These perception sensors can employ various technologies, such as ultrasound, radar, lidar, and cameras, while localization sensors include inertial measurement units, odometers, and GPS satellite geopositioning systems.Object characterization allows us to understand the environment around the vehicle and classify objects by type, such as vehicles (motorcycle, car, truck, bicycle, etc.), pedestrians, infrastructure, and signage. This allows us to focus solely on vehicles as the target. Data from these sensors, particularly when combined with mapping, also allows us to identify road geometry (slopes, curves, etc.). Additionally, navigation information can provide contextual details based on the vehicle's perceived positioning: area type (urban, suburban, rural), road type (highway, city, motorway, etc.), speed limit (which can be combined with sign recognition), traffic information, and geometric information (slope, curve, number of lanes, etc.).The kinematic attributes Att_Cible of the identified target are then transmitted to the distance control CD block which also consumes driver data DC as input, namely the speed regulation selected by the driver and the predetermined follow time, default of 2 seconds, chosen by the driver, which is also translated in terms of predetermined follow setpoint distance chosen by the driver as a function of the speed of the vehicle ego by means of a table for example.

[0019] Based on this information, the distance control CD block generates an output longitudinal speed command Vc allowing the vehicle to automatically maintain the predetermined following distance with the preceding vehicle in the same traffic lane, the vehicle automatically modulating its speed to maintain this distance.This speed command Vc, as well as the vehicle's speed Vm, obtained in particular from measurements by sensors C, notably wheel speed sensors, whose measurements are averaged and associated with a Kalman filter, are then sent as input to the speed control CV block which generates an acceleration command Ac at output. After feedback with the vehicle's acceleration data Am, obtained by sensors C, notably by inertial measurement unit or based on the measured wheel speed, as soon as the road is clear in front of the vehicle, the vehicle's acceleration is automatically increased to reach the speed selected by the driver. The actuators A (engine, brakes...)These systems are controlled in tandem by the torque command Cc generated at the output of the torque control block CC, which is a function of the consolidated acceleration command at the feedback output, thus assisting the driver in their driving task. However, this system only considers one target at a time, the one present in front of the ego in its lane, which makes it very sensitive, particularly to vehicle insertions between the ego and the vehicle in front, as well as to lane changes by the target vehicle, disrupting the smoothness of the driving with sudden irregularities in the commands.

[0020] We are also familiar with document FR2912981 concerning a method for automatically piloting a motor vehicle, including an ACC system, which aims to improve the smoothness of vehicle behavior to increase user comfort. However, this method becomes very resource-intensive when dealing with multiple targets because it requires numerous calculations to determine the dynamic components to be monitored by the ACC.

[0021] An ACC system that takes into account multiple vehicles in front of or to the sides of the ACC vehicle is described in EP2658763A1.

[0022] One of the aims of the invention is to remedy at least some of the drawbacks of the prior art by providing a method for assisting the driving of an ego vehicle moving on a traffic lane, comprising: a first step of identifying traffic surrounding the ego vehicle on the ego vehicle's traffic lane and on at least one adjacent parallel lane in the same direction of travel, a second step of determining a virtual barycentric target, with calculation of a position of the virtual barycentric target, a velocity of the virtual barycentric target, and an acceleration of the virtual barycentric target; a third step of calculating a longitudinal speed setpoint of the ego vehicle, an acceleration setpoint and a torque setpoint, said longitudinal speed setpoint being a function of the position of the virtual barycentric target, the velocity of the virtual barycentric target, and the acceleration of the virtual barycentric target.Thanks to the invention, it is possible to anticipate the trajectory of vehicles located in surrounding traffic without modifying the ACC control loop itself, but only its inputs, and to take into account multiple targets, including in the absence of a target in the lane of the ego vehicle, while improving the fluidity of the ACC control loop.

[0023] According to an advantageous feature, the surrounding traffic includes at least two target vehicles preceding the ego vehicle traveling in its lane of travel or in the adjacent parallel lane in the same direction of travel, which allows not only lateral vehicles to be taken into account but also a vehicle that would slow down in front of the vehicle preceding the ego.

[0024] According to another advantageous feature, the first identification step includes a sub-step of detection of each of the at least two target vehicles with for each target vehicle determination at output of a position of the target vehicle relative to the ego vehicle, a speed of the target vehicle, an acceleration of the target vehicle, and in particular determination of a trajectory of the target vehicle, which allows, whether the position is relative or absolute, to anticipate safety distances by taking into account the predicted trajectories.

[0025] According to another advantageous feature, the second step of determining the virtual barycentric target consumes as inputs a preselected control speed, a predetermined tracking setpoint distance, and a result from the identification step, which makes it possible to build the virtual barycentric target on the basis of readily available information, and taking into account control speed and tracking setpoint distance customizable by the driver.

[0026] Advantageously, the second step in determining the virtual barycentric target includes a filtering step so as to select only certain targets based on their speed or spacing in terms of time.

[0027] According to another advantageous feature, at least one of the target vehicles is located in the traffic lane of the ego vehicle, it is possible to take into account the phenomena of slowdowns by the preceding vehicles, but also of one of these vehicles leaving the lane preceding the ego.

[0028] The advantage linked to the characteristic that less than one of the target vehicles is located in an adjacent lane and seeks to merge into the lane of the ego vehicle, is to allow the consideration of anticipatory merge phenomena.

[0029] According to another advantageous feature, the second step of determining the virtual barycentric target uses a target change prediction coefficient determined for each target, which allows the barycenter to be weighted dynamically and thus gain fluidity in ACC regulation.

[0030] According to another advantageous feature, the target change coefficient is a function of a relative lateral distance between a trajectory of the ego vehicle and at least one target vehicle, or a function of a relative lateral distance between a center of the lane on which the ego vehicle travels and at least one target vehicle, which makes it possible to overcome the curvature of the road.

[0031] According to another advantageous feature, the target change coefficient is a function of an estimated intersection time of an estimated trajectory of the target vehicle with an estimated trajectory of the ego vehicle, which allows weighting using other information given by the surrounding traffic identification module.

[0032] Advantageously, the target change coefficient is a function of the tracking time of a pre-target vehicle, which allows for a smooth consideration of a vehicle traveling in front of the vehicle preceding the ego that would slow down.

[0033] According to another advantageous feature, the target change coefficient is a function of a stiffness coefficient, thus ensuring ease of development because this stiffness coefficient is unique for all targets in insertion.

[0034] The invention also relates to a system for an ego motor vehicle moving on a traffic lane, comprising: a module for identifying traffic surrounding the ego vehicle on the ego vehicle's lane of travel and on at least one adjacent parallel lane in the same direction of travel; a module for determining a virtual barycentric target, with calculation of a position of the virtual barycentric target, a velocity of the virtual barycentric target, and an acceleration of the virtual barycentric target; a module for calculating a longitudinal speed setpoint for the ego vehicle, an acceleration setpoint, and a torque setpoint, said longitudinal speed setpoint being a function of the position of the virtual barycentric target, the velocity of the virtual barycentric target, and the acceleration of the virtual barycentric target, which presents advantages similar to those of the method using means commonly available on vehicles such as radar, camera.

[0035] The invention also relates to an ego motor vehicle comprising a powertrain, acceleration and braking means and including a driving assistance system according to the invention, which allows simple adaptation on vehicles equipped with ACC devices whether or not they are autonomous.

[0036] Other objects, features and advantages of the invention will become apparent from the following description, given solely by way of non-limiting example, and made with reference to the accompanying drawings in which: [ Fig.1 ] there figure 1 which has already been mentioned, schematically illustrates a classic ACC device according to the state of the art, and, [ Fig.2 ] there figure 2 schematically represents a driver assistance system according to the invention, and, [ Fig. 3 ] there figure 3 illustrates the application of the process according to the invention, and, [ Fig. 4 ] there figure 4 represents an illustration of relative lateral distance in a curve in a use case of the invention, and, [ Fig. 5 ] there figure 5 represents an illustration of trajectory prediction in a use case according to the invention, and, [ Fig. 6 ] there figure 6 represents another use case of the invention.

[0037] Throughout this text, directions and orientations are designated with reference to a direct orthonormal XYZ coordinate system commonly used in automotive design, where X denotes the longitudinal direction of the vehicle, oriented in the direction of travel, Y is the direction transverse to the vehicle, oriented to the left, and Z is the vertical direction oriented upwards. The terms "front" and "rear" refer to the normal forward direction of travel of the vehicle. Throughout this description, the term "substantially" means that a slight deviation from a specified nominal quantity is permissible; for example, "substantially constant" means that a deviation of approximately 5% is permissible within the scope of the invention. For clarity, identical or similar elements are identified by identical reference numerals in all figures.

[0038] We have schematically represented on the figure 2 an embodiment of an adaptive speed control system 1 for an EGO motor vehicle, forming part of a more comprehensive driver assistance system according to one aspect of the invention. The elements of system 1 are identical to the elements constituting the DISP device of the figure 1 they carry the same references.

[0039] In this driver assistance system according to the invention, sensors C, such as perception sensors, measure the dynamics of the vehicle and perceive the environment. These sensors C not only provide information on the vehicle's speed and acceleration, as well as the position, speed, and acceleration of objects in the environment, but also predict the trajectory of these objects. Indeed, the determination of environmental information (Env), which includes the kinematic attributes such as lane presence, position, speed, and relative acceleration of surrounding vehicles, by the sensors C, is required to correctly predict their trajectories. This environmental information is, for example, provided in a reference frame of the vehicle positioned at the rear axle, but any other positioning of the reference frame is possible.Similarly, environmental information regarding the position of other vehicles is often based on rear axle detection. All this environmental information (Env), derived from the fusion of data from sensors C, is indexed object by object and generated as output by what we will call an identification module (not shown) for traffic surrounding the vehicle ego in the vehicle ego's lane and in adjacent parallel lanes in the same direction of travel. This information is then sent as input to the CBV module for determining a virtual center-of-center target. The identification module fuses the aforementioned information from sensors C with information from light signals, such as the turn signals of surrounding vehicles.A change of positioning reference frame can then be made upstream or in the CBV module to determine distances not axle to axle but bumper to bumper as illustrated on the . figures 3 and following.

[0040] The CBV module also takes as input the driver's DC data, namely the speed selected by the driver and the predetermined following time chosen by the driver, which is also translated into a predetermined following distance chosen by the driver. Based on these inputs, the CBV module determines a virtual center-of-center target in the ego vehicle's lane, and calculates its position, speed, and acceleration. This CBV module can take into account information from various objects, select objects located ahead of the ego vehicle, whether in its lane or to the side, as targets, and deduce target vehicles for which predicted trajectories have been predetermined. This allows the module to anticipate the movements of target vehicles and react as a human driver would.To filter out small oscillations due to measurement errors, a filter can be added to the CBV module so that it only considers targets for which the relative lateral velocity with respect to the ego exceeds a threshold. This CBV module is placed upstream of the control loops of the actuator subsystems A (engine, brakes, etc.) and thus provides the control commands like a conventional ACC target without requiring modification of the ACC logic, thereby simplifying its integration.Indeed, these kinematic attributes Att_CBV of the virtual barycentric target, namely its presence in the lane, position, velocity, and relative acceleration in dynamics, generated at the output of the CBV module for determining a virtual barycentric target, are sent as input to the BCD loop for calculating dynamic commands. These commands include the longitudinal velocity command Vc of the vehicle ego and the acceleration command Ac, said longitudinal velocity command being a function of the virtual target's position, velocity, and acceleration. The BCD loop for calculating dynamic commands includes: The distance control CD block, which takes as input the kinematic attributes Att_CBV of the virtual barycentric target as well as the driver data DC, namely the regulation speed selected by the driver and the predetermined following time chosen by the driver, which is also translated in terms of the predetermined following setpoint distance chosen by the driver, and which generates as output the longitudinal speed setpoint Vc, which corresponds to the desired speed regulation quantity allowing to meet the driver's expectations in terms of safe distance in complete fluidity; the speed control CV block, which, as before, takes as input the speed setpoint Vc, as well as the vehicle speed ego Vm, and which generates as output the acceleration setpoint Ac; the torque control CC block, which, as before, takes as input the consolidated acceleration command at the feedback output.and which generates at output the torque setpoint Cc controlling the actuators A, this torque setpoint Cc allowing the wheels to be driven in such a way as to follow the speed setpoint Vc.

[0041] The CBV module for determining the single virtual center-of-center target takes into account surrounding lateral vehicles that could, in the near future, merge and become the target for the ACC to follow. This CBV module also aims to ensure normal ACC regulation when no vehicle is detected in front or to the side, or when only one target vehicle is detected in the ego vehicle's lane. Advantageously, this process can also be applied to a vehicle V0 that leaves the ego vehicle's lane to change lanes, a phenomenon known as a cut-out. Once this vehicle is merged into the ego vehicle's lateral lane, it would then become a vehicle Vx or would no longer be indexed if it leaves the ego vehicle's approach range, delimited by the sensor perception distances and / or a predetermined distance from the ego vehicle, notably based on the driver's customizable follow distance setting.

[0042] In a straight line, as illustrated in figure 3 The ego EGO vehicle perceives two target vehicles, as determined by the surrounding traffic identification module: a vehicle V0 in front of it in its lane, and a vehicle V1 slightly ahead, located laterally to its left in the left lane, with its right turn signal illuminated to indicate it will move into the ego vehicle's lane. The three lanes in the same direction of travel are represented by short dashed lines, and a solid line represents their separation from a potential oncoming lane. Although in this example the target vehicle V1 is to the left of the ego EGO in the direction of travel, alternatively, the target vehicle V1 could be to the right of the ego EGO and attempt to merge into the ego vehicle's lane, for example, by activating its left turn signal.The virtual barycentric target G therefore corresponds here to the barycenter G of a system ( A, a ) ( B , b ) with a + b ≠ 0, a and b being weighting coefficients, and for any point O taken as origin , we have: . OG → = 1 a + b a OA → + b OB →

[0043] The surrounding traffic identification module provides, in the frame linked to the ego vehicle, here centered on its rear axle but for which another choice could be made, to the CBV module for determining the virtual barycentric target G the geometric references corresponding to points A and B which are the positions of vehicles V0 and V1, and the CBV module performs a change of frame so as to link the new frame of the ego vehicle to the front of the bumper of the ego vehicle:

[0044] X 0 , which is the distance of the first target vehicle, specifically its rear bumper, in the direct approach field of the ego vehicle, in the relative frame of reference linked to the ego vehicle.

[0045] X 1 , is the distance of a second target vehicle, in particular its front bumper, which risks entering the approach field of the ego vehicle, in the relative frame linked to the ego vehicle.

[0046] Ẋ 0 and Ẋ 1 are the respective relative speeds of the first vehicle and the second vehicle.

[0047] Similarly Ẍ 0 and Ẍ 1 are the respective relative accelerations of the first vehicle and the second vehicle.

[0048] The virtual barycentric target approach aims to anticipate target changes by adding a dynamic offset o_d to the tracking setpoint distance d_s_c provided by the driver and applied to the target vehicle V0 in the lane. The sum of the dynamic offset distances o_d and the tracking setpoint distance d_s_c constitutes the inter-vehicle distance to be maintained between the front of the ego vehicle EGO and the rear of the target vehicle V0 sharing the same lane. The offset o_d is obtained by projecting the virtual barycentric target G onto the ego's trajectory. The offset o_d, along with the position information of the virtual barycentric target G and other kinematic attributes Att_CBV, is then transmitted to the CD module. The weighting coefficients a and b of the barycenter are chosen as corresponding target change prediction coefficients. These coefficients can be obtained in two distinct ways.

[0049] The first method is based on the lateral position of the merging vehicle V1. Thus, the closer the lateral target vehicle V1 is to the lane of the target vehicle EGO, the more likely it is considered that it will merge into that lane. Preferably, rather than the vehicle's ordinate Y1, it is better to use the lateral distance Y1' relative to the trajectory and / or the lane center of the target vehicle EGO, which are here combined and represented by long dashed lines. This eliminates the interference caused by the orientation of the target vehicle EGO within its lane, which is a tangent multiplied by the distance to the target (5° at 100 meters = 8 meters of lateral error). The method can therefore be used on curves.

[0050] There figure 4 The system visually indicates the error made on a curve. The values ​​of Y1 and Y2 correspond to the ordinates of the endpoints, for example, the center of the bumpers, the rear ends of the lateral target vehicles V1 and V2 in the relative coordinate system linked to the ego vehicle. The values ​​Y1' and Y2' correspond to the lateral distances from the lane center of the ego vehicle to the lateral target vehicles V1 and V2. Regarding the lateral distances, Y1' > Y2', while with the ordinates, Y1 < Y2. The lateral distance values ​​are provided by the surrounding traffic identification module, as is the position of the targets. We therefore obtain weighting coefficients a and b as functions of the lateral distance, specifically a as a function of Y1' and b as a function of the inverse of Y1'.

[0051] The second method, illustrated in figure 5 , consists of using the estimated intersection time with the trajectory of the ego EGO vehicle as the target change prediction coefficient, corresponding here to 3 seconds T+3. The estimated intersection time with the trajectory of the ego EGO is based on the trajectory prediction, both being provided by the surrounding traffic identification module as well as the position of the targets.

[0052] For the remainder we will use prediction coefficients based on the lateral position of the vehicle V1 inserting itself, but the application would be the same with the second method.

[0053] The weighting coefficients a and b preferentially involve a stiffness coefficient k, or tuning coefficient, notably in the following way: a = Y1'; b = k / Y1'

[0054] The stiffness coefficient k is thus placed at the level of the quantity X1, which represents the vehicle V1 that is inserted. Behaviorally, this coefficient defines the strength of the target's consideration in the calculation of the virtual target; the larger k is, the more the ego vehicle will anticipate. Since the problem is symmetrical along the x-axis, the coefficient k for a target on the right and the coefficient k for a target on the left will be identical, allowing the method to be applied in both right-hand and left-hand drive countries without specific modifications. The choice to use inversely proportional weighting coefficients a, b (a=1 / b) maintains the proportions modulo the stiffness coefficient, which is essential for the balance of the center of gravity, its representativeness, proportionality, and facilitates the management of the boundaries.We then obtain the equations for the three quantities positioning, velocity, acceleration of the barycenter G of the target vehicles V0, V1: . X G = Y 1 ′ . X 0 + k Y 1 ′ X 1 Y 1 ′ + k Y 1 ′ X ˙ G = Y 1 ′ . X ˙ 0 + k Y 1 ′ X ˙ 1 Y 1 ′ + k Y 1 ′ et X ¨ G = Y 1 ′ . X ¨ 0 + k Y 1 ′ X ¨ 1 Y 1 ′ + k Y 1 ′

[0055] These quantities are calculated at each time step, so the centroid G dynamically takes into account the variations in target distances, speeds, and accelerations throughout the entire maneuver. The values ​​of Y1' and k / Y1' are not considered in the derivation here because they are dimensionless coefficients, and if Y1' corresponds to a lateral distance, it could also be an intersection time, as previously stated. It is also possible to calculate XG and then differentiate it to obtain the other speed and acceleration quantities. Furthermore, the fact that the weighting coefficients a and b are a function of the lateral positioning of the lateral target V1 or the intersection time allows the dynamics of the insertion to be taken into account. The more the target vehicle V1 merges into the lane of the ego vehicle EGO, the more the parameters of the merging vehicle V1 are considered.Advantageously, only a single stiffness coefficient k is needed for the three equations, which greatly simplifies the development process. This stiffness coefficient k can be adapted according to the vehicle's speed ego to achieve different behavior depending on the situation (free-flowing highways or congested urban ring roads); its value can notably be in the range ]0;10] and is preferably 1.

[0056] Advantageously, the method, by creating a virtual barycentric G-target with a distance, velocity, and acceleration, is not intrusive on the conventional control loops of the ACC. Therefore, the method of the invention does not modify the powertrain control settings, whether thermal or electric, nor does it modify the braking system settings. Furthermore, advantageously, the method has a spatial representation that allows for easy visualization during its development, as already illustrated in figure 3 which graphically shows the calculation carried out to realize the barycenter G of the positions of the target vehicles V0, V1. Physically, the application of the process amounts to adding an offset o_d to the tracking setpoint distance d_s_c.

[0057] Once vehicle V1 has entered the lane of the ego EGO vehicle between the ego EGO vehicle and vehicle V0, it then becomes indexed as vehicle V0.

[0058] Advantageously, as already mentioned, the method can take into account a target located to the right as well as to the left, and in particular allows for the consideration of multiple targets, whether they are located to the right and / or left in front of the ego EGO vehicle. Similar to the case of inserting a target vehicle V1 between the ego EGO vehicle and the preceding vehicle V0, the method thus applies to three or more targets, as illustrated. figure 6 The virtual barycentric target G then corresponds to the barycenter G of a system (A, a), (B, b), (C, c) with a + b + c ≠ 0, a, b and c being weighting coefficients, and for any point O taken as the origin, we have: OG → = 1 a + b + c a OA → + b OB → + c OC →

[0059] Which gives: X G = Y 1 ′ . Y 2 ′ . X 0 + k Y 1 ′ X 1 + k Y 2 ′ X 2 Y 1 ′ . Y 2 ′ + k Y 1 ′ + k Y 2 ′ − k . NbTm

[0060] With NbTm The number of missing lateral targets to be able to apply the process, including in the case of missing targets, while maintaining homogeneity of dynamic behavior. Furthermore, it is necessary to define default values ​​when one or more targets are missing. Thus, in the event of no vehicle in the EGO X lane, 0 is taken as equal to the d_s_c tracking distance setpoint. Ẋ 0 is taken to be equal to the cruise control speed selected by the driver. Ẍ 0 is taken as equal to 0, similarly in the absence of a lateral Vi target X i , Ẋ i , Ẍ i , are taken to be equal to 0 and Y i' taken equal to 1. Therefore, when the target vehicles V1 and V2 are missing, we have X G = X 0 .

[0061] The challenge for the ACC target is maintaining the correct following time relative to the preceding target. Therefore, it is also in the interest of the ego vehicle to maintain an additional safety distance if the following time between the two vehicles ahead in its lane is dangerous. The method thus allows monitoring not only of lateral target vehicle merges but also the behavior of the vehicle preceding the vehicle the ego vehicle is following. Indeed, the method allows application not only for multiple targets (V1, V2) attempting to merge in front of the ego vehicle (EGO) but also takes into account the vehicle (VP) preceding the vehicle (V0) followed by the ego vehicle (EGO). As represented in figure 6 In its lane, the ego vehicle EGO follows vehicle V0, which itself follows vehicle VP, which we will call the target vehicle. In the side lanes are vehicle V1, located to the front left of the ego vehicle, and vehicle V2, located to the front right of the ego vehicle. If we initially focus only on the ego vehicle EGO, the preceding vehicle V0, and the target vehicle VP, we can write XP as the distance between the ego vehicle EGO and the target vehicle VP. Ẋ P, which is its time derivative, and kP, which is the stiffness coefficient in the case of a pre-target, with the weighting coefficients a=XP / Ẋ P and b=k P. Ẋ P / XP, the single stiffness coefficient kP being placed at the level of the quantity XP: X G = X P X ˙ P . X 0 + k P . X ˙ P X P X P X P X ˙ P + k P . X ˙ P X P = X P X ˙ P . X 0 + k P . X ˙ P X P X ˙ P + k P . X ˙ P X P

[0062] Here, the weighting coefficients a and b depend on the tracking time of the target VP, physically representing the intersection time of the trajectories in longitudinal coordinates. Therefore, both targets V0 and VP are considered at all times. To prevent the ego vehicle from experiencing variations with each change in speed of the target VP, filtering based on a threshold speed is preferentially performed. This ensures that the target VP is only considered when its relative speed with respect to V0 is below a certain threshold. Thus, the target VP will only be considered if the difference between the target speed and the ego vehicle's speed is less than a predetermined threshold, this threshold being, for example, a few km / h, in order to filter out small oscillations and only consider targets that are approaching each other.This threshold, which will be a fine-tuning parameter, can optionally take into account the relative distance of the target to assimilate it to a tracking time threshold. In other words, it performs a filtering process by considering the target VP based on its speed, according to the tracking time settings entered by the driver, so that the target VP is only considered if the time between VP and EGO is less than a predetermined threshold. The value of the stiffness coefficient kP preferentially belongs to the range ]0; 10], knowing that the higher it is, the earlier the braking occurs. Its setting can therefore depend on the driving mode chosen by the driver, with, for example, a higher value in sport mode than in city mode.And in order to take into account the absence of the target vehicle VP, an indicator F Pabs is used as before, indicating the absence of the target vehicle when it is 1 or the presence of the target vehicle when it is 0, and default values ​​are taken so that when the target vehicle is missing: . X G = X 0: X G = X P X ˙ P . X 0 + k P . X ˙ P − k P . F Pabs X P X ˙ P + k P . X ˙ P X P − k P . F Pabs X ˙ G = X P X ˙ P . X ˙ 0 + k P . X ˙ P X P X ˙ P − k P . F Pabs X P X ˙ P + k P . X ˙ P X P − k P . F Pabs X ¨ G = X P X ˙ P . X ¨ 0 + k P . X ˙ P X P X ¨ P − k P . F Pabs X P X ˙ P + k P . X ˙ P X P − k P . F Pabs

[0063] Preferably, XG is calculated and then derived to obtain the other velocity and acceleration quantities. The process can also take into account four (or more) targets as shown in the figure 6 : V1, V2, V0 and VP with default values ​​X 0 is taken equal to the tracking setpoint distance d_s_c , Ẋ 0 is taken to be equal to the cruise control speed selected by the driver. Ẍ 0 is taken as equal to 0, similarly in the absence of a lateral Vi target X i , Ẋ i , Ẍ l , are taken to be equal to 0 with NbTm the number of missing lateral targets and in case of absence of a target F pabs = 1, so that when the lateral target vehicles are missing we obtain X G = X 0: X G = Y ′ 1 . Y ′ 2 ⋅ X P X ˙ P . X 0 + k Y 1 ′ X 1 + k Y 2 ′ X 2 + k P . X ˙ P − k P . F Pabs Y 1 ′ . Y 2 ′ ⋅ X P X ˙ P + k Y 1 ′ + k Y 2 ′ k P . X ˙ P X P − k . NbTm + k P . F Pabs X ˙ G = Y 1 ′ . Y 2 ′ . X P X ˙ P . X ˙ 0 + k Y 1 ′ X ˙ 1 + k Y 2 ′ X ˙ 2 + k P . X ˙ P X P X ˙ P − k P . F Pabs Y 1 ′ . Y 2 ′ . X P X ˙ P + k Y 1 ′ + k Y 2 ′ + k P . X ˙ P X P − k . NbTm + k P . F Pabs X ¨ G = Y 1 ′ . Y 2 ′ . X P X ˙ P . X ¨ 0 + k Y 1 ′ X ¨ 1 + k Y 2 ′ X ¨ 2 + k P . X ˙ P X P X ¨ P − k P . F Pabs Y 1 ′ . Y 2 ′ ⋅ X P X ˙ P + k Y 1 ′ + k Y 2 ′ + k P . X ˙ P X P − k . NbTm + k P . F Pabs

[0064] Preferably, XG is calculated and then derived to obtain the other velocity and acceleration quantities. More targets, particularly lateral ones, could be taken into account if the perception sensors allow it.

[0065] Thanks to this process, the ACC system will behave more smoothly when changing targets, and by anticipating the movements of target vehicles, safety is increased. The behavior of the EGO vehicle will thus more closely resemble that of a human driver and will be fluid in traffic.

Claims

1. Driver assistance method for an ego vehicle (EGO) travelling in a traffic lane, comprising: - a first step of identifying traffic surrounding the ego vehicle in the traffic lane of the ego vehicle and in at least one adjacent parallel lane in the same traffic direction, characterized in that it further comprises: - a second step of determining a virtual barycentric target (G), with calculation of a position of the virtual barycentric target, a speed of the virtual barycentric target and an acceleration of the virtual barycentric target; - a third step of calculating a longitudinal speed setpoint (Vc) of the ego vehicle, an acceleration setpoint (Ac) and a torque setpoint (Cc), said longitudinal speed setpoint (Vc) being a function of the position of the virtual barycentric target (G), the speed of the virtual barycentric target (G) and the acceleration of the virtual barycentric target (G).

2. Driver assistance method according to the preceding claim for an ego vehicle (EGO), characterized in that the surrounding traffic includes at least two target vehicles (V0, V1, V2, VP) preceding the ego vehicle (EGO) travelling in its traffic lane or in the adjacent parallel lane in the same traffic direction.

3. Driver assistance method according to the preceding claim for an ego vehicle (1), characterized in that the first, identification step includes a sub-step of detection of each of the at least two target vehicles (V0, V1, V2, VP) with for each target vehicle determination as output of a position of the target vehicle relative to the ego vehicle (EGO), a speed of the target vehicle and an acceleration of the target vehicle, and in particular determination of a trajectory of the target vehicle.

4. Driver assistance method according to any one of the preceding claims for an ego vehicle (EGO), characterized in that the second step of determination of the virtual barycentric target consumes as input a preselected control speed, a predetermined following distance setpoint (d_s_c) and a result of the identification step.

5. Driver assistance method according to any one of the preceding claims for an ego vehicle (EGO), characterized in that the second step of determination of the virtual barycentric target includes a filtering step.

6. Driver assistance method according to any one of Claims 2 to 5 for an ego vehicle (EGO), characterized in that at least one of the target vehicles (V1, V2) is situated in an adjacent lane and seeks to merge into the traffic lane of the ego vehicle (EGO).

7. Driver assistance method according to any one of the preceding claims for an ego vehicle (EGO), characterized in that the second step of determination of the virtual barycentric target uses a change of target prediction coefficient determined for each target.

8. Driver assistance method according to the preceding claim for an ego vehicle (EGO), characterized in that the change of target coefficient is a function of a relative lateral distance (Y1', Y2') between a trajectory of the ego vehicle (EGO) and at least one target vehicle, or a function of a relative lateral distance between a centre of the lane in which the ego vehicle (EGO) is travelling and at least one target vehicle (V1, V2).

9. Driver assistance method according to Claim 7 for an ego vehicle (EGO), characterized in that the change of target coefficient is a function of an estimated time of intersection of an estimated trajectory of the target vehicle with an estimated trajectory of the ego vehicle (EGO).

10. Driver assistance method according to Claim 7 for an ego vehicle (EGO), characterized in that the change of target coefficient is a function of the following time of a pre-target vehicle (VP).

11. Driver assistance method according to any one of Claims 7 to 10 for an ego vehicle (1), characterized in that the change of target coefficient is a function of a coefficient (k, kP) of stiffness.

12. Driver assistance system (1) for an ego automobile vehicle (EGO) moving in a traffic lane, comprising: - a module for identification of traffic surrounding the ego vehicle in the traffic lane of the ego vehicle and in at least one adjacent parallel lane in the same traffic direction, characterized in that it further comprises: - a module (CBV) for determination of a virtual barycentric target, with calculation of a position of the virtual barycentric target, a speed of the virtual barycentric target and an acceleration of the virtual barycentric target; - a module (CD, CV, CC) for calculation of a longitudinal speed setpoint (Vc) of the ego vehicle (EGO), an acceleration setpoint (Ac) and a torque setpoint (Cc), said longitudinal speed setpoint (Vc) being a function of the position of the virtual barycentric target (G), the speed of the virtual barycentric target (G) and the acceleration of the virtual barycentric target (G).

13. Ego automobile vehicle (EGO) including a powertrain and acceleration and braking means, characterized in that it includes a driver assistance system (1) according to the preceding claim.