Methods for controlling autonomous vehicles
The system optimizes autonomous vehicle trajectories by locally modifying maps and kinematic profiles to adapt to dynamic traffic scenarios, enhancing responsiveness and safety in real-world driving conditions.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- AMPERE SAS
- Filing Date
- 2022-04-29
- Publication Date
- 2026-07-29
AI Technical Summary
Existing autonomous vehicle control systems are limited to predefined traffic scenarios and lack the ability to adapt to dynamic changes in the environment, leading to inefficiencies and suboptimal performance in real-world driving conditions.
The system utilizes perception means to detect situations within the perception limit of the vehicle and optimizes the trajectory and kinematic profile by locally modifying the map and/or kinematic parameters, such as increasing speed, acceleration, and jerk thresholds, to adapt to dynamic traffic scenarios.
Enhances the responsiveness and adaptability of autonomous vehicles to dynamic traffic conditions, improving driving comfort and safety by replicating human driving behaviors and optimizing trajectories in real-time.
Smart Images
Figure 0007897266000002 
Figure 0007897266000003 
Figure 0007897266000004
Abstract
Description
[Overview of the project]
[0001] The present invention relates to a method for controlling an autonomous vehicle. The present invention further relates to a device for controlling an autonomous vehicle. The present invention also relates to a computer program that implements the above-described method. Finally, the present invention relates to a storage medium in which such a program is stored.
[0002] Autonomous vehicle technology essentially uses high-resolution maps associated with speed profiles that satisfy the lateral and longitudinal constraint parameters of the autonomous vehicle to predefine trajectories corresponding to different autonomous driving scenarios within a given traffic zone.
[0003] For example, when an autonomous vehicle approaches an X-shaped intersection, the potential trajectories for approaching the intersection (turning right, turning left, or going straight) are already pre-mapped, just as the associated velocity profile is.
[0004] However, these technologies have the drawback of being inherently limited to traffic scenarios predefined from maps.
[0005] U.S. Patent Publication 9511767(B1) discloses a method for controlling an autonomous vehicle using data from a perception system to adapt the autonomous vehicle's trajectory in response to the behavior of surrounding vehicles. However, this solution has drawbacks.
[0006] The object of the present invention is to provide a device and method for controlling an autonomous vehicle that overcomes the aforementioned drawbacks and improves upon devices and methods for controlling autonomous vehicles known from the prior art. In particular, the object of the present invention is to provide a device and method that is simple, reliable, and improves autonomous driving.
[0007] For this purpose, the present invention relates to a method for controlling an autonomous vehicle comprising at least one perception means and a first map containing stored digital data representing the actual infrastructure of the environment of the autonomous vehicle. The control method comprises - defining a first trajectory of the autonomous vehicle and a first associated kinematic profile; - detecting, from the data transmitted by at least one perception means, a situation to be optimized on a given segment of the first trajectory, the given segment being located over a distance smaller than the maximum detection distance of the at least one perception means in front of the autonomous vehicle, detecting the situation to be optimized; - optimizing at least one element from the first map and the first kinematic profile, optimizing the first map includes determining a local map on a given segment, and / or optimizing the first kinematic profile includes determining a second kinematic profile associated with the given segment, at least one parameter of the second kinematic profile being different from the parameters of the first kinematic profile; optimizing at least one element; - controlling the movement of the autonomous vehicle on a given segment taking into account the local map and / or the second kinematic profile and including.
[0008] Determining the local map can in particular make it possible, by means of Bézier curves, for a new curve of the given segment to be determined between the entry point and the exit point of the given segment, the new curve being able to have a continuity G2 with the first trajectory at the entry point and the exit point.
[0009] Determining a second kinematic profile associated with a given segment can include calculating a maximum collision - free speed threshold from data resulting from at least one perception means, and the second kinematic profile can have a maximum speed below the maximum collision - free speed threshold.
[0010] The maximum collision - free speed threshold can be calculated by solving an equation in which the maximum collision - free speed threshold is equal to the square root of the product of the braking value required to reach the maximum collision - free speed threshold and twice the maximum detection distance of at least one perception means.
[0011] The step of detecting a situation to be optimized includes the autonomous vehicle measuring a dwell time of remaining in the situation to be optimized and comparing the dwell time with a triggering threshold associated with the situation, and the optimization step can be executed when the dwell time is greater than the triggering threshold.
[0012] A set of predefined situations can be stored in an electronic memory, and the step of detecting a situation to be optimized can include comparing a perception resulting from at least one perception means with the set of predefined situations.
[0013] The set of predefined situations can include negotiating a roundabout with two empty lanes and / or entering a roundabout in dense traffic conditions and / or two traffic lanes merging into a single lane and / or crossing a congested junction and / or overtaking a very slow vehicle on a two - way road with a single lane in each direction.
[0014] The second kinematic profile can have a maximum speed at least 20% higher, and / or a maximum acceleration at least 200% higher, and / or a jerk at least 1,000% higher than the first kinematic profile.
[0015] The present invention also relates to a device for controlling an autonomous vehicle, the autonomous vehicle being equipped with a motion actuator. The device comprises hardware and / or software elements that implement the method defined above, in particular hardware and / or software elements designed to implement the method according to the present invention, and / or the device comprises means for implementing the method defined above.
[0016] The present invention further relates to an autonomous vehicle comprising the control device described in the prior claims.
[0017] The present invention also relates to a computer program product comprising program code instructions stored on a computer-readable medium for implementing the steps of the method defined above when the program runs on a computer. The present invention also relates to a computer program product that is downloadable from a communication network and / or stored on a computer-readable data medium and / or can be executed by a computer, wherein the computer program product comprises instructions that cause the computer to implement the method defined above when the program is executed by the computer.
[0018] The present invention also relates to a computer-readable data storage medium for storing a computer program comprising program code instructions for implementing the method defined above. The present invention also relates to a computer-readable storage medium comprising instructions, wherein when the instructions are executed by a computer, the computer causes the computer to implement the method defined above.
[0019] The present invention further relates to signals in a data medium for transmitting the computer program product defined above.
[0020] The attached drawings show, as an example, one embodiment of a control device according to the present invention and an embodiment for carrying out a control method according to the present invention. [Brief explanation of the drawing]
[0021] [Figure 1] This figure shows one embodiment of an autonomous vehicle equipped with a control device. [Figure 2] This figure shows an embodiment of the present invention for determining the kinematic profile of an autonomous vehicle. [Figure 3] This is a flowchart of an embodiment for executing a control method according to one embodiment of the present invention. [Figure 4] This figure shows a first example for implementing a control method according to an embodiment of the present invention. [Figure 5] This figure shows a second example for implementing a control method according to an embodiment of the present invention. [Modes for carrying out the invention]
[0022] The autonomous vehicle 100 according to embodiments of the present invention may be any type of autonomous vehicle, in particular a passenger car, a utility vehicle, a truck, or even a public transport vehicle such as a bus or shuttle.
[0023] The autonomous vehicle 100 includes a control system 10 and a motion actuator 4.
[0024] The motion actuator 4 forms part of the chassis of the autonomous vehicle 100. The motion actuator 4 includes, in particular, an engine torque actuator, a brake actuator, and a rotary actuator for the steering wheel. The motion actuator 4 receives commands from the control system 10 to implement the movement of the autonomous vehicle along a trajectory determined by the control system 10.
[0025] The control system 10 mainly comprises the following elements: - A perceptual system 2 comprising a perceptual means 21 and a module 22 for processing the electronic horizon. - Electronic memory 3, and - A computing unit 5 comprising a microprocessor 1, a local electronic memory 6, and a communication interface 7 that enables the microprocessor 1 to communicate with a perception system 2 and a motion actuator 4.
[0026] The perception means 21 may comprise all or part of the following devices, namely a set of cameras and / or lidar and / or radar. Alternatively or additionally, the perception means 21 may comprise a vehicle-to-vehicle communication system (V2V system) or a vehicle-to-anything communication system (V2X system) that enables vehicles to exchange information with each other, with infrastructure and with pedestrians.
[0027] Advantageously, the perception means 21 allows the environment located 360° around the autonomous vehicle 100 to be perceived.
[0028] In a preferred embodiment, the sensing means 21 comprises five cameras and one LiDAR. Images from the cameras are captured periodically and synchronously and aggregated with data resulting from one rotation of the LiDAR.
[0029] Based on the data generated from the perceptual means 21, module 22 for processing the electronic horizon constructs a structured perception of the driving scene.
[0030] Periodically, especially each time data is received, module 22 determines the maximum capacity of the perception system 2, i.e., the maximum detection distance DLIM of the perception means 21, which varies according to the driving scenario. For example, when the autonomous vehicle 100 is traveling on a highway, the distance DLIM is greater than when it is traveling on a traffic circular intersection, typically called a "roundabout".
[0031] Distance DLIM also depends on other parameters such as weather conditions, brightness, the presence of road infrastructure elements (e.g., tunnels), or surrounding traffic (e.g., trucks).
[0032] Throughout the remainder of this specification, the term “roundabout” refers to a circular traffic intersection where vehicles traveling on the intersection have priority over vehicles entering the intersection. Furthermore, the maximum detection distance of a perceptual means is called the “perception limit distance (DLIM)”.
[0033] The perceptual limit distance (DLIM) can be determined from the perceptual limits of each perceptual means. In one embodiment, the perceptual limit distance (DLIM) is the minimum value between the perceptual limit for the image generated by the camera and the perceptual limit for the data generated by the lidar.
[0034] The perceptual limit of an image generated by a camera can be defined as the maximum distance at which the height of any object of interest (vehicle, pedestrian) is greater than a minimum threshold, e.g., 25 pixels. The magnitude of this first limit is several tens of meters, for example, 40 meters under normal visibility conditions, especially with regard to weather.
[0035] The perceptual limit of data generated by LiDAR can be defined as the maximum distance at which the number of laser points reaching the object of interest (vehicle, pedestrian) exceeds a minimum threshold.
[0036] In the perceptual zone, defined by a circle centered on the autonomous vehicle and using a radius equal to the perceptual limit distance DLIM, module 22 for processing the electronic horizon constructs an occupied grid of the driving scene. The point cloud resulting from the lidar is discretized to some extent into cells. Each cell in the grid is determined to be empty, occupied, or undetermined. This determination is based on processing the height of the points in the cell, particularly by applying minimum and maximum height thresholds for points in the cell.
[0037] The occupied grid defines a collision-free zone around the autonomous vehicle 100's position, as well as a three-dimensional representation of obstacles located within the perception zone.
[0038] The data generated from module 22 for processing the electronic horizon provides a structured perception of driving scenarios on the perception zone, particularly the perception of traffic around the autonomous vehicle 100, with respect to the traffic density and location of objects of interest (vehicles, pedestrians).
[0039] The control system 10 further comprises an electronic memory 3 in which a set of digital data representing the actual infrastructure of the environment of the autonomous vehicle 100 is stored. This digital data generally pertains to all elements that can be used to infer choices regarding driving the autonomous vehicle 100. These elements, in particular, pertain to the layout and limits of navigable road sections, the number of lanes in the road sections, applicable speed limits in the road sections, the types and locations of ground markings, the presence and location of traffic signs, traffic signals, and roundabouts.
[0040] This digital data will be simply referred to as "Map M1" below.
[0041] In embodiments of the present invention, computer 1 enables the execution of software comprising the following modules. - Module 11 for determining the trajectory, in cooperation with perception system 2 and map 3. - Module 12, which works in conjunction with perceptual system 2 to detect situations that should be optimized. - Optimization module 13, which works in conjunction with perceptual system 2. - A module 14 for controlling the movement of the autonomous vehicle, which works in conjunction with the motion actuator 4.
[0042] A configuration for implementing a method for controlling an autonomous vehicle is described below with reference to Figure 3. This method includes steps E1 to E4.
[0043] In the first step E1, the planned trajectory T1 is determined between the starting point and the destination point of the autonomous vehicle 100. For this purpose, the map M1 stored in the electronic memory 3 and the data generated from the perception system 2 are used in particular.
[0044] Throughout the remainder of this specification, the term “trajectory” is used to specify the temporal change of a state vector that defines the characteristics of the movement of the autonomous vehicle 100. In a preferred embodiment, the state vector comprises position, in particular the coordinates x, y, longitudinal and transverse velocities and / or longitudinal and transverse accelerations and / or yaw rate and / or jerk. Throughout the remainder of this specification, the term “position” is used to specify either the x, y coordinates of the state vector or the state vector as a whole.
[0045] The planned orbit T1 is preferably divided into orbital segments, which will also be referred to as “segments” for the remainder of this specification.
[0046] Advantageously, each segment is defined such that when the autonomous vehicle 100 enters the segment, it is possible for it to be fully positioned within the perception zone of the autonomous vehicle 100. In particular, the length of the segments is smaller than the perception limit distance DLIM.
[0047] The term “kinematic profile” is used throughout the remainder of this specification to specify the general velocity, acceleration, and jerk profiles implemented by a vehicle. To some extent, the kinematic profile acts as a template for calculating the velocity, acceleration, and jerk implemented in trajectory calculations. The template defined by the kinematic profile sets limits on velocity, acceleration, and jerk.
[0048] The kinematic profile takes into account constraints imposed by road infrastructure, namely the road layout, number of lanes, width of each lane, presence of intersections and / or traffic signals, and speed limits. The kinematic profile is also determined according to the technical characteristics of the autonomous vehicle. The kinematic profile is further defined based on driving comfort criteria, particularly related to maximum speed and / or lateral acceleration and / or jerk thresholds.
[0049] In one embodiment, the kinematic profile may be defined according to a method for limiting the jerk (where the jerk is the derivative of acceleration) shown in Figure 2.
[0050] Figure 2 is a graph with three curves, J(t), a(t), and v(t), representing the temporal changes in jerk, acceleration, and velocity of the autonomous vehicle 100, respectively.
[0051] Curve J(t) is a step curve, where time is divided into fixed-duration intervals, and the jerk value is constant across each interval. The duration of the intervals depends on the vehicle's acceleration limits and the distance available for acceleration or braking.
[0052] In the illustrated example, Jerk can assume three discrete values: a value of 0, a maximum value Jmax, and a minimum value -Jmax.
[0053] The value of Jmax is small to promote user comfort in the autonomous vehicle 100. In one embodiment, the value of Jmax is 1 m / s -3 It will be set to this.
[0054] The curve J(t) therefore determines the temporal change of acceleration a(t). - Curve a(t) consists of linear segments whose slope is determined by the value of Jmax and the duration of the time interval. - The curve a(t) changes between the acceleration value Amax and the acceleration value -Amax.
[0055] Curve a(t) determines the temporal change of velocity v(t) between zero velocity and maximum velocity Vmax, where Vmax can be determined by the velocity limit for the orbital segment.
[0056] This method is repeated on each track segment to define a track T1 that follows the road infrastructure and is as comfortable as possible without slowing down traffic. Thus, for each track segment, the applied velocity, acceleration, and jerk are determined according to the kinematic profile and map M1.
[0057] Throughout the remainder of this specification, orbital T1 is defined as having N segments S1, ..., S N It is considered to be divided into the first kinematic profile P1 i Each segment S i In relation to this, the first kinematic profile P1 i Each of these is defined according to the first map M1, and segment S i Each of these is a first map M1 and a first profile P1 associated with the first map M1. i It is defined accordingly.
[0058] Therefore, the term “track segment” also specifies the curve of the track, and similarly the velocity, acceleration, and jerk values applied to move the vehicle along the curve of the segment.
[0059] Therefore, assuming two separate points C and D, the two segments connecting point C to point D are considered different if their respective curves are different, or if the velocity, acceleration, or jerk values applied to move a vehicle along either of the curves of the segments are different. Thus, the trajectory T1, divided into trajectory segments, is stored in the local memory 6 of the computing unit 5.
[0060] Next, step E2 follows. In step E2, the intention is to detect the situation to be optimized on segment S of track T1 located in front of autonomous vehicle 100 i and segment S i is completely located within the perception limit DLIM of the perception means 21.
[0061] In other words, step E2 determines segment S of track T1 where the situation to be optimized is located. Segment S i is located in front of the vehicle, which means that segment S i forms a part of track T1 that the autonomous vehicle has not yet passed through. i In the described embodiment, it is assumed that the situation to be optimized is limited to a separate segment. In an alternative embodiment, it can be considered that the situation to be optimized can extend over several consecutive track segments. However, all segments must be within the perception limit DLIM.
[0062] The situation to be optimized is a traffic scenario where the movement of the autonomous vehicle is obstructed by the first track T1, particularly by the limits of the first kinematic profile P1 related to track T1, V1max and / or A1max and / or J1max. For example, in a dense traffic situation, autonomous vehicle 100 may remain stuck at the entrance of a roundabout for an extremely long time if the kinematic profile related to its track does not allow it to accelerate sufficiently to enter the traffic.
[0063]
[0064] In other words, the kinematic profile P1 implemented by trajectory T1 was defined according to driving comfort criteria that limit the responsiveness of the autonomous vehicle. Profile P1 allows the autonomous vehicle to manage most situations under optimal safety and comfort conditions. However, profile P1 may prove unsuitable in some situations. This unsuitability, for example, may cause the vehicle to significantly slow down when it is unable to overtake a truck in front of it, or when it is kept stationary for an extended period, such as when it is unable to enter a roundabout due to traffic density.
[0065] Other situations that should be optimized may relate not to the first kinematic profile P1, but to a first map M1 representing the actual infrastructure of the traffic environment for the autonomous vehicle 100. In some cases, the first map M1 induces a specific trajectory for the vehicle that does not correspond to the normal behavior of a human driver and does not optimize driving. This occurs, for example, on a trajectory segment crossing a two-lane roundabout where no other vehicles are proceeding. The first trajectory T1 determines movement on the outer lane of the roundabout, which requires a significant reduction in the autonomous vehicle's speed before it enters the roundabout. Thus, trajectory T1 is defined with respect to map M1 representing a roundabout with two lanes. However, when a human driver approaches a multi-lane roundabout where no other vehicles are moving, the human driver adapts the trajectory of their vehicle to this situation; that is, the human driver anticipates reducing the curvature of their trajectory by using two lanes, and therefore approaches the roundabout at a considerably higher speed than the human driver would use if they were traveling only on the outer lane of the roundabout. The adaptations implemented by human drivers are almost equivalent to modifying the road infrastructure, particularly by assuming that the roundabout has only one lane.
[0066] In the configuration for implementing this method, data received from the perception system 2, particularly data from module 22 for processing the electronic horizon, is compared with a predefined set of situations stored in local electronic memory 6.
[0067] Throughout the remainder of this specification, - The data generated from module 22 for processing the electronic horizon is called "perceptual data". - The set of predefined situations stored in local electronic memory 6 is called the "situation table". - Data included in the status table that relates to predefined statuses are called "status parameters."
[0068] The predefined situations correspond to the situations that should be optimized. These may include the following: - Throughout the remainder of this specification, the term "empty roundabout situation" refers to the entrance of a roundabout with two empty lanes, and / or - Throughout the remainder of this Specification, a “congested roundabout situation” refers to a long stop when entering a roundabout in congested traffic conditions, and / or - Merging two traffic lanes into a single lane under congested traffic conditions, and / or - Long stops at congested intersections, and / or - To remain behind an extremely slow vehicle for an extended period of time on a two-way road with a single lane in each direction.
[0069] Each of the predefined situations may be associated with one or more situation parameters, including road infrastructure elements and / or traffic density and / or triggering thresholds.
[0070] Therefore, comparing perceptual data with all predefined situations may include comparing the road infrastructure elements taken by trajectory T1 with the road infrastructure elements associated with each of the predefined situations. These road infrastructure elements may include, for example, multi-lane roundabouts, or X-shaped intersections or junctions between two lanes, or two-way roads with one lane in each direction.
[0071] Furthermore, comparing perceptual data with predefined situations may also include comparing the traffic density resulting from the perceptual data with density thresholds associated with each of the predefined situations. Density thresholds associated with predefined situations could, for example, be a minimum density threshold for a congested roundabout situation. Density thresholds could also be, for example, a maximum threshold for an empty roundabout situation.
[0072] Furthermore, the situation parameter may include a “triggering” threshold corresponding to the minimum time to remain in the said situation. The minimum time to remain in the situation varies according to a predefined situation. For example, the minimum time to remain in the situation may be 0 for an empty roundabout situation. For a congested roundabout situation, the minimum time to remain in the situation may be on the order of one minute or more.
[0073] Other situational parameters not described herein may be considered.
[0074] When all situational parameters related to this situation are examined, especially when compared with perceptual data, the situation that should be optimized is identified.
[0075] If no situation requiring optimization is detected, step E4 proceeds, which involves controlling the movement of the autonomous vehicle 100.
[0076] When a situation requiring optimization is detected, step E3 follows, which involves optimizing the trajectory.
[0077] The trajectory is based on the first map M1 and / or the first kinematic profile P1, specifically segment S where the situation to be optimized is located. i Kinematic profile P1 related to i It will be optimized by correcting it.
[0078] Advantageously, the situation table associates at least one predefined optimization method with each predefined situation.
[0079] The predefined optimization methods are: - Map correction, or - Kinematic profile correction, or - Map and kinematic profile corrections It is possible.
[0080] Regarding empty roundabout situations, the situation table can associate the empty roundabout situation with a map modification that involves creating a local map M2 that converts a two-lane roundabout to a one-lane roundabout. Optionally, the situation table can associate map and kinematic profile modifications with this situation under several traffic conditions.
[0081] In the case of a congested roundabout situation, the situation table modifies the kinematic profile and creates a new trajectory segment S'. i It can be associated with the calculation of segment S, and that curve is segment S i It is superimposed on the curve, kinematic profile P1 i Kinematic Profile P2 with a higher maximum velocity and / or acceleration and / or jerk threshold than that of the other model. i Implement it.
[0082] In a preferred embodiment, the second kinematic profile P2 i Determining the second kinematic profile P2 involves calculating the maximum collision-free speed threshold VCOL based on data from the perceptual means 2. i This is defined to include speeds below the maximum collision-free speed threshold VCOL.
[0083] The threshold VCOL is the maximum speed at which the autonomous vehicle can move while having the potential to stop in order to avoid colliding with an obstacle detected by the perception system 2.
[0084] The threshold VCOL is calculated from the occupied grid of the driving scene defined by module 21 for processing the electronic horizon. The driving scene is used to determine which obstacles, in particular, of interest such as other vehicles or pedestrians, are closest to the autonomous vehicle 100.
[0085] In a simplified embodiment, the threshold VCOL can be calculated by solving the following equation Math1. [Math1] TIFF0007897266000001.tif5170 Here, - A is acceleration, and more specifically, the braking value required to reach velocity VCOL. - DLIM is the perceptual limit distance.
[0086] Second kinematic profile P2 i To determine this, the previously described method for limiting the jerk may be used. In this case, the increase in the responsiveness of the autonomous vehicle 100 corresponds to the second profile P2 i The maximum speed threshold of the first profile P1 iThis is achieved by setting Vmax2 to a value greater than the maximum velocity threshold Vmax1 and less than the maximum collision-free velocity VCOL. Advantageously, the maximum acceleration and / or jerk thresholds Amax2 and Jmax2 of the second profile are set to the same value as the first profile P1, respectively. i It is greater than the Amax1 and Jmax1 thresholds.
[0087] Local map M2 and / or second kinematic profile P2 i After determining the second orbital segment S', this method determines the second orbital segment S'. i , or optimized segment S' i Returning to step E1 to determine the orbit in order to calculate the first orbital segment S i Replace it.
[0088] Optimized segment S' i This is the local map M2 and / or the second kinematic profile P2. i It is determined to conform to the requirements.
[0089] Optimized segment S' i In particular, when map M1 is not modified, the first segment S i They may have the same curve. However, those segments differ with respect to the velocity and / or acceleration and / or jerk they implement.
[0090] Optimized segment S' i is a given segment S i Connect the input and output points of the optimized segment S'. i This is advantageous because it involves continuity G2, i.e., the optimized segment S'. i The continuity of curvature between various arcs forming the optimized segment S' at point A. i The continuity of curvature with the preceding orbital segment, and the optimized segment S' at point B. iIt is determined to have continuity, which is also the continuity of curvature with the subsequent orbital segment.
[0091] In one embodiment, the optimized segment S' i It can be defined by a fifth-degree polynomial, particularly using a Bézier curve.
[0092] Next comes step E4. In the fourth step E4, the movement of the autonomous vehicle is controlled according to the trajectory determined in the preceding step, which may be an optimized segment S'. i It is equipped with.
[0093] The trajectory is transmitted to the control law of the autonomous vehicle 100 in order to be converted into a command order sent to the motion actuator 4 of the autonomous vehicle 100.
[0094] Two examples of implementing the control method are illustrated in Figures 4 and 5. Note that the position of the autonomous vehicle 100 in these figures represents step E4, in which the vehicle moves along an optimized trajectory.
[0095] Figure 4 shows a first example for implementing this method. At instantaneous T, the autonomous vehicle 100 moves along segment S of trajectory T1. i-1 As it proceeds, the trajectory T1 is calculated according to the map M1 and kinematic profile P1. The following segment S i It is perfectly located within the perceptual zone defined by the perceptual limit distance DLIM calculated at instantaneous T. The subsequent segment S i This implements movement on the outer lane of a roundabout. If there is no traffic on the roundabout, an empty roundabout situation is detected in step E2. In step E3, map M1 is locally modified by local map M2 to replace the two-lane roundabout with a one-lane roundabout. Then the new trajectory S' i However, in step E1, the new orbit S' is calculated using the local map M2.i Segment S i The entrance point A is segment S i It connects to exit point B. In step E4, the autonomous vehicle 100 moves segment S' between point A and point B. i It is controlled to move upwards.
[0096] Figure 5 shows a second example for implementing this method. In this second example, the following segment S i The system determines movement on the outer lane of the roundabout. Traffic on the roundabout is congested, and the autonomous vehicle 100 must stop at the entrance to the roundabout. When the given duration of time during which the autonomous vehicle has been stationary at the entrance to the roundabout ends, in step E2, the congested roundabout situation is detected. In step E3, a new kinematic profile P2 i However, it is defined to facilitate the vehicle entering the roundabout. Next, kinematic profile P2 i In step E1, a new segment S' connects point A and point B. i Used to determine segment S i and S' i It follows the same curve but does not implement the same velocity and / or acceleration and / or jerk. In step E4, the autonomous vehicle 100 moves segment S' between point A and point B. i It is controlled to move upwards.
[0097] Finally, the control method according to the present invention uses perceptual means to detect a set of traffic scenarios and optimize the trajectory of the autonomous vehicle according to the identified traffic scenario.
[0098] The trajectory optimization according to the present invention implements two optimization levers that can be used independently of each other or simultaneously. The first optimization lever involves locally modifying a map representing the infrastructure of an identified traffic scenario. The second optimization lever involves temporarily increasing the responsiveness of an autonomous vehicle by increasing the maximum speed and / or acceleration and / or jerk thresholds applied in the trajectory of the autonomous vehicle.
[0099] These optimization levers are determined according to the surrounding traffic to avoid any collisions. Map modifications are preferably performed in the absence of surrounding traffic. Increased vehicle responsiveness is performed under dense traffic conditions, taking into account the maximum collision-free speed calculated from electronic horizon data. Thus, trajectory optimization is performed under safe conditions.
[0100] In addition to reliability in terms of road safety, the present invention has several advantages. Firstly, the present invention reduces the risk of autonomous vehicles being in congested situations due to high traffic density. By periodically increasing the responsiveness of the autonomous vehicle, this allows the autonomous vehicle to easily enter dense traffic and, once the entry operation is complete, return to an optimal level of driving comfort. Thus, the present invention promotes traffic fluidity and driving comfort by adapting the driving style of the autonomous vehicle to the surrounding traffic. More generally, in a given set of situations, the present invention allows the autonomous vehicle to replicate the optimal choices of a human driver faced with these situations.
[0101] Therefore, the present invention enables autonomous vehicles to adapt their driving responsiveness in response to perceptual data relating to static data (infrastructure) and dynamic data (particularly traffic) of the driving scene. The driving fluidity of autonomous vehicles is thus improved while respecting driving safety, and particularly in congested situations, it is avoided.
Claims
1. A control method performed by a computer for controlling an autonomous vehicle (100), comprising at least one perceptual means (21) and a first map (M1) containing stored digital data representing the actual infrastructure of the environment of the autonomous vehicle (100), wherein the control method is - Step (E1) of defining a first trajectory (T1) of the autonomous vehicle (100) and a related first kinematic profile (P1), - Step (E2) of detecting a situation to be optimized on a given segment (S) of the first trajectory (T1) from data transmitted by the at least one sensing means (21), wherein the given segment (S) is located in front of the autonomous vehicle (100) at a distance smaller than the maximum detection distance (DLIM) of the at least one sensing means (21), - A step (E3) of optimizing at least one element from the first map (M1) and the first kinematic profile (P1), Optimizing the first map includes determining a local map (M2) on the given segment (S), and / or Optimizing the first kinematic profile (P1) includes determining a second kinematic profile (P2) related to the given segment (S), wherein at least one parameter of the second kinematic profile (P2) is different from the parameters of the first kinematic profile (P1). Step (E3), - Step (E4) of controlling the movement of the autonomous vehicle (100) on the given segment (S) taking into account the local map (M2) and / or the second kinematic profile (P2) A control method characterized by including
2. The control method according to claim 1, characterized in that determining a local map (M2) enables a new curve of the given segment (S) to be determined between an entry point (A) and an exit point (B) of the given segment (S), and the new curve has continuity G2 with the first trajectory (T1) at the entry point (A) and the exit point (B).
3. The control method according to claim 2, characterized in that determining the local map (M2) makes it possible to determine the new curve by the Bézier curve of the given segment (S) between the entry point (A) and the exit point (B) of the given segment (S).
4. The control method according to claim 1, characterized in that determining a second kinematic profile (P2) related to the given segment (S) includes calculating a maximum collision-free speed threshold (VCOL) from data generated from the at least one sensing means (21), and the second kinematic profile (P2) has a maximum speed that is below the maximum collision-free speed threshold (VCOL).
5. The control method according to claim 4, characterized in that the maximum collision-free speed threshold (VCOL) is calculated by solving an equation such that the maximum collision-free speed threshold (VCOL) is equal to the square root of the product of the braking value required to reach the maximum collision-free speed threshold (VCOL) and twice the maximum detection distance (DLIM) of the at least one sensing means (21).
6. The control method according to claim 1, characterized in that the step (E2) of detecting a situation to be optimized includes measuring the retention time during which the autonomous vehicle (100) remains in a situation to be optimized and comparing the retention time with a triggering threshold related to the situation, and the optimization step (E3) is performed when the retention time is greater than the triggering threshold.
7. The control method according to claim 1, characterized in that a set of predefined situations is stored in an electronic memory (6), and the step (E2) of detecting a situation to be optimized includes comparing a perception arising from the at least one perceptual means (21) with the set of predefined situations.
8. The control method according to claim 7, characterized in that the set of predefined situations includes negotiating a roundabout having two empty lanes and / or entering a roundabout in congested traffic conditions and / or two traffic lanes merging into a single lane and / or crossing a congested junction and / or overtaking an extremely slow vehicle on a two-way road having a single lane in each direction.
9. The control method according to claim 1, characterized in that the second kinematic profile (P2) has a maximum velocity at least 20% higher, and / or a maximum acceleration at least 200% higher, and / or a jerk at least 1,000% higher than the first kinematic profile (P1).
10. A control device (10) for controlling an autonomous vehicle (100), wherein the autonomous vehicle is equipped with a motion actuator (4), and the control device comprises hardware and / or software elements (1, 2, 3, 5, 6, 7, 11, 12, 13, 14, 21, 22) for implementing the control method described in any one of claims 1 to 9, or hardware and / or software elements (1, 2, 3, 5, 6, 7, 21, 22) designed to implement the control method described in any one of claims 1 to 9, and / or means for implementing the control method described in any one of claims 1 to 9.
11. An autonomous vehicle (100) characterized by comprising the control device (10) described in claim 10.
12. A computer program product comprising a program including program code instructions stored on a computer-readable medium, wherein the program code instructions are for performing the steps of the control method described in any one of claims 1 to 9 when the program is executed on a computer program product which is downloadable from a computer or a communication network and / or stored on a computer-readable data medium and / or can be executed by a computer, and the computer program product comprises instructions that cause the computer to perform the control method described in any one of claims 1 to 9 when the program is executed by the computer.
13. A computer-readable data storage medium for storing a computer program comprising program code instructions for carrying out the control method described in claim 1 or claim 12, or a computer-readable storage medium comprising instructions, wherein when the instructions are executed by a computer, the computer causes the computer to carry out the control method described in claim 1.