Method for controlling a motor vehicle lighting system

Abstract

The invention relates to a method for controlling a lighting system (3) of a motor vehicle (1) having a system (2) for detecting objects, comprising the following steps: (E1) defining a set (Gi) of types of at least one detectable object; (E2) acquiring, by means of the detection system, a data set (Si) relating to the position (Pi,j,k) of a plurality of objects (Oi,j,k) belonging to a type (Ti,j) of the set; (E4, E51, E52) determining a lighting model (Mi) associated with the set and defining at least one zone (Ai,l) called initial detection zone and a light pattern (Pi,l) called initial light pattern for the light beam (Fi) emitted in the initial detection zone; and (E6) controlling the lighting system so as to emit a light beam having the initial light pattern in the initial detection zone of the lighting model.

Description

Technical Field

[0001] This invention relates to the field of motor vehicle lighting. More specifically, this invention relates to lighting systems for motor vehicles. Background Technology

[0002] Modern motor vehicles are increasingly equipped with systems for partial or full autonomous driving. This type of system is used to replace the human driver of the vehicle under certain conditions, particularly speed or environmental conditions, either only for a portion of its journey or for the entire journey. For this purpose, among other things, the autonomous driving system controls all or some of the various components of the motor vehicle that may affect its trajectory or speed, and in particular controls the steering, braking, and engine or transmission components.

[0003] In order to autonomously implement this control without endangering the lives of vehicle occupants or other road users, the vehicle is equipped with a collection of sensors and one or more computers capable of processing the data collected by these sensors to estimate the environment in which the vehicle is driving. Therefore, the autonomous driving system controls the aforementioned components based on route instructions and environmental assessments to guide passengers to their destination while ensuring the safety of passengers and others.

[0004] The set of sensors available in a vehicle typically includes cameras capable of capturing images of all or part of a road scene. This type of sensor is valuable given the high image resolution and acquisition frequency it can provide. However, such sensors have significant drawbacks, particularly their relationship to the lighting of the road scene. In practice, the road scene must be adequately illuminated to allow for the detection of objects within the scene using image processing software employed by one or more computers in the autonomous driving system. In insufficient lighting, objects may go undetected, a situation particularly detrimental if the object is a road user or obstacle on which the vehicle is approaching.

[0005] Therefore, there is a need for lighting that can maximize the probability of detecting objects on the road based on images of the road scene captured by the vehicle's cameras.

[0006] Currently, although motor vehicles are typically equipped with road lighting systems, usually including a pair of headlights, these lighting systems emit beams of light, and the area of ​​these beams emitting on the road, as well as the luminosity within these areas, is used to help the driver perceive objects. On the other hand, these beams are by no means optimized for cameras, and their emission areas and / or their luminosity within these areas may be insufficient or unsuitable to allow the detection of objects in the images captured by the camera. Summary of the Invention

[0007] Therefore, the present invention falls into this context and is intended to meet the cited need by proposing a solution capable of generating road illumination from a motor vehicle, which differs from road illumination obtained using existing lighting beams, and enables the maximization of the probability that objects on the road can be detected based on images of the road scene captured by the vehicle's camera.

[0008] For these purposes, one aspect of the present invention is a method for controlling a lighting system of a motor vehicle equipped with an object detection system, the object detection system including a system for acquiring images of all or part of the vehicle's environment, the method comprising the following steps:

[0009] a. A set of types for at least one object detected by the detection system of the motor vehicle;

[0010] b. The detection system acquires a dataset relating to the positions of multiple objects belonging to the type of the set in the environment of the vehicle;

[0011] c. Based on the dataset, determine a lighting model associated with the set, the lighting model defining at least one region, referred to as an initial detection region, associated with the set of object types and addressable by the lighting system, and a luminance, referred to as an initial luminance, for a light beam emitted by the lighting system in the initial detection region associated with the set;

[0012] d. Control the lighting system based on the determined lighting model so as to emit a light beam with the initial luminous intensity in the initial detection area of ​​the lighting model.

[0013] Therefore, the present invention proposes collecting data related to the positions of objects on the road, which are classified into at least one set of object types, and specifically into a predefined set of multiple object types. This data enables the description of at least one area in which any new object belonging to one of these sets or types, to be detected by a vehicle's detection system, is likely to be present. However, each of these sets of object types may require illumination characteristics specific to that set, particularly due to the ability of these types of objects to reflect their received light back to the detection system, or due to the ability of these types of objects to contrast with the rest of the road scene based on their received light. Therefore, a luminance can be defined for each set of object types, maximizing the probability that the detection system will actually detect that type of object. Thus, illumination that can be emitted by the illumination system can be segmented into beams, each beam emitted in an initial detection area within the initial detection area, its own luminance dedicated to the type of object likely to be present in that area. Therefore, it can be understood that the areas and dedicated luminance are thus intended to fully support the image acquisition system, rather than for the driver of the vehicle. Therefore, these beams are the "default" beams emitted before any detection is performed by the detection system. Then, each detection of an object in the initial detection area performed by the detection system can result in a modification of the beams emitted in that area, for example, for the purpose of tracking the object or not dazzling the object.

[0014] For example, the image acquisition system may be a camera capable of capturing images of the road scene in front of or behind a motor vehicle, or, as a variation, one or more cameras capable of capturing images of the road scene around the entire motor vehicle. Where applicable, the detection system may include one or more processing units designed to perform image processing algorithms on the images acquired by the image acquisition system to detect objects in the images, particularly objects of the type mentioned above. If desired, the detection system may include one or more additional sensors, particularly laser scanners, radar or infrared sensors, and possibly processing units designed to perform data fusion algorithms on data from the image acquisition system and the sensors or these other sensors.

[0015] Advantageously, it is possible to pre-collect datasets related to the location of objects under daytime conditions.

[0016] In one embodiment of the present invention, the dataset related to the position of the object acquired in the acquisition step includes: for each object, the position of the object, referred to as the initial position, when the object is detected by the detection system.

[0017] Preferably:

[0018] a. The limiting steps include limiting the set of multiple individual object types;

[0019] b. The data acquisition step includes acquiring a dataset for each set related to the location of multiple types of objects belonging to the set in the vehicle environment;

[0020] c. The determination step includes determining, based on each dataset, an illumination model associated with the set associated with that dataset, each model defining at least one region, referred to as an initial detection region, that is associated with a set of object types and can be addressed by the illumination system, and a luminance, referred to as an initial luminance, for a light beam emitted by the illumination system in the initial detection region associated with that set.

[0021] Where applicable, the control steps include controlling the lighting system based on the determined lighting model to emit multiple light beams, particularly multiple light beams emitted simultaneously, each light beam having an initial luminance in an initial detection region of one of these lighting models. Thus, the set of light beams forms a segmented overall light beam. The set of object types is understood to specifically refer to a set of at least one object type, particularly a set of multiple object types having substantially the same or similar lighting requirements, reflectance coefficients, dynamic behavior, and / or geometric characteristics. The set of object types may, for example, include:

[0022] a. Various types of traffic signs and traffic lights;

[0023] b. All types of road users, especially pedestrians, bicycles, and vehicles; all types of animals;

[0024] c. Various types of ground markings and obstacles that may be reached by vehicles within a time frame less than a given threshold (e.g., 2 seconds).

[0025] Advantageously, the step of determining the model includes: for each type of object in the set, modeling a region, referred to as a first detection region, based on the dataset, the region encompassing all initial positions of objects of that type. Where applicable, the initial detection region is determined based on the first detection regions of all object types in the set. Preferably, the step of determining the model may include: for each type of object in the set, modeling a region, referred to as a first detection region, based on the dataset associated with the set, the region encompassing all initial positions of objects of that type. Where applicable, each initial detection region is determined based on the first detection regions of all object types in the same set. For example, the initial detection region or each initial detection region may be formed by a combination of all first detection regions of all object types in the same set.

[0026] In one embodiment of the invention, each step of modeling a first detection region for the type of object implements a machine learning algorithm, enabling the determination of the first detection region based on the initial position of the object of the type of object. For example, the machine learning algorithm may include, but is not limited to, learning algorithms trained in a supervised or unsupervised manner, such as linear or nonlinear regression, Naive Bayes classifiers, etc. Support vector machine (machineàvecteur support) or neural network (réseau de neurones), K-means algorithm (un algorithme du type K-moyennes).

[0027] For example, in the case of a collection of multiple different object types, a machine learning algorithm can be trained to determine a first detection region for each object type based on multiple datasets (each dataset including the initial positions of multiple objects of a type belonging to one set in the set) in the vehicle's environment, such that the initial detection regions do not intersect, and each initial detection region is formed by a combination of all the first detection regions of the same set of object types.

[0028] According to a non-limiting example, a machine learning algorithm can be trained to determine the boundaries of a region for each object type, such that the probability of detecting an object of said object type within it is greater than a given threshold, and / or the probability of detecting an object of a type other than said object type within it is less than a given threshold. Where applicable, each threshold can be different for each object type.

[0029] Advantageously, in the step of determining the model, the initial luminosity of the light beam is determined based on the type of at least one object in the set of object types. Preferably, in the step of determining the model, the initial luminosity of the light beam is determined based on a first detection region for each object type in the set of object types, and particularly based on the location of each first detection region in the environment of the motor vehicle.

[0030] In an exemplary embodiment of the invention, the method includes the step of providing at least one range of values ​​for parameters related to the behavior of the motor vehicle or to the environment. Where applicable, the step of determining the lighting model associated with the set is the step of determining a lighting model associated with the set that is variable based on the values ​​of the parameters.

[0031] For example, parameters related to the behavior of a motor vehicle could be its speed, and / or its trajectory, and / or its yaw. Parameters related to the environment of a motor vehicle could be weather conditions and / or road profiles, and particularly road curvature and / or road gradient, and / or data regarding the vehicle's location, particularly GPS (Global Positioning System) data.

[0032] A variable lighting model is understood to refer to a lighting model in which the initial detection area in the vehicle's environment has a shape, size, and / or position that varies based on the values ​​of the parameters, and / or the initial luminance of the lighting model varies based on the values ​​of the parameters. In other words, a variable lighting model defines multiple initial detection areas and / or initial luminances associated with a set of the same object type, and each initial detection area and / or initial luminance is associated with a given value within the range of values ​​of the parameters.

[0033] Advantageously, the step of determining the model includes: for each object type in the set and for each value of the range of values ​​for the parameter, the step of modeling a first detection region for the object type based on the dataset, the first detection region covering all initial positions of objects of the object type whose parameter has the value when the initial position is acquired. Where applicable, each initial detection region associated with the set of object types is determined based on the first detection regions of all object types in the set associated with the same value of the parameter.

[0034] In one exemplary embodiment of the present invention:

[0035] a. The limiting step includes limiting at least three sets of objects of a certain type, the at least three sets including a first set containing at least objects of the ground marking type, a second set containing at least objects of the road user type, and a third set containing at least objects of the traffic sign type;

[0036] b. The determination step includes determining three lighting models, each associated with a set in the set, the three lighting models including a first lighting model associated with the first set, a second lighting model associated with the second set, and a third lighting model associated with the third set; and

[0037] c. The step of controlling the lighting system includes controlling the lighting system based on a determined lighting model, so as to simultaneously emit a first light beam having an initial luminance of the first lighting model in an initial detection region of the first model, emit a second light beam having an initial luminance of the second lighting model in an initial detection region of the second model, and emit a third light beam having an initial luminance of the third lighting model in an initial detection region of the third model.

[0038] In this example, the initial detection region determined for the first model can be the bottom region, the initial detection region determined for the second model can be the central region, and the initial detection region determined for the third model can be the top region.

[0039] Advantageously, the method further includes the following steps:

[0040] a. The vehicle's object detection system detects objects of a given type from a set of object types;

[0041] b. Control the lighting system to modify the beam based on the type of object being detected.

[0042] According to the present invention, the light beam has an initial luminosity in the initial detection region suitable for assisting the object detection system in detecting the appearance of an object of a given type. However, motor vehicles and / or the object being detected can move, causing the object to move within the reference frame of the image acquisition system. While the initial luminosity may be suitable during the initial detection of the object, it may subsequently become unsuitable due to the movement. Therefore, this feature enables the initial luminosity to be adapted to the type of the object and its possible movement, allowing the detection performance of the object detection system to be maintained after the initial detection of the object. Where applicable, the step of detecting an object of a given type may include a sub-step of estimating the position of the object.

[0043] Advantageously, the steps of controlling the lighting system include: generating a region in the light beam flush with the detected object, the region having a luminance adapted to the type of the detected object; and moving the region based on the movement of the detected object in a reference frame of the image acquisition system. "A region with adapted luminance" is understood to refer to a region whose size, shape, position in the road scene, and / or luminance is adapted to the type of the detected object. For example, in the case of detecting a "motor vehicle" type object, the region could be a region centered on the detected vehicle and whose luminance is less than a given glare threshold. In the case of detecting a "pedestrian" type object, the region could be a region centered on the detected pedestrian and whose luminance is greater than a given detection threshold.

[0044] In one embodiment of the invention, the motor vehicle is equipped with a system for partial or full autonomous driving. Where applicable, the implementation of the step of controlling the lighting system is conditional upon activation of the autonomous driving system, and the method includes the following steps:

[0045] a. The occupants of the vehicle receive an instruction to retract manual control of the motor vehicle;

[0046] b. Control the lighting system to emit at least one predetermined illumination and / or signal indication beam.

[0047] The predetermined illumination and / or signal indication beam can be, for example, a predetermined near beam, a dark beam, or a predetermined far beam. Advantageously, the control step may include a sub-step of turning off the beam with initial luminosity in the initial detection area.

[0048] Another subject of the present invention is a motor vehicle, the motor vehicle comprising: an object detection system including a system for acquiring images of all or part of the environment of the vehicle; a lighting system; a system for partially or fully autonomous driving; and a controller for the lighting system, the controller being designed to implement control steps according to the method of the present invention.

[0049] Another subject of the present invention is a lighting system for a motor vehicle according to the invention.

[0050] Advantageously, the lighting system includes at least one lighting module and a controller, the at least one lighting module being capable of emitting pixelated light beams, the controller being capable of receiving instructions for emitting a given light function, and the controller being designed to control the lighting module to emit pixelated light beams with defined characteristics based on the instructions.

[0051] According to an exemplary embodiment of the invention, the illumination module is designed such that the pixelated beam is a beam comprising multiple pixels, for example, 500 pixels of size between 0.05° and 0.3°, distributed across multiple rows and columns, such as 20 rows and 25 columns. For example, the illumination module may include multiple basic light sources and optical devices designed to emit the pixelated beam together. Where applicable, the controller may be designed to selectively control each basic light source of the illumination module such that the light source emits a basic beam of light from one pixel of the pixels forming the pixelated beam. A light source is understood to refer to any light source that may be associated with an electro-optical element capable of being selectively activated and controlled to emit a basic beam of light whose intensity is controllable. This may in particular be a light-emitting semiconductor chip, a light-emitting element of a monolithic pixelated light-emitting diode, part of a light-converting element capable of being excited by a light source, or a light source associated with a liquid crystal or micromirror. Attached Figure Description

[0052] The invention will now be described using examples that are merely illustrative and do not in any way limit the scope of the invention, and reference will be made to the accompanying drawings, in which, in each figure:

[0053] [ Figure 1 A method for controlling a lighting system for a motor vehicle according to an embodiment of the present invention is illustrated schematically and in part.

[0054] [ Figure 2 The diagram schematically and partially illustrates a motor vehicle according to an exemplary embodiment of the present invention;

[0055] [ Figure 3 The illustration schematically and partially shows the methods used for implementation. Figure 1 The dataset for the method;

[0056] [ Figure 4 [Illustratively and partially showing] Figure 1 The implementation steps of the method;

[0057] [ Figure 5 [Illustratively and partially showing] Figure 1 The implementation steps of the method;

[0058] [ Figure 6 [Illustratively and partially showing] Figure 1 The implementation steps of the method;

[0059] [ Figure 7 [Illustratively and partially showing] Figure 1 The implementation steps of the method; and

[0060] [ Figure 8[Illustratively and partially showing] Figure 1 The implementation method of the steps of the method. Detailed Implementation

[0061] In the following description, unless otherwise stated, elements that are structurally or functionally consistent and appear in each figure retain the same reference numerals.

[0062] Figure 1 A method for controlling a lighting system 3 of a motor vehicle 1 according to an embodiment of the present invention is described.

[0063] Figure 2 The vehicle 1 shown includes an object detection system 2. The detection system 2 includes an image acquisition system 21.

[0064] The system 21 includes a camera capable of capturing images of the entire road scene surrounding the motor vehicle 1. The detection system 2 also includes a processing unit (not shown) designed to perform image processing algorithms on the images captured by the camera 21 in order to detect objects in the images.

[0065] The motor vehicle 1 includes a lighting system 3, which includes a plurality of lighting modules 31 to 36, each of which is capable of emitting a pixelated beam of light in a given direction, so that the lighting system 3 can illuminate the entire road around the motor vehicle 1.

[0066] The motor vehicle 1 includes a controller for the lighting system 3, which is capable of selectively controlling each of the lighting modules 31 to 36 and selectively controlling each pixel of the pixelated light beam emitted by these lighting modules 31 to 36.

[0067] The motor vehicle 1 includes a system for fully autonomous driving, which is designed to control the steering components, braking components, and engine or transmission components of the motor vehicle when the motor vehicle is in autonomous driving mode, in particular based on objects detected by the processing unit of the detection system 2 in images acquired by the camera 21.

[0068] In the remainder of the description, [ Figure 1 The method will be used to control lighting modules 31 and 32, and will be combined with [ Figure 3 ]to[ Figure 8 To describe [ Figure 1 The method, Figure 3 ]to[ Figure 8Each shows the road scene in front of the vehicle, such as what can be seen by camera 21 and what can be illuminated by lighting modules 31 and 32. It should be understood that the method can also be applied to the road scenes on the sides and rear of the vehicle by controlling lighting modules 33 to 36.

[0069] In step E1, the sets G1 to G2, which predefine the types of multiple objects, are... N Each set Gi will contain one or more objects of type T. i，j Grouping them into groups. In the described example, step E1 is simplified by limiting the following sets: type T1 of objects grouping traffic signs, and the first set G1 of objects grouping pedestrians and vehicles into separate groups. 2，1 and T 2，2 The second set G2; and the type T that groups together obstacles and ground markings that can be reached by a vehicle in less than two seconds. 3，1 The third set G3. In the diagram, type T 1，1 Objects will be represented by squares, type T 2，1 Objects will be represented by circles, type T 2，2 The object will be represented by a triangle, and type T 3，1 Objects will be represented by stars.

[0070] In step E2, multiple datasets S1 to S2 are collected. N Dataset S i Each data P i，j，k This represents the values ​​estimated by the vehicle detection system and belonging to set G. i Type T i，j Object O i , j The set of positions P, where k is the location of the detection system, which is similar to detection system 2 and includes a camera similar to camera 21. i，j，k The object O i，j，k All positions are grouped together, starting from the object's initial position P. i，j，k (0) Until the final position, the initial position is estimated when the object is detected by the detection system in the camera's field of view, and the final position is estimated at the last moment before the object disappears from the camera's field of view.

[0071] [ Figure 3 This illustrates a simplified example of datasets S1 to S3 associated with sets G1 to G3, with the initial position P of the data in these datasets. i，j，k The road scene projected onto the front of motor vehicles.

[0072] For each data P in the set representing the location of the object... i，j，kEach dataset S i It also includes the speed V of the motor vehicle when estimating the set of positions of the object. i，j，k .

[0073] In the preliminary step E1', parallel to the limiting step E1, multiple velocity ranges ΔV1 to ΔV are defined. M .

[0074] In step E3, the dataset S1 to S N Each dataset in the dataset is divided into multiple subsets S. 1，1 To S N，M If in the object O i，j，k initial position P i，j，k (0) The speed of the motor vehicle V i，k (0) in the range ΔV I Inside, the dataset S i Each data P i，j，k Assigned to subset S i，l In other words, subset S i，l Includes its type T i，j Belonging to set G i And its initial velocity V i，j，k (0) in the range ΔV I Object O inside i，j，k All initial positions P i，j，k (0).

[0075] In step E4, for each object of each set Gi, the type T is determined. i，j And for each speed range ΔV I The region Z of the first detection area, which is referred to as the type of the object. i，j、l Modeling is performed. This region Z... i，j，l The type of object T i，j Object O i，j，k All initial positions P i，j，k (0), and its initial velocity V i，j，k (0) in the range ΔV I Inside.

[0076] For these purposes, a support vector machine has been pre-trained to determine the boundaries of a region for each label under supervision and based on multiple points labeled with different labels and located in space, such that the number of points labeled with that label and presented in the region is greater than a given threshold, and the number of points labeled with a different label and presented in the region is less than a given threshold.

[0077] In step E4, then provide ΔV for the same range at the input of the previously trained support vector machine. ISubdataset S i，l Each subset of the dataset, along with the type for each object and the threshold for each range, is used to determine the object type T. i，j The first detection area Z i，j、l Therefore, each region Z i，j，l The type of object T i，j Object O i，j，k initial position P i，j，k (0), and its initial velocity V i，j，k (0) in the range ΔV I Inside. Also note that each region Z... i，j，l This is thus modeled by a neural network, so that when the initial velocity V i，j，k (0) in the range ΔV I The type T of the object was detected within the time. i，j Object O i，j，k The probability is highest when the initial velocity V is [missing information]. i，j，k (0) in the range ΔV I The internal time detects a type T different from the object. i，j Object O of type i，j，k The probability is the lowest.

[0078] In step E51, by combining those belonging to the same set G i The type of object T i，j The first detection area Z i，j、l Determine the initial detection region A i，l .

[0079] thus,[ Figure 4 The dataset S shows an initial speed between 90 km / h and 130 km / h. 1，1 S 2，1 and S 3，1 [ Figure 4 It also shows the type T determined at the end of step E51. 1，1 T 2，1 T 2，2 and T 3，1 and the region A determined at the end of step E52. 1，1 A 2，1 and A 3，1 Related region Z 1，1，1 Z 2，1，1 Z 2，2，1 and Z 3，1，1 .

[0080] [ Figure 5 The dataset S, showing initial speeds between 50 km / h and 90 km / h, is also presented. 1，2S 2，2 and S 3，2 [ Figure 5 It also shows the type T determined at the end of step E51. 1，1 T 2，1 T 2，2 and T 3，1 and the regions A1, 2, and A determined at the end of step E52. 2，2 and A 3，2 Related region Z 1，1，2 Z 2，1，2 Z 2，2，2 and Z 3，1，2 .

[0081] [ Figure 6 The dataset S, representing initial speeds between 0 and 50 km / h, is also shown. 1，3 S2, S3 and S 3，3 [ Figure 6 It also shows the type T determined at the end of step E51. 1，1 T 2，1 T 2，2 and T 3，1 and the region A determined at the end of step E52. 1，3 A 2，3 and A 3，3 Related region Z 1，1，3 Z 2，1，3 Z 2，2，3 and Z 3，1，3 .

[0082] Area A associated with the set of traffic signs G1 1、1 A 1、2 and A 1、3 It is the area located more in the upper part of the road scene, region A, which is associated with the set of road users G2. 2、1 A 2、2 and A 2、3 It is the area located more in the center of the road scene, and the area A associated with the collection of objects G3 in the vehicle's directly navigable space. 3、1、 A 3，2 and A 3，3 It is the area located more in the lower part of the road scene. It can be seen that, with the same set G... i Associated initial detection region A i，l Its shape, size, and position in space change based on its initial velocity.

[0083] Each initial detection area A i，l The following regions in space belong to the set G associated with those regions.i Type T i，j The probability that the object can be detected by the detection system 2 based on the image acquired by the camera 21 is particularly high.

[0084] In step E52, for those belonging to the same set G i The type of object T i，j Each initial detection region A i，l Determine the initial photometric value P i，l This allows us to take into account the set G. i To improve the detection performance of detection system 2 in the case of the type of object. Determine the initial photometric value P. i，l This may include: determining the area to be detected by the lighting system 3 in the initial detection zone A. i，l The minimum, average, and / or maximum light intensity of the emitted light beam; or the determination of the light intensity used by the illumination system 3 in the initial detection area A. i，l The light intensity of multiple pixels, multiple groups of pixels, or even all pixels of the emitted light beam.

[0085] For example, for region A 3，1 A 3，2 and A 3，3 The light emitted by illumination modules 31 and 32 is substantially parallel to the ground. Therefore, the retroreflection of this light to camera 21 will not be very strong, and thus the average light intensity of the beams emitted in these areas must be high to allow for the detection of markers or obstacles in these areas. For area A... 2、1 A 2、2 and A 3、3 The light emitted by lighting modules 31 and 32 will be substantially perpendicular to the road user. Therefore, this light will be satisfactorily reflected to camera 21, such that the average light intensity of the beam emitted in these areas can be lower than that emitted in area A. 3，1 A 3，2 and A 3，3 The average light intensity of the beam in region A. 1，1 A 1，2 and A 1，3 The light emitted by illumination modules 31 and 32 will be substantially perpendicular to the traffic sign. Since traffic signs are typically equipped with a reflective coating, the light will be reflected back in an amplified form. Therefore, the average light intensity of the beam emitted in these areas must be low to avoid saturating the sensor of camera 21.

[0086] At the end of step E52, for ΔV1 to ΔV M All ranges and for the same set G i Initial detection area A i，l and initial luminosity P i，lThe set forms with the set G i Associated lighting model M i .

[0087] It should be noted that this is used for sets G1 to G... N Determine these lighting models M1 to M N Steps E1 to E52 are generated by a computer unit including a memory and a processor, the memory storing the set G1 to G2 defined in steps E1 and E1'. N and the speed range ΔV1 to ΔV M and datasets S1 to S N The processor is capable of performing these steps. The computer unit is separated from vehicle 1, therefore steps E1 to E52 are executed before subsequent steps. At the end of step E52, models M1 to M... N The image is loaded into the memory of the controller of the lighting system 3, for example, in the form of an image, where each pixel represents a pixel for a pixelated beam emitted by modules 31 and 32, and the gray level of the image pixels represents the light intensity setpoint of the basic beam of pixels that can be emitted by these modules 31 and 32 to form a pixelated beam.

[0088] In step E6, when the motor vehicle 1 is in autonomous driving mode, the lighting modules 31 and 32 of the lighting system 3 are controlled by the controller to emit multiple beams F in front of the vehicle. i The overall beam F is formed, and each beam F i Conforms to model M1 to M N One of the models. Because the speed of the motor vehicle is in the range ΔV... I Within a range, therefore each beam F i With initial luminosity P i，l Launched in initial detection area A i，l In the middle. These beams F1 to F N It is the default beam emitted when no object is detected on the road.

[0089] [ Figure 7 The image shows a road scene illuminated by beams F1, F2, and F3 simultaneously emitted by lighting modules 31 and 32, which together form a segmented overall beam F. Figure 7 In the example, the motor vehicle travels at a speed between 50 km / h and 90 km / h.

[0090] Steps E7 and E8, which will now be described, relate to the adaptation of the segmented global beam F performed after object O is detected, while step E9 relates to the vehicle switching from autonomous driving mode to manual driving mode.

[0091] In step E7, object O1 is detected by detection system 2, and the detection system 2 classifies object O1 as belonging to set G. i Type T 2，1 Another object O2 is detected by detection system 2, and this detection system 2 classifies the other object O2 into type T belonging to set G2. 2，2 .like[ Figure 7 As shown in the image, object O1 is a motor vehicle, and object O2 is a pedestrian. These objects are located in the initial detection area A. 2，2 Therefore, objects O1 and O2 are illuminated by beam F2, and the luminosity P of beam F2 is... 2，2 This enables the detection performance of detection system 2 for these types of objects to be improved.

[0092] In step E8, after detecting object O, the controller controls the lighting system 3 to generate region B in the beam, which is centered on object O and has a luminous intensity suitable for the type of object O. In the described example, after detecting objects O1 and O2, the controller controls modules 31 and 32 to generate a low-intensity region B1 centered on object O1 and an over-intensity region B2 centered on object O2 in the beam F2. Region B1 allows the detection system 2 to continue detecting vehicle O1 and the movement of vehicle 1 while the vehicle is moving, without dazzling the possible driver of the vehicle. Region B2 allows the detection system 2 to continue detecting pedestrian O2 while vehicle 1 is moving. Therefore, while these objects O1 and O2 are moving in the field of view of camera 21, regions B1 and B2 remain centered on these objects O1 and O2, and the estimation of the position of these objects O1 and O2 at a given moment allows the controller to move regions B1 and B2 at the next moment, as […]. Figure 8 As shown, this continues until objects O1 and O2 leave the camera's field of view. At the end of step E8, the controller of the lighting system controls modules 31 and 32 to make beam F2 conform to the default lighting model M2.

[0093] In step E9, when the autonomous driving system receives an instruction I to retract manual control of the vehicle 1, the controller controls the lighting system, and specifically controls lighting modules 31 and 32, to gradually convert the overall beam F into a specified low beam or low beam LB. If the autonomous driving system receives an instruction to switch the vehicle 1 to autonomous mode, the controller controls the lighting system 3 to emit F1, F2, and F3, respectively, conforming to models M1, M2, and M3, using lighting modules 31 and 32.

[0094] The above description clearly explains how the present invention achieves its set objectives, and particularly by proposing a method for controlling a lighting system for a motor vehicle, wherein data related to the location of objects classified according to their type enables the description of at least one region in which any new object belonging to one of these types is likely to exist, and wherein luminance is defined to maximize the probability that an object of that type will be actually detected by the detection system of the motor vehicle. With the aid of the present invention, the light beam emitted by the lighting system is thus used for an image acquisition system that fully supports the detection system.

[0095] In no event should the invention be considered limited to the embodiments specifically described in this document, and in particular extends to any equivalent apparatus and any technically feasible combination of such apparatuses. In particular, types of detection systems different from the described detection system are conceivable, and systems combining image acquisition systems with other types of sensors, such as detecting and estimating the position of objects on a road through multi-sensor data fusion, are conceivable. Types of objects different from those described are also conceivable. Other examples of methods for modeling the first detection region are also conceivable, and in particular, specific types of machine learning algorithms different from those described. Modeling the first detection region based on parameters different from vehicle speed is also conceivable.

Claims

1. A method for controlling a lighting system (3) of a motor vehicle (1), the motor vehicle being equipped with an object detection system (2), the object detection system comprising an image acquisition system (21) for acquiring images of all or part of the environment of the vehicle, the method comprising the steps of: a. (El) defines a set (G of types of at least one object to be detected by the detection system of the motor vehicle i ); b. (E2) the detection system acquires a dataset (S i ) related to positions (P i，j，k ) in the environment of the vehicle of a plurality of objects (O i，j，k ) belonging to a type (T i，j ) of the set; c. (E4, E51, E52) Determine the lighting model (M) associated with the set based on the dataset. i The lighting model defines at least one region (A) that is associated with a set of object types and can be addressed by the lighting system, and is referred to as the initial detection region. i，l ) and a light beam (F) emitted by the illumination system in the initial detection area associated with the set. i The initial luminosity (P) is called the initial luminosity. i，l ) light intensity; d. (E6) Controlling the lighting system based on the determined lighting model to emit a light beam with the initial luminance in the initial detection area of ​​the lighting model. Among them, the data collected in the acquisition step (E2) is related to the object (O). i，j，k The location (P) i，j，k The dataset (S) related to this i This includes: for each object, the position of the object, referred to as the initial position (P), when the object is detected by the detection system (2). i，j，k (0)), and Among them, the lighting model (M) is determined. i Steps (E4, E51, E52) of the above include: for the set (G) i The type of each object (T) i，j Based on the dataset (S) i The region (Z) is designated as the first detection region for the type of the object. i，j，l The modeling steps (E4), the region (Z) i，j，l ) encompasses objects of the type described (O) i，j，k All initial positions (P) i，j，k (0)), and wherein the initial detection region (A) is determined based on a first detection region of all types of objects in the set. i，l ).

2. The method according to claim 1, wherein, For the type of object (T) i，j The first detection area (Z) i，j，l Each step of the modeling process (E4) implements a machine learning algorithm, enabling the modeling of objects based on their type (O). i，j，k The initial position (P) i，j，k (0)) to determine the first detection area.

3. The method according to claim 1 or 2, wherein, In determining the lighting model (M) i In steps (E4, E51, E52), based on the set of types of the object (G) i At least one type of object in ) (T i，j ) to determine the beam (F i The initial luminosity (P) of ) i，l ).

4. The method of claim 3, wherein the method includes providing parameters (V) related to the behavior of the motor vehicle (1) or the environment. i，j，k (0)) at least one value range (ΔV) L The lighting model is variable based on the values ​​of the parameters.

5. The method according to claim 1 or 2, wherein: a. The limiting step (E1) includes limiting the type of object (T) 1，1 T 2，1 T 2，2 T 3，1 The at least three sets (G1, G2, G3) comprise a first set (G1) containing at least objects of the ground marking type, a second set (G2) containing at least objects of the road user type, and a third set (G3) containing at least objects of the traffic sign type. b. The determination steps (E4, E51, E52) include determining three lighting models (M1, M2, M3), each lighting model being associated with a set in the set, the three lighting models including a first lighting model associated with the first set, a second lighting model associated with the second set, and a third lighting model associated with the third set; and c. Step (E6) of controlling the lighting system (3) includes controlling the lighting system based on the determined lighting model in order to emit light into the initial detection area (A) of the first lighting model. 1，2 The initial luminosity (P) of the first illumination model in the ) 1，2 The first beam (F1) is emitted into the initial detection area (A) of the second illumination model. 2，2 The initial luminosity (P) of the second illumination model in ) 2，2 The second beam (F2) and the initial detection area (A) emitted in the third illumination model. 3，2 The initial luminance (P) of the third illumination model in the ) 3，2 The third beam (F3).

6. The method according to claim 1 or 2, further comprising the following step: a. (E7) The object detection system (2) of the vehicle (1) detects the types of the objects from the set (G) i Detecting a given type (T) in ) i，j Object (O); b. (E8) Control the lighting system (3) to modify the beam (F) based on the type of object detected. i ).

7. The method according to claim 6, wherein, Step (E8) of controlling the lighting system (3) includes: in the beam (F i The step of generating a region (B) that is flush with the detected object (O) and has a region adapted to the type (T) of the detected object. i，j The steps of measuring the luminosity of the detected object in the reference frame of the image acquisition system (21) and moving the region based on the movement of the detected object in the reference frame of the image acquisition system (21).

8. The method according to claim 1 or 2, wherein the motor vehicle (1) is equipped with a system for partially or fully autonomous driving, wherein, The implementation of step (E6) of controlling the lighting system (3) is conditional upon activation of the autonomous driving system, and the method includes the following steps: a. The occupants of the vehicle receive an instruction (I) for retracting manual control of the motor vehicle; b. (E9) Control the lighting system to emit at least one predetermined specified lighting and / or signal indicator beam (LB).

9. A motor vehicle (1), said motor vehicle comprising: The object detection system (2) includes a system (21) for acquiring images of all or part of the environment of the vehicle. Lighting system (3); system for partially or fully autonomous driving; and controller for said lighting system, said controller being designed to implement control steps (E6) of the method according to any one of claims 1-8.

Images