METHOD FOR DETERMINING THE TYPE OF LANE USED BY A MOTOR VEHICLE

Description

TECHNICAL FIELD OF THE INVENTION

[0001] The present invention relates generally to motor vehicles.

[0002] It relates more specifically to a process for determining the types of roads on which the motor vehicle has traveled. STATE OF THE ART

[0003] We know from document WO2018011500 a method for determining whether the road used by a motor vehicle is in good condition or not.

[0004] While satisfactory in itself for adjusting vehicle suspension in real time or choosing routes with good road conditions, this method does not provide as much information as one might want about the roads used by motor vehicles. In particular, it does not allow for precise knowledge of how each motor vehicle is being used.

[0005] This usage could be determined by exploiting data from vehicle navigation systems (e.g., GPS systems), but this exploitation is not desired here as it is considered too invasive with regard to the personal data of motor vehicle users.

[0006] Neural network methods such as those in US20020128751 require a significant number of runs and impose a fixed choice of method parameters and environments to be categorized, without easy evolution since all the learning would then have to be restarted. PRESENTATION OF THE INVENTION

[0007] In order to remedy the aforementioned drawback of the state of the art, the present invention proposes to determine the use made of the vehicle in a non-invasive manner, that is to say without exploiting any personal data of the users of the motor vehicle.

[0008] More specifically, the invention proposes a method for determining the type of track taken by a motor vehicle, comprising the following steps: acquisition by a computer, during at least one working interval, of the values ​​of at least two parameters which are taken from measurements made by sensors fitted to the motor vehicle and which are related to the dynamics of the vehicle, determination by the computer, for each parameter and for the same sampling period included in each of said working intervals, of a probability coefficient that the vehicle is on one type of road rather than another, depending on the values ​​taken by the parameter during said sampling period, and deduction by the computer, depending on the probability coefficients determined, of the type of road used by the motor vehicle during said sampling period.

[0009] Thus, thanks to the invention, we exploit the values ​​of parameters which are generally already measured on motor vehicles (speed, engine speed, steering wheel angle...) to deduce, on the basis of a statistical and probabilistic method, the type of road probably taken by the motor vehicle.

[0010] This method requires only one preliminary road test operation to determine thresholds for each parameter. Below these thresholds, it can be statistically observed that the test vehicles generally operated on a certain type of road, and above them, it can be statistically observed that the test vehicles generally operated on a different type of road. The resulting thresholds are then valid for all passenger vehicles.

[0011] The method according to the invention then exploits these thresholds and the data measured by motor vehicles to determine the probability that the vehicle is on one type of lane rather than another.

[0012] The method proposed here is therefore easy to implement. It requires no additional technical equipment on the vehicles, and in particular no geolocation chip. It is non-invasive with regard to the personal data of vehicle users, and its results are easily usable in real time or in after-sales service, to adjust vehicle settings or to improve a vehicle model based on its usage.

[0013] Other advantageous and non-limiting features of the process according to the invention, taken individually or in all technically possible combinations, are as follows: The type of road is selected from at least motorways, urban roads, country roads, and possibly also mountain roads; the values ​​of the parameters are averaged over their working intervals, and the averages obtained are each compared to at least a predetermined threshold to deduce the probability coefficients; a first parameter being the longitudinal speed of the motor vehicle, a filtering operation is planned for the average of the values ​​of the first parameter during which, if, for a working interval, the average is less than a determined threshold, the type of road for each sampling period included in said working interval is determined according to the type of road deduced for the sampling period that follows or precedes said working interval; the working intervals used for the acquisition of the values ​​of the two parameters only partially coincide;The working intervals are intervals of distance or time; the acquisition and determination steps are carried out with at least five distinct parameters, and preferably with an odd number of distinct parameters, for example, seven; the parameters are chosen from: a longitudinal speed of the motor vehicle, a longitudinal acceleration of the motor vehicle, an angular velocity of an element of a steering system of the motor vehicle, a torque developed by an engine equipping the motor vehicle, a gradient of the road traveled by the motor vehicle, a duration during which the engine speed is below a predetermined threshold, and a number of turns encountered by the vehicle; the values ​​of said parameters are derived solely from measurements taken by internal sensors of the vehicle; said internal sensors communicate exclusively with internal components of the motor vehicle.

[0014] The invention also relates to a use of the aforementioned determination method, in which, the deduction step being implemented during the driving of the motor vehicle, a subsequent operation is provided during said driving of adjusting at least one internal setting of the motor vehicle according to the type of track determined.

[0015] It also relates to the use of the aforementioned determination method, in which, the vehicle having made several journeys, a subsequent operation is planned to determine the main use of the motor vehicle or to determine the wear of at least one wear part of the motor vehicle according to the types of roads detected on said several journeys.

[0016] Of course, the different features, variants and embodiments of the invention can be combined with each other in various ways as long as they are not incompatible or mutually exclusive. DETAILED DESCRIPTION OF THE INVENTION

[0017] The description that follows, with regard to the attached drawings, given by way of non-limiting examples, will make it clear what the invention consists of and how it can be carried out.

[0018] Regarding the attached drawings: [ Fig. 1 ] is an example of a matrix generated as part of the process according to the invention; [ Fig. 2 ] is a first type of logic diagram used in the process according to the invention; [ Fig. 3 ] is a second type of logic diagram used in the process according to the invention; and [ Fig. 4 ] is a third type of logic diagram used in the process according to the invention.

[0019] The present invention relates to the exploitation of internal data acquired by a motor vehicle.

[0020] Such a motor vehicle may take the form of a car, a truck or a motorcycle, for example.

[0021] The example considered here is a car with four wheels, a chassis, an engine and a steering system.

[0022] The engine can be either of the internal combustion and / or electric type.

[0023] The steering system includes a steering wheel and a set of shafts and possibly actuators that allow the steering wheels of the vehicle to be turned according to the angle of the steering wheel.

[0024] This motor vehicle is also typically equipped with a computer including a processor, memory, and an input / output interface.

[0025] Thanks to its interface, the computer is adapted to communicate with different components of the vehicle, via a CAN type data bus (from the English "Controller Area Network").

[0026] The computer can thus receive input signals from various internal sensors fitted to the vehicle.

[0027] The term "internal sensor" refers to sensors that are integrated into the vehicle and communicate (preferably) only with the vehicle's various control units and the diagnostic port. A GPS chip or a mobile phone chip cannot therefore be considered an internal sensor. Advantageously, these internal sensors measure one-dimensional physical parameters.

[0028] The input signals considered here are all related to the vehicle's dynamics, that is, its movement. This can include data relating to the engine, the steering system, the wheels, the shape of the road, the vehicle's position, speed, acceleration or acceleration derivative, atmospheric pressure, and so on.

[0029] In the embodiment presented here, the calculator receives several input data, namely: the longitudinal speed of the motor vehicle, measured for example by means of a wheel angular speed sensor, the lateral acceleration experienced by the motor vehicle in the plane of the road, in a direction orthogonal to the direction of travel of the vehicle (measured for example by means of an accelerometer), the angular position of the steering wheel (measured for example by means of a steering wheel angular position sensor), the atmospheric pressure (measured for example by means of a pressure sensor), the speed of the internal combustion engine (measured for example by means of a crankshaft angular speed sensor), and the position or pressure exerted on the accelerator pedal (measured for example by means of a sensor placed on this pedal).

[0030] The memory stores information used in the process described below.

[0031] In particular, it stores several basic database registers, whose matrix-like architectures will be detailed later in this presentation.

[0032] It also stores a computer application, consisting of computer programs including instructions whose execution by the processor allows the computer to implement the process described below.

[0033] The objective of this process is to determine, for example on part of a journey or on an entire journey or even on several journeys, the types of roads used by the motor vehicle.

[0034] In the embodiment presented here, we want to know the distribution of driving done on motorway, mountain road, in built-up area and on other roads (hereafter referred to as country roads) by a motor vehicle.

[0035] To do this, in a first step, the computer acquires the aforementioned input data via internal sensors.

[0036] This acquisition is performed at a given frequency, for example 1 Hz, called the sampling frequency. The sampling period δt will then be defined as a period of time (here, one second) between two successive acquisitions of the input data.

[0037] The calculator deduces, based on the measured and acquired input data, the values ​​of several parameters P1, P2, P3, P4, P5, P6, and P7. There are at least two of these parameters. Preferably, they are an odd number to minimize the risk of having two sets of parameters of the same size that would produce contradictory results. In this case, there are seven parameters: the longitudinal speed P 1 of the motor vehicle, the longitudinal acceleration P 2 of the motor vehicle, the angular speed P 3 of the flywheel, the torque P 4 developed by the engine, the slope P 5 of the road taken by the motor vehicle, the duration P 6 during which the engine speed is below a predetermined threshold, and the number of turns P 7 encountered by the vehicle.

[0038] The longitudinal speed P1 is directly measured by an internal sensor, the angular speed P3 of the flywheel is derived from the value of the angle at the flywheel measured by the sensor of the same name, and the duration P6 during which the engine speed is below a predetermined threshold is calculated by a counter using the sampling frequency and the measurement of the engine speed from, for example, a tachometer.

[0039] The longitudinal acceleration P2 could be measured by an accelerometer. However, here it is preferentially obtained by differentiating the longitudinal velocity of the vehicle with respect to time.

[0040] The torque P4 developed by the engine is determined by an engine control unit and passes through the CAN type data bus; it is, for example, a function of the engine speed and the position of the accelerator pedal.

[0041] The slope P 5 of the road used by the motor vehicle is obtained as a function of the atmospheric pressure differences measured over time.

[0042] The number of P7 turns taken by the vehicle is determined based on the lateral acceleration experienced by the vehicle and measured by an accelerometer.

[0043] Once it has obtained the values ​​of these seven parameters, the computer (or another computer that is not necessarily on board the vehicle) will attempt to determine, for each parameter, the type of lane on which the vehicle was located.

[0044] In the case of an on-board computer, calculations are performed internally within the vehicle and can allow for real-time modification of certain vehicle parameters (suspension settings, power steering assistance, etc.). Conversely, in the case of a remote-controlled computer, calculations can be performed after driving is complete, for example, to estimate tire wear or particulate filter clogging, based on the types of roads driven.

[0045] The idea is that each parameter will allow the calculation of a probability coefficient that the vehicle is on one type of lane rather than another. These probability coefficients can then be combined to determine with greater reliability the type of lane the vehicle is using.

[0046] To determine these probability coefficients, the calculator will generate seven matrices, each associated with one of the seven parameters.

[0047] It operates in essentially the same way for the majority of these parameters, which can be summarized as follows.

[0048] First, the computer divides the route into work intervals. These can be time or distance intervals. Each work interval preferably has a duration strictly greater than the sampling period δt (this duration is preferably at least ten times greater than the sampling period). Preferably, the value of the work intervals is not unique for all parameters but specific to each parameter or groups of parameters.

[0049] The calculator then calculates an average of the parameter during this working interval. It then compares this average with one or more predefined thresholds.

[0050] At this stage, it can be specified that the thresholds used can be developed following a campaign of tests on different types of known roads.

[0051] The result of this comparison will then allow him to define the most probable type of lane that was used by the vehicle during this work interval.

[0052] Each matrix will then be constructed by assigning the value 1 to the column associated with the most likely type of route taken, and assigning the value 0 to the other columns.

[0053] As shown by figure 1 Each matrix (associated with a parameter) here has four columns: a first column T1 associated with a road in an urban area, a second column T2 associated with a mountain road, a third column T3 associated with a rural road, and a fourth column T4 associated with a motorway. The number of columns and their type are not limited; other types of roads could also be included, for example, circuits or unpaved roads (tracks).

[0054] Each matrix has as many rows as there are sampling periods δt considered.

[0055] So, if the result of the aforementioned comparison indicates, for example, that the type of road is "motorway" during a working interval of, for example, 100 sampling periods δt, the computer is programmed to complete the matrix corresponding to the parameter considered with one hundred new identical rows in which the coefficients in column T4 are all set to one and in which all other coefficients are set to zero.

[0056] The values ​​thus assigned to the columns of the matrices form probability coefficients equal to 0 or 1.

[0057] We can now describe in detail how the seven matrices associated with the seven parameters considered are precisely obtained.

[0058] In the embodiment presented here, the measured parameter values ​​are used retrospectively, that is, after the vehicle has completed one or more full journeys, and these journeys are then analyzed by the computer. For the sake of simplicity, we will assume that the vehicle has completed only one journey.

[0059] The first parameter considered is the longitudinal speed P1 of the motor vehicle, values ​​of which were successively measured at successive and regular time steps.

[0060] This parameter allows us to differentiate between urban roads and highways from other types of roads. Indeed, the speed limit will be very high on highways and very low on urban roads.

[0061] We can describe how the values ​​of this parameter are used by referring to the figure 2 .

[0062] As this shows figure 2 , during a first step (here represented by the rectangles referenced Eb1 and Eb2), the computer acquires the successive values ​​taken by the longitudinal speed P 1 of the motor vehicle along the route.

[0063] During a second step Eb3, the computer divides the path into several working intervals Δ 1.

[0064] Here, these working intervals Δ1 do not have regular time or distance intervals. Rather, they are divided according to the second derivative of the velocity with respect to time (which is preferably filtered by a first- or second-order, or even third-order, low-pass filter to remove frequencies above 0.01 Hz). Thus, each time this filtered second derivative reaches zero, corresponding to an inflection point of the velocity, a new working interval Δ1 is created.

[0065] Each working interval Δ 1 therefore corresponds to a variable number of sampling periods δt.

[0066] This segmentation solution ensures that the speed within each working interval Δ 1 remains substantially constant and that the average which will then be calculated will have a usable meaning.

[0067] During a third step Eb4, the computer then calculates the arithmetic mean, within each working interval Δ 1, of the longitudinal speed P 1 of the motor vehicle.

[0068] At the end of this step, the computer performs an average speed filtering operation. If one of them is less than or equal to a predetermined speed threshold (which is preferably between 1 and 10 km / h), the computer assigns the same type of road to this interval as to the following interval (or previous one if it does not exist).

[0069] Thus, in the case of a stop at a motorway rest area, the time interval corresponding to the break (zero average) will be assigned the type "motorway", since the following interval will be considered.

[0070] This situation can indeed be considered equivalent to a prolonged vehicle stop. If this filtering operation were not carried out, there would be a risk that a stop at a motorway rest area would be considered as driving in a city.

[0071] Then, during steps Eb5 and Eb6, the calculator compares the value of the average speed calculated within the working interval Δ 1 with two predefined thresholds (a minimum speed threshold and a maximum speed threshold).

[0072] As an example, the minimum speed threshold could be around 50 km / h, and the maximum speed threshold could be between 80 and 90 km / h.

[0073] If the average longitudinal speed P1 is below the minimum speed threshold, the vehicle can be assumed to be in an urban area. Therefore, during step Eb7, the computer assigns the value 1 to column T1 (roads in urban areas) and the value 0 to all rows of the speed matrix corresponding to sampling periods δt within the working interval Δ1.

[0074] If the average longitudinal speed value P1 is above the maximum speed threshold, it can be assumed that the vehicle is on a motorway. Then, during step Eb9, the computer assigns the value 1 to column T4 (motorway) and the value 0 to the other columns in the same rows of the speed matrix.

[0075] If the average longitudinal speed value P1 is between the minimum and maximum speed thresholds, it can be assumed that the vehicle is on a mountain road or a country road. Then, during step Eb8, the computer assigns the value 1 to columns T2 and T3 of the speed matrix, and the value 0 to the other columns.

[0076] The second parameter considered is the longitudinal acceleration P2 of the motor vehicle.

[0077] This parameter allows us to differentiate between highways and rural, mountain, and urban roads. Indeed, longitudinal acceleration P2 on highways is on average lower than elsewhere.

[0078] We can describe how the values ​​of this parameter are used by referring again to the figure 2 .

[0079] As this shows figure 2 In the first step Eb1, the computer measures successive values ​​of the longitudinal speed of the motor vehicle along the route, then it derives this longitudinal speed as a function of time. In the second step Eb2, it acquires the time.

[0080] During a third step Eb3, the computer divides the path into several working intervals Δ 2.

[0081] Here, these work intervals Δ2 are regular time intervals of at least one minute, so that the average subsequently derived from them is representative of the portion of the journey traveled by the vehicle. The work interval Δ2 is preferably chosen to be two minutes.

[0082] During a fourth step Eb4, the computer then calculates the arithmetic mean, within each working interval Δ 2, of the longitudinal acceleration P 2 of the motor vehicle.

[0083] Then, during steps Eb5 and Eb6, the calculator compares the average value of the longitudinal acceleration P2 with two predefined thresholds (a minimum acceleration threshold and a maximum acceleration threshold).

[0084] The minimum acceleration threshold is for example between 0.1 and 0.4 m / s², while the maximum acceleration threshold is between 0.6 and 1 m / s².

[0085] If the average value of the longitudinal acceleration P2 is less than the minimum acceleration threshold, it can be assumed that the vehicle is on a motorway. Then, during a step Eb7, the computer assigns to all rows of the acceleration matrix that correspond to sampling periods δt within the working interval Δ2, the value 1 to column T4 (motorways) and the value 0 to the other columns.

[0086] If the average longitudinal acceleration value P2 is greater than the maximum acceleration threshold, it can be assumed that the vehicle is on a country road or a mountain road. Therefore, during step Eb9, the computer assigns the value 1 to columns T2 and T3 and the value 0 to the other columns.

[0087] If the average value of the longitudinal acceleration P2 is between the minimum and maximum acceleration thresholds, it can be assumed that the vehicle is on a built-up or rural road. Therefore, during step Eb8, the computer assigns the value 1 to columns T1 and T3 and the value 0 to the other columns.

[0088] The third parameter considered is the rotational speed P3 of the steering wheel of the motor vehicle.

[0089] This parameter allows us to differentiate between highways and rural, mountain, and urban roads. Indeed, the steering wheel rotation speed P3 on highways is on average lower than elsewhere.

[0090] We can describe how the values ​​of this parameter are used by referring again to the figure 2 .

[0091] As this shows figure 2 , during a first stage Eb1, the computer acquires the successive values ​​taken by the longitudinal speed of the motor vehicle along the route.

[0092] During a second stage Eb2, the computer acquires the values ​​successively taken by the steering wheel angle along the route.

[0093] During a third step Eb3, the computer divides the path into several working intervals Δ 3.

[0094] Here, these working intervals Δ3 are regular distance intervals of at least 100 meters, so that the average subsequently derived from them is representative of the presence or absence of turning zones on the portion of the route traveled by the vehicle. The distance interval is preferably chosen to be one kilometer.

[0095] In practice, the journey under consideration is divided into working intervals Δ 3 by integrating the measured speed values ​​to obtain distance intervals.

[0096] During a third step Eb4, the computer differentiates with respect to time the "steering wheel angle" function in order to obtain the rotation speed P 3 of the steering wheel at each sampling period δt.

[0097] Then, the computer calculates the arithmetic mean, within each working interval Δ 3 of the journey, of the rotation speed P 3 of the steering wheel of the motor vehicle.

[0098] During steps Eb5 and Eb6, the computer compares the average value of the rotational speed P3 of the flywheel with two predefined thresholds (a minimum angular velocity threshold and a maximum angular velocity threshold).

[0099] The minimum angular velocity threshold is, for example, approximately equal to 1° / s while the maximum angular velocity threshold is around 5° / s.

[0100] If the average value of the rotational speed P3 is less than the minimum angular velocity threshold, it can be assumed that the vehicle is on a motorway. Then, during a step Eb7, the computer assigns to all rows of the angular velocity matrix that correspond to sampling periods δt within the working interval Δ3, the value 1 to column T4 (motorways) and the value 0 to the other columns.

[0101] If the average rotational speed P3 exceeds the maximum angular velocity threshold, it can be assumed that the vehicle is in an urban area or on a mountain road. Therefore, during step Eb9, the computer assigns the value 1 to columns T1 and T2 and the value 0 to the other columns.

[0102] If the average value of the rotational speed P3 is between the minimum angular velocity threshold and the maximum angular velocity threshold, it can be assumed that the vehicle is on a mountain or country road. Then, during step Eb8, the computer assigns the value 1 to columns T2 and T3 and the value 0 to the other columns.

[0103] The fourth parameter considered is the torque P4 developed by the motor vehicle's engine.

[0104] This parameter allows us to differentiate between country roads and urban roads, and motorways and mountain roads. Indeed, the torque delivered by the engine in mountainous areas or on motorways will, on average, be higher than elsewhere.

[0105] We can describe how the values ​​of this parameter are used by referring to the figure 4 .

[0106] As shown by figure 4 , during a first step Ea1, the computer acquires the successive values ​​taken by the longitudinal speed of the motor vehicle along the route.

[0107] In a second step, Ea2, the computer acquires the successive values ​​of the torque P4 along the path. The torque values ​​considered here are not absolute values, expressed in Nm, but rather relative values ​​with respect to the maximum torque the engine can develop (in practice, these are percentages of maximum torque). In this way, the same algorithm can be used on vehicles equipped with different engines.

[0108] During a third step Ea3, the computer divides the path into several working intervals Δ 4.

[0109] Here, these operating intervals Δ4 are irregular, chosen so that the torque does not vary excessively within each interval. A new interval is thus defined each time the longitudinal speed reaches at least one predefined longitudinal speed value. Preferably, a new interval is defined each time the longitudinal speed reaches 20 km / h and 90 km / h.

[0110] During a fourth step Ea4, the calculator calculates the arithmetic mean, within each working interval Δ 4, of the representative values ​​of the torque.

[0111] Then, during a fifth step Ea5, for each working interval Δ 4, the calculator compares the average value of the torque P 4 with a predefined representative torque threshold.

[0112] This threshold can, for example, be chosen between 15 and 30% of the maximum torque that the engine can develop.

[0113] If the average torque value P4 is below the torque threshold, it can be assumed that the vehicle is in urban areas or on country roads. Then, during step Ea6, the computer assigns the value 1 to columns T1 and T3 of the torque matrix, and the value 0 to all rows of the torque matrix that correspond to sampling periods δt within the working interval Δ4.

[0114] Conversely, if the average value of the P4 torque is above the torque threshold, it can be assumed that the vehicle is on a mountain road or a motorway. Therefore, during step Ea7, the computer assigns the value 1 to columns T2 and T4 and the value 0 to the other columns.

[0115] The fifth parameter considered is the slope P 5 of the road taken by the motor vehicle.

[0116] This parameter allows us to distinguish mountain roads from other types of roads. Indeed, the average gradient of roads in mountainous areas is higher than elsewhere.

[0117] We can describe how the values ​​of this parameter are used by referring to the figure 4 .

[0118] As shown by figure 4 , during a first step Ea1, the computer acquires the different values ​​of the longitudinal speed of the motor vehicle along the route.

[0119] During a second stage Ea2, the computer acquires the different values ​​taken by the atmospheric pressure along the route.

[0120] During a third step Ea3, the computer divides the path into several working intervals Δ5.

[0121] Here, these working intervals Δ 5 are regular distance intervals, chosen so that the results are representative of the type of road used by the vehicle.

[0122] This involves intervals of more than 100 meters, and preferably intervals of 1 km. For example, the slope is calculated for 100m sections and averaged over a 1km interval. The route is then divided into working intervals Δ5 by integrating the measured speed values ​​to obtain distance intervals.

[0123] During a fourth step Ea4, the calculator calculates the arithmetic mean, within each working interval Δ 5, of the slope values ​​P 5.

[0124] To do this, the computer determines, based on atmospheric pressure, the altitude at which the vehicle was located at least at the beginning and end of the working interval Δ 5. Here, this altitude is determined more frequently: it is determined every 100 meters. More precisely, it is the difference in altitude that is determined between the beginning and end of each 100-meter segment.

[0125] Here, the calculation of the altitude f(t) as a function of the atmospheric pressure p(t) is performed using the following equation: f t = 288.15 0.0065 . 1 − p t 1013.25 1 5.255

[0126] This altitude calculation is approximate because it would need to incorporate temperature and meteorological data to obtain a more precise value. However, this is of little importance since only the altitude variation is used here.

[0127] This altitude f(t) is then filtered using a low-pass filter of order 1 or 2, or even 3, in order to remove frequencies above 0.02 Hz.

[0128] The calculation of the slope g(t) every 100 meters of distance traveled is then performed as a function of this filtered altitude ff(t) using the following equation: g t = 100 . f f t i − f f t i + 1 100 2 − f f t i − f f t i + 1 2

[0129] In this equation, the variable ti corresponds to the instant at the beginning of the 100-meter interval, and the variable t i+1 corresponds to the instant at the end of this 100-meter interval.

[0130] The arithmetic mean of the slope values ​​P 5, within each working interval Δ 5 of 1 km length, is then calculated as a function of the ten slope values ​​g(t) obtained.

[0131] Then, during a step Ea5, the calculator compares the average value of the slope P 5 with a predefined slope threshold (here of the order of 5%).

[0132] If the average slope value P5 is below the slope threshold, it can be assumed that the vehicle is in an urban area, on a country road, or on a motorway. Then, during step Ea6, the computer assigns the value 1 to columns T1, T3, and T4, and the value 0 to the last column, for all rows of the slope matrix that correspond to sampling periods δt within the working interval Δ5.

[0133] Conversely, if the average slope value P5 exceeds the slope threshold, it can be assumed that the vehicle is on a mountain road. Therefore, during step Ea7, the computer assigns the value 1 to column T2 and the value 0 to the other columns.

[0134] The sixth parameter considered is the duration P6 during which the vehicle is at zero or near-zero speed (traffic lights, stop signs, yield signs, etc.), taking into account, for example, the engine speed. Here, we will consider the vehicle to be at zero or near-zero speed when the engine is idling (if it is an internal combustion engine) or practically stopped (if it is an electric motor).

[0135] This parameter allows us to distinguish between roads in urban areas (on which the vehicle is forced to stop, due to stop signs, traffic lights and various other events) and other types of roads.

[0136] We can describe how the values ​​of this parameter are used by referring to the figure 4 .

[0137] As shown by figure 4 , during a first step Ea1, the computer acquires the successive values ​​taken by the longitudinal speed of the motor vehicle along the route.

[0138] During a second stage Ea2, the computer acquires the successive values ​​taken by the engine speed and compares it to the critical speed representative of the vehicle stopping related to the technology.

[0139] During a third step Ea3, the computer divides the path into several working intervals Δ 6.

[0140] Here, these working intervals Δ 6 are regular distance intervals, chosen so that the results are representative of the type of road used by the vehicle.

[0141] This involves intervals of more than 100 meters, and preferably intervals of 1 km in length. The route is then divided into working intervals Δ 6 by integrating the measured speed values ​​to obtain distance intervals.

[0142] During a fourth step Ea4, the computer calculates the arithmetic mean, within each working interval Δ 6, of the durations during which the engine speed was below a speed threshold (here 2000 revolutions per minute if it is an internal combustion engine and a few hundred revolutions per minute if it is an electric motor) and during which the longitudinal speed of the vehicle was below another threshold (here 10 km / h).

[0143] It should be noted here that the use of a longitudinal speed threshold makes it possible in particular to apply the process described here to vehicles equipped with an option which allows the engine to be switched off or put into idle when the vehicle is on a steep descent and at high speed, which helps to reduce its consumption.

[0144] Then, during a step Ea5, the calculator compares the average duration value P 6 with a predefined duration threshold.

[0145] If the average duration value P6 is greater than the duration threshold, it can be assumed that the vehicle is in an urban area or similar. Then, during a step Ea7, the computer assigns to all rows of the duration matrix that correspond to sampling periods δt within the working interval Δ6, the value 1 in column T1 and the value 0 in the other columns.

[0146] Conversely, if the average duration value P6 is below the duration threshold, it can be assumed that the vehicle is on a country road, a mountain road, or a motorway. Therefore, during step Ea6, the computer assigns the value 1 to columns T2, T3, and T4, and the value 0 to column T1.

[0147] The seventh parameter considered is the number P7 of turns made by the vehicle.

[0148] This parameter allows highways to be distinguished from other types of roads.

[0149] We can describe how the values ​​of this parameter are used by referring to the figure 3 .

[0150] As shown by figure 3 , during a first stage Ec1, the computer acquires the successive values ​​taken by the longitudinal speed of the motor vehicle along the route.

[0151] During a second stage Ec2, the computer acquires the successive values ​​taken by the lateral acceleration undergone by the motor vehicle.

[0152] During a third step Ec3, the computer divides the path into several working intervals Δ 7.

[0153] Here, these working intervals Δ 7 are regular distance intervals, chosen so that the results are representative of the type of road used by the vehicle.

[0154] This involves intervals of more than 100 meters, and preferably intervals of 1.5 km. The route is then divided into working intervals Δ 7 by integrating the measured speed values ​​to obtain distance intervals.

[0155] During a fourth step Ec4, the calculator calculates, within each working interval Δ 7, the number of times this lateral acceleration has exceeded a predetermined threshold, here chosen to be equal to 0.2 times the gravitational constant (which can be written as 0.2G).

[0156] Then, during steps Ec5, Ec6, Ec7, the calculator compares this number with three predefined thresholds, including a minimum threshold, an intermediate threshold, and a maximum threshold.

[0157] For example, the minimum threshold could be equal to 1, the intermediate threshold could be equal to 5, and the maximum threshold could be equal to 10.

[0158] If the number is less than the minimum threshold, it can be assumed that the vehicle is on a motorway. Then, during a step Ec8, the computer assigns to all rows of the turn matrix that correspond to sampling periods δt within the working interval Δ7, the value 1 in column T4 and the value 0 in the other columns.

[0159] If the number is between the minimum and intermediate thresholds, it can be assumed that the vehicle is on a country road. Then, during an Ec9 step, the computer assigns the value 1 to column T3 and the value 0 to the other columns.

[0160] If the number is between the intermediate threshold and the maximum threshold, it can be assumed that the vehicle is on a country road or a mountain road. Then, during an Ec10 step, the computer assigns the value 1 to columns T2 and T3 and the value 0 to the other columns.

[0161] Finally, if the number is higher than the maximum threshold, it can be assumed that the vehicle is in an urban area or on a mountain road. Then, during an Ec11 step, the computer assigns the value 1 to columns T1 and T2 and the value 0 to the other columns.

[0162] Thus the calculator is able to establish seven matrices for the seven parameters, whose respective rows are associated with corresponding sampling periods δt, and whose columns are associated with the same four types of channels considered here.

[0163] The seven matrices are then combined so that the type of lane taken by the vehicle on each sampling period δt is determined according to the values ​​taken by the seven parameters P1, P2, P3, P4, P5, P6, P7.

[0164] We could assign greater weight to some of these parameters. However, here, the decision was made to give each parameter the same weight. Indeed, a parameter might prove to be a better criterion for determining the type of track in a given situation, but a poor criterion in other situations. Therefore, giving each parameter the same weight is the solution that yields the best results.

[0165] Combining the seven matrices simply involves summing them. The result of this sum is a final matrix of the same dimensions as the other seven, with 4 columns corresponding to the four types of traffic lanes, and a number of rows equal to the number of sampling periods δt considered.

[0166] All the coefficients of this final matrix are then probability coefficients which take the form of natural numbers between 0 and 7, inclusive.

[0167] For each sampling period δt, i.e. for each row of the final matrix, the calculator selects the column(s) where the coefficient is highest.

[0168] If the highest probability coefficient of the row under consideration is found only in one column, the type of road associated with that column and that sampling period δt is selected.

[0169] In the case where several columns have the same highest probability coefficient, the result "indeterminate" is assigned to the type of lane of this sampling period δt.

[0170] This yields a vector whose rows correspond to the considered sampling periods δt and whose coefficients indicate the associated road type. This coefficient can thus be equal to: 0 if the type of road is undetermined, 1 if the type of road is an urban road, 2 if the type of road is a mountain road, 3 if the type of road is a country road, 4 if the type of road is a motorway.

[0171] Preferably, this vector can be corrected using two filtering operations.

[0172] The first filtering operation is as follows.

[0173] If a sampling period δt corresponds to a channel of indeterminate type and the preceding and following sampling periods δt are of the same type, which is not indeterminate, then this latter type of channel is assigned to this sampling period δt.

[0174] As an example, we can consider three successive sampling periods δt. If the first and third sampling periods δt correspond to a road of type "highway" and the second sampling period δt corresponds to a road of type "undetermined", then the road type "highway" is assigned to the second sampling period δt.

[0175] It should be noted that this situation generally occurs when the vehicle passes over an interchange, between two highway lanes, or when slowing down in a work zone.

[0176] The second filtering operation is as follows.

[0177] If a first type of channel assigned to several successive sampling periods δt (of total duration less than one minute) is different from a second type of channel assigned to the sampling periods preceding and following these successive sampling periods δt, then the second type of channel is assigned to these successive sampling periods δt.

[0178] As an example, we can consider a hundred successive sampling periods δt. If the first 40 and last 40 sampling periods δt are associated with a road type "highway", while the middle 20 sampling periods δt are associated with a road type "country road", we modify the association of these middle 20 sampling periods δt so as to assign them the road type "highway".

[0179] It should be noted that this situation also occurs when the vehicle passes over an interchange, between two lanes of a motorway, or when slowing down in a construction zone...

[0180] The result of these filtering operations will be a vector allowing us to know the percentage of driving carried out on each type of lane by the motor vehicle during its last journey.

[0181] At this stage, it can be noted that the higher the number of parameters considered, the more reliable the results will be.

[0182] The use of seven parameters thus makes it possible to dilute an error in qualifying a type of track made due to the value of one of the seven parameters, thanks to the other six parameters.

[0183] For example, in the case of a mountain descent, the torque parameter might incorrectly classify this section of the journey as a "country road". In this case, this error will be compensated for by the fact that the parameters relating to the number of turns, steering angle, gradient, and longitudinal acceleration will correctly identify this part of the journey as a mountain road.

[0184] The number of parameters considered will preferably be odd to reduce the number of chances of obtaining the "indeterminate" qualification for the types of routes.

[0185] It should also be noted that the working intervals used for the seven parameters are different from one another. Thus, at least one working interval is different from the others. Preferably, the same working interval will not be used for more than three different parameters.

[0186] In this way, the measurements used to construct the seven matrices will have been carried out during periods offset from each other, which will provide more reliable results.

[0187] Indeed, for example, driving in the city at a steady 50 km / h without any turns for 1 kilometer could be interpreted as highway driving because all the parameters correspond to this profile, except for the average speed and average torque. However, having operating intervals that do not coincide in time will solve this problem, since an operating interval associated with one parameter will be linked to other operating intervals that will have started earlier for some other parameters (before the vehicle reaches that straight street) or later for other parameters (after the vehicle has left that straight street).

[0188] We can now give several industrial applications of this method of determining the type of track taken by a motor vehicle.

[0189] The first application is in after-sales service. When vehicles are brought back to dealerships for servicing, it's possible to determine the percentage of driving on each type of road for each car. This information can then be used to better adapt future car models to their most common usage patterns. In other words, this will allow for the development of vehicles that effectively meet the needs of future customers.

[0190] Furthermore, as an example, it is possible to adapt tire pressure, shock absorbers, steering ratio, and even audio system settings to the driving conditions most commonly experienced by users of the vehicle model in question.

[0191] Furthermore, it is possible to better predict the wear and tear of vehicle parts in general (by vehicle model) or specifically (vehicle by vehicle). This will allow users to be informed of the best time to have their vehicle serviced, to change their tires, brake pads, etc.

[0192] It is also possible to use the results of this process in real time. Thus, when the vehicle is moving and it is detected that it is on a particular type of road (for example, on a mountain road), the shock absorbers and steering ratio can also be adapted to this type of road.

[0193] The present invention is in no way limited to the embodiment described and represented, but a person skilled in the art will be able to make any variation in accordance with the invention.

[0194] Thus, when constructing each matrix, we could plan to assign values ​​other than 0 and 1 to the probability coefficients. For example, when the parameter does not allow us to determine whether the type of road is more of a mountain road or a country road, we could assign coefficients equal to 0.5 to columns T2 and T3.

[0195] According to another embodiment of the invention, parameters based on information obtained from sensors other than those mentioned above could be used. For example, information obtained from the cameras installed in the motor vehicle could be used. This could include, for instance, the maximum speed indicated on a road sign, or any other information readable on a road sign.

Claims

1. Method for determining a type of track (T1, T2, T3, T4) used by a motor vehicle, comprising the steps of: - acquisition, for at least two parameters (P1, P2, P3, P4, P5, P6, P7), and for each parameter during at least one respective working interval (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7), of the values of said parameters (P1, P2, P3, P4, P5, P6, P7) which are derived from measurements taken by sensors fitted to the motor vehicle and which relate to the vehicle dynamics, - determination by a computer, for each parameter (P1, P2, P3, P4, P5, P6, P7) and for all sampling periods (δt) within each of the at least one working interval (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7) associated with said parameter, of a coefficient of probability that the vehicle is on one type of track (T1, T2, T3, T4) rather than another, on the basis of the values assumed by the parameter (P1, P2, P3, P4, P5, P6, P7) during said working interval (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7), and - deduction by the calculator, on the basis of the determined probability coefficients, of the type of track (T1, T2, T3, T4) taken by the motor vehicle during one same sampling period (δt) belonging to working intervals (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7) each associated with a respective one of said parameters (P1, P2, P3, P4, P5, P6, P7), wherein the working intervals (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7) used for the acquisition of the values of said at least two parameters (P1, P2, P3, P4, P5, P6, P7) coincide with one another only in part.

2. Determination method according to the preceding claim, wherein the type of track (T1, T2, T3, T4) is selected from at least motorways (T4), urban roads (T1), country roads (T3) and optionally also mountain roads (T4).

3. Determination method according to one of the preceding claims, wherein the values of the parameters (P1, P2, P3, P4, P5, P6, P7) are averaged over their working intervals (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7), and the averages obtained are each compared with at least one predetermined threshold to deduce the probability coefficients therefrom.

4. Determination method according to the preceding claim, wherein, with a first of the parameters being the longitudinal speed (P1) of the motor vehicle, a filtering operation is provided during which operation if, for one working interval, the average is lower than a predetermined threshold, the probability coefficient for each sampling period (δt) included in said working interval is determined on the basis of the probability coefficient determined for the sampling period following or preceding said working interval.

5. Determination method according to one of the preceding claims, wherein the working intervals (Δ1, Δ2, Δ3, Δ4, Δ5, Δ6, Δ7) are distance or time intervals.

6. Determination method according to one of the preceding claims, wherein the acquisition and determination steps are carried out with at least five distinct parameters (P1, P2, P3, P4, P5, P6, P7), and preferably with an odd number, for example equal to seven, of distinct parameters (P1, P2, P3, P4, P5, P6, P7).

7. Determination method according to one of the preceding claims, wherein the parameters (P1, P2, P3, P4, P5, P6, P7) are chosen from: - a longitudinal speed (P1) of the motor vehicle, - a longitudinal acceleration (P2) of the motor vehicle, - an angular velocity (P3) of a component of a steering system of the motor vehicle, - a torque (P4) developed by an engine fitted to the motor vehicle, - a gradient (P5) of the track taken by the motor vehicle, - a period (P6) during which the engine-speed of said engine is lower than a predetermined threshold, and - a number of bends (P7) encountered by the vehicle.

8. Use of the determination method according to one of Claims 1 to 7, wherein, with the deduction step being carried out while the motor vehicle is being driven, a subsequent operation of, while said vehicle is being driven, adjusting at least one internal setting of the motor vehicle according to the determined track type (T1, T2, T3, T4) is provided.

9. Use of the determination method according to one of Claims 1 to 7, wherein, once the vehicle has made several journeys, a subsequent operation of determining the main use of the motor vehicle or of determining the wear of at least one wearing part of the motor vehicle according to the track types (T1, T2, T3, T4) detected on said plurality of journeys is provided.