Method for securely evaluating cartographic data

Abstract

The invention relates to a method for securely evaluating cartographic information or data acquired by a sensor by a moving vehicle including at least one sensor and a navigation system, the navigation system being configured to acquire cartographic information with their associated cartographic positions, the method including: - a plurality of acquisitions by the navigation system, at different acquisition instants, of one cartographic datum among the cartographic information; - determining, for each acquired cartographic datum, a quantity representative of the cartographic datum at a location of the vehicle substantially equal to the cartographic position associated with the cartographic datum; - determining, for each acquired cartographic datum, a differential between the cartographic data and the corresponding quantities; - determining a confidence index by comparing at least part of the differentials; and - generating at least one control datum of the vehicle at least according to the confidence index.

Description

[0001] Description

[0002] Title: Method for secure evaluation of cartographic data

[0003] Technical field

[0004] The present invention relates to a method for securely evaluating cartographic data by a moving vehicle and a vehicle comprising a computer program product configured to implement a method for securely evaluating cartographic information by a moving vehicle.

[0005] Prior art

[0006] Current automotive navigation systems generally integrate external map information. This information comes in particular from variable resolution maps ("SDmap" or "HDmap") stored on remote servers and accessible via an internet connection, or from other vehicles via a communication system allowing vehicles to exchange information with each other, for example via V2V ("Vehicle to Vehicle") or V2X ("Vehicle to Everything") protocol.

[0007] Usually, this information allows for better visibility or information on the vehicle's environment, particularly the environment in front of the vehicle, and is therefore more commonly referred to as electronic horizon information (otherwise known as "eHorizon") of the vehicle. It makes it possible to supplement and / or anticipate information from the vehicle's sensors, for example a camera, radar, lidar or sonar.

[0008] This data includes, for example, map information relating to the characteristics of a road, the characteristics of a predetermined route, road signs, and / or external conditions.

[0009] Taking this information into account allows for anticipation of the situations that the vehicle will encounter, and thus an adaptation in anticipation of the dynamic behavior of the vehicle.

[0010] However, the electronic horizon information from external information or directly from other vehicles is of limited reliability due to the fact that it is notably conveyed through an internet network, or wifi for certain V2X configurations, which constitutes a vulnerability, particularly in terms of cybersecurity. Indeed, the information from the eHorizon is transported to the vehicle through technologies that constitute a possible attack surface from a cybersecurity point of view, which poses a problem of trust in this information and introduces a risk regarding the use of this information as data for controlling the vehicle's actuations. This vulnerability prevents their use to directly and securely control the vehicle's actuators.The term "actuator" refers to a device configured to control, alone or in combination with other actuators, the dynamic behavior of said vehicle according to instructions determined by an algorithm, in particular by influencing the brakes, the engine, the steering and / or the speed of the wheels. It is indeed essential to control the actuators with reliable data.

[0011] Application FR3133443A1 discloses a method for evaluating and / or validating map data associated with a position on the road for controlling a driving assistance device (otherwise known as "ADAS"), in particular, its activation, deactivation or reconfiguration, or for sending an alert to the user concerning the validity of the map data provided by the navigation device. In this context, the map data is not used for direct control of the vehicle's actuators. The user information allows the latter to know that he cannot trust the data sent by the navigation system and the control of the driving assistance device allows deactivation or reconfiguration of the driving assistance if the data is not reliable.

[0012] Today, the control of actuators, in particular by the chassis system (otherwise called "Chassis Domain") or the micro-powertrain (otherwise called "Powertrain Domain"), is essentially reactive, in the sense that the information taken into account by said systems to generate the control data is previously calculated or generated by means of direct physical measurements ("sensing") or indirect ("virtual sensing") of the vehicle. Taking into account at least partially the cartographic information provided by the navigation system to have proactive control of the actuators, that is to say in an anticipated manner with respect to the occurrence of a future event on the predicted path of the vehicle, would be interesting in that it would make it possible to anticipate the behavior of the vehicle and to act on the actuators in anticipation to improve the reaction of the vehicle to the event.For example, if a vehicle is traveling on a road and receives map information relating to the value of the friction coefficient of the road downstream of the vehicle, it could be possible to anticipate the drift of said vehicle following a sudden reduction in grip by modifying the instructions of one or more actuators of the vehicle's wheels, in particular a yaw instruction. This improves the vehicle's reaction and the driver's comfort.

[0013] There is therefore a need to assess or validate the security of mapping information from external sources and received in a vehicle in a sufficiently reliable and continuous manner over time, in particular to enable the use of mapping information in the control of actuators, in particular by the chassis system or the micro-powertrain.

[0014] Statement of the invention

[0015] The invention meets this need, according to a first of its aspects, using a method for securely evaluating cartographic information or data acquired by a sensor by a moving vehicle comprising at least one sensor and a navigation system, the navigation system being configured to acquire from the outside, at an acquisition time, the location of the vehicle at the acquisition time and cartographic information with their associated cartographic positions, the method comprising: a plurality of acquisitions by the navigation system, at different acquisition times, of a cartographic datum from among the cartographic information with the cartographic position associated with said cartographic datum, the cartographic positions associated with said acquired cartographic data being geographically different from each other, the determination, for each acquired cartographic datum,of a quantity representative of the cartographic data from at least one measurement by the at least one sensor at a location of the vehicle substantially equal to the cartographic position associated with said acquired cartographic data, the determination, for each acquired cartographic data, of a differential between the acquired cartographic data and the corresponding quantities determined by measurement, the determination of a confidence index by comparison of at least part of the determined differentials, the generation of at least one vehicle control data at least as a function of the determined confidence index.,

[0016] By "associated map position", it is understood that the navigation system acquires the map position corresponding to the position of each map information on a predetermined path. The navigation system can acquire a map position per acquired map information or acquire a map position for a plurality of acquired map information located at the same location on a path predetermined as being the most probable. The map position can comprise the geographical location of a point where the map information is located at the time of acquisition relative to a fixed or mobile reference point. This geographical location can be determined relative to a fixed system, in particular a geodetic system, for example the WGS 84 system associated with the GPS positioning system.Alternatively, it is determined relative to a mobile system, in particular relative to the vehicle, the geographical location then being the position at the time of acquisition of the cartographic information(s) acquired relative to the vehicle. The cartographic position may include GPS coordinates, latitude, longitude and altitude coordinates, coordinates relative to the vehicle, a distance relative to the vehicle on the most probable path or any other means making it possible to position the cartographic data on a map or relative to the vehicle.

[0017] By "acquisition" we mean that at an acquisition time, the navigation system receives a data stream containing the map information and the associated map positions from a remote server through an internet network, in particular a cloud, or from another vehicle.

[0018] By "representative quantity" is meant a quantity determined from a measurement of at least one sensor which, when the cartographic data is accurate and reliable, has a relationship with the acquired cartographic data. The relationship is in particular an equality, a fixed difference or a proportionality.

[0019] By "substantially equal" is meant that the location of the vehicle, when the determination of the representative quantity is carried out, is equal to the map position of the acquired data with a predefined acceptable margin of error. The acceptable margin of error may correspond to an error in the measurement time less than or equal to 0.5 s, better still less than or equal to 0.1 s. The acceptable margin of error may correspond to a distance less than or equal to 50 m, better still less than or equal to 10 m, even better still less than or equal to 5 m, even better still less than or equal to 2 m, for example substantially equal to 1 m.

[0020] By "differential" we mean a real relative or absolute difference observed between the value of the cartographic data and that of the corresponding measured representative quantity.

[0021] By means of the invention, it is possible, by comparing the map data from sources external to the vehicle over several acquisitions and corresponding quantities from measurements in the vehicle, to assess whether the map data and the measured data are consistent with each other or not and thus to establish whether confidence can be attributed to this map data or to the sensor measurements according to which the use of this data presents a security to be determined.

[0022] If the map data is considered reliable, the confidence index can provide information on the confidence to be given to the vehicle sensor measurements. It can then inform the user about a sensor failure.

[0023] In the case where the sensor measurements are considered reliable, the confidence index can provide information on the confidence to be given to the map data and potentially also to the received map information by extending the confidence to all the information received with an acquisition in which a map data has been established as reliable. Indeed, confidence in these acquired map data can result in confidence in the data flow and therefore in any map information acquired simultaneously with the map data. It is then possible to have reliable information that can be used to at least partially generate control data in the vehicle.Conversely, it is also possible to establish that externally sourced map data is unreliable and thus limit the control data to data from the vehicle for the sake of user safety.

[0024] Making this determination on a plurality of cartographic data at different times makes it possible to reinforce security and to have a reliable confidence index. It is then possible to evaluate the confidence on a time window between the first acquisition time and the last acquisition time and no longer on an acquisition of a data at a single time. The comparison of several acquisitions makes it possible in particular to detect differences in time which could reflect a corruption of the cartographic information or a specific error of a cartographic data. In addition, the comparison is no longer limited to equality, but it is possible to take into account a different context by determining a correlation equation different from the equality between the cartographic data and the corresponding representative quantities. This is not possible with a single differential.

[0025] The invention thus makes it possible to significantly limit the risks linked to a possible cyberattack or a failure of cartographic information or a failure of a sensor without, however, excluding cartographic data presenting a correlation linked to different contexts. This results in improved user security. The level of confidence in the operation of the system is also reinforced.

[0026] Since the electronic horizon or "eHorizon" is a function already present on motor vehicles, notably used for controlling F ADAS and determining the confidence index using already on-board sensors, the invention does not require any significant additional software or technical complexity. The cost of the proposed solution is therefore relatively low.

[0027] Such a method is thus particularly suitable for the use of proactive algorithms - unlike state-of-the-art algorithms which are essentially reactive - making it possible in particular to adapt the parameters of a regulation algorithm of one or more actuators by distributing the instructions differently to the actuators according to acquired mapping information. A system defined as such is thus made more efficient.

[0028] Preferably, the method comprises the navigation system determining the most probable path of the vehicle based on user input data, information acquired from one or more sensors of the vehicle and / or the location of the vehicle determined by the navigation system. The most probable path of the vehicle may be a route validated by the user or a route corresponding to a journey determined as being the most probable in view of the location of the vehicle and one or more data of the vehicle or the user.

[0029] The acquired map information is associated with a map position on or along the determined most probable path.

[0030] The navigation and location system preferably has a GNSS receiver, for example GPS type, and allows the driver to enter a route for the vehicle and / or the latter to know its position in real time.

[0031] The navigation system may acquire the map data and / or map information from a remote server through an internet network, including a cloud, or from another vehicle, for example by V2V (“Vehicle to Vehicle”) or V2X (“Vehicle to Everything”) protocol.

[0032] Information

[0033] The cartographic information may include: one or more characteristics of a road, in particular the coefficient of adhesion and / or the roughness of the road, the type of road, and / or one or more characteristics of a predetermined route, in particular the curvature of a bend, the slope of the road, the presence of obstacles on the road, in particular the presence of speed bumps, and / or road signs, in particular the presence and identification of road signs or lines on the ground, and / or external conditions, in particular weather or temperatures, or localized information on the maximum vehicle size or the maximum vehicle mass authorized on a road.

[0034] Data

[0035] Preferably, the map data provides information relating to a physical or structural property of the road on the most probable path determined by the navigation system. The map data may be selected from any of the aforementioned map information.

[0036] Preferably, it is chosen from information comprising a numerical value, preferably substantially continuous over a path. It can be chosen from the curvature of a bend comprising, for example, a value of the local curvature angle of the road, the slope of a road comprising a value of the local inclination angle of the road, the road adhesion coefficient comprising a value of the local adhesion coefficient of the road, the road roughness coefficient comprising a value of the local roughness coefficient of the road. Having a numerical value allows for more precise characterization of the confidence. The fact that the data is substantially continuous over the path allows for sampling of the acquisitions and easy detection of an erroneous acquisition that would have to be discarded.

[0037] The cartographic data may include the numerical value and one or more labels identifying the nature of the numerical value. The label(s) may characterize the unit of the numerical value and / or the nature of the cartographic data, for example a coefficient of adhesion, roughness, a slope, an angle of curvature.

[0038] Alternatively, the cartographic data indicates the presence of road signs, in particular traffic signs, road lines, obstacles, or even characteristic places in the environment (for example, historical monuments or rivers). The characteristic data may then include a Boolean, in particular characterizing the presence or absence, and / or a numerical value, for example indicating the speed limit, and one or more labels, in particular characterizing the nature of the cartographic data, for example the type of sign identified, the type of line on the ground, the nature of an obstacle and / or the name of a characteristic place in the environment.

[0039] The method may comprise, at each acquisition, the acquisition of a plurality of different cartographic data from among the aforementioned cartographic information, each associated with a cartographic position, said cartographic positions being identical or not. The method may then determine for each cartographic data a confidence index on a part of the plurality of acquisitions and the generation of the control data according to the different confidence indices determined, in particular the taking into account of the cartographic information only when a confidence for all the cartographic data is established, or even in particular the taking into account only the cartographic information for which a confidence for the corresponding cartographic data has been established, the cartographic information for which the corresponding cartographic data are not trusted being not taken into account.

[0040] The representative quantity may be a value derived from a measurement by one or more sensors on board the vehicle. The value may be obtained by a direct measurement from a vehicle sensor or deduced from one or more measurements from one or more sensors.

[0041] The representative quantity can be determined using a single sensor.

[0042] The method may include determining the distance on the most likely path between the vehicle at the acquisition time and the map position associated with the map data acquired at the acquisition time. Alternatively, this information is acquired directly with the map data for each acquired map data.

[0043] The determination of the representative quantity for each acquired map data may include the determination, as a function of the speed of the vehicle at the acquisition time, of the measurement time so that the location of the vehicle at the measurement time is substantially equal to the map position associated with the map data. The measurement time may be determined at least from the distance on the most probable path between the vehicle at the acquisition time and the map position associated with the map data acquired at the acquisition time and the speed of the vehicle at the acquisition time. The determination of the measurement time may also be done dynamically by integrating any variations in the speed of the vehicle between the acquisition time and the measurement time to improve said determination.

[0044] Determining the time of measurement of the representative quantity by calculation allows for a relatively precise and reliable process. In practice, since the vehicle's geolocation is not available continuously, but periodically according to a given geolocation update frequency, this method ensures that the measurement is carried out at the best time, for example between two acquisitions by the navigation system.

[0045] Alternatively, the determination of the measurement time can be done by acquiring the location of the vehicle periodically and triggering the measurement when the location of the vehicle identified by the periodic acquisition is closest to the map position acquired with the map data.

[0046] The acquisition of map data can be successive in time or based on the vehicle's position on its route. Each acquisition can be triggered automatically or manually.

[0047] Map information can be acquired by the navigation system at a predefined acquisition frequency, particularly when the vehicle is in motion.

[0048] The acquisitions of the cartographic data may be sampled in time or on the positioning of the vehicle on its route at a predetermined sampling frequency identical to or different from the frequency of acquisition of the cartographic information, in particular at a sampling frequency corresponding to an acquisition of the cartographic data every N acquisitions of cartographic information by the navigation system, N being an integer. The acquisition times may be spaced apart by a substantially constant duration or are spaced apart by a duration corresponding to a substantially constant distance traveled by the vehicle.Alternatively, the method may include searching for the map data each time map information is received and acquiring the map data if it is present at a map position at a distance from the vehicle on the most likely path less than or equal to a predetermined distance and if it has not been previously acquired at a previous acquisition time.

[0049] The acquired map positions are at a distance and / or travel time relative to the vehicle that is not zero at the corresponding acquisition time. Preferably, the distance to the vehicle is less than or equal to a predetermined vehicle distance, in particular less than or equal to 500m, better still less than or equal to 100m and / or greater than or equal to 1m, better still greater than or equal to 5m.

[0050] The distance, particularly along the most likely path, between the acquired map positions and the vehicle locations at the corresponding acquisition time may be substantially constant. Alternatively, said distance is not constant.

[0051] The determination of the confidence index can be done on at least a part of the cartographic data acquired over a predetermined acquisition duration, and / or a predefined number of acquisitions. The acquisition duration can be predetermined according to the cartographic data. The number of acquisitions can be predefined according to the cartographic data, in particular according to the frequency of appearance of the cartographic data in the case of cartographic data characteristic of a discrete event.

[0052] Preferably, the determination of the confidence index is done at an evaluation time and takes into account at least the cartographic data acquired at the acquisition time upstream in the time closest to the evaluation time.

[0053] Sliding window

[0054] Preferably, the determination of the confidence index is carried out at an evaluation time by comparing at least a portion of the differentials, better all the differentials, corresponding to at least a portion of the cartographic data acquired upstream of the evaluation time over a time window of a predetermined acquisition duration ending at the evaluation time or corresponding to a predetermined number of successive cartographic data acquisitions going back in time from the evaluation time.This makes it possible to determine the confidence index over a sliding time window upstream of the evaluation instant corresponding to a constant predefined duration or a constant number of acquisitions and thus to have an up-to-date confidence index over a predetermined number of cartographic data to be both sufficiently large to allow stability of the determination and at the same time sufficiently low to limit the risk of error in the evaluation of confidence.

[0055] The comparison of the differentials may include the determination, from the acquired cartographic data which allowed the determination of the confidence index and the corresponding representative quantities, of a polynomial correlation equation between the cartographic data and the representative quantity of the cartographic data of a rank lower than a predefined rank and of maximum correlation index with the acquired cartographic data and the corresponding representative quantities which were used to determine the polynomial correlation equation.

[0056] Preferably, the determined correlation polynomial equation is of rank less than or equal to 1, that is to say is a proportionality between the cartographic data and the quantity representative of said cartographic data and / or a constant shift between the cartographic data and the quantity representative of said cartographic data. By "proportionality", we understand a relationship of the type y=cx, y being the representative quantity, x the acquired cartographic data and c a proportionality factor, c can take any value, in particular be substantially equal to 1. By "constant shift", we understand a relationship of the type y=x+b, y being the representative quantity, x the acquired cartographic data and b a constant value.

[0057] The method may include assigning the confidence index based on the correlation index, including assigning a confidence index greater than the threshold confidence value or equal to a discrete confidence value when the correlation index is greater than a threshold correlation value. For example, it is possible to assign a value of 0 to the confidence index when the correlation index is less than a threshold correlation value and a value of 1 to the confidence index when the correlation index is greater than a threshold correlation value. This makes it possible to determine the existence of a stable relationship between the acquired cartographic data and the corresponding representative quantity and, if such a stable relationship exists, to consider that there is a good level of confidence. This also makes it possible to be independent of the context.Indeed, a context different from that corresponding to that of the acquired cartographic data can have an influence on the value determined for the representative quantity which can result in the existence of a polynomial equation between the acquired cartographic data and the representative quantity. The value of a cartographic data can vary according to certain contextual parameters, in particular rain, temperature, ice, humidity, or even the height of snow on the road, the method according to the invention makes it possible to generalize the validity of cartographic data provided without context, or with a context different from that in which the vehicle is moving at the time of acquisition of the cartographic data.

[0058] Contextualization

[0059] The method may include automatically correcting the map data acquired by the navigation system based on the determined correlation equation to contextualize the acquired map data if the correlation index is greater than a predetermined threshold correlation value by applying the determined correlation equation to it. The control data may be generated at least based on the acquired map data after contextualization when the confidence index is within a predetermined value range or at a predetermined discrete value.

[0060] Preferably, the determination of the correlation index is carried out at a correlation time by comparing at least a portion of the differentials, better all the differentials, corresponding to at least a portion of the cartographic data acquired upstream of the correlation time over a time window of a predetermined acquisition duration ending at the correlation time or corresponding to a predetermined number of successive cartographic data acquisitions going back in time from the correlation time.This makes it possible to determine the correlation index over a sliding time window upstream of the correlation instant corresponding to a constant predefined duration or a constant number of acquisitions and thus to have an up-to-date correlation index over a predetermined number of cartographic data to be both sufficiently large to allow stability of the determination and at the same time sufficiently low to limit the risk of error in the evaluation of the correlation. The correlation instant may be identical to the evaluation instant and / or the predefined correlation duration may be identical to the acquisition duration or the number of cartographic data considered identical.

[0061] Preferably, the correlation index is defined such that the smaller the deviations between the map data acquired after contextualization and the corresponding representative quantities, the larger the index, and vice versa. Thus, the correlation index characterizes the quality of the correlation model determined from the polynomial correlation equation.

[0062] The method may comprise determining a correlation index associated with the or each cartographic data item acquired after determining the polynomial correlation equation corresponding to the correlation of the or each cartographic data item acquired after determining the polynomial correlation equation and the corresponding representative quantity with the polynomial correlation equation. This index may be determined from the difference between the cartographic data item acquired after contextualization and the corresponding representative quantity. Such an index makes it possible to characterize for each cartographic data item the compliance or non-compliance with the determined polynomial equation. Preferably, the correlation index is defined such that the smaller the difference between the cartographic data item acquired after contextualization and the corresponding representative quantity, the larger the index, and vice versa.Thus, the correlation index characterizes the quality of the contextualization carried out from the polynomial correlation equation.

[0063] The method may comprise determining a new polynomial correlation equation if the correlation index is lower than the correlation threshold value over a period greater than a predetermined period, in particular the acquisition period, or over a number of successive acquisitions greater than a predetermined number, or if the correlation indices of the map data acquired over a period greater than the predetermined period, in particular the acquisition period, or over a number of successive acquisitions greater than the predetermined number are all lower than a threshold index value.

[0064] The method may comprise, as long as the correlation index is greater than the threshold correlation value or the correlation indices of the acquired cartographic data are greater than a threshold index value, the use of the cartographic data acquired after contextualization, in particular by enriching a database using this cartographic data or generating control data at least as a function of the cartographic data acquired after contextualization, the control data being in particular input data of a user information device or of an actuator of the vehicle.

[0065] The method may comprise determining context data from one or more vehicle sensors or external data and associating the context data with the map data acquired after contextualization and / or with the determined correlation equation, the context data including in particular the presence of rain or not and / or its intensity, the temperature, the humidity level, the wind speed and / or the dew point.

[0066] The method may include sending contextualization information to a database, in particular internal or external, the contextualization information comprising the context data and / or the combination of the cartographic data acquired after contextualization and the determined correlation equation. Preferably, the method for determining the polynomial equation is chosen from polynomial interpolation, in particular Lagrangian, and polynomial regression. In particular, the polynomial regression may be carried out by the least squares method, maximum likelihood, by Bayesian inference, or by machine learning methods such as, for example, support vector machines (SVM).

[0067] Preferably, the correlation index is chosen from the Bravais-Pearson coefficient, the Spearman rho coefficient, the Goodman and Kruskal gamma coefficient, or the Kendall tau coefficient.

[0068] The threshold correlation value may depend on contextual parameters.

[0069] Preferably, the threshold correlation value depends on the type of map data. For example, the threshold correlation value associated with map data relating to temperature is different from that associated with map data relating to the road adhesion coefficient.

[0070] The threshold correlation value may depend on the number of map data sampled to determine the correlation polynomial equation.

[0071] Confidence index

[0072] The confidence index can be a numerical value between two different numbers, for example between 0 and 1.

[0073] Alternatively, the confidence index may be discrete, in particular of Boolean type or defined by a discrete number of different values ​​each corresponding to different confidence levels, in particular at least three different values.

[0074] The generation of at least one control data may be at least based on a comparison of the confidence index with at least one threshold confidence value.

[0075] The generation of at least one control data may be at least based on a comparison of the confidence index with several different threshold confidence values.

[0076] The method may comprise generating a control data item as a function of a security level of the control data item from among a plurality of different security levels and the confidence index, a value range of the confidence index being assigned to each security level to determine whether or not to take into account one or more pieces of cartographic information in generating the control data item.For example, the control data may be associated with a first security level corresponding to a first discrete value or a first value range of the confidence index and a second security level corresponding to a second discrete value or a second value range of the confidence index, the first security level being associated with control data for which taking into account erroneous data has a negligible impact on the security of users and the second security level being associated with control data for which taking into account erroneous data presents a risk to the security of users. This makes it possible to reduce the risks for users by authorizing the use of potentially erroneous data when the impact of the erroneous data on security is negligible and by refusing the use of potentially erroneous data if its use presents a risk to the user.Trust levels may be a function of user security and / or user comfort in taking the data into account.

[0077] Preferably, the confidence index is determined or updated periodically based on successive determinations of the differentials.

[0078] The confidence index update frequency may be different from the map data sampling frequency, in particular the map data sampling frequency is a multiple of the confidence index update frequency.

[0079] In particular, the confidence index can be determined or updated after a predefined number of acquisitions of the cartographic data at different successive acquisition times, in particular greater than 2. For example, the confidence index can be updated following the calculation of 5 successive differentials.

[0080] The confidence index can be determined from an error function taking as a parameter at least the differentials between the acquired cartographic data, in particular after contextualization, i.e. the cartographic data acquired after contextualization, and the corresponding representative quantities, the error function also taking in particular as a parameter said acquired cartographic data, in particular after contextualization, i.e. the cartographic data acquired after contextualization. For example, the error function is chosen from the variance (or "L2"), the mean square error, the absolute norm error (or "L1"), the maximum error or their combination.

[0081] The error function may be relative. By "relative" it is meant that the error function makes it possible to quantify the deviations observed between the acquired cartographic data, in particular after contextualization where appropriate, and the corresponding representative quantities relative to the values ​​of said acquired cartographic data, in particular after contextualization where appropriate.

[0082] Preferably, the threshold confidence value(s) depend on the type of map data acquired. For example, the threshold confidence value(s) associated with map data or a temperature sensor are different from those associated with map data or a road grip coefficient sensor.

[0083] The threshold confidence value(s) may depend on the number of map data sampled to determine the confidence index value.

[0084] The threshold confidence value(s) preferably depend on contextual parameters, including determined context data, e.g., weather conditions. For example, the threshold confidence value(s) may be different depending on whether the weather is sunny, rainy, or snowy.

[0085] Order data

[0086] Preferably, in the case where the confidence index is characteristic of the confidence that can be had in the cartographic data, the measurement data of the sensor being considered reliable, the control data is generated as a function of at least one of the acquired cartographic information if the confidence index is greater than or equal to a threshold confidence value. The control data can be generated as a function of at least one quantity representative of cartographic information if the confidence index is less than the threshold confidence value.

[0087] The acquired cartographic information taken into account to generate the control data may be the cartographic data. In this case, the control data may be generated based on the acquired cartographic data, in particular after contextualization where appropriate, if the confidence index is greater than or equal to a threshold confidence value. Alternatively, the at least one of the acquired cartographic information taken into account to generate the control data may be different from the cartographic data. For example, if the confidence index associated with cartographic data relating to the roughness coefficient is greater than the threshold confidence value, the control data may be established from cartographic information grouping the temperature cartographic data, as well as other cartographic data relating to atmospheric pressure, or to the roughness coefficient.

[0088] Preferably, in the case where the confidence index is characteristic of the confidence that can be had in the measurements of a sensor, the corresponding cartographic data being considered reliable, the control data can be generated as a function of at least one quantity resulting from the measurement of the sensor if the confidence index is greater than or equal to the threshold confidence value. The control data can be generated as a function of at least one of the acquired cartographic information if the confidence index is less than a threshold confidence value.

[0089] Preferably, the control data is sent to a vehicle system, in particular an actuator of the vehicle, said actuator being configured to control alone or in combination with other actuators the dynamic behavior of said vehicle, in particular by influencing the brakes, the engine, the steering and / or the speed of the wheels, depending at least on the input data.

[0090] Alternatively or in combination, the control data may be sent to a vehicle alert device, said alert device being configured to alert a user at least when the confidence index is below the threshold confidence value.

[0091] The method may comprise generating a plurality of vehicle control data to control different systems of the vehicle, in particular different actuators, a vehicle warning device, the ADAS system, in particular for controlling the chassis, in particular the suspensions, the torque setpoint motorization, in particular the thermal or electric engine, the wheel motors, the brakes, or even noise cancellation systems inside the vehicle (from the English "active noise control" or ANC). Safety sense

[0092] Preferably, when a cartographic datum has a differential between the acquired cartographic datum, in particular after contextualization where applicable, and the representative quantity less than a threshold value and when taking into account in the generation of the control datum the value of an acquired cartographic datum, in particular after contextualization where applicable, is less secure than taking into account the value of the measurement of the representative quantity, the confidence index is determined as being less than the threshold confidence value temporarily at least for a predetermined duration, in particular at least until the next acquisition of the cartographic datum.This can also be done on the basis of the correlation index when it is lower than a threshold index value and when taking into account in the generation of the control data the value of a cartographic data item acquired after contextualization is less secure than taking into account the value of the measurement of the representative quantity, the confidence index is determined as being lower than the threshold confidence value temporarily at least for a predetermined duration, in particular at least until the next acquisition of the cartographic data item. This makes it possible, punctually on a specific erroneous cartographic data item, to discard or retain the erroneous cartographic data item depending on the risks that it may generate if it is taken into account.

[0093] Preferably, when a cartographic datum has a differential between the acquired cartographic datum, in particular after contextualization where applicable, and the representative quantity less than a threshold value and when taking into account in the generation of the control datum the value of an acquired cartographic datum, in particular after contextualization where applicable, is more secure than taking into account the value of the measurement of the representative quantity, the confidence index is determined as being greater than the threshold confidence value temporarily at least for a predetermined duration, in particular at least until the next acquisition of the cartographic datum.This can also be done on the basis of the correlation index when it is lower than a threshold index value and when taking into account in the generation of the control data the value of said cartographic data acquired after contextualization is more secure than taking into account the value of the representative quantity, the confidence index is temporarily determined as being higher than the threshold confidence value at least for a predetermined duration, in particular at least until the next acquisition of the cartographic data.

[0094] By "safe", it is meant that taking into account acquired cartographic data, in particular after contextualization where appropriate, in the generation of control data may be associated with a certain level of risk depending in particular on the type of cartographic data and the value of the value of said data and optionally on a context. For example, taking into account a first cartographic data relating to the road's grip coefficient with a numerical value of 0.9 is less safe than taking into account a second data of the same type with a numerical value of 0.5. Indeed, a relatively low grip coefficient exposes the vehicle to more risks of loss of grip than with a relatively higher grip coefficient and the vehicle takes control data that are safer for the user in the case of a lower grip coefficient.Thus, applying a low grip coefficient value determined incorrectly instead of a higher grip coefficient causes the vehicle to adopt a safe posture corresponding to the low grip, which may lead to a temporary loss of performance of the vehicle but does not risk endangering the user. Conversely, incorrectly taking into account high grip in the case of lower actual grip may cause the vehicle to adopt a performance posture generating, in the case of lower actual grip, a greater risk for users.

[0095] Preferably, the predetermined duration is conditioned by a predefined minimum number of measurements of representative quantities to be carried out, in particular one or two to update the confidence index and / or the polynomial correlation equation, the confidence index and / or the polynomial correlation equation being in particular updated following said measurements.

[0096] Discarding of clearly erroneous data

[0097] The comparison of the differentials may include the removal of a clearly erroneous data item in the determination of the confidence index and / or the search for the polynomial correlation equation having a maximum correlation index with the data and representative quantities taken to search for the equation. Preferably, a data item is considered clearly erroneous if the differential corresponding to the difference between the acquired cartographic data item, in particular after contextualization where applicable, and the associated representative quantity is greater in absolute value than a predetermined error threshold and the closest acquired cartographic data item, in particular after contextualization where applicable, have differentials corresponding to the difference between the acquired cartographic data item, in particular after contextualization where applicable, and the associated representative quantity lower in absolute value than a predetermined error threshold.Thus, initially, the cartographic data is identified as at risk and the following acquisition makes it possible to discriminate the erroneous nature of the acquisition.

[0098] Removing obviously erroneous data allows for a more accurate and stable process.

[0099] The impact of a relatively large differential is reduced or even completely cancelled out, this difference being able to be caused by measurement noise from at least one sensor, or by a clear and isolated error in a map datum.

[0100] The at least one sensor of the vehicle may be of the external type, and carry out measurements relating to physical parameters, objects or infrastructures external to the vehicle.

[0101] Conversely, the at least one sensor may be of the internal type and carry out measurements relating to physical parameters or objects internal to the vehicle, in particular tire pressure, the angle of rotation of the wheels or even the pressure in the passenger compartment.

[0102] The at least one sensor is preferably chosen from odometers, gyrometers, accelerometers, thermocouples, RTDs, thermistors, barometers, cameras, sonars, radars or even Lidars or a combination of these sensors.

[0103] Actuator

[0104] Preferably, the actuator is configured to control alone or in combination with other actuators the dynamic behavior of said vehicle, in particular by influencing the brakes, the engine, the steering and / or the speed of the wheels, as a function at least of the control data. The actuator can be configured to control the chassis, in particular the suspensions, the torque setpoint motorization, in particular the thermal or electric engine, the wheel motors, the brakes, or even the noise cancellation systems inside the vehicle (from the English "active noise control" or ANC).

[0105] Vehicle

[0106] The invention also relates to a vehicle equipped with at least one sensor and a navigation system allowing the acquisition at an acquisition time of cartographic information with their associated cartographic position and the location of the vehicle at the acquisition time, said vehicle further comprising a computer program product configured to, when the vehicle is moving: locate the vehicle and acquire cartographic information, carry out a plurality of acquisitions by the navigation system, at different acquisition times, of a cartographic datum from among the cartographic information and of the cartographic position associated with said cartographic datum, the acquired cartographic positions associated with said acquired cartographic data being geographically different from each other, determine, for each acquired cartographic datum,a quantity representative of the cartographic data from at least one representative measurement by the at least one sensor at a location of the vehicle calculated to be substantially equal to the associated cartographic position at which said cartographic data is acquired, determining, for each acquired cartographic data, a differential between the acquired cartographic data and the corresponding quantities, determining a confidence index based on the comparison of at least part of the differentials of the cartographic data. generating control data based on the comparison of the confidence index with a predetermined threshold confidence value.,

[0107] The features described above in connection with the method also apply alone or in combination to the above vehicle.

[0108] Brief description of the drawings The invention may be better understood by reading the detailed description which follows, non-limiting examples of its embodiment, and by examining the attached drawing, in which:

[0109] [Fig 1] Figure 1 illustrates, schematically and partially, an example of equipment of a vehicle suitable for implementing the method according to the invention.

[0110] [Fig 2] Figure 2 is a block diagram illustrating steps of an exemplary method for evaluating map data according to the invention.

[0111] [Fig 3] Figure 3 is a diagram illustrating steps of acquiring cartographic data and measuring corresponding representative quantities according to the invention.

[0112] [Fig 4] Figure 4 is a block diagram illustrating steps of an exemplary method of contextualizing cartographic data according to the invention.

[0113] [Fig 5] Figure 5 schematically and partially illustrates different samples of cartographic data and corresponding measurements of quantities according to the invention.

[0114] Detailed description

[0115] Figure 1 illustrates an example of a motor vehicle 1 in top view having equipment suitable for implementing the method according to the invention.

[0116] Vehicle 1 can be thermal, for example gasoline, diesel, gas, hydrogen, or hybrid, or even electric.

[0117] The vehicle 1 may include different sensors 3.

[0118] Vehicle 1 is also equipped with a computer 5, which receives data from the various sensors 3.

[0119] The computer 5 comprises one or more processors executing one or more programs allowing the implementation of the method according to the invention.

[0120] The computer 5 may be composed of all or part of an on-board computer system comprising, for example, in addition to the aforementioned processor(s), at least one RAM memory, at least one ROM memory used to save the applications of the vehicle 1, one or more possible digital-to-analog converters as well as one or more input / output interfaces used to communicate with the various sensors 3. The computer 5 may access an on-board memory storing a plurality of map data, as well as data measured by the sensors 3.

[0121] The calculator 5 can be connected to any type of interface allowing information to be presented to the user of the vehicle 1.

[0122] The vehicle 1 further comprises a navigation and location system 7, preferably with a GNSS receiver, for example of the GPS type, enabling in particular the driver to enter a route for the vehicle 1 and / or the latter to know its position in real time.

[0123] The vehicle 1 may be configured to access a remote server via any type of communication means, for example a 4G or 5G network or other.

[0124] The remote server may include a plurality of map information.

[0125] The server is accessible by a set of several vehicles 1 according to the invention, the connection of each of the vehicles 1 to the server being able to be carried out via a protocol having a certain level of security. The level of security may be sufficient for the use of the map information in certain areas. However, for use in the chassis system or the micro-powertrain, the level of security of the map information is often not sufficient. For this, the invention relates in particular to a method for evaluating the map information illustrated in Figure 2.

[0126] In the example considered, the cartographic data relating to the road adhesion coefficient are used, but the method can be generalized to other types of data, such as the road roughness coefficient, the external temperature or pressure, the curvature of a bend, the slope of a road, or to any other type of cartographic data measurable directly or indirectly by a sensor 3 of the vehicle 1.

[0127] A plurality of acquisitions of cartographic data of road adhesion coefficient ai, sa, ...a n with their associated map position Pai, Pd2, . ..Pdn at different acquisition times Ti, T? ... T n is carried out by the navigation system 7 during a first acquisition stage 20.

[0128] In Figure 2, n successive moments of acquisition of a cartographic data are represented by Ti, T? ... T n. Sampling of the acquisition of road adhesion coefficient values ​​ai, ai, ...a n is thus carried out on N acquisitions. The number of acquisitions is chosen in particular according to the cartographic data. The n acquisition times Ti, T2 ... T n can be spaced regularly over time and cover an acquisition period (T n -Ti) predetermined.

[0129] At this acquisition stage 20, for each adhesion coefficient value acquired by the navigation system 7, a measurement of the adhesion coefficient mi, m2, . ..m n is carried out by a sensor 3 of the vehicle 1, in this case a sensor 3 located at the level of the wheels of said vehicle 1. The measurement of the representative quantity can be direct by measuring the quantity with a sensor 3 or deduced from data acquired by one or more sensors 3 of the vehicle 1.

[0130] Road adhesion coefficient measurements mi, m2, ...m n are carried out at n locations of the vehicle P v i', P V 2', ...Pvn' substantially equal to the cartographic positions of the cartographic data acquired according to the principle detailed below in connection with figure 3.

[0131] The method then comprises a comparison step 24 comprising the determination of differentials between each adhesion coefficient value ai, a2, ...a n and corresponding sensor measurement 3 mi, m2, ...m n .

[0132] For example, for an acquired adhesion coefficient value a of 0.90 and a corresponding measurement m carried out by sensor 3 of vehicle 1 of 0.87, the differential d determined in absolute value |am| is 0.03.

[0133] Step 24 of comparing the differentials di, d2, . ..d nmay include the removal of clearly erroneous data, so that it is not taken into account in determining the confidence index I c This may be the case, for example, if a differential in absolute value is greater than a first predetermined error threshold d s i, and that the absolute value differentials of the map data acquired at least at the closest acquisition times are less than a second error threshold d S 2, less than the first error threshold d s i. Alternatively, this may also be the case when a value of the cartographic data or representative quantity is clearly erroneous or aberrant in view of the acquired close values, without it being necessary to calculate a deviation. This makes it possible to exclude a cartographic data which would have an aberrant value compared to the other close values. In step 26, a confidence index I cis determined from the differentials di, di, ...d n determined in the previous step 24.

[0134] In the case of a substantially identical context between that in which the cartographic data was when it was recorded in the database and that in which the vehicle 1 is located, equality of the values ​​acquired by the navigation system 7 via the cartographic data and the values ​​measured by the sensor 3 of the vehicle 1 is sought.

[0135] In this case, the confidence index I c is determined from the differentials di, d2, ... d n determined in the previous step 24 and the values ​​of the road adhesion coefficients acquired in step 20. For example for n=3, noting A = [ai, ai, as] the list of acquired adhesion coefficient values ​​and D = [di, d?, ds] the list of respective successive differentials, the confidence index I c can be defined by Ic = 1- f(D, A) with f the relative error function in absolute norm, or

[0136] For A = [0.90; 0.89; 0.80] and D = [0.03; -0.03; 0.05], the calculation gives I c = 0.96.

[0137] The Confidence Index I c is then compared to a threshold confidence value I cs at step 28.

[0138] The threshold confidence value I cs retained can be equal to 0.95 to limit the confidence to the values ​​of adhesion coefficients within the error limit of 5%. Thus, if the confidence index I c is greater than this value, the data is considered to be trusted, as is the case in the example above. Conversely, if the confidence index I c is below this threshold value, the data are not of sufficient confidence.

[0139] In the case of a different context between that in which the cartographic data was when it was recorded in the database and that in which vehicle 1 is located, the determination of the confidence index I c is no longer done by equality. Indeed, a more significant gap can be observed between the values ​​acquired through cartographic data and the representative quantities determined by the measurements. In most cases, the context will lead to a constant or proportional gap between the ZI values ​​acquired through cartographic data and the representative quantities determined by the measurements. This gap can substantially follow a known rank 1 correlation equation or one that can be determined from the comparison of the different differentials di, d?, ...d n .

[0140] Thus, step 26 may include the determination of a polynomial correlation equation, for example of rank 1, from at least part of the differentials di, d2, ...d n The equation can be of the form y=cx+b, x being the cartographic data, y being the corresponding representative quantity, c being a proportionality factor and b a constant value.

[0141] The polynomial equation is for example determined from a polynomial regression.

[0142] The determined polynomial equation is the one with a maximum correlation index I r with the cartographic data and representative quantities taken into account, for example the Bravais-Pearson correlation index.

[0143] Figure 5 illustrates three examples of correlations observed between cartographic data, for example the adhesion coefficient, and representative quantities measured by a sensor 3 of the vehicle 1. The graphs illustrate on the ordinate the value of the adhesion coefficient and on the abscissa the position of the vehicle 1 corresponding to the data (the cartographic position for the cartographic data and the location of the vehicle 1 at the time of determination of the representative quantity). In these three examples, seven values ​​of acquired cartographic data are represented jointly with the seven values ​​of representative quantities measured by the sensor 3 of the vehicle 1.

[0144] The acquired map data are represented on the dotted line curve, and the measurements of representative quantities on the solid line curve.

[0145] The graph in Figure 5a) shows a constant deviation b between the acquired map data and the corresponding representative quantity measurements. The determined polynomial equation is thus of degree 1 and of type y=x+b.

[0146] The graph in Figure 5b) shows a proportional deviation of proportionality coefficient c between the acquired cartographic data and the corresponding measurements of representative quantities. The determined polynomial equation is thus of degree 1 and of type y=cx. For these first two cases, there exists a polynomial equation of rank 1 presenting a very good correlation with the cartographic data and representative quantities and therefore a good confidence in the cartographic data.

[0147] The graph in Figure 5 c) shows a sample of acquired map data and corresponding representative quantity measurements without rank 1 polynomial equation with a good correlation index I r determinable. The confidence index I c corresponding to such a sample of cartographic data is low.

[0148] In this case, the confidence index I c may be dependent on the correlation index I r determined. It may be greater than a threshold confidence value I cs or equal to a discrete value representative of a confidence I ci in the case where the correlation index I r is greater than a threshold correlation value I rs .

[0149] For example, if the determined correlation polynomial equation has a correlation index I r of 0.9 and that the threshold correlation value I rsis 0.85, the map data are considered to be trustworthy and the confidence index I c is significant of confidence in the map data.

[0150] Conversely, the confidence index I c is determined to be less than the threshold confidence value I cs or equal to a discrete value representing an absence of confidence I c o in the case where the correlation index I r is less than a threshold correlation value I rs -

[0151] For example, if the polynomial correlation equation with maximum correlation index has a correlation index I r of 0.5 and the threshold correlation value is 0.85, the correlation between the adhesion coefficient values ​​and the corresponding measurements is considered poor, the map data is not considered reliable and the confidence index I cis significant of a lack of confidence.

[0152] Confidence can be determined from the analysis of differentials between the cartographic data acquired after contextualization, corresponding to the acquired cartographic data corrected from the polynomial correlation equation, and the representative quantities by seeking equality of equivalent values. The case is then equivalent to that described previously in the case where the contexts were identical. The method comprises in step 28 the comparison of the confidence index I c with one or more threshold confidence values. This can be done by comparison with the threshold confidence value I cs or by comparison to discrete values ​​among a plurality of discrete values ​​representing different security levels [I c o, Here]-

[0153] A filtering step 30 is then carried out, comprising the generation of control data for the vehicle 1, in particular for an actuator of the vehicle 1, based on the comparison of step 28.

[0154] If the map data ai, a2, ... a n are trusted, the control data can be established based on the grip value data acquired by the navigation system 7 relating to positions located in front of the vehicle 1, for which a measurement of a representative quantity by the wheel sensor 3 has not yet been carried out. Said data can, for example, inform of a drop in the value of the road grip coefficient to 0.5.

[0155] If necessary, the map data used for the order data can be contextualized upstream by applying the determined correlation equation to it.

[0156] Furthermore, good confidence in the acquired map data may indicate good general confidence in the map information acquired by the navigation system 7 at the times of acquisition and therefore the absence of cybersecurity risk. It is then possible, as long as the confidence index I c is greater than the threshold confidence value, to use the acquired cartographic information on information other than that of the cartographic data. In this context, it is possible to use cartographic information other than that for determining confidence, for example on road signs or the route layout, from a confidence index I cdetermined on the adhesion coefficient. It is also possible to contextualize other cartographic information, in particular that of roughness or road layout according to the method described below, or to validate the measurements made by a sensor 3 of the vehicle 1 from cartographic information according to the previous method by applying the confidence index I c to the representative quantities to detect a malfunction of a sensor 3, the map information being considered reliable. Finally, the method ends with a control step 32 during which the vehicle 1 is controlled from the control data, for example the actuator influences the dynamic behavior of the vehicle 1 in an anticipated manner, for example here on the brakes of the vehicle 1 on the basis of the information considered reliable of the drop in road grip.

[0157] If the map data ai, ai, ... a nare not trusted, the control data can be established based on measurements of representative quantities by wheel sensor 3 only.

[0158] The method may include saving or sending acquired map data after contextualization to a remote server when the data is established as being trusted. This sending of contextualized data may be accompanied by context information.

[0159] Preferably, the confidence index I c and the correlation index I rare respectively evaluated periodically according to different or identical update frequencies, based on a new sampling of cartographic data acquired at later times. Some of the data may overlap between the two samplings. The confidence in said cartographic data, and in the correlation model of said data, are thus likely to evolve over time. Alternatively, the update of the confidence index I c and / or the correlation index I r may be carried out non-periodically, in particular when a mismatch between the polynomial correlation equation and the cartographic and measurement data is identified over a time greater than a predetermined time or a predetermined number of acquisitions. Preferably, the determination of the confidence index I c and / or correlation I ris done from a predetermined number of acquired map data directly upstream of the determination, i.e. directly preceding the determination, or from a time window of predetermined evaluation duration ending at the time of evaluation of the confidence index I c or correlation I r. This allows a determination on a sliding window with the evaluation time always being up to date. As explained previously, the number of cartographic data to be taken into account for this evaluation can be fixed or variable depending on whether one or more data are discarded because they are aberrant or less secure in their consideration as explained below. When a polynomial correlation equation has already been determined previously, a correlation index can then be determined for each subsequent acquisition. Such a correlation index establishes the correlation between the acquired cartographic data and the representative quantity from the previously determined polynomial correlation equation. It can correspond to a differential between the cartographic data acquired after contextualization and the corresponding representative quantity.Such a correlation index can help identify erroneous cartographic data as mentioned above or identify a decorrelation with respect to the correlation polynomial equation. A decorrelation on several successive cartographic data can lead to an update of the correlation polynomial equation or of the confidence index I. c .

[0160] The method may include taking into account or not taking into account cartographic data, in particular the adhesion coefficient, acquired for the actuation of the vehicle 1, if taking into account or not taking into account the data exposes the vehicle 1 to a safety risk compared to the cartographic data acquired directly previously. The safety risk may be directly derived from the value of the adhesion coefficient of the road announced by the cartographic data, for example 1.2, whereas the last measurement corresponded to a value of 0.6. In this first case, it may be dangerous to take into account the adhesion coefficient value of 1.2 which may give the vehicle 1 dangerous behavior if the road has an adhesion coefficient of 0.6 as previously. We therefore choose not to take into account the value of the cartographic data punctually while viewing the following cartographic data.Conversely, if taking the map data into account is safer, we can decide not to discard it. If the trend is confirmed in either case with the following data, we can then review the correlation equation.

[0161] All of the examples described above are based on the presence of a single threshold confidence or correlation value. Alternatively, the method may comprise a plurality of threshold confidence or correlation values ​​and the determination of a security level associated with each of the ranges of values ​​defined by the threshold confidence values. It is then possible to assign to each control data item a confidence level necessary for taking into account the acquired map data, in particular after contextualization where appropriate, in the control of the vehicle 1. Alternatively, it is possible to assign to each control data item a threshold confidence or correlation value, the control data items having different security criticalities.Some control data may then have a lower confidence requirement because it is not safety critical and others a higher confidence requirement because it is safety critical. Figure 3 schematically illustrates a vehicle 1 moving from left to right along a road represented by a first axis 36 and the acquisition of the map data and the determination of the corresponding representative quantity.

[0162] The first axis 36 illustrates the operation of the navigation system 7, in particular the different times and positions of acquisition of the cartographic data relating to the road grip coefficient, as well as the cartographic positions of said data.

[0163] A second equivalent axis 38 is represented below the first axis 36, illustrating the operation of the measurement sensor(s) 3 to determine the corresponding representative quantities, in particular the different positions and times of measurements by the sensor 3 of the vehicle 1 of the quantity representative of the road grip coefficient.

[0164] Vehicle 1 acquires position Pv n a road grip coefficient value geolocated at a map position Pd n located in front of him.

[0165] A calculation is carried out, in particular from the location of vehicle 1, its speed and its relative distance to point Pd n , in order to determine a measurement time Tn' where vehicle 1 will reach a point Pv n ' corresponding substantially to the position Pd nacquired. This determination of the measurement point by the sensor(s) 3 to determine the representative quantity can take into account the speed of the vehicle 1 at the acquisition time T n and can refine the determination with the speed of vehicle 1 between time T n and measurement.

[0166] Similarly, a new road grip coefficient data is acquired at time T n+i at position Pv n+ i, the data being located at a position Pdn+i and the measurement being carried out at a time T n+ i' when vehicle 1 has reached a position Pvn+i' corresponding substantially to position Pd n+i acquired. The distance r n separating the acquisition point from the Pv data n and the location of the map data at position Pd n is preferably less than a predefined threshold distance r threshold. Similarly, the distance r n+iis less than r seu ii. They may be identical. Indeed, the map data acquired for the determination or updating of confidence may be at a constant distance from vehicle 1 at the time of acquisition. Alternatively, these distances are not identical. They may in particular depend on the occurrence of an event corresponding to the map data.

[0167] The distance r n+ i' between the acquired map position and the location of vehicle 1 at the next acquisition time is preferably greater than or equal to 0.

[0168] In particular, in the case where r n+ i' = 0, the measurement of the quantity representative of the value of the adhesion coefficient relative to the geolocated data at Pd n is carried out at approximately the same time as the acquisition of the geolocated data at Pd n+ i.

[0169] Alternatively, the measurement of the quantity representative of the value of the adhesion coefficient relative to the geolocated data at Pd n is carried out after the acquisition of the geolocated data at Pd n +i. In this case, the point Pd n is preferably located after Pv n+i and before Pd n+ i.

[0170] An example of a process for contextualizing map data is shown in Figure 4.

[0171] In the example considered, the cartographic data to be contextualized are considered reliable. They may for example relate to the roughness coefficient of the road, but the method is generalizable to other types of data, such as the road grip coefficient, the road slope, the external temperature or pressure, the curvature of a bend, or to any other type of cartographic data and measurable directly or indirectly by one or more sensors 3 of the vehicle 1.

[0172] A plurality of acquisitions of cartographic data to be contextualized, here of road roughness coefficient, is carried out by the navigation system 7 during a first step 40.

[0173] In Figure 4, u times of acquisition of a cartographic data to be contextualized ki, i, ... k u successive are represented by Tl, T2 ... Tu. The sampling of the acquisition of the road roughness coefficient values ​​is thus carried out on u acquisitions. For each acquired road roughness coefficient value ki, k2, ... k u , a measure of a representative quantity vi, V2, ...v u is carried out by the sensor 3 of the vehicle 1, in this case a sensor 3 located at the level of the wheels of said vehicle 1.

[0174] The road roughness coefficient measurements are carried out at different times according to the same principle as in the case of the evaluation of the cartographic data described previously.

[0175] The method then comprises a step 44 of determining differentials between each acquired roughness coefficient value ki, k2, ... k u and the corresponding representative quantities vi, V2, ...v u .

[0176] In step 46, a rank 1 polynomial correlation equation with a correlation index I r maximum is determined from the differentials determined in the previous step 44 and the road roughness coefficient values ​​acquired in step 40 and / or the corresponding representative quantities vi, V2, ...v u .

[0177] Determination of the correlation index I rmaximum and the associated correlation model can for example be carried out by the least squares method.

[0178] Such a polynomial correlation equation provides an approximation of the roughness coefficient data by the values ​​measured directly by the sensor 3. Thus for a value v of the roughness coefficient of the road measured by the sensor 3 of the vehicle 1, the value of the acquired cartographic data k is estimated by a polynomial equation of rank 1, for example here of the form k = c'v + b', c' being a coefficient of proportionality and b' being a constant representing a fixed deviation.

[0179] The correlation index I r is compared to a threshold correlation index I rs at step 48.

[0180] In a first scenario, the correlation index I r determined is greater than the threshold correlation index I rs. The map data 50 representing the road roughness coefficient can then be contextualized with the determined correlation polynomial equation.

[0181] Step 50 may also comprise the generation of control data for the vehicle 1, in particular to an actuator of the vehicle 1, as a function of contextualized data of corrected road roughness coefficients relating to positions located in front of the vehicle 1, for which a measurement of a representative quantity by the wheel sensor 3 has not yet been carried out. Said data may, for example, provide information on an increase or decrease in the road roughness coefficient.

[0182] Finally, the method ends with a control step 52 during which the vehicle 1 is controlled from the control data, for example the actuator influences the dynamic behavior of the vehicle 1 in an anticipated manner, for example here on the brakes of the vehicle 1 on the basis of the contextualization of the data considered reliable, informing of the decrease or increase in the roughness coefficient of the road.

[0183] The method may include recording or sending contextualized data to a remote server when the confidence index I c and the correlation index I r are respectively greater than a threshold confidence value I cs and to a threshold correlation value I rs- The method may also include recording context data from data of the vehicle 1 or from outside the vehicle 1, for example temperature, humidity, wind, rain or ice. This data may be transmitted with the acquired map data, in particular after contextualization where appropriate, to record the context of acquisition of said data.

[0184] In the example considered, the correlation index I r is evaluated periodically according to a predetermined update frequency, based on a new sampling of map data acquired at later times. The confidence in the correlation model of said data is thus likely to evolve over time. Alternatively, the update of the correlation index I r can be performed non-periodically, especially when a decorrelation with the model is determined.

[0185] The method may comprise, for any data subsequent to the acquisition of a polynomial correlation equation, the determination of a correlation index between the acquired cartographic data, the corresponding determined characteristic quantity and the polynomial equation. Such an index may be determined as a function of a differential between the acquired cartographic data, in particular after contextualization where appropriate, on the basis of the polynomial correlation equation and the corresponding determined characteristic quantity. The method may comprise the removal of a cartographic data whose correlation index is below a predetermined threshold and / or for which a less secure command in the control of the vehicle 1 is carried out if it is taken into account.The method may comprise the determination of a new polynomial correlation equation when a plurality of successive cartographic data in a pre-established quantity are decorrelated with the determined equation, that is to say present an index characteristic of a decorrelation, in particular below a predetermined threshold.

[0186] The invention applies to any technical field and to any system. It can be generalized to the validation of future cartographic data from data measured in the present, or to the confirmation of measured data using cartographic data. Thus, it is possible to carry out the secure evaluation of measured data from corresponding cartographic data considered reliable and acquired securely.

[0187] Alternatively, the acquisition of map data can be carried out in groups (or "batches"). For example, a predefined number of map data can be acquired simultaneously at each acquisition.

Claims

Claims 1. Method for securely evaluating cartographic information or data acquired by a sensor (3) by a moving vehicle (1) comprising at least one sensor (3) and a navigation system (7), the navigation system (7) being configured to acquire from the outside, at an acquisition time, the location of the vehicle (1) at the acquisition time and cartographic information with their associated cartographic positions, the method comprising: a plurality of acquisitions (20, 40) by the navigation system (7), at acquisition times (Ti, ... T n ) different, from a cartographic data (ai,... a n ) among the map information with the map position (Pdi, . .. Pd n ) associated with said cartographic data (ai,... a n ), the map positions (Pdi, ... Pd n ) associated with said acquired cartographic data (ai,... a n) being geographically different from each other, the determination (20), for each acquired cartographic data, of a representative quantity (mi, ... m n ) of the map data from at least one measurement (mi, ... m n ) by at least one sensor (3) at a location (Pv, ... Pv n ') substantially of the vehicle (1) substantially equal to the map position (Pdi, . .. Pd n ) associated with said acquired cartographic data (ai,... a n ), the determination (24), for each acquired cartographic data (ai,... a n ), of a differential (di, ...d n ) between the acquired map data (ai,... a n ) and the representative quantities (mi, ... m n ) corresponding, the determination of a confidence index I c (26) by comparison of at least part of the differentials (di, .. .d n) determined, the generation (50) of at least one vehicle control data (1) at least as a function of the confidence index I c determined.

2. Method according to claim 1, in which the control data is generated according to at least one of the acquired map information if the confidence index I c is greater than a threshold confidence value I cs and at least one quantity representative of at least one of the acquired cartographic information if the confidence index I c is less than the threshold confidence value I cs .

3. Method according to claim 2, in which the control data is sent to a system of the vehicle (1), in particular an actuator of the vehicle (1), said actuator being configured to control alone or in combination with other actuators the dynamic behavior of said vehicle (1), in particular by influencing the brakes, the engine, the steering and / or the speed of the wheels, as a function at least of the input data.

4. Method according to any one of the preceding claims, in which the cartographic data is chosen from the curvature of a bend comprising for example a value of the local curvature angle of the road, the slope of a road comprising a value of the local inclination angle of the road, the adhesion coefficient of the road comprising a value of the local adhesion coefficient of the road, the roughness coefficient of the road comprising a value of the local roughness coefficient of the road, the type of road, localized information on the maximum size of the vehicle or the maximum mass of the vehicle authorized on a road.

5. Method according to any one of the preceding claims, in which the cartographic information is acquired by the navigation system (7) at a predefined acquisition frequency, the acquisitions of the cartographic data being sampled in time or on the positioning of the vehicle (1) on its path at a predetermined sampling frequency identical to or different from the frequency of acquisition of the cartographic information, in particular at a sampling frequency corresponding to an acquisition of the cartographic data every N acquisitions of cartographic information by the navigation system (7), N being an integer.

6. Method according to any one of the preceding claims, in which the map positions (Pdi, ... Pd n ) acquired are at a distance and / or a travel time relative to the vehicle (1) not zero at the time of acquisition (Ti, ... T n) corresponding, in particular the distance, between the acquired map positions (Pdi, ... Pdn) and vehicle locations (Pvi, ... Pv n ) at the acquisition time (Ti, ... T n ) corresponding being substantially constant.

7. Method according to any one of the preceding claims, in which the determination of the representative quantity (mi, ... m n ) for each acquired map data (ai,... a n ) includes the determination, as a function of the speed of the vehicle at the time of acquisition (Ti, .. . T n ), of the measurement time (Ti', . .. T n ') so that the location of the vehicle at the measurement time (Pvi', . .. Pv n ') is substantially equal to the map position (Pdi, ... Pd n ) associated with the cartographic data (ai,. .. a n ), the measurement time (Tf, ... T n') being in particular determined at least from the distance on the most probable path between the vehicle (1) at the acquisition time (Ti, ... T n ) and the map position (Pdi, ... Pd n ) associated with the map data acquired at the time of acquisition and the speed of the vehicle at the time of acquisition.

8. Method according to any one of the preceding claims, in which the determination of the confidence index I c (26) is carried out at an evaluation time by comparing at least part of the differentials (di, . ..d n ), better all the differentials, corresponding to at least part of the acquired cartographic data (ai,. .. a n) upstream of the evaluation time over a time window of a predetermined acquisition duration ending at the evaluation time or corresponding to a predetermined number of successive map data acquisitions going back in time from the evaluation time.

9. Method according to any one of the preceding claims, in which the comparison (24) of the differentials (di, .. ,d n ) includes the determination (26), from the acquired cartographic data (ai,... a n ) having allowed the determination of the confidence index I c and corresponding representative quantities (mi, . .. m n ), of a polynomial correlation equation between the cartographic data and the quantity representative of the cartographic data of a rank lower than a predefined rank, in particular lower than or equal to rank 1, and of maximum correlation index I rwith the acquired map data (ai,... a n ) and the corresponding representative quantities (mi, ... m n ) used to determine the polynomial correlation equation and includes the attribution of the confidence index I c depending on the maximum correlation index I r , in particular the allocation of a confidence index I c greater than a threshold confidence value I cs or equal to a discrete confidence value when the maximum correlation index I r is greater than a threshold correlation value I rs .

10. Method according to the preceding claim, comprising the determination of a new polynomial correlation equation if the correlation index Ir is less than the correlation threshold value I rs over a period greater than a predetermined period, in particular the acquisition period, or over a number of acquisitions successive greater than a predetermined number or if correlation indices of the cartographic data acquired over a period greater than the predetermined period, in particular the acquisition period, or over a number of successive acquisitions greater than the predetermined number are all less than a threshold index value, the correlation indices associated with the or each cartographic data item acquired after the determination of the polynomial correlation equation corresponding to the correlation of the or each cartographic data item acquired after the determination of the polynomial correlation equation and the corresponding representative quantity with the polynomial correlation equation.

11. Method according to any one of claims 9 and 10, in which the comparison (24) of the differentials (di, ...d n) involves the exclusion of a clearly erroneous data item in the determination of the confidence index (26) and / or the search for the polynomial correlation equation (26) having a maximum correlation index I r with the data and representative quantities taken to find the equation.

12. Method according to one of the preceding claims, in which, when an acquired map data (ai,... a n ) presents a differential between the acquired cartographic data (ai,... a n ), in particular after contextualization where appropriate, and the representative quantity (mi, ... m n ) less than a threshold value and when taking into account in the generation of the control data the value of an acquired cartographic data (ai,... a n ), especially after contextualization where appropriate, is less secure than taking into account the value of the representative quantity (mi, ... m n), the confidence index I c is determined to be less than the threshold confidence value I cs temporarily at least for a predetermined duration, in particular at least until the next acquisition of the map data.

13. Method according to any one of the preceding claims, comprising the determination by the navigation system (7) of the most probable path of the vehicle (1) as a function of user input data, information acquired from one or more sensors (3) of the vehicle (1) and / or the location of the vehicle (1) determined by the navigation system (7).

14. Method according to any one of the preceding claims, comprising the generation of a control data item (30) as a function of a security level of the control data item among a plurality of different security levels and of the security index confidence I c, a range of values ​​of the confidence index I c being assigned to each security level to determine whether or not to take into account one or more pieces of cartographic information in the generation of the control data (30).

15. Vehicle equipped with at least one sensor (3) and a navigation system (7) allowing the acquisition at an acquisition time of cartographic information with their associated cartographic position and the location of the vehicle (1) at the acquisition time, said vehicle (1) further comprising a computer program product configured to, when the vehicle (1) is moving: locate the vehicle (1) and acquire cartographic information, carry out a plurality of acquisitions (20, 40) by the navigation system (7), at acquisition times (Ti, ... T n ) different, from a cartographic data (ai,. .. a n) among the map information and the map position (Pdi, ... Pd n ) associated with said cartographic data (ai,... a n ), the acquired map positions (Pdi, ... Pd n ) associated with said acquired cartographic data (ai,... a n ) being geographically different from each other, determine (20), for each acquired cartographic data, a quantity representative of the cartographic data from at least one representative measurement (mi, ... m n ) by at least one sensor (3) at a location of the vehicle (1) (Pvi', ... Pv n ') calculated to be substantially equal to the map position (Pdi, . .. Pd n ) associated with which said cartographic data is acquired (ai,. .. a n ), determine (24), for each acquired cartographic data (ai,... a n ), a differential (di, ...d n ) between the acquired map data (ai,... an ) and the sizes (mi, ... m n ) corresponding, determine a confidence index I c (26) based on the comparison of at least a portion of the differentials (24) of the map data. generate a control data item (30) based on the comparison (28) of the confidence index I c at a threshold confidence value I cs predetermined.

Images