FR3138534A1 - Method and device for controlling a vehicle driving assistance system based on a level of wear of the brake pads - Google Patents

Abstract

The present invention relates to a method and a device for monitoring a first vehicle (10) based on a wear category or a wear level of the brake pads of the first vehicle (10). The wear category is determined by classification of data representative of the dynamic behavior of the vehicle (10), data representative of the environment (1) of the first vehicle (10) and data representative of a change of the brake pads on the basis of a model learned during a learning phase from dynamic behavior data of a set of second vehicles (11), data from their environment and data from changes in their brake pads. Figure for abstract: Figure 1

Description

Method and device for controlling a vehicle driving assistance system based on a level of wear of the brake pads

The present invention relates to methods and devices for controlling a vehicle, in particular an autonomous or semi-autonomous vehicle, as a function of the wear of the brake pads of the vehicle. The present invention also relates to methods and devices for controlling at least one driver assistance system, called an ADAS (Advanced Driver-Assistance System) system, of a vehicle. The present invention also relates to a method and device for determining or predicting a level of wear or a category of wear of the brake pads of a vehicle.

Technological background

Contemporary vehicles are made up of a large number of components or organs, each susceptible to breakdown, defect or wear.

When the driver of a vehicle is faced with wear or a defect in a vehicle component such as the vehicle's brake pads, the vehicle's behavior may be affected, posing a risk to the safety of the vehicle and its occupants.

One of the problems encountered by the driver is that the defect or wear of the affected part is not easily detectable by the driver. To establish the diagnosis, it is often necessary to have the vehicle checked by a professional.

To limit the risks of breakdown or defect, to measure the level of wear and prevent associated problems, preventive maintenance operations of vehicles are planned. However, these maintenance operations do not guarantee that a vehicle component will encounter a defect or suffer an unforeseen breakdown or wear after the fact.

Such monitoring through the implementation of regular preventive maintenance operations is also synonymous with costs and inconvenience caused by the immobilization of the vehicle during the maintenance action.

Summary of the present invention

An object of the present invention is to solve at least one of the problems of the technological background described above.

Another object of the present invention is to prevent the risks associated with wear or failure of the brake pads of a vehicle.

Another object of the present invention is to improve the prediction of wear of the brake pads of a vehicle.

According to a first aspect, the present invention relates to a method of controlling a vehicle, the method being implemented by at least one processor and comprising the following steps:

- reception of first data representative of dynamic behavior of the vehicle from a set of sensors embedded in the vehicle, the set of sensors comprising at least one sensor;

- reception of second data representative of an environment in which the vehicle is circulating;

- receipt of third data representative of a change of a set of brake pads of the vehicle;

- determination of fourth data representative of a level of wear of the set of brake pads of the vehicle as a function of the first and second data and determination of fifth data representative of a driving mode of the vehicle as a function of the first data;

- determination of information representative of a wear category of the set of brake pads by classification of a set of data comprising the third data, the fourth data and the fifth data from a classification model learned in a so-called learning phase, the wear category being associated with a class of a set of classes at the output of the classification model;

- vehicle control based on information representative of a wear category.

The use of data representative of the dynamic behavior of the vehicle and on vehicle environment data acquired in real time to cross them with a previously learned classification model makes it possible to determine or predict a level of wear of one or more brake pads of the vehicle in real time. Determining the wear of the vehicle's brake pads makes it possible to adapt the control of the vehicle to reduce the risks associated with the level of wear, for example by controlling one or more ADAS systems of the vehicle to adapt the driving mode to the level of wear and/or by alerting the driver.

This helps to improve the safety of the vehicle and its passengers and/or to improve the life of the brake pads by adapting the vehicle's behavior to the determined level of wear.

According to one variant, the vehicle control comprises at least one of the following steps:

- control of at least one driving assistance system, known as ADAS system, of the vehicle according to the wear category; and/or

- rendering of a first message representative of the wear category via a human-machine interface embedded in the vehicle; and/or

- rendering of a second message representative of at least one recommendation based on the wear category via the human-machine interface on board the vehicle; and/or

- determination of a route for the vehicle based on the wear category.

According to another variant, the ADAS system belongs to a set of ADAS systems comprising:

- an electronic stability control system, known as the ESC system;

- an anti-lock braking system, known as the ABS system;

- a speed regulation system;

- a trajectory control system;

- a route control system;

- a system for controlling at least one component of a vehicle's powertrain.

According to yet another variant, the first data belongs to a data set comprising:

- representative acceleration data;

- representative yaw rate data;

- representative speed data;

- representative steering angle data;

- data representative of the use of a vehicle brake pedal;

- data representative of the use of a vehicle accelerator pedal;

- representative braking distance data;

- representative deceleration data;

- data representative of the sound generated during braking.

According to an additional variant, the second data belong to a data set comprising:

- data representative of meteorological conditions;

- data representative of the condition and/or level of road maintenance;

- data representative of road type;

- representative slope or elevation data;

- data representing the location of the vehicle;

- map data.

In another variant, the classification is implemented via a k-means partitioning method.

In an additional variant, the classification is implemented by a neural network.

According to an additional variant, the neural network is a network of the recurrent network type with short and long term memory, called LSTM network.

According to a second aspect, the present invention relates to a device for controlling a vehicle, the device comprising a memory associated with a processor configured for implementing the steps of the method according to the first aspect of the present invention.

According to a third aspect, the present invention relates to a vehicle, for example of the automobile type, comprising a device as described above according to the second aspect of the present invention.

According to a fourth aspect, the present invention relates to a computer program which comprises instructions adapted for the execution of the steps of the method according to the first aspect of the present invention, in particular when the computer program is executed by at least one processor.

Such a computer program may use any programming language, and may be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form.

According to a fifth aspect, the present invention relates to a computer-readable recording medium on which is recorded a computer program comprising instructions for carrying out the steps of the method according to the first aspect of the present invention.

On the one hand, the recording medium may be any entity or device capable of storing the program. For example, the medium may include a storage medium, such as a ROM memory, a CD-ROM or a microelectronic circuit type ROM memory, or a magnetic recording medium or a hard disk.

On the other hand, this recording medium may also be a transmissible medium such as an electrical or optical signal, such a signal being able to be conveyed via an electrical or optical cable, by conventional or hertzian radio or by self-directed laser beam or by other means. The computer program according to the present invention may in particular be downloaded from a network such as the Internet.

Alternatively, the recording medium may be an integrated circuit in which the computer program is incorporated, the integrated circuit being adapted to perform or to be used in performing the method in question.

Brief description of the figures

Other characteristics and advantages of the present invention will emerge from the description of the particular and non-limiting exemplary embodiments of the present invention below, with reference to the appended figures 1 to 5, in which:

schematically illustrates a communication environment of a vehicle, according to a particular exemplary embodiment of the present invention;

schematically illustrates a grouping or partitioning of data obtained in the environment of the in a predetermined number of groups, according to a particular and non-limiting exemplary embodiment of the present invention;

schematically illustrates a short-term and long-term memory cell according to a particular and non-limiting exemplary embodiment of the present invention;

schematically illustrates a device 4 configured to control at least one ADAS system of a vehicle of the , according to a particular and non-limiting exemplary embodiment of the present invention; and

illustrates a flowchart of the different steps of a method for controlling at least one ADAS system of a vehicle of the , according to a particular and non-limiting exemplary embodiment of the present invention.

Description of examples of implementation

A method and device for controlling a vehicle will now be described in the following with joint reference to Figures 1 to 5. The same elements are identified with the same reference signs throughout the description which follows.

According to a particular and non-limiting example of embodiment of the invention, the control of a vehicle, for example an autonomous or semi-autonomous vehicle, i.e. a vehicle traveling under the control, at least partially, of one or more ADAS systems, is implemented according to the wear of the brake pads. The wear of the brake pads is determined by categorizing it in real time by classifying a set of data.

For this purpose, first data representative of the dynamic behavior of the vehicle are received or collected from a set of sensors embedded in the vehicle. Second data representative of the environment in which the vehicle is moving are also received or collected. Third data representative of the change of the brake pads, for example on a specific date, are also received or collected.

The first and/or second data are processed to estimate a level of wear of the brake pads and to determine a driving mode or style of the vehicle (for example corresponding to a so-called sporty mode, a so-called calm mode, a so-called regular mode, etc.).

The set of data used to determine the category representing the wear of the brake pads advantageously comprises the third data, fourth data representative of the estimation of the wear of the brake pads and fifth data representative of the determined driving mode.

Worn brake pads result in longer braking distances and reduced vehicle braking efficiency, which reduces vehicle safety.

The classification is for example implemented in a machine learning method, for example implemented with a neural network. Such a method makes it possible to estimate brake pad wear more reliably and in real time based on current data from the vehicle and its environment, based on a model learned from the same data from a set of other vehicles.

Such an advantage is obtained by determining, during a learning phase, the correlation between the data collected from a set of vehicles and representing the actual use made of these vehicles with a set of levels or categories of wear of the brake pads of these same vehicles.

Predicting the level of brake pad wear makes it possible to anticipate a problem and potentially correct it before it appears by controlling the ADAS system(s) controlling the vehicle, improving the prevention of the defect or a risk for the vehicle and its passengers.

There schematically illustrates a communication environment 1 of a first vehicle 10, according to a particular and non-limiting exemplary embodiment of the present invention.

The first vehicle 10 corresponds for example to a vehicle with a thermal engine, with electric motor(s) or even a hybrid vehicle with a thermal engine and one or more electric motors. The vehicle 10 thus corresponds for example to a land vehicle, for example an automobile, a truck, a coach, a motorcycle.

The first vehicle 10 also advantageously carries a communication system configured to communicate with one or more remote devices 101 via a wireless communication network infrastructure. The remote device 101 advantageously corresponds to a device configured to process data, for example data stored in the memory of the remote device 101 and/or data received from the vehicle 10. The remote device 101 corresponds for example to a server of the “cloud” 100.

The communication system of the first vehicle 10 comprises, for example, one or more communication antennas connected to a telematic control unit, called a TCU (Telematic Control Unit), itself connected to one or more computers of the on-board system of the first vehicle 10. The antenna(s), the TCU and the computer(s) form, for example, a multiplexed architecture for carrying out various services useful for the proper operation of the first vehicle 10 and for assisting the driver and/or passengers of the first vehicle 10 in controlling the first vehicle 10 and/or for establishing a diagnosis on the operation of one or more components of the first vehicle 10. The computer(s) and the TCU communicate and exchange data with each other via one or more computer buses, for example a communication bus of the CAN (Controller Area Network) or CAN FD (Controller Area Network Flexible Data-Rate) type. "Flexible Data Rate Controller Area Network"), FlexRay (according to ISO 17458) or Ethernet (according to ISO/IEC 802-3).

The mobile communication infrastructure enabling wireless data communication between the first vehicle 10 and the remote device 101 comprises, for example, one or more communication devices 110 of the relay antenna (cellular network) or roadside unit, known as UBR, type. In a communication mode using such a network architecture, the data are, for example, transmitted by the first vehicle 10 to the remote device 101 of the “cloud” 100 via a relay antenna 110 (the antenna 110 being, for example, connected to the “cloud” 100 via a wired link).

The wireless communication system allowing the exchange of data between the first vehicle 10 and the remote device 101 corresponds for example to:

- a vehicle-to-infrastructure (V2I) communication system, for example based on the 3GPP LTE-V or IEEE 802.11p ITS G5 standards; or

- a cellular network type communication system, for example an LTE (Long-Term Evolution) or LTE-Advanced (also called LTE 3G, 4G or 5G) type network; or

- a Wifi type communication system according to IEEE 802.11, for example according to IEEE 802.11n or IEEE 802.11ac.

The first vehicle 10 corresponds for example to a vehicle traveling under the total supervision of a driver or traveling in an autonomous or semi-autonomous mode. The first vehicle 10 travels according to a level of autonomy equal to 0 or according to a level of autonomy ranging from 1 to 5 for example, according to the scale defined by the American federal agency which has established 5 levels of autonomy ranging from 1 to 5, level 0 corresponding to a vehicle having no autonomy, the driving of which is under the total supervision of the driver, level 1 corresponding to a vehicle with a minimum level of autonomy, the driving of which is under the supervision of the driver with minimal assistance from an ADAS system, and level 5 corresponding to a completely autonomous vehicle.

The 5 levels of autonomy of the classification of the federal agency responsible for road safety are:

- level 0: no automation, the vehicle driver has full control over the vehicle's main functions (engine, accelerator, steering, brakes);

- level 1: driver assistance, automation is active for certain vehicle functions, the driver retaining overall control over the vehicle's driving; cruise control is part of this level, as are other aids such as ABS (anti-lock braking system) or ESP (programmed electro-stabilizer);

- level 2: automation of combined functions, the control of at least two main functions is combined in the automation to replace the driver in certain situations; for example, adaptive cruise control combined with lane centering allows a vehicle to be classified as level 2, as does automatic parking assistance (from the English “Park assist”);

- level 3: limited autonomous driving, the driver can hand over complete control of the vehicle to the automated system which will then be in charge of critical safety functions; autonomous driving can however only take place in certain specific environmental and traffic conditions (only on motorways for example);

- level 4: fully autonomous driving under certain conditions, the vehicle is designed to ensure all critical safety functions on its own over a complete journey; the driver provides a destination or navigation instructions but is not required to make himself available to take back control of the vehicle;

- level 5: completely autonomous driving without driver assistance in all circumstances.

According to a particular exemplary embodiment, the first vehicle 10 travels in a semi-autonomous or autonomous mode, that is to say with a level of autonomy greater than or equal to 2 according to the classification above.

The first vehicle 10 advantageously incorporates one or more ADAS systems for driving in autonomous or semi-autonomous mode. Examples of such ADAS systems are provided for illustration purposes in a non-exhaustive manner below:

- anti-lock braking system, known as ABS (from the German “Antiblockiersystem”);

- electronic stability control system, known as ESC (Electronic Stability Control), DSC (Dynamic Stability Control) or ESP (Electronic Stability Program);

- lane keeping assistance system such as a lane keeping assistance system, known as LKA (Lane-Keeping Assist) system;

- cruise control system such as the adaptive cruise control system, known as ACC (from the English “Adaptive Cruise Control”) and/or the predictive cruise control system, known as PCC (from the English “Predictive Cruise Control”);

- control system for one or more components of the drive train of the first vehicle 10.

A process for controlling the first vehicle 10 is advantageously implemented by the first vehicle 10, that is to say by a computer or a combination of computers of the on-board system of the first vehicle 10, for example by the computer(s) responsible for controlling the ADAS system(s) of the first vehicle 10 when the latter is traveling in an autonomous or semi-autonomous driving mode.

Such a process of controlling the first vehicle 10 comprises operations of a data classification process, which is part, for example, of a process of predicting or determining the category representing the wear of the brake pads of the first vehicle 10.

The classification and/or prediction process is for example implemented by the first vehicle 10 or by a system comprising the first vehicle 10 (i.e. by one or more devices embedded in the vehicle 10, for example one or more computers) and the remote device 101 connected in communication with the first vehicle 10 via a wireless connection or link.

The process of predicting/determining the level of wear and/or the category representing the level of wear of the brake pads advantageously comprises two phases, each of these phases comprising one or more operations.

The first phase corresponds to a so-called learning or training phase of one or more prediction models (and/or one or more classification models) and the second phase corresponds to a so-called production or prediction phase based on the model(s) learned in the learning phase and data feeding the model(s) learned.

The first phase and the second phase are for example implemented by the remote device 101.

According to an alternative embodiment, the first phase is implemented by the remote device 101 (or by a server outside the “cloud”, for example a server hosted in a center (for example a design office) suitable for carrying out the learning) and the second phase by the first vehicle 10, for example by a computer of the on-board system of the first vehicle 10.

According to another embodiment, the learning is implemented in the first phase and the parameters of the prediction model(s) are refined in real time from the data relating to the first vehicle 10 received or collected during the second phase.

Learning phase

The learning implemented in the first phase advantageously corresponds to supervised learning from a set of data associated with a set or group of second vehicles 11 comprising for example a few tens, a few hundreds, thousands or tens/hundreds of thousands of second vehicles. According to an alternative embodiment, the learning implemented in the first phase corresponds to unsupervised learning from the set of data associated with the set of second vehicles.

According to a particular example, the second vehicles included in the set of second vehicles 11 all have the same configuration or a similar configuration, that is to say that they all correspond to the same type of vehicle (for example the same series of a particular model of vehicle) with identical or similar components or organs (same version for example).

In a first operation of the learning phase, the remote device 101 collects a set of data representative of the dynamic behavior of each second vehicle 11, this data being for example emitted by each of the second vehicles of the set 11 as the second vehicles move. The data are for example transmitted by each second vehicle via the wireless network infrastructure described above to the remote device 101. The data correspond for example to some or all of the vehicle signals emitted by sensors embedded in each second vehicle and/or by sensors (for example one or more gyroscopes and/or accelerometers) of mobile communication devices 103 (for example a tablet, a smartphone) embedded in the second vehicles 11 and for example connected in communication with the second vehicle (for example by Bluetooth® or Wifi®) which embeds them. These data are for example transmitted continuously at a determined frequency (for example every second, hour, day, etc.) or at each variation of a quantity measured by a sensor or a computer embedded in each second vehicle. The data reported to the remote device 101 correspond for example to the data of a determined list and stored in the memory of a computer of the embedded system of each second vehicle which makes a selection of the data to be transmitted among the set of data or physical quantities that each second vehicle detects or measures.

The data representative of the dynamic behavior of each second vehicle 11 include for example all or part of the following data:

- representative acceleration data, for example longitudinal, lateral and/or vertical acceleration obtained for example from an inertial unit of the vehicle;

- representative yaw rate data;

- representative speed data (e.g. longitudinal and/or lateral), measured by an odometer for example;

- representative steering angle data measured by a steering angle sensor;

- data representative of use of the brake pedal (for example frequency of use, intensity of pressure exerted on the pedal during each braking, number of uses of the pedal and/or level of braking (gentle or hard braking for example, corresponding to the speed of pressing the pedal during braking)), obtained from one or more sensors associated with the pedal and/or from a computer receiving the commands from the pedal;

- data representative of use of the accelerator pedal (for example frequency of use, intensity of pressure exerted on the pedal during each acceleration, number of uses of the pedal and/or level of acceleration (gentle or strong or sudden acceleration for example, corresponding to the speed of pressing the pedal during acceleration)), obtained from one or more sensors associated with the pedal and/or from a computer receiving the pedal commands;

- data representative of braking distance when the brake pedal is actuated, these data representing the distance travelled when the brake pedal is actuated until the vehicle speed falls below a determined threshold (for example a threshold equal to 0, 2 or 5 km/h);

- data representative of the deceleration of the vehicle, for example the intensity of the deceleration, which is for example represented by a set of levels (very strong, strong, normal or weak);

- data representative of sound or noise acquired by one or more microphones embedded in the vehicle or in a mobile communication device present in the vehicle, in particular during the vehicle braking sequence, the sounds or noises captured making it possible to obtain information on the condition of the brake pads.

These data are for example acquired or measured by accelerometers, microphones, gyroscopic sensors integrated in each second vehicle or integrated in a mobile communication device embedded in each second vehicle. These sensors make it possible to measure the movements, accelerations, and variations in the trajectory of the vehicle according to each direction of a three-dimensional reference associated with the vehicle, for example.

In a second operation of the learning phase, the remote device 101 collects a set of data representative of the environment in which each second vehicle 11 circulates. These data are for example acquired by one or more sensors (for example a camera) embedded in the second vehicles, by one or more sensors arranged in the environment. According to another example, all or part of these data are received from one or more data processing or storage devices (for example a server) connected in communication to the remote device 101, for example a weather data server, a map data server, a road maintenance data server, etc.

The data representative of the environment in which the second vehicles 11 circulate include, for example, all or part of the following data:

- data representative of weather conditions (rain, snow, ice, temperature, sunshine, etc.);

- data representative of the condition and/or level of road maintenance (presence of holes, potholes, quality of the coating, etc.);

- data representative of the type of road, for example expressway, motorway, urban road (a road in urban areas having, for example, a higher density of intersections, for example roundabouts and/or a higher density of speed bumps than other types of roads);

- map data;

- data representing the location of the vehicle, this data being for example obtained from a receiver of a geolocation system (for example of the GPS type) embedded in the vehicle and transmitted by the vehicle.

In a third operation, the remote device 101 collects data representative of a change of the brake pads. This data includes, for example, information on the date on which one or more brake pads were changed or replaced and/or the type of brake pads fitted to the vehicle to replace the old brake pads.

These data are for example obtained from one or more servers 102 associated with after-sales services provided by one or more companies providing repairs to vehicles 11 (car garage for example), which provide the updating of a database relating to vehicles that have undergone a maintenance operation (preventive or curative), in particular in relation to the brake pad(s) of each second vehicle 11. The server(s) 102 transmit for example a set of data identifying each second vehicle for which one or more brake pads have been changed.

The server(s) 102 also provide, for example, data on the level of wear observed on the second vehicles during maintenance operations on the second vehicles, whether or not the brake pads are changed.

The data obtained include, for example, for each second vehicle and for each maintenance carried out on the brake pads, the date on which the measurement or maintenance was carried out. The data set also includes, for example, an identifier for each second vehicle, making it possible to determine one or more of the following information:

- the type of vehicle;

- the vehicle model;

- the make of the vehicle;

- the hardware and software configuration of the vehicle.

The identifier of each second vehicle corresponds, for example, to the so-called VIN number (from the English “Vehicle Identification Number”).

The collection of data obtained in the first, second and/or third operations is, for example, implemented over a specific time interval corresponding, for example, to several months or several years of use of the second vehicles, for example in real time.

These data relating to the set of second vehicles during the determined time interval are advantageously stored in the memory of the remote device 101 or of a remote device dedicated to the collection and storage of these data and connected to the remote device 101, for example via a wired link.

According to an alternative embodiment, information representative of the time instant at which each data item was measured, transmitted or received is associated with said data item. In other words, the data is time-stamped.

In a fourth operation of the learning phase, the remote device 101 processes the data obtained during the first, second and third operations.

The processing comprises a cross-referencing of the first and second data with, for example, the data obtained from the server 102 on the condition of the pads determined during maintenance operations for the generation of a model for predicting the level of wear of the brake pads.

At the output of the brake pad wear level prediction model, fourth data representative of the predicted or estimated brake pad wear level are obtained.

The processing also includes processing the first data for generating a driving mode or style prediction model, such a mode belonging for example to a set of driving modes comprising 3, 4, 5 or more modes and representing the driving style of the vehicle (for example sporty, aggressive, calm, economical, standard style, etc.).

At the output of the vehicle driving mode prediction model, fifth data representative of the vehicle driving mode are obtained.

A data set formed from the third, fourth and fifth data is thus obtained. The data in this data set are processed and/or cross-referenced to generate a set of classes, each class in the set being associated with a brake pad wear category representing a specific degree or level of wear. The classes or groups obtained are said to be labelled, i.e. a label is associated with each class or group. For example, if 5 classes or groups are generated when classifying the data in the data set, 5 wear categories will also be generated (for example from 0 to 4), with for example level 0 corresponding to the lowest wear category (corresponding for example to a new condition of the brake pads) and 4 to the highest or maximum wear category (corresponding to a very degraded/worn condition of the brake pads). According to another example, if 3 classes are generated, 3 wear categories will be generated with for example: a minimum level, a standard level and a maximum level.

Each prediction model as well as the model for classifying data into different wear categories are generated or trained by classifying the different data processed, according to any methods known to those skilled in the art.

The generation of the different prediction/classification models are for example learned by a neural network, the parameters or characteristics (from the English “features”) of the brake pad wear prediction model being learned in a first layer or in a first set of layers of the network, the parameters or characteristics of the driving mode prediction model being learned in a second layer or in a second set of layers of the network and the parameters or characteristics of the prediction/classification model into wear categories being learned in a third layer or in a third set of layers of the network.

The classification implemented corresponds for example to a classification via a k-means partition method. This k-means partition method is for example associated with a method called “Elbow” in order to group the different profiles obtained into groups or classes of different wear categories.

The k-means partitioning method is described in relation to the The principles of such a method associated with the “Elbow” method are known to those skilled in the art and for example described in the document entitled “Determining the appropriate cluster number using elbow method for K-means algorithm”, published by H. Humaira and R. Rasyidah in January 2018.

Such a classification/prediction process is advantageously implemented by a neural network implementing a machine learning method, also called a machine learning method, for example a deep learning method.

According to a particular and non-limiting example, the neural network corresponds to a recurrent neural network with short and long term memory, called a LSTM network (from the English "Long Short Term Memory"), an example of such an LSTM network being described in the document "Long short-term memory" published by Sepp Hochreiter and Jürgen Schmidhuber, in Neural Computation. 9 (8): 1735 1780.doi:10.1162/neco.1997.9.8. 1735. PMID 9377276. An LSTM-type neural network consists of interconnected cells (neurons) interacting non-linearly, an example of which is described with regard to the . The output of a cell is a nonlinear combination of its inputs. A neural network of the LSTM type consists of interconnected cells (neurons) interacting nonlinearly. The output of a cell is a nonlinear combination of its inputs.

When the prediction or classification phase is implemented by the first vehicle 10, the parameters defining each prediction or classification model obtained in the fourth operation are transmitted by the remote device 101 to the first vehicle 10 which stores them in memory associated with a computer of the on-board system for the implementation by this computer of the prediction/classification phase.

Such a learning phase thus makes it possible to cross-reference data from (or associated with) second vehicles with data obtained from after-sales services (data on the state of wear of the brake pads and/or on the changing of these brake pads).

Prediction phase

In a first operation of the prediction phase, the device in charge of the prediction phase (i.e. the remote device 101 or one or more computers of the on-board system of the first vehicle 10) collects a set of first data representative of the dynamic behavior of the first vehicle 10. These first data correspond for example to the data representative of the dynamic behavior of the second vehicles having been used in the learning phase.

The first data are for example transmitted by the first vehicle 10 to the remote device 101 via the wireless network infrastructure over time, for example at regular intervals or when a new measurement is carried out by one of the sensors acquiring the first data.

The first data thus includes all or part of the following data, for example:

- representative acceleration data, for example longitudinal, lateral and/or vertical acceleration; and/or

- representative yaw rate data; and/or

- representative speed data (e.g. longitudinal and/or lateral); and/or

- representative steering angle data; and/or

- representative data on brake pedal usage; and/or

- data representative of the use of the accelerator pedal; and/or

- data given representative of braking distance; and/or

- representative deceleration data; and/or

- data representative of sound or noise generated during braking.

When the prediction phase is implemented by the first vehicle 10, the first data is stored in a memory associated with the computer of the on-board system of the first vehicle 10, or in a server of the “cloud” to which the first vehicle 10 is connected.

According to an alternative embodiment, information representative of the time instant at which each first data item was measured or acquired or transmitted is associated with said first data item. In other words, each first data item is time-stamped.

In a second operation of the prediction phase, the device in charge of the prediction phase (i.e. the remote device 101 or one or more computers of the on-board system of the first vehicle 10) collects a set of second data representative of the environment in which the first vehicle 10 is traveling or evolving. These second data correspond for example to the environmental data obtained for the second vehicles 11, in a similar or identical manner, which were used in the learning phase.

The second data thus includes, for example, all or part of the following data:

- data representative of meteorological conditions; and/or

- data representative of the condition and/or level of road maintenance; and/or

- data representative of road type; and/or

- map data; and/or

- location data of the first vehicle.

In a third operation of the prediction phase, third data representative of a change or replacement of the brake pads of the first vehicle 10 are received or obtained, for example from the server 102.

The collection of the first data, the second data and the third data is thus similar or identical to the collection of the dynamic behavior data of the second vehicles 11, the environmental data of the second vehicles 11 and the data on the maintenance of the brake pads of the second vehicles 11 described with regard to the learning phase.

In a fourth operation of the prediction phase, the first and second data are used to determine fourth data representative of a level of wear of the brake pads of the first vehicle 10. These fourth data are for example obtained from the wear prediction model learned during the learning phase.

These fourth data correspond for example to a wear level, the number of levels being for example equal to 3, 4 or 5.

In another example, these fourth data correspond to a value between, for example, -10 and 0 (-10 for very significant or maximum wear and 0 for no wear) or between 0 and 10.

In yet another example, these fourth data correspond to a percentage of efficiency, that is, a value between 0 and 1, with 1 representing maximum efficiency and minimum wear and 0 representing minimum efficiency and maximum wear.

In a fifth operation of the prediction phase, the first data are used to determine fifth data representative of a driving mode of the first vehicle 10. These fourth data are for example obtained from the driving mode prediction model learned during the learning phase.

These fifth data correspond for example to an identifier among a set of identifiers (for example coded on 2 or 4 bits depending on the number of identifiers), each identifier corresponding to a specific driving mode.

In a sixth operation, a data set formed of the third, fourth and fifth data is used as input to the prediction/classification model learned or trained during the learning phase and making it possible to identify a brake pad wear category for the first vehicle 10. This or these models produces information representative of a prediction of the brake pad wear category to be associated with the brake pads of the first vehicle 10.

The determination of the wear category of the brake pads of the first vehicle 10 is for example implemented using the so-called k-means method, which is for example implemented by a neural network, for example of the LSTM type. The LSTM type neural network makes it possible to predict the wear category of the brake pads of the first vehicle 10 in real time from the first, second and third data collected.

The first, second and third data are thus processed to determine the probability of belonging to each class of the plurality of classes generated during the learning phase. The brake pad wear category determined for the first vehicle 10 corresponds to the wear category associated with the class for which the probability of belonging is the highest.

According to an alternative embodiment, the fourth, fifth and sixth operations form a single operation implemented by different layers of the neural network, the output data of a layer (or a set of layers) feeding the following layer (or set of layers) as input until the classification of the set of data comprising the third, fourth and fifth data is obtained.

In a seventh operation, the first vehicle 10 is checked according to the result of the classification, that is to say according to the wear category determined in the previous operation.

According to a first exemplary embodiment, the control of the first vehicle 10 comprises the control of one or more ADAS systems of the first vehicle 10 according to the wear category of the brake pads.

Control of these ADAS systems makes it possible to adapt the driving mode of the first vehicle 10 and/or its route depending on the wear category.

For example, when the wear category indicates a significant or maximum level of wear (for example above a threshold), a route favoring major roads (motorways and expressways) will be calculated or determined to avoid roads requiring more braking (roads in built-up areas for example) likely to wear out the brake pads more quickly.

The route will then be proposed to the driver of the first vehicle 10, for example via an on-board navigation system. According to a variant, when the first vehicle 10 is traveling in an autonomous mode greater than or equal to 4, the first vehicle 10 automatically follows the route calculated under the control of the ADAS systems implemented for such a function.

In another example, the ABS system and/or the ESP system will be adapted, for example by switching to a brake pad-saving mode when the level of wear is significant, i.e. above a threshold.

In yet another example, the speed of the first vehicle will be adapted to reduce speed and limit accelerations/decelerations when the wear level is high. This is for example achieved by controlling the cruise control system of the first vehicle 10.

According to a second exemplary embodiment, the control of the first vehicle 10 comprises the rendering of a first message and/or a second message in the passenger compartment of the first vehicle 10. The rendering comprises for example the display of a text or an icon and/or the generation of a sound.

The first message is for example representative of the wear category determined to alert the driver of the first vehicle 10, in particular when the wear is significant, requiring a maintenance operation to, for example, replace one or more brake pads.

The second message is for example representative of one or more recommendations intended for the driver of the first vehicle 10, for example to recommend that he change the brake pads, adapt his driving to reduce the risks linked to a level of wear above a threshold, etc.

The first message and/or the second message are for example displayed on a display screen of the first vehicle 10. The display of a message is for example implemented through an HMI (Human Machine Interface) controlled by the IVI (In-Vehicle Infotainment) computer of the first vehicle 10.

The first exemplary embodiment and the second exemplary embodiment described above with regard to the seventh operation are, for example, implemented independently of one another or combined, i.e. implemented in parallel.

There schematically illustrates a grouping of dynamic behavior data and environmental data into a predetermined number K of groups according to a particular and non-limiting exemplary embodiment of the present invention.

According to this example of implementation, a method called “k-means” is used.

The "k-means" method groups data sets (represented by squares) into k (here k=3) groups called "clusters" in English. To do this, each set is represented in a multidimensional space (here in a two-dimensional space). Each group GRi is represented in this space by a centroid Ci (here an average) whose initial value is a random value. On the , each centroid Ci is represented by a circle 201, 202, 203. For each added dataset, a distance Di is calculated between a point in space representing this set and each centroid Ci. The dataset is then added to the group GRi whose distance Di is minimal. The centroid Ci of the group Ci in which the dataset has just been added is then updated taking into account the added dataset.

Alternatively, multiple clusters of data sets are computed for different K numbers and one of these clusters is retained.

For example, the conserved clustering is the one that minimizes standard deviations between the data sets of the groups in that cluster and the centroid associated with those groups.

There schematically illustrates a cell, denoted 'c', with short-term and long-term memory according to a particular and non-limiting exemplary embodiment of the present invention.

The idea associated with a long-term short-term memory neural network, called LSTM network, is that each cell is linked to two internal states; one called hidden state ( ) and the other says cell state ( ) which plays the role of memory. The passage from an input state from a cell at a time instant (t-1) to an output state at a following time instant t is done by transfer at constant gain and equal to 1. In this way the errors propagate to the previous steps without gradient disappearance phenomenon. The state of the cell can be modified through an input gate which allows or blocks the update. The same goes for a control door if the cell state is communicated at the cell output. Cells can also use a gate allowing the cell state to be reset.

As an example, considering the following initial values: And , and considering an activation function of these sigmoid type cells and hyperbolic tangent , the states of the reset, output and update gates as well as the hidden, cell and output states are given by:

in which o is an operator symbolizing the Hadamard matrix product (term-by-term product) and , , , are weights (parameters) and are offset values (parameters).

Alternatively, other functions can be used.

During the learning phase, the neural network learns to produce a prediction of the wear category (and/or wear level and/or driving mode) from a set of data (first, second, third, fourth and/or fifth data) collected.

A classic neural network learning algorithm is then used for learning. The principle of this learning algorithm is that the neural network learns the values of the cell parameters by backpropagation of the error gradient, i.e. the algorithm progressively calculates the parameter values in order to minimize a cost function for each of the short-term and long-term memory cells starting from the outputs of the neural network.

According to a particular and non-limiting example of embodiment, the cost function is a precision function given by:

According to a particular and non-limiting example of embodiment, the cost function is a function of type F1 given by:

This cost function can be minimized using a classical gradient descent algorithm such as the supervised gradient descent algorithm ADAM ( Kingma & Ba (2015) Adam: A Method for Stochastic Optimization. Proceedings of the International Conference for Learning Representations, San Diego (ICLR) ).

During the production phase, the neural network is identical to the neural network in the learning phase, except that the input of the neural network is no longer supplied with data from the second vehicles 11 but with the first, second and third data relating to the first vehicle 10 and its environment. The neural network is then capable of reproducing a prediction of the wear category of the brake pads as explained above which corresponds to the wear category associated with the group GRi to which the data set belongs.

Alternatively, the LSTM-type neural network has a single cell.

Alternatively, the LSTM-type neural network has multiple cells.

According to a particular and non-limiting exemplary embodiment, the neural network is stored in a memory or a database of the first vehicle 10.

There schematically illustrates a device 4 configured to control one or more ADAS systems of a vehicle, for example the first vehicle 10, according to a particular and non-limiting exemplary embodiment of the present invention.

According to another exemplary embodiment, the device 4 is configured to determine or predict a wear category of the brake pads of a vehicle, for example of the first vehicle 10.

Such a device 4 is configured for implementing the learning phase and the production phase of a machine learning method.

According to yet another example, the device 4 is configured both to control one or more ADAS systems of a vehicle and to determine or predict a wear category of the brake pads of this vehicle.

The device 4 is for example configured for the implementation of the operations described with regard to figures 1, 2, and 3 and/or the steps of the method described with regard to the . Examples of such a device 4 include, but are not limited to, a computer, a laptop, a server, a mobile phone, a tablet, or on-board electronic equipment such as an on-board computer of a vehicle or an electronic calculator such as an ECU (“Electronic Control Unit”). The elements of the device 4, individually or in combination, may be integrated into a single integrated circuit, into several integrated circuits, and/or into discrete components. The device 4 may be implemented in the form of electronic circuits or software (or computer) modules or a combination of electronic circuits and software modules.

The device 4 comprises one (or more) processor(s) 40 configured to execute instructions for carrying out the steps of the method and/or for executing the instructions of the software(s) embedded in the device 4. The processor 40 may include integrated memory, an input/output interface, and various circuits known to those skilled in the art. The device 4 further comprises at least one memory 41 corresponding for example to a volatile and/or non-volatile memory and/or comprises a memory storage device which may comprise volatile and/or non-volatile memory, such as EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic or optical disk.

The computer code of the embedded software(s) comprising the instructions to be loaded and executed by the processor is for example stored in memory 41.

According to a particular and non-limiting exemplary embodiment, the device 4 comprises a block 42 of interface elements for communicating with external devices, for example a remote server or the “cloud”, other nodes of the ad hoc network. The interface elements of the block 42 comprise one or more of the following interfaces:

- RF radio frequency interface, for example Wi-Fi® type (according to IEEE 802.11), for example in the 2.4 or 5 GHz frequency bands, or Bluetooth® type (according to IEEE 802.15.1), in the 2.4 GHz frequency band, or Sigfox type using UBN (Ultra Narrow Band) radio technology, or LoRa in the 868 MHz frequency band, LTE (Long-Term Evolution), LTE-Advanced;

- USB interface (from the English “Universal Serial Bus” or “Universal Serial Bus” in French);

- HDMI interface (from the English “High Definition Multimedia Interface”);

- LIN interface (from the English “Local Interconnect Network”).

According to another particular and non-limiting exemplary embodiment, the device 4 comprises a communication interface 43 which makes it possible to establish communication with other devices (such as other computers of the on-board system) via a communication channel 430. The communication interface 43 corresponds for example to a transmitter configured to transmit and receive information and/or data via the communication channel 430. The communication interface 43 corresponds for example to a wired network of the CAN (Controller Area Network), CAN FD (Controller Area Network Flexible Data-Rate), FlexRay (standardized by the ISO 17458 standard) or Ethernet (standardized by the ISO/IEC 802-3 standard).

According to a particular and non-limiting exemplary embodiment, the device 4 can provide output signals to one or more external devices, such as a display screen, touch-sensitive or not, one or more speakers and/or other peripherals (projection system) via respective output interfaces.

There illustrates a flowchart of the different steps of a method for controlling a vehicle, for example the first vehicle 10, according to a particular and non-limiting exemplary embodiment of the present invention. The method is for example implemented by a device on board the vehicle, for example by the device 4 of the , or by a system comprising the on-board device connected in wireless communication to a remote device such as a server or computer.

In a first step 51, first data representative of a dynamic behavior of the vehicle from a set of sensors embedded in the vehicle are received, the set of sensors comprising at least one sensor.

In a second step 52, second data representative of an environment in which the vehicle is traveling are received.

In a third step 53, third data representative of a change of a set of brake pads of the vehicle are received.

In a fourth step 54, fourth data representative of a level of wear of the set of brake pads of the vehicle are determined as a function of the first and second data and fifth data representative of a driving mode of the vehicle are determined as a function of the first data.

In a fifth step 55, information representative of a wear category of the set of brake pads is determined by classifying a set of data comprising the third data, the fourth data and the fifth data from a classification model learned in a so-called learning phase, the wear category being associated with a class of a set of classes at the output of the classification model.

In a sixth step 56, the vehicle is checked according to information representative of a wear category.

Alternatively, the variants and examples of the operations described in relation to Figures 1, 2 and 3 apply to the steps of the method of the .

Of course, the present invention is not limited to the exemplary embodiments described above but extends to a method for predicting a wear category or a wear level of the brake pads of a vehicle, which would include secondary steps, without thereby departing from the scope of the present invention. The same would apply to a device configured for implementing such a method.

The present invention also relates to a vehicle, for example an automobile, comprising the device of the .

The present invention also relates to a system comprising a vehicle and one or more remote server-type devices, the vehicle being connected in wireless communication to the remote device(s).

Claims (10)

Method for controlling a vehicle (10), said method being implemented by at least one processor and comprising the following steps:
- reception (51) of first data representative of a dynamic behavior of said vehicle (10) from a set of sensors embedded in said vehicle (10), said set comprising at least one sensor;
- reception (52) of second data representative of an environment (1) in which said vehicle (10) is circulating;
- reception (53) of third data representative of a change of a set of brake pads of said vehicle (10);
- determination (54) of fourth data representative of a level of wear of said set of brake pads of said vehicle (10) as a function of said first and second data and determination of fifth data representative of a driving mode of said vehicle (10) as a function of said first data;
- determination (55) of information representative of a wear category of said set of brake pads by classification of a set of data comprising said third data, said fourth data and said fifth data from a classification model learned in a so-called learning phase, said wear category being associated with a class of a set of classes at the output of said classification model;
- control (56) of said vehicle (10) as a function of said information representative of a wear category. The method of claim 1, wherein said control of said vehicle comprises at least one of the following steps:
- control of at least one driving assistance system, called ADAS system, of said vehicle according to said wear category; and/or
- rendering of a first message representative of said wear category via a human-machine interface embedded in said vehicle; and/or
- rendering of a second message representative of at least one recommendation based on said wear category via said human-machine interface embedded in said vehicle; and/or
- determining a route for said vehicle (10) based on said wear category. The method of claim 2, wherein said at least one ADAS system belongs to a set of ADAS systems comprising:
- an electronic stability control system, known as the ESC system;
- an anti-lock braking system, known as the ABS system;
- a speed regulation system;
- a trajectory control system;
- a route control system;
- a system for controlling at least one component of a powertrain of said vehicle. Method according to one of claims 1 to 3, for which said first data belong to a set of data comprising:
- representative acceleration data;
- representative yaw rate data;
- representative speed data;
- representative steering angle data;
- data representative of the use of a brake pedal of said vehicle;
- data representative of the use of an accelerator pedal of said vehicle;
- representative braking distance data;
- representative deceleration data;
- data representative of sound generated during braking,
and for which said second data belong to a data set comprising:
- data representative of meteorological conditions;
- data representative of the condition and/or level of road maintenance;
- data representative of road type;
- representative slope or elevation data;
- data representing the location of said vehicle;
- map data. Method according to one of claims 1 to 4, for which said classification is implemented via a k-means partition method. Method according to claim 5, wherein said classification is implemented by a neural network. Method according to claim 6, for which said neural network is a network of the recurrent network type with short and long term memory, called LSTM network. Computer program comprising instructions for implementing the method according to any one of the preceding claims, when these instructions are executed by a processor. Device (4) for controlling a vehicle, said device (4) comprising a memory (41) associated with at least one processor (40) configured for implementing the steps of the method according to any one of claims 1 to 7. Vehicle (10) comprising the device (4) according to claim 9.