Method for detecting a type of road on which an equipped vehicle is travelling

Abstract

A detection method for detecting a road type on which a vehicle is travelling, combining a phase of determining a road type and a phase of learning to identify a class of road type, integrating an assessment of the driving comfort on the road as experienced by a user, for a vehicle of a given model and equipped in a given way.

Description

Description Title of the invention: Method for detecting the type of road traveled by an equipped vehicle Technical field of the invention

[0001] The invention relates to a method for detecting, in particular intelligently detecting, the type of road traveled by a vehicle, especially a vehicle in motion. The invention also relates to a system for detecting a road traveled. Furthermore, the invention relates to a vehicle equipped with such a system and to a computer program for implementing the steps of the detection method. Prior art

[0002] In the automotive industry, it is common practice to equip vehicles with electronically controlled suspension systems. Typically, such systems require knowledge of the condition and / or characteristics of the road traveled by an equipped vehicle in order to optimize the system's operation based on the specific characteristics of the road.

[0003] To this end, it is possible to implement methods for detecting poor road conditions and / or obstacles. Such methods rely on detection systems that perform explicit calculations of intermediate variables and characteristics, or "features." Therefore, it is necessary to fine-tune and adjust the detection system to achieve an acceptable level of accuracy. Another drawback of such detection systems is that they are not designed to account for the variability in driver perception or comfort from one type or model of vehicle to another, particularly due to differences in their equipment, for a given type of poor road condition or obstacle.Consequently, a road detected as poor for one vehicle will also be incorrectly considered poor for a different vehicle better suited to that type of road, for example, due to specific equipment such as more appropriate suspension systems. Conversely, a road identified as good, or passable, could be perceived as poor and uncomfortable depending on the type of vehicle used. Such methods therefore suffer from a lack of portability, making them unsuitable for application to a range of different vehicles.

[0004] The present invention falls within this context and aims to provide a detection method suitable for a wide range of vehicles with distinct characteristics, for example private vehicles, commercial vehicles or others, and at a lower cost.

[0005] The present invention relates to a method for detecting a type of road traveled. by a vehicle, in particular a vehicle equipped with a variable suspension system, the method being characterized in that it comprises a phase of determining a type of road travelled comprising: - at least one acquisition step, in real time or for a predetermined duration, of data relating to the vehicle and / or a route traveled by the vehicle implemented through a detection module comprising at least one sensor and / or at least one means of detection, such as a camera; - a step of estimating a probability rate of belonging of the route traveled to at least one class of road type, including bad roads, predefined according to said acquired data; - a stage of classifying the road travelled into one of the road type classes according to the estimated probability rates of belonging and of extracting at least one level of comfort felt by a user previously associated with each road type class according to at least one vehicle-specific equipment (10) capable of influencing the comfort felt by a user during the driving phase.

[0006] Specifically, the process is an intelligent detection process.

[0007] In particular, the method according to the invention may include, prior to the execution of the determination phase, at least one supervised learning phase, specific to the equipped vehicle, of the identification of at least one predefined road type class and of the association of at least one level of comfort felt by a user to each road type class according to at least one vehicle-specific equipment capable of influencing the comfort felt by a user during the driving phase, the learning phase being implemented via a computer, a human-machine interface and / or a memory element.

[0008] In addition, the process may include a step of adjusting at least one auxiliary system fitted to the vehicle, in particular a variable suspension system, so as to correct a level of comfort felt by the user, such a step being implemented when the level of comfort associated with a class of road type determined during the determination phase is less than a predefined adapted comfort level threshold and / or when the determined class of road type is included in a predefined subset of classes identified as requiring the execution of an adjustment step.

[0009] The process may also include a step of updating a map fitted to the vehicle so as to indicate a specific road type class.

[0010] According to one example of an embodiment of the process according to the invention, the classification step can be implemented only when the estimated probability rate is greater than a predetermined threshold of 60%.

[0011] According to one example of how the process is executed, the estimation step can be performed by a network capable of processing time series of data and being trained during a learning phase, in particular by a convolutional neural network.

[0012] The data acquisition stage can be implemented by: - detection and / or estimation of at least one characteristic relating to a portion of the road traveled located upstream of the vehicle, in particular a height of a road profile; and / or - measurement by at least one sensor of at least one characteristic relating to the vehicle in driving situation on said road travelled, in particular suspension travel, wheel speed, longitudinal speed of the vehicle and / or acceleration of the vehicle.

[0013] Also, the determination phase may include, subsequent to the estimation stage, a validation stage of the estimated probability rate, the validation stage being carried out on the basis of measurements, by at least one sensor, of at least one characteristic relating to the vehicle in a driving situation on said road traveled, in particular a suspension travel, a wheel speed, a longitudinal speed of the vehicle and / or an acceleration of the vehicle.

[0014] The present invention also relates to a system for detecting the type of road traveled by a vehicle in motion. The system comprises hardware and / or software elements implementing the method described above. The hardware elements include at least one data processing unit, a module for detecting at least one characteristic of a road traveled, and a control module for at least one auxiliary system fitted to the vehicle, in particular a variable suspension system. Specifically, the detection system may be an intelligent detection system.

[0015] The invention further relates to a motor vehicle comprising at least one auxiliary system capable of affecting a user's perceived level of comfort, such as a variable suspension system, and a detection system according to the invention, capable of controlling said auxiliary system.

[0016] The invention can further be extended to a learning method for a detection method according to the invention or for a detection system according to the invention, comprising at least one supervised learning phase, specific to the equipped vehicle, of the identification of at least one predefined road type class and of the association of at least one level of comfort felt by a user to each road type class according to at least one vehicle-specific equipment capable of influencing the comfort felt by a user during driving, the learning phase being implemented via a computer, a human-machine interface and / or a memory element.

[0017] The invention further extends to a computer program product comprising ins- Program code instructions stored on a computer-readable medium to implement the steps of the process according to the invention when said program is running on a computer. Alternatively, the invention can be extended to a computer program product downloadable from a communication network and / or stored on a computer-readable data medium.

[0018] Also, the invention relates to a computer-readable data recording medium on which is recorded a computer program comprising program code instructions for implementing the detection method according to the invention, or a computer-readable recording medium comprising instructions which, when executed by a computer, cause the computer to implement said detection method.

[0019] The invention also relates to a signal from a data carrier, carrying the computer program product according to the invention.

[0020] Further details, features and advantages will become clearer upon reading the detailed description given below, which is indicative and not exhaustive, in relation to the various implementation examples illustrated in the following figures:

[0021] Figure [Eig.l] is a schematic representation of an embodiment of a detection system according to the invention.

[0022] Figure [Eig.2] is a general flowchart of an example of the execution of a method for detecting a type of road.

[0023] Figure 3 is a schematic representation of an example of the execution of a data acquisition step.

[0024] Figure 4 is a flowchart detailing a second example of an execution of the process according to the invention. Detailed description

[0025] Figure 1 schematically illustrates an example of an embodiment of a detection system 1, in particular an intelligent detection system, for a type of road traveled by a vehicle 10. Such a system is intended for a vehicle, in particular a motor vehicle. The vehicle 10 may be, for example, a passenger car, a commercial vehicle, a truck, or a bus. In particular, the vehicle in question may be a connected and / or autonomous vehicle. Figures 2 to 4 illustrate examples and variants of an embodiment of a detection method, in particular an intelligent detection method, for a type of road traveled by such a vehicle 10. The detection method may also be considered as a method of operating the detection system 1 according to the invention or as a method of operating the motor vehicle 10 equipped with such a detection system 1.

[0026] Here, "route travelled" refers to a road or a section of a road, a traffic lane or roadway at which or from which vehicle 10 is traveling in real time.

[0027] The vehicle 10 includes the detection system 1, which comprises hardware and / or software elements capable of implementing, or designed to implement, the detection method according to the invention, as further described below.

[0028] The said hardware elements comprise at least one data processing unit 2, a detection module 3 for at least one characteristic relating to a route travelled and a control module 4 for at least one auxiliary system 5 equipping the vehicle 10, such as a variable suspension system 51, a driving assistance system 52 or a longitudinal speed management system 53.

[0029] The processing unit 2 comprises at least one computer 21 with hardware and software resources, specifically at least one processor, or microprocessor, cooperating with memory elements 6 of the detection system 1 and / or the vehicle 10. This computer 21 is capable of executing instructions for the implementation of a computer program. In particular, the processing unit 2 can perform at least one other function within the vehicle 10. For example, the computer 21 can be the on-board computer of the vehicle 10, or, more generally, any computer 21 installed in the vehicle 10 permanently or temporarily.

[0030] The detection module 3 can be configured to detect and / or estimate at least one characteristic of a portion of the road traveled upstream of the vehicle 10, such as a road profile height. Here, "road profile height" refers to the height of a point on the road relative to the vehicle. Specifically, at least one such characteristic can be detected and / or estimated based on images of said road portion.

[0031] For example, the detection module 3 may include at least one optical detection means 31, such as a camera, scanning laser, or lidar, capable of providing signals representative of an image of the road traveled, which can then be transmitted to the processing unit 2 for processing and / or archiving. Alternatively, or in combination, the detection means 31 may be acoustic, for example, using ultrasound. Preferably, such a detection means 31 may be mounted at the front of the vehicle 10 or on a roof of the vehicle 10 in order to acquire data relating to a section of road located upstream of the vehicle 10.

[0032] Alternatively or additionally, the detection module 3 may include at least one sensor 32 fitted to the vehicle 10 and capable of measuring at least one characteristic specific to the vehicle 10's operation, for example, relating to the vehicle itself or to a system formed by the vehicle and the road being traveled. Such a characteristic may, but is not limited to, relate to suspension travel, wheel rotation speed, and / or the vehicle's longitudinal speed.

[0033] The control module 4 is configured to control at least one auxiliary system 5 fitted to the vehicle 10. Such an auxiliary system 5 may be a variable suspension system 51, a driver assistance system 52, a longitudinal speed management system, and / or a mapping system, such as a location device 7 fitted to the vehicle 10. The control module 4 is specifically capable of switching on, switching off, modifying, and / or adjusting all or part of the parameters defined by F for at least one of the aforementioned auxiliary systems 5. Advantageously, the control module 4 may be configured to control a plurality of auxiliary systems fitted to the vehicle 10.

[0034] Optionally, the detection system according to the invention may also include all or part of the following options: - a means of localization 7; - a communication module 8.

[0035] The location device 7 can, in particular, enable the location of the vehicle 10. It incorporates, for example, an approximate location system for the vehicle 10 and / or high-definition mapping of the road infrastructure. Specifically, the approximate location of the vehicle 10 can be provided by a GPS-type system (Global Positioning System), which allows the extraction of information concerning the road infrastructure around the approximate position of the vehicle 10 from a mapping database. Alternatively, or in addition, the location device can be a location system embedded in the vehicle 10.

[0036] The communication module 8 enables the vehicle 10 to transmit and / or receive data from a database, a remote server, one or more connected devices such as a mobile phone, or from a vehicle itself equipped with a communication module 8 and a system according to the invention. Such communication is established via a low-frequency or high-frequency wireless link. For example, it could be a wireless link based on cellular, Wi-Fi, or Bluetooth technologies. As further explained below, the communication module 8 can be used to extract data relating to the external environment, including weather or traffic conditions, or to transmit information about the route traveled in order to update mapping data.

[0037] The detection system 1 may, in addition, include a memory element 6 or a storage space constituting a recording medium readable by a computer or by the calculator 21 of the processing unit 2 comprising instructions which, when executed by the computer or the calculator 21, cause it to implement the process as described below.

[0038] An embodiment of the method for detecting a type of road traveled by a vehicle 10 in motion is described below with reference to Figures 2 to 4. Generally, the detection method 100 according to the invention comprises a determination phase PI of a type of road traveled by the vehicle and, prior to the execution of the determination phase PI, at least one supervised learning phase P2, specific to the equipped vehicle 10, associating in particular a level of comfort perceived by a user with different classes of road types, including poor roads. Such a learning phase P2 can also be defined as relating to a training method for a detection method and / or a detection system according to the invention.

[0039] Throughout the description below, a road type, or a road type class, may, but is not limited to, relate to: - one or more types of poor roads likely to cause pumping movements of the body when the vehicle 10 is moving on a road where the spectral amplitude density of vertical stresses is maximum at low frequencies; - one or more types of poor roads likely to generate wheel jolts or bounces when the vehicle 10 is moving on a road where the spectral amplitude density of vertical stresses is maximum at high frequencies; - paved roads or mountain paths; - roads with poor quality surfaces, for example due to bad weather such as precipitation, frost, or wear and tear; - roads including obstacles, such as speed bumps, level crossings, potholes or ridges, depending on their impact on comfort, for example depending on their dimensions, likely to cause vibrations or shocks in the vehicle 10, in particular presenting a high frequency of said obstacles, i.e. a degraded road with a short distance between obstacles.

[0040] The PI determination phase includes at least one acquisition step El, either in real time or for a predetermined duration, of data relating to a route traveled by the vehicle 10. According to an example of an embodiment not detailed in [Fig. 2], the PI determination phase may include a plurality of acquisition steps relating to similar or different types of data. For example, this data may relate to one or more characteristics of the route traveled and / or the vehicle. Alternatively, the acquisition step El may include the acquisition of different, distinct types of data. The acquisition step El is implemented via the detection module 3 as described previously.

[0041] In order to classify the route traveled into at least one class of For a predetermined type of road, such as a poor road type, the acquisition step El can be carried out in such a way as to include a detection and / or estimation substep El i of at least one feature relating to a portion of the road being traveled located upstream of the vehicle 10 and / or a measurement substep E12 of at least one feature relating to the vehicle 10 while driving on said road being traveled. In other words, the acquisition step El can implement data acquisition in such a way as to anticipate features relating to a future road portion included in the road being traveled and / or in such a way as to react to a road portion on which the vehicle 10 is moving, that is to say, with which at least one of the wheels of the vehicle 10 is in contact.

[0042] The detected and / or estimated data El 1 based on the portion of the road traveled upstream of the vehicle 10 may, for example, relate to characteristics such as the height of a road profile relative to the vehicle, in particular to a wheel of the vehicle, or any other information from road images. For example, such data may be obtained by means of an optical detection device, such as a camera.

[0043] Optionally, but preferably, the detected and / or estimated data El 1 based on the road segment upstream of vehicle 10 can be associated with a confidence index a representing the degree of reliability of the detection and / or estimation performed. Specifically, such a confidence index a can be defined based on detected weather or road traffic conditions and / or the type of road surface, which may affect the detection and / or estimation. For example, the confidence index a is defined according to whether the road is paved, covered with snow, ice, leaves, or sand. Advantageously, the use of a confidence index makes it possible to discard data defined as unreliable, which could bias the road type classification, for example, by discarding data with a confidence index below a threshold a. minpredefined. For example, the confidence index can be defined by the processing unit 2 and / or by the detection module 3.

[0044] The measured data E12 may, but are not limited to, be obtained by at least one sensor 32 and relate to characteristics such as suspension or wheel travel, for example, wheel center travel, wheel speed, longitudinal speed of the vehicle 10, and / or acceleration of the vehicle 10. At least one sensor 32 may be included in the sensing module 3. Alternatively, the sensor 32 may be an on-board sensor specific to one of the auxiliary systems 5 equipping the vehicle 10 and configured to transmit data to said sensing module 3. Such sensors 32 may include an accelerometer, a travel sensor, a deformation sensor associated with one of the wheels, and / or a sensor of pressure.

[0045] Preferably, the method according to the invention is at least carried out in anticipation, by estimation and / or detection El 1 of data relating to the portion of road located upstream of the vehicle 10. As further explained below, with reference to [Fig.4], the implementation of a measurement substep, in addition to a detection and / or estimation substep El i, can improve the detection and classification of the road traveled by means of a validation step.

[0046] Regardless of the type of data acquisition performed, i.e., anticipatory or reactive, the detection and / or estimation (E1) of data, on the one hand, and the measurement of data (E12), on the other hand, are each carried out at at least one point on the road. Preferably, data acquisition is performed at at least two points, for example, a point located on two lines representing a virtual trajectory of the vehicle's front wheels.

[0047] Figure 3 illustrates a specific implementation of the acquisition step 1E1, in which data acquisition is performed in advance using an optical detection method to allow for the estimation of the profile heights of the road section upstream of vehicle 10 through image processing. In the illustrated example, but not limited to, data is collected at a plurality of points positioned in front of each of the front wheels of vehicle 10. Specifically, data is collected at ten points for each front wheel. Preferably, these points are spaced at a fixed frequency for each front wheel so that data acquisition is not negatively affected by the longitudinal speed of vehicle 10. The data for each of the twenty measurement points, ten for each front wheel, are transmitted to the processing unit 2.

[0048] For each of the aforementioned points, the estimated value of the profile height of the road segment is associated with a confidence index a x of said estimate. As explained above, such an index can be defined by the detection module 3 and / or the processing unit 2. This results in a dataset comprising forty incoming signals that represent the heights and confidence indices. x from the different points of the section of road located upstream of the two front wheels of vehicle 10.

[0049] Such a principle applies mutatis mutandis to the "by reaction" operation described above, according to which the data are directly obtained by E12 measurements representative of a set formed by the route traveled and the vehicle 10, with the difference that no confidence index is then included.

[0050] The processing unit 2 then performs an estimation step E2 of at least one probability rate of the route traveled belonging to at least one type class of A predefined route, including a bad route, is determined based on the acquired data. As explained previously, when a confidence index is associated with the acquired data, such a probability rate is also defined based on that data.

[0051] Membership rate estimation can be performed by any type of network capable of carrying out a classification task on time series data that includes a prior supervised P2 learning phase. The network receives as input the data acquired during the acquisition phase and provides as output estimates of the membership rate of the route traveled to different predefined road type classes, corresponding to a probability rate that the route traveled is similar to, or of the same type as, a given road type class.In other words, such a network is first trained during a supervised P2 learning phase, as further detailed below, in order to perform the identification of different classes of route types on the basis of different data types and / or data ranges, and then can subsequently be run by processing unit 2 during process implementation in order to perform the classification of the route traveled into at least one of the classes identified during the learning phase.

[0052] In one example, the probability rates of belonging to different classes of road types can be calculated by a neural network, specifically a convolutional neural network (CNN). This type of model can be used to perform supervised classification of time series data. This approach is advantageously simple to implement and ensures time-translation invariance of the processing. In another, non-limiting example, the probability rates of belonging to different classes of road types can be calculated by a recurrent neural network (RNN).

[0053] Based on the probability rates of belonging, processing unit 2 implements a classification step E3 of the road traveled within a road type class and extracts at least one level of comfort experienced by a user previously associated, for example during at least one supervised learning phase P2 as described below, with each road type class according to at least one vehicle-specific piece of equipment 10 capable of influencing the comfort experienced by a user while driving. "Equipment" is understood, by way of non-limiting examples, to include features likely to affect the level of comfort experienced by a user, such as a suspension system or tires.

[0054] For example, the route traveled is classified into a predefined route type class based on the highest estimated probability rate obtained. Specifically, the E3 classification step can only be implemented when the estimated probability rate is greater than a predetermined threshold of 50%, or even 60%. of 75% in order to ensure the accuracy of the classification. The processing unit 2 then extracts at least one level of comfort felt by a user associated with the detected road type class, i.e. previously defined, by calculation implemented by the processing unit 2 or by manual intervention by a user or technician, and recorded on the memory elements 6 of the detection system 1 and / or the vehicle 10.

[0055] According to a non-limiting example of implementation, applied to a vehicle 10 in a driving situation as illustrated in [Fig. 3], at each calculation step the processing unit 2 forms, as input to the network, specifically the neural network, a sliding data window. In this case, such a window includes the last forty values ​​relating to road profile heights and confidence indices. The neural network performs an inference to obtain a probability distribution for the different road classes, including poor road conditions, and then the road type is determined by selecting the class with the highest probability rate.

[0056] The network used to extract features, or characteristics, relating to route type classes based on acquired raw data time series and to define probability rates can be defined, in a non-limiting way, by a basic block comprising: - a layer, or filter, of Convolution ID; then - a ReLU layer, for Rectified Linear Units, corresponding to a correction layer including the ReLU activation function; then - a MaxPooling layer, corresponding to a pooling layer; then - a batch normalization layer, or "Batch Normalization".

[0057] Such a block can be repeated several times, for example twice, to obtain a sequence of three blocks. According to a non-limiting embodiment, these three basic blocks can be assembled with different filter sizes and used as input for a fully connected layer to produce a probability distribution of route type class at the output. For example, the filter sizes used in the convolution layers can be defined as follows when the three basic blocks are executed successively: 128, 256, 128. Similarly, the kernel sizes of the convolutions can be: 5, 3, 3. For the execution of the convolutions of each block, the padding applied to the input allows the output to maintain the same time dimension, and the step size used is preferably 1.

[0058] Following the implementation of the basic blocks, a softmax layer can be implemented. It provides a probability rate of route membership for the route traveled. each of the possible classes, the sum of these probability rates amounting to 1. In a conventional manner, the softmax function takes as input a vector of dimension K: z=(zi ...z K ) and is defined according to:

[0060] In this example implementation, the network, specifically the convolutional neural network, receives as input a 40x40 matrix containing a time window with 40 time steps representing the 40 data signals described previously, relating to the profile height of the road segment upstream of vehicle 10 and the confidence indices. The network outputs a distribution of the probability rates for the road belonging to each of the predefined classes of possible road types established during the P2 training phase. The E3 classification step, as described above, is then executed, for example, so that the final classification of the road traveled corresponds to the class for which the probability rate obtained as the network output is maximum.

[0061] The training of the network, for example a convolutional neural network, is carried out, as described above, during at least one supervised P2 learning phase, implemented prior to the PI determination phase as described above. "Supervised" means that such a P2 learning phase requires, at least partially, human intervention.

[0062] The P2 learning phase aims, on the one hand, to train the network, in other words the model, so that it is able to identify a predefined route type class corresponding to a route traveled when the PI determination phase is executed, and, on the other hand, to train the model so that it is able to associate with the different route type classes a level of comfort felt by the user.

[0063] The P2 learning phase can be implemented via a computer 21, a human-machine interface, and / or a memory element. Once trained, the model can easily be implemented in a computer 21 of the vehicle 10, for example in the processing unit 2, during the implementation of the PL determination phase.

[0064] Training the network to identify a route type includes an "inference" phase, carried out according to a principle similar to that described in the PI determination phase, using data for which the route type class is previously known, for example, through manual, i.e., supervised, labeling of a training database; therefore, the previous description applies mutatis mutandis. To build the training database, data records relating to different route types are created on the based on one or more types of data relating to characteristics specific to the route traveled and / or the vehicle, as previously explained with reference to acquisition step El.

[0065] These records, for which the model provides the desired processing result, are submitted to the model to train it to identify—that is, to perform the estimation and classification steps described above—a road type based on detected and / or estimated data (i.e., in anticipation) and / or measured data (i.e., in reaction). The different road type classes are thus defined according to at least one data type, specifically based on defined value ranges for at least one data type, such as road profile height. Then, based on the class or the distribution of probability rates for the road belonging to each of the predefined possible road type classes determined by the model, an estimate of the model's error is calculated.

[0066] In the case of networks, particularly convolutional neural networks, a loss or cost function, such as a cross-entropy function, can be used. Optionally, but preferably, weighting can be applied to such a loss function to prevent overfitting, or to enable artificially accelerated learning if the recordings of the different route types are not evenly distributed in terms of the number of examples per class.

[0067] Network training also aims to identify and / or associate at least one estimated comfort level, representative of user comfort and the perceived experience of the road traveled, with the different road type classes. Such training can be performed based on at least one criterion, which can be objective or subjective. Advantageously, training can be carried out based on a plurality of objective and / or subjective criteria. The different road type classes can be associated with at least one criterion, in particular with at least one defined range of values ​​for the at least one criterion considered.

[0068] For example, an "objective criterion" is defined as criteria based on values ​​measured, estimated, and / or calculated by detection devices and / or sensors 32 installed in the vehicle 10. For example, an objective criterion may relate to transmissibility, that is, the ratio between the displacement of the sprung mass and the road load, which allows for the evaluation of the vibration isolation characteristic of a given suspension. A criterion may also be defined based on the wheel travel relative to the body when passing disturbances or obstacles, and in particular the reaching of compression stops or relaxations that severely penalize comfort, or a criterion on the effective value, or RMS, of the vertical acceleration at the center of gravity of the car.

[0069] When training is based on objective criteria, it can advantageously be carried out automatically and is defined specifically according to the equipment specific to vehicle 10, or to a model of vehicle 10, so that the assessment of the perceived comfort associated with the type of road identified is defined appropriately in relation to the vehicle 10 equipped with the detection system and its equipment, such as its suspension system, and not in a generic way.

[0070] Conversely, a "subjective criterion" refers to learning that is carried out in a supervised manner by a user or expert who can, for example, assign a predefined comfort level to each of the different road type classes based on the vehicle's equipment, such as the suspension system. The user or expert is then able to subjectively classify the type of road encountered based on their perception of the vehicle's response and the resulting passenger comfort when driving on that road.

[0071] Similar to what has been explained above, during comfort-related training, a loss function and, optionally, a weighting function, can be implemented.

[0072] The P2 learning phase can thus be advantageously adapted according to the need and the different types of roads can thus be classified by the user or the expert according to what is desirable to detect, i.e. according to the equipment of the vehicle 10 and / or according to the auxiliary systems 5 that can be controlled, in particular the suspension or shock absorber system, and according to a desired level of comfort defined by the user or the expert.

[0073] During this P2 learning phase, the quality and consistency of the training are essential to ensure accurate and efficient identification of the different types. In this application, the method according to the invention advantageously achieves an overall accuracy of approximately 98% with a weighted Fl score of 97%.

[0074] In the case described above, at least one learning phase P2 is implemented prior to the execution of the PL determination phase. A plurality of detection phases can be executed following such a learning phase P2, as illustrated by the dotted box P2'. For example, the learning phase P2 can be carried out prior to the vehicle being put into service or during vehicle maintenance or servicing.

[0075] Optionally, the method according to the invention may include an additional learning phase P2', implemented after the execution of the PI determination phase of a road type. Such an additional learning phase P2 is executed similarly to that described above and may be, for example, implemented by a user to refine or adapt the training of identification and / or comfort level associated with different types of road to their preferences or needs.

[0076] Thus, as illustrated in [Fig. 2], under driving conditions, once the road type class specific to the road being traveled has been identified—that is, at the end of the E3 classification step of the PI determination phase—the processing unit can extract, for example from a memory element, a comfort level associated with the identified road type class. This level is then representative of the comfort level felt, in real time, by the user as the vehicle travels on the road. The detection system 1 can then initiate, via the control module 4, an adjustment step E4 of at least one auxiliary system 5 equipping the vehicle 10 in order to correct the comfort level felt by the user.

[0077] The adjustment step E4 can be implemented when the comfort level associated with the identified road type is below a predefined threshold for an appropriate comfort level, for example, when the determined road type is a poor road associated with a comfort level deemed unsuitable during the learning phase. Such a threshold can be defined beforehand by an expert and / or adjusted by a user. Alternatively, the adjustment step can be implemented when the road type class identified during the PI determination phase falls within a subset of classes previously identified, for example, during the learning phase, as requiring correction of perceived comfort. For example, such a road type could be a "poor road."

[0078] Here, "correction" means that when the estimated level of perceived comfort is below the required threshold, it is necessary to adapt at least one system (see Annex 5) to increase user comfort. This can be achieved by adapting the suspension system to make it softer or stiffer, or by adjusting the vehicle's longitudinal speed, for example, by reducing it.

[0079] For example, when controlling a variable suspension or shock absorber system, the damping force can be modified, notably increased, in various proportions when an obstacle, speed bump, or low-frequency road imperfection is detected, with these proportions being adapted to the type of obstacle or road imperfection detected. Conversely, the damping force can be decreased when passing over a high-frequency road imperfection or experiencing a minor impact.

[0080] As previously explained, the Annex 5 system in question may be a variable suspension system 51, a longitudinal speed control system, and / or a driver assistance system 52. It should be noted that driver assistance functions may include, in particular, collision warnings, driver information or alerts, emergency braking, and emergency steering. The correction made may then affect one or more of said systems.

[0081] As mentioned above, advantageously, road type classes, particularly for poor roads, can be adapted according to the intended use of the vehicle and / or the detection system, i.e., for example, according to the Annex 5 system referred to and / or according to a vehicle operating mode 10 from among commonly available modes such as a "sport" or "comfort" driving mode. The method according to the invention thus enables, at the moment the road type of the road being traveled is detected, the triggering of a control strategy for the auxiliary systems, such as the variable suspension system 51, specifically adapted to the road type determined during the PI determination phase and to the vehicle 10 equipped in order to ensure optimal user comfort.

[0082] Additionally or alternatively, once the road type class specific to the road being traveled has been identified, the detection system 1 can initiate an update step E5 of the mapping data installed on the vehicle 10 to populate the road type, including poor road conditions, determined during the PI determination phase. In this respect, during the PI determination phase, for example during the acquisition step E1 and / or the classification step E3, the acquired data can be combined with GPS data. Optionally, this data can also be associated with a timestamp, particularly to ensure appropriate mapping when the determined road type has a temporary component, for example, due to temporary obstacles or the presence of inclement weather such as rain, snow, or ice that may affect the road class.

[0083] The road type class determined for the route traveled, along with GPS data and optionally a timestamp, can then be transmitted to other vehicles, a database, and / or remote servers via the communication module 8 to make this data accessible to other users, including other users of a similar or identical vehicle model 10. For example, it is then possible to associate, in real time, a GPS position and optionally a timestamp corresponding to an entry and exit point on the detected road type. This data can also optionally be associated with the direction of travel of vehicle 10 on the route traveled and / or the longitudinal speed of vehicle 10.

[0084] Once this information has been transmitted to remote databases or servers, it may be subjected to additional processing, including filtering, grouping and / or validating the information before making it available to other users.

[0085] Figure 4 illustrates a second embodiment of the method according to the invention, in the PI determination phase includes, subsequent to the estimation step E2, a validation step E6 of the estimated probability rate based on measurements, by at least one sensor 32, of at least one characteristic relating to the vehicle 10 while driving on the road being traveled. In such an execution mode, the estimation step E2 as described above can be implemented in anticipation, based on estimated and / or detected data relating to a portion of the road being traveled located upstream of the vehicle 10, or in reaction, based on measured data.

[0086] The validation step E6 can be executed simultaneously and / or subsequently to the estimation step E2. The validation step E6 is executed in a similar manner to that described above with regard to the processing of measured data during the estimation step E2.

[0087] The E3 classification step of the route within one of the predefined road type classes can then be executed conditionally only when the results of the E2 estimation step and the E6 validation step allow for the identification of the same road type. In other words, the E3 classification step is implemented only when the probability rates of belonging to a road type class obtained during the E2 estimation step and the E6 validation step are such that they allow for the identification of the same road type class.

[0088] Alternatively, the validation step E6 can be performed in parallel with the estimation step E2 and the classification step E3, with the validation step E6 reproducing said estimation steps E2 and classification steps E3 on the basis of data measured for the purpose of validation. The method according to the invention then comprises two distinct classification steps, a primary one E3, as described with reference to [Fig. 2], and the other included in the validation step E6. A classification of the route traveled is then recorded, for example on a memory element 6, and used for the adjustment E4 of an auxiliary system only when the two classification steps result in the identification of the same road type class. Such a principle advantageously allows for further optimization of the accuracy of the method according to the invention.

[0089] The present invention thus proposes a method for detecting the type of road traveled by a vehicle, combining a phase for determining the road type and a phase for learning to identify a class of road type, incorporating an evaluation of the road comfort experienced by a user for a vehicle of a given model and equipment. The method according to the invention can advantageously be transposed to any type of vehicle at a lower cost, as the execution of the learning phase makes it possible to obtain a detection method adapted to the characteristics of the vehicle, in particular its model and / or equipment likely to impact road comfort. The invention thus makes it possible to obtain a performance of The detection of road types, particularly poor roads, is refined and adapted to the vehicle being equipped, while ensuring the invention's portability to a wide range of vehicles. The method according to the invention thus allows for more tailored control of the vehicle's auxiliary systems related to user comfort.

[0090] The present invention cannot, however, be limited to the means and configurations described and illustrated herein and it also extends to any equivalent means or configuration and to any technically operative combination of such means insofar as they ultimately fulfill the functionalities described and illustrated in this document.

Claims

Demands

1. A method for detecting (100) the type of road traveled by a vehicle (10), in particular a vehicle (10) equipped with a variable suspension system (51), the method being characterized in that it comprises a determination phase (PI) of the type of road traveled comprising: - at least one acquisition step (El), in real time or for a predetermined duration, of data relating to the vehicle and / or a route travelled by the vehicle (10) implemented through a detection module (3) comprising at least one sensor (32) and / or at least one detection means (31), such as a camera; - an estimation step (E2) of a probability rate of belonging of the road traveled to at least one class of road type, including bad roads, predefined according to said acquired data; - a classification step (E3) of the road travelled into one of the road type classes according to the estimated probability rates of belonging and extraction of at least one level of comfort felt by a user previously associated with each road type class according to at least one vehicle-specific equipment (10) capable of influencing the comfort felt by a user during driving.

2. A detection method (100) according to the preceding claim, further comprising, prior to the execution of the determination phase (PI), at least one supervised learning phase (P2), specific to the vehicle (10) equipped, of the identification of at least one predefined road type class and of the association of at least one level of comfort felt by a user to each road type class according to at least one vehicle-specific equipment (10) capable of influencing the comfort felt by a user during the driving phase, the learning phase (P2) being implemented via a computer (21), a human-machine interface and / or a memory element.

3. A detection method (100) according to any one of the preceding claims, further comprising an adjustment step (E4) of at least one auxiliary system (5) equipping the vehicle (10), in particular a variable suspension system (51), so as to correct a level of comfort perceived by the user, such a step being implemented when the level of comfort associated with a determined road type class during of the determination phase (PI) is below a predefined adapted comfort level threshold and / or when the determined road type class is included in a predefined subset of classes identified as requiring the execution of an adjustment step (E4).

4. A detection method (100) according to any one of the preceding claims, further comprising an update step (E5) of a map equipping the vehicle (10) so as to provide information on a specific road type class.

5. A detection method (100) according to any one of the preceding claims, wherein the classification step (E3) is implemented only when the estimated probability rate is greater than a predetermined threshold of 60%.

6. A detection method (100) according to any one of the preceding claims, wherein the estimation step (E2) is executed by a network capable of processing time series of data and of being trained during a learning phase (P2), in particular by a convolutional neural network.

7. A detection method (100) according to any one of the preceding claims, wherein the data acquisition step (El) is implemented by - detection and / or estimation of at least one characteristic relating to a portion of the road travelled located upstream of the vehicle (10), in particular a height of a road profile; and / or - measurement by at least one sensor (32) of at least one characteristic relating to the vehicle (10) in driving situation on said road travelled, in particular suspension travel, wheel speed, longitudinal speed of the vehicle (10) and / or acceleration of the vehicle (10).

8. A detection method (100) according to any one of the preceding claims, wherein the determination phase (PI) comprises, subsequent to the estimation step (E2), a validation step (E6) of the estimated probability rate, the validation step (E6) being carried out on the basis of measurements, by at least one sensor (32), of at least one characteristic relating to the vehicle (10) in a driving situation on said road traveled, in particular a suspension travel, a wheel speed, a longitudinal speed of the vehicle (10) and / or an acceleration of the vehicle (10).

9. A detection system (1) for a type of road traveled by a vehicle (10) in a driving situation, the system comprising hardware and / or software elements implementing the method according to any one of claims 1 to 8, the hardware elements comprising at least one data processing unit (2), a detection module (3) for at least one characteristic relating to a road traveled and a control module (4) for at least one auxiliary system (5) equipping the vehicle (10), in particular a variable suspension system (51).

10. Motor vehicle (10) comprising at least one auxiliary system (5) capable of affecting a user's perceived level of comfort, such as a variable suspension system (51), and a detection system (1) according to claim 9, capable of controlling said auxiliary system (5).

11. Product computer program comprising program code instructions recorded on a computer-readable medium to carry out the steps of the process according to any one of claims 1 to 8 when said program is run on a computer.

Images