System and method for predicting the trajectory of a vehicle

Description

[0001] The present invention relates to the field of devices and methods for predicting a path and / or trajectory of a motor vehicle, and computer programs intended for implementing such methods.

[0002] In particular, the present invention relates to the field of driving aids for a motor vehicle, and in particular assistance for the driver to activate or not activate the driving aid systems.

[0003] Motor vehicles are now equipped with advanced driver assistance systems, known as "advanced driver assistance systems", or ADAS in English terms, which are becoming increasingly efficient. The purpose of an advanced driver assistance system is to enable autonomous driving of a motor vehicle, i.e. without intervention from the driver, or in sharing with the driver of the vehicle in order to keep the vehicle in its lane and / or slow down its speed. In particular, an advanced driver assistance system can be used to anticipate a path and / or trajectory of a motor vehicle. In the present application, the path of a vehicle will be considered as a geometric shape corresponding to the progression of the vehicle between a starting point and an arrival point. The trajectory of a vehicle will be considered as the temporal evolution of the position of the vehicle between the starting point and the arrival point.

[0004] So-called "autonomous" or partially assisted driving vehicles require a rich model of the vehicle's environment to enable an algorithm to make decisions. This is made possible by various proprioceptive sensors, such as an accelerometer, a gyrometer, etc., and exteroceptive sensors, such as cameras, radars, lidars, ultrasounds, etc., as well as data fusion processes configured to process the received information and calculate the state (position, speed, acceleration, yaw, etc.) of the vehicle and surrounding objects.

[0005] In such vehicles, it is thus essential to predict the movements of the ego vehicle as well as other mobile objects present in the close environment of said vehicle and likely to become obstacles when the trajectory of the ego vehicle interferes with the trajectory of an object.

[0006] We know, for example, of lane keeping assist systems, known as "lane keeping assist", with the acronym "LKA" in Anglo-Saxon terms, which allow the vehicle to automatically reposition itself in its lane, or of systems known as "lane change assist", with the acronym "LCA" in Anglo-Saxon terms, which allow the vehicle to change lanes.

[0007] Other examples of driving assistance are known, such as the so-called "automatic emergency steering" assistance systems, acronym "AES" in Anglo-Saxon terms, capable of detecting obstacles and carrying out emergency steering, the prediction of the trajectories of moving objects makes it possible to analyze the potential risk of collision, or the so-called "adaptive cruise control" assistance systems, acronym "ACC" in Anglo-Saxon terms, capable of regulating speed and automatically maintaining a safe distance from the vehicle in front.

[0008] Document EP 3 056 405 - A1 discloses a method for controlling the lane change of an ego vehicle as a function of a first distance between two vehicles adjacent to the ego vehicle as a target for the lane change of said ego vehicle.

[0009] Document WO2019 / 204053A1 describes a method with similar drawbacks.

[0010] The main drawback of these driver assistance systems lies in the lack of smoothness in their response. Indeed, it is particularly difficult to predict the behavior of vehicles in the lanes adjacent to the ego vehicle, so the decision-making time of these assistance systems is particularly long.

[0011] There is therefore a need to optimize the lane change of an ego vehicle from a main lane to an adjacent lane.

[0012] In view of the above, the invention aims to enable the prediction of a motor vehicle trajectory while overcoming the aforementioned drawbacks.

[0013] The present invention relates to a method for predicting the trajectory of an ego vehicle traveling on a main lane in which a lane change of the ego vehicle from the main lane to an adjacent lane is determined based on an estimation of the dynamic behavior of a group of vehicles traveling on the adjacent lane. Said group of vehicles comprises at least one main vehicle located near the ego vehicle and a secondary vehicle located behind said ego vehicle.

[0014] Thus, the traffic evolution corresponding to the dynamic behavior of vehicles on the lane adjacent to the lane of the ego vehicle is predicted in order to make a lane change decision.

[0015] The prediction of the trajectory of the ego vehicle is determined based on the behavior of a group of vehicles present on the lane adjacent to the traffic lane of the ego vehicle.

[0016] According to the invention, position, orientation and speed information of the ego vehicle and the vehicles of the vehicle group are collected and a dynamic model is established by pairs of consecutive vehicles traveling on the adjacent lane according to the collected information. This makes it possible to obtain an accurate dynamic model for each of the adjacent vehicles according to the data of the nearby vehicles on the same traffic lane.

[0017] The dynamic model for pairs of consecutive vehicles is obtained by determining a second-order transfer function corresponding to the behavior of the ego vehicle with respect to the considered pair of consecutive vehicles using an autoregressive exogenous model, "autoregressive exogenous model", acronym "ARX" in English terms, to obtain the dynamic data of a pair of consecutive vehicles. The behavior of the ego vehicle depends on its longitudinal model and its longitudinal controller

[0018] The position, orientation and speed of vehicles are obtained in particular by the various proprioceptive and exteroceptive sensors of an ego vehicle perception system.

[0019] For example, the established dynamic models are validated by comparing an error of the exogenous auto autoregressive calculation model with a threshold value depending on the actual speed of each vehicle at a given time and the speed of said vehicle at a previous time.

[0020] According to a subsequent step, the movement of adjacent vehicles is predicted based on the validated dynamic model and the initial position of said vehicles, the movement of the ego vehicle is predicted based on the prediction of the movement of adjacent vehicles and information from proprioceptive sensors of the ego vehicle and the lane change of the ego vehicle is determined based on said predictions of the movement of the vehicle and adjacent vehicles, an overall trajectory of the ego vehicle and, for example, information from a map of the road on which the ego vehicle is traveling.

[0021] The ego vehicle lane change determination step allows for the assessment of an adequate time window for the ego vehicle to safely change lanes.

[0022] The lane change instruction from the ego vehicle is then transmitted to a lane change execution module of the ego vehicle.

[0023] In the event that there is no possibility for the ego vehicle to change lanes, the ego vehicle is informed of the need to modify its parameters, such as its speed.

[0024] Advantageously, the steps of the method are repeated until a lane change possibility is found.

[0025] According to a second aspect, the invention relates to a system for predicting the trajectory of an ego vehicle traveling on a main lane configured to determine a lane change of the ego vehicle from the main lane to an adjacent lane based on an estimation of the dynamic behavior of a group of vehicles traveling on the adjacent lane, said group of vehicles comprising at least one main vehicle located near the ego vehicle and a secondary vehicle located behind said ego vehicle.

[0026] According to the invention, the system comprises: a module for collecting or retrieving information on the position, orientation and speed of the ego vehicle and the vehicles in the vehicle group; a module for estimating a dynamic model for each pair of consecutive vehicles traveling on the adjacent lane based on the information collected. This makes it possible to obtain an accurate dynamic model for each of the adjacent vehicles based on the data from nearby vehicles on the same lane.

[0027] The dynamic model for pairs of consecutive vehicles is, for example, obtained by determining a second-order transfer function corresponding to the behavior of the ego vehicle with respect to the considered pair of consecutive vehicles using an autoregressive exogenous model, "autoregressive exogenous model", acronym "ARX" in English terms, to obtain the dynamic data of a pair of consecutive vehicles. The behavior of the ego vehicle depends on its longitudinal model and its longitudinal controller.

[0028] The position, orientation and speed of vehicles are obtained in particular by the various proprioceptive and exteroceptive sensors of an ego vehicle perception system.

[0029] Advantageously, the system includes: a module for validating the dynamic models established by comparing an error in the exogenous auto-autoregressive calculation model with a threshold value depending on the actual speed of each vehicle at a given time and the speed of said ego vehicle at a previous time; a module for predicting the movement of adjacent vehicles based on the validated dynamic model and the initial position of said vehicles; a module for predicting the movement of the ego vehicle based on the prediction of the movement of adjacent vehicles and information from proprioceptive sensors of the ego vehicle; and a module for determining the lane change of the ego vehicle based on said predictions of the movement of the ego vehicle and adjacent vehicles, an overall trajectory of the ego vehicle and, for example, information from a map of the road on which the ego vehicle is traveling.

[0030] According to another aspect, the invention relates to an ego motor vehicle comprising a perception system and a system for predicting the trajectory of the ego vehicle as described above.

[0031] Other aims, characteristics and advantages of the invention will appear on reading the following description, given solely by way of non-limiting example, and made with reference to the appended drawings in which: [ Fig 1 ] is a schematic view of two adjacent lanes on which an ego motor vehicle and a plurality of adjacent vehicles are traveling, the ego vehicle comprising a trajectory prediction system according to one embodiment of the invention; [ Fig 2 ] schematically represents the system for predicting the trajectory of an ego vehicle according to an embodiment of the figure 1 ; And [ Fig 3 ] illustrates a flowchart of a method for predicting the trajectory of an ego vehicle according to an embodiment of the invention implemented by the system of the figure 1 .

[0032] On the figure 1 , two adjacent traffic lanes 1, 2 are shown very schematically, on which motor vehicles travel in the same direction.

[0033] As illustrated, a motor vehicle ego 10 is traveling on the first traffic lane 1 and four vehicles 3, 4, 5, 6 are traveling on the lane 2 adjacent to the first lane.

[0034] Vehicles 3, 4, 5, 6 form a group of vehicles 7 traveling on the adjacent lane 2 to the traffic lane 1 of vehicle ego 10.

[0035] The motor vehicle ego 10 comprises a system 11 for perceiving the environment of said vehicle configured to detect the group 7 of vehicles traveling in the adjacent lane 2.

[0036] Generally, the group of vehicles 7 traveling on the adjacent lane 2 comprises at least one main vehicle located in the immediate vicinity of the ego vehicle 10 and at least one secondary vehicle located behind the ego vehicle 10.

[0037] The perception system 11 makes it possible to detect the main vehicle 6 located on the adjacent lane 2, in the immediate vicinity of the ego vehicle 10.

[0038] The perception system 11 comprises various proprioceptive sensors, such as an accelerometer, a gyrometer, etc., and exteroceptive sensors, such as cameras, radars, lidars, ultrasounds, etc., as well as data fusion methods configured to process the information received and calculate the state (position, speed, acceleration, yaw, etc.) of the ego vehicle 10 and the surrounding objects 7.

[0039] Detecting the behavior of a vehicle of the main vehicle 6 makes it possible to determine a prediction of the positions of the secondary vehicles 3, 4, 5 located behind the main vehicle 6, in the direction of movement of the vehicles.

[0040] The speed oscillation of the main vehicle is propagated to the secondary vehicles, so that the greater the number of secondary vehicles, the greater the prediction capacity of a lane change time window of the ego vehicle 10 will be.

[0041] The prediction makes it possible to create a time window for the change of vehicle ego 10 based on all the vehicles circulating on the lane adjacent to vehicle ego 10.

[0042] The ego vehicle comprises a system 12 for predicting the trajectory of said ego vehicle configured to issue a lane change instruction for the ego vehicle based on an estimate of the behavior of the vehicles traveling in the adjacent lane 2.

[0043] As illustrated in detail on the figure 2 , the system 12 for predicting the trajectory of the ego vehicle 10 comprises a module 13 for determining the state of the vehicles 3, 4, 5, 6 traveling on the lane 2 adjacent to the ego vehicle 10.

[0044] For this purpose, the module 13 comprises a module 13a for determining the position P and the orientation O of the vehicles 3, 4, 5, 6 traveling on the adjacent lane 2 to the ego vehicle 10, and a module 13b for determining the speed V of the vehicles 3, 4, 5, 6 traveling on the adjacent lane 2 to the ego vehicle 10. The position, P, the orientation O and the speed V of the adjacent vehicles are obtained in particular by the various proprioceptive and exteroceptive sensors of the perception system 11 of the ego vehicle 10.

[0045] The system 12 for predicting the trajectory of the ego vehicle 10 further comprises a module 14 for estimating a dynamic model of the vehicles traveling on the adjacent lane 2. The module 14 is configured to establish a dynamic model for each pair of consecutive vehicles traveling on the adjacent lane 2. This makes it possible to obtain an accurate dynamic model for each of the adjacent vehicles based on the data from the vehicles in the vicinity on the same traffic lane.

[0046] The vehicle dynamics model estimation module 14 is configured to determine a second-order transfer function corresponding to the behavior of a vehicle relative to adjacent vehicles. The behavior of the ego vehicle depends on its longitudinal model and its longitudinal controller.

[0047] Module 14 uses an autoregressive exogenous model, or ARX, to obtain dynamic data from a pair of consecutive vehicles.

[0048] The exogenous autoregressive calculation model is written according to the following equation: A z . y t = B z . u t − nk + e t

[0049] With: z, a time offset, nk, a delay, u(t), an input data, here the speed of the previous secondary vehicle, y(t) an output data, here the speed of the main vehicle, e(t), an error value, and A(z) and B(z) second order polynomials.

[0050] The polynomials A(z) and B(z) are written according to the following equations: A z = 1 + a 1 . z − 1 + a 2 . z − 2 B z = b 1 + b 2 . z − 1 + b 3 . z − 2

[0051] The estimated dynamic models are then validated in a dynamic model validation module 16 configured to validate the dynamic models as a function of the actual speed V(t) at a time t and the previous speed V(t-1) at a previous time t-1.

[0052] The system 12 for predicting the trajectory of the ego vehicle 10 further comprises a module 18 for predicting the movement of adjacent vehicles as a function of the validated dynamic model and their initial position and a module 20 for predicting the movement of the ego vehicle 10 as a function of the prediction of the movement provided by the module 18 and information coming from proprioceptive sensors C of the ego vehicle 10.

[0053] The system 12 for predicting the trajectory of the ego vehicle 10 further comprises a module 22 for determining the lane change of the ego vehicle 10 as a function of the predictions of the movement of the ego vehicle and of the adjacent vehicles and of the overall trajectory T of the ego vehicle 10 and, by way of non-limiting example, of information originating from a Cart map of the road on which the ego vehicle is traveling.

[0054] The lane change determination module 22 of the ego vehicle 10 is configured to evaluate an adequate time window allowing the ego vehicle to change lanes safely.

[0055] The lane change instruction from the vehicle ego 10 is transmitted to a module 24 for executing the lane change.

[0056] In the case where there is no possibility for the ego vehicle to change lanes, the lane change determination module 22 can be configured to inform the trajectory prediction module 20 of the ego vehicle 10, in particular with a view to modifying its parameters, such as in particular its speed.

[0057] As illustrated on the figure 3 , the method 50 for predicting the trajectory of the ego vehicle 10 comprises a step 51 for determining the state of the vehicles 3, 4, 5, 6 traveling on the lane 2 adjacent to the ego vehicle 10.

[0058] The determination step 51 makes it possible to collect or recover the information on the position P, the orientation O and the speed of the ego vehicle 10 and the vehicles 3, 4, 5, 6 traveling on the adjacent lane 2 to the ego vehicle 10. The position, P, the orientation O and the speed V of the adjacent vehicles are obtained in particular by the various proprioceptive and exteroceptive sensors of the perception system 11 of the ego vehicle 10.

[0059] The method 50 for predicting the trajectory of the ego vehicle 10 further comprises a step 52 for estimating a dynamic model of the vehicles traveling on the adjacent lane 2. During this step 52, a dynamic model is established for each pair of consecutive vehicles traveling on the adjacent lane 2. This makes it possible to obtain a precise dynamic model for each of the adjacent vehicles based on the data from the vehicles in the vicinity on the same traffic lane.

[0060] The dynamic model for a pair of consecutive vehicles is obtained by determining a second-order transfer function corresponding to the behavior of a vehicle with respect to adjacent vehicles using an autoregressive exogenous model, acronym "ARX" in English terms, to obtain the dynamic data of a pair of consecutive vehicles. The behavior of the ego vehicle depends on its longitudinal model and its longitudinal controller. The autoregressive exogenous model is explained with reference to equations Math1 to Math3 above.

[0061] The estimated dynamic models are then validated in a dynamic model validation step 53 configured to validate the dynamic models as a function of the actual vehicle speed V(t) at a time t and the previous vehicle speed V(t-1) at a previous time t-1. For example, to validate a dynamic model, the error of the ARX calculation model is compared with a threshold value. If the error is less than said threshold value, the dynamic model is validated.

[0062] The method 50 for predicting the trajectory of the ego vehicle 10 further comprises a step 54 for predicting the movement of adjacent vehicles based on the dynamic model validated in step 53 and their initial position.

[0063] The method 50 for predicting the trajectory of the ego vehicle 10 further comprises a step 55 for predicting the movement of the ego vehicle 10 based on the prediction of the movement provided in step 53 and information from proprioceptive sensors C of the ego vehicle 10.

[0064] The lane change of the ego vehicle 10 is then determined in step 56 as a function of the predictions of the movement of the ego vehicle and the adjacent vehicles and of the overall trajectory T of the ego vehicle 10 and of information coming from a mapping Cart of the road on which the ego vehicle is traveling.

[0065] Step 56 of determining the lane change of the ego vehicle 10 makes it possible to evaluate an adequate time window allowing the ego vehicle to change lanes safely.

[0066] The lane change instruction of the ego vehicle 10 is transmitted, in step 57, to a module 24 for executing the lane change of the ego vehicle 10.

[0067] In the event that there is no possibility for the ego vehicle to change lanes, the ego vehicle 10 is informed of the need to modify its parameters, such as its speed.

[0068] The method 50 of predicting the trajectory of an ego vehicle is repeated until a lane change possibility is found.

[0069] Thanks to the invention, the evolution of traffic on a lane adjacent to the traffic lane of the ego vehicle is predicted reliably and in real time, and makes it possible to predict a position of the ego vehicle after changing from the main lane to the adjacent lane.

[0070] Furthermore, the invention makes it possible to take into account the constraints of the road on which the ego vehicle is traveling.

[0071] Such a system and method for predicting the trajectory of an ego vehicle makes it possible to smooth traffic flow, without endangering the safety of the ego vehicle and surrounding vehicles.

Claims

1. Method (50) for predicting the trajectory of an ego vehicle (10) travelling in a main lane (1), wherein a lane change by the ego vehicle from the main lane (1) to an adjacent lane (2) is determined according to an estimate of the dynamic behaviour of a group of vehicles travelling in the adjacent lane (2), said group (7) of vehicles comprising at least one main vehicle (6) located in immediate proximity to the ego vehicle (10) and one secondary vehicle (5) located to the rear of said ego vehicle, behind the main vehicle (6) in the direction of travel, wherein position (P), orientation (O) and speed (V) information of the ego vehicle (10) and of the vehicles (3, 4, 5, 6) of the group (7) of vehicles is gathered, characterized in that a dynamic model per pair of consecutive vehicles travelling in the adjacent lane (2) is established according to the gathered information, the dynamic model per pair of consecutive vehicles being acquired by determining a second-order transfer function corresponding to the behaviour of the ego vehicle relative to each considered pair of consecutive vehicles using an autoregressive exogenous computation model.

2. Method (50) according to Claim 1, wherein the established dynamic models are confirmed by comparing an error (e(t)) of the autoregressive exogenous computation model with a threshold value dependent on the actual speed (V(t)) of each vehicle at an instant (t) and on the speed (V(t-1)) of said vehicle at a previous instant (t-1).

3. Method (50) according to Claim 2, wherein the movement of the adjacent vehicles is predicted according to the confirmed dynamic model and the initial position of said vehicles, the movement of the ego vehicle (10) is predicted according to the prediction of the movement of the adjacent vehicles and information originating from proprioceptive sensors (C) of the ego vehicle (10) and the lane change by the ego vehicle (10) is determined according to said predictions of the movement of the ego vehicle and of the adjacent vehicles and an overall trajectory (T) of the ego vehicle (10).

4. Method (50) according to any one of the preceding claims, wherein the steps of the method (50) are repeated until a lane change possibility is found.

5. System (12) for predicting the trajectory of an ego vehicle (10) travelling in a main lane (1) configured to determine a lane change by the ego vehicle (10) from the main lane (1) to an adjacent lane (2) according to an estimate of the dynamic behaviour of a group of vehicles travelling in the adjacent lane (2), said group (7) of vehicles comprising at least one main vehicle (6) located in immediate proximity to the ego vehicle (10) and one secondary vehicle (5) located to the rear of said ego vehicle, behind the main vehicle (6) in the direction of travel, the system comprising a module (13) for gathering the position (P), orientation (O) and speed (V) information of the ego vehicle (10) and of the vehicles (3, 4, 5, 6) of the group (7) of vehicles, characterized in that it comprises a module (14) for estimating a dynamic model per pair of consecutive vehicles travelling in the adjacent lane (2) according to the gathered information, the dynamic model per pair of consecutive vehicles being acquired by determining a second-order transfer function corresponding to the behaviour of the ego vehicle relative to each considered pair of consecutive vehicles using an autoregressive exogenous computation model.

6. System (12) according to Claim 5, comprising: - a module (16) for confirming the established dynamic models by comparing an error (e(t)) of the autoregressive exogenous computation model with a threshold value dependent on the actual speed (V(t)) of each vehicle at an instant (t) and on the previous speed (V(t-1)) at a previous instant (t-1); - a module (18) for predicting the movement of the adjacent vehicles according to the confirmed dynamic model and the initial position of said vehicles; - a module (20) for predicting the movement of the ego vehicle (10) according to the prediction of the movement of the adjacent vehicles and information originating from proprioceptive sensors (C) of the ego vehicle (10); and - a module (22) for determining the lane change by the ego vehicle (10) according to said predictions of the movement of the (ego) vehicle and of the adjacent vehicles and an overall trajectory (T) of the ego vehicle (10).

7. Ego motor vehicle (10) comprising a system (11) for perceiving and a system (12) for predicting the trajectory of the ego vehicle according to either one of Claims 5 and 6.