Method for characterizing the posture of a mobile object moving along a traffic axis

Abstract

The invention relates to a method for characterizing the posture of a motor vehicle moving on a road (30) comprising several traffic lanes (V0, V1), the method comprising steps of: - acquiring data characterizing edge lines (31, 32, 33) delimiting each traffic lane, - projecting a critical point (Pc) attached to said moving object onto curves representative of said edge lines deduced from the acquired data, and - selecting the traffic lane on which the critical point is located, taking into account the projections of the critical point. Figure for abstract: Fig.2

Description

Title of the invention: Method for characterizing a posture of a mobile object moving on a circulation axis Technical field of the invention

[0001] The present invention relates generally to the control of mobile objects.

[0002] It applies more particularly to cars and other motorized vehicles traveling on roads, but could also be applied to other fields such as robotics.

[0003] The invention relates to a method for characterizing a posture of a mobile object moving on a traffic axis comprising several traffic lanes.

[0004] The invention also relates to a method for controlling a motor vehicle and a motor vehicle suitable for implementing such a method. State of the art

[0005] In an effort to make motor vehicles safer, they are currently being equipped with driving assistance systems and even highly automated driving systems.

[0006] These are typically lane centering systems (better known by the English acronym LCA for “Lane Centering Assist”), adaptive cruise control systems, etc.

[0007] Many of these systems need, in order to function, to know on which traffic lane the motor vehicle (EGO) is driving, if the vehicle is changing traffic lanes or if it is about to change...

[0008] Similarly, it is desirable for these systems to know which traffic lanes the surrounding vehicles are traveling on.

[0009] For this purpose, a lane assignment method is known from document EP2826687. This method proposes using a sensor, such as a RADAR, to determine the positions of surrounding vehicles relative to the EGO vehicle. The history of the successive positions of surrounding vehicles is more precisely used to determine their trajectories. It is thus possible to know, when their trajectories overlap, whether or not surrounding vehicles are using the same traffic lane as the EGO vehicle, or a lane further to the right or further to the left.

[0010] This method has the disadvantage of not being very reliable in the case where traffic is not very dense. Typically, if the EGO vehicle is following a vehicle and these two vehicles are alone in their environment, the change of traffic lane of the preceding vehicle will only be detectable late. Presentation of the invention

[0011] In order to overcome the aforementioned drawback of the state of the art, the present invention proposes a solution which is independent of the density of vehicles circulating on the road and which makes it possible to determine on which traffic lanes the EGO vehicle and the surrounding vehicles are located.

[0012] More particularly, according to the invention, a method is proposed for characterizing a posture of a mobile object (for example the EGO vehicle or a surrounding vehicle) moving on a traffic axis comprising several traffic lanes, the method comprising steps implemented by a computer: - acquisition of data characterizing the edge lines delimiting each traffic lane, - projection of a critical point attached to said moving object onto curves representative of said edge lines, which representative curves are deduced from the acquired data, and - selection of the traffic lane on which the critical point is located, taking into account the projections of the critical point.

[0013] Preferably, these steps are followed by the calculation of the time required for the vehicle to cross one of the two edge lines which border this traffic lane and / or the position of the zone in which the vehicle will cross this edge line.

[0014] Thus, thanks to the invention, the shape of the lane edge lines is used to know on which lane each object is located (EGO vehicle, surrounding vehicle, etc.). This method is therefore not based on the past positions of objects but rather on instantaneous data, which makes it more robust.

[0015] An advantage of the invention is that it does not generate any additional cost since it uses data already used in vehicles for other purposes.

[0016] The use of a critical point whose position can be selected taking into account the use that will be made of the characterization of the posture of the vehicle also allows the method to be particularly adaptable.

[0017] For example, it will be possible to detect a change of lane very early on by using a critical point located on one side of the vehicle.

[0018] As another example, it will be possible to display on a screen visible to the driver the general position of the vehicle on the road, using for this a critical point centered on the vehicle (i.e. located on the longitudinal axis of the vehicle).

[0019] Another advantage of the proposed solution is that it can be easily evaluated by the driver of the vehicle. To do this, the driver will be able to compare the data perceived by the computer and displayed on the screen with the reality on the ground.

[0020] Other advantageous and non-limiting characteristics of the method according to the invention, taken individually or in all technically possible combinations, are the following: - the selected traffic lane is the one for which the product of two scalar products is negative, each scalar product being calculated for one of the two curves representing the two edge lines delimiting this traffic lane between, on the one hand, the vector formed by the critical point and the projection of the critical point on said curve and a vector normal to said curve at the level of the projection of the critical point - the projection and selection steps are repeated with at least one other critical point attached to said moving object; - if two traffic lanes are selected, it is determined that the moving object changes traffic lanes; - the acquisition, projection and selection steps being carried out at a current time, it is planned to acquire which traffic lane was selected at a time preceding the current time, and to determine whether the mobile object has changed traffic lane by comparing the traffic lane selected at the previous time with the traffic lane selected at the current time; - it is intended to determine the width of the selected traffic lane taking into account the projections of the critical point, and to validate the selection of the selected traffic lane only if said width is greater than a predetermined width of the moving object; - the representative curves of said edge lines having limited lengths, if the projection of the critical point on one of the representative curves does not exist, it is planned to extend said representative curve; - it is intended to determine an estimated time before the moving object crosses at least one of the edge lines bordering the preselected traffic lane; - said duration is estimated by modeling a future trajectory of the moving object and determining the intersection between the future trajectory and the edge line, preferably using a Newton-Raphson method;

[0021] The invention also relates to a method for controlling a motor vehicle traveling on a road comprising several traffic lanes, the method comprising: - an operation of characterizing the posture of the motor vehicle by means of a characterization method as mentioned above, and - an automatic steering operation of the motor vehicle depending on the selected traffic lane, for example to maintain or center the motor vehicle in the selected traffic lane.

[0022] The invention also relates to a motor vehicle comprising data acquisition means characterizing edge lines delimiting each traffic lane of a road taken by the motor vehicle, and a computer programmed to implement a characterization or control method such as aforementioned.

[0023] Of course, the various features, variants and embodiments of the invention may be combined with each other in various combinations to the extent that they are not incompatible or mutually exclusive. Detailed description of the invention

[0024] The description which follows with reference to the appended drawings, given as non-limiting examples, will make it clear what the invention consists of and how it can be implemented.

[0025] In the attached drawings:

[0026] [Fig.l] is a schematic view of a motor vehicle suitable for implementing a method in accordance with the present invention;

[0027] [Fig.2] is a schematic top view of the motor vehicle of [Fig.l], on which traffic lane edge lines appear;

[0028] [Fig.3] is a block diagram illustrating the different steps of a method according to the present invention.

[0029] In [Fig.l], a motor vehicle 10 is shown which is suitable for implementing the invention.

[0030] This is a car. Alternatively, it could be another type of vehicle (truck, motorcycle, etc.).

[0031] Here, this vehicle 10 conventionally comprises a passenger compartment in which there are in particular a seat for the driver 20 of the vehicle, a dashboard with a display screen, and a steering wheel 12.

[0032] This vehicle 10 comprises a powertrain, a braking system and a steering system for turning the vehicle (not visible in the figure). Conventionally, the steering system comprises an electronically controllable power steering actuator, the powertrain comprises an electronically controllable engine control actuator, and the braking system comprises an electronically controllable braking actuator.

[0033] The vehicle 10 further comprises an electronic and / or computer processing unit (hereinafter called computer 11) comprising at least one microprocessor or microcontroller, at least one memory and input and output interfaces.

[0034] Thanks to its input interfaces, the computer 11 is adapted to receive different input data, which come from sensors or third-party computers.

[0035] Among these sensors, at least one sensor is provided for detecting the edges of the road, for example: - a front camera, and / or - a RADAR or LIDAR remote sensor.

[0036] Thanks to its output interfaces, the computer 11 is adapted to control the display screen, the power steering actuator, the engine control actuator, and the braking actuator.

[0037] Thus, the computer 11 is adapted to implement driving assistance functions and / or automated driving functions (in which the vehicle can move through traffic autonomously, without intervention from the driver).

[0038] Thanks to its memory, the computer 11 stores a computer application, consisting of computer programs comprising instructions whose execution by the computer allows the implementation of the method described below.

[0039] In [Fig.2], the motor vehicle 10 is shown seen from above, while it is using a traffic lane V0 of a road.

[0040] Two traffic lanes V0, V1 are represented here, namely the one taken by the motor vehicle 10 and a lane further to the left. Of course, the invention can be applied to all possible road configurations.

[0041] In this [Fig.2], we observe that each traffic lane V0, VI is delimited by two lane edge lines, namely left edge lines 33 and right edge lines 31 (here, the terms right and left are used considering the direction of advance of the vehicle), and an intermediate edge line 32.

[0042] The roadside edge lines can be formed by road markings affixed to the road (solid white line, broken line, etc.), by safety barriers, by raised shoulders or by any other distinctive element.

[0043] In this [Fig.2], a reference point (X, Y) attached to the motor vehicle 10 has also been shown, which will be the one considered for the calculations developed below. This reference point comprises an abscissa which extends along the longitudinal axis of the vehicle, towards the front, and an ordinate which extends towards the left of the vehicle. Its origin is located for example at the level of the center of gravity of the motor vehicle 10 or at the level of the center of its rear axle.

[0044] In [Fig. 3], the method that the computer 11 is adapted to implement in order to characterize the posture of the vehicle 10 is illustrated.

[0045] Posture means, in particular, position. This term “posture” can also refer to other concepts, for example, the orientation of the vehicle relative to the lane or whether or not the vehicle is changing lanes.

[0046] The idea here consists mainly of assigning a traffic lane to the vehicle 10, that is to say of identifying on which traffic lane the vehicle 10 is located, this traffic lane being hereinafter called “lane taken”.

[0047] The method according to the invention comprises several successive steps, all of these steps being repeated in a loop at regular intervals so as to regularly update update the identification of the route taken.

[0048] The detailed steps of this method are as follows.

[0049] In this example, we can consider that the method is implemented at a time T, and that it was already implemented at previous times (T-1, T-2, etc.).

[0050] During the first step SI, the computer 11 attempts to acquire information relating to the road 30 and the motor vehicle 10.

[0051] It first attempts to acquire information characterizing the three trackside lines 31, 32, 33.

[0052] This information is provided directly by a sensor (for example in the form of distance between the line and the sensor).

[0053] Preferably, only the information measured by the vehicle sensors is used here to detect these lines. No information received from road infrastructures or third-party vehicles is used, for example.

[0054] This information can be obtained directly (as measured) or be subjected to a data fusion procedure (or any other procedure, typically data association) in order to improve its reliability.

[0055] Here, each track edge line is modeled by the computer in analytical form.

[0056] As an example, one could provide that these lines are modeled in a polynomial form. However, here, they are modeled in the form of clothoids, this form being preferred because it provides better results (for an identical number of coefficients taken into account).

[0057] Such a clothoid is defined as having a curvature c which depends on its length, which can be written:

[0058] [Math.l] c(ï) = c0 + c^l

[0059] In this equation: - it is therefore the equation of the clothoid, -1 is the curvilinear abscissa of the arc of the clothoid, which varies between 0 inclusive and L inclusive, - L is the length of the arc of the spiral, - c0 is the curvature at the origin of the arc, and - Ci is the rate of change of curvature of the arc.

[0060] In other words, a clothoid can be defined by six coefficients which are: - x0, the abscissa of the origin point M of the clothoid in the (X, Y) frame, - y0 the ordinate of this point of origin M, - Wq which is the heading angle of the clothoid at the origin of the arc, -L, - c0, and - Here.

[0061] In the following, each clothoid will be defined by the set C of these six coefficients, which can be noted as follows: C (x0, y0, c0, cb L).

[0062] It will be noted that the point of origin M of the clothoid will have a position which will depend on that of the vehicle. Indeed, it will be the point of the edge line which is the one closest to the vehicle which was detected by the sensors.

[0063] For information, the heading angle along the clothoid can be written as follows:

[0064] [Math.2] T(Z) = 0 + / ^ / ^^ c{>1 + -J1

[0065] The Cartesian coordinates (x, y) of a point located along the clothoid can be written as follows:

[0066] [Math.3] r / x ( Z ) - x0 + J Ocos(^ ( u )}.du y (0 = .>'o + Josin(T (u)).du

[0067] During this first step E0, it could be envisaged that the computer 11 receives the sets C characterizing the track edge lines from one of the sensors on board the vehicle (for example from the camera if the latter is equipped with a processor adapted to carry out the operations necessary for determining the sought coefficients). However, preferably, the computer 11 receives data recorded by several of the vehicle's sensors (camera and RADAR and / or LIDAR), merges this data, and deduces the sought coefficients therefrom.

[0068] During step E0, in addition to the aforementioned coefficients, the computer 11 receives the three variance-covariance matrices (hereinafter called covariance matrices) associated respectively with the coefficients of the three track edge lines 31, 32, 33. Each covariance matrix makes it possible to take into account the uncertainty in the detection of each track edge line by the sensor(s).

[0069] These matrices are for example provided by the manufacturer of the sensors used. Alternatively, they can be determined during a sensor test campaign and recorded in the computer memory. The coefficients of these matrices are preferably invariable.

[0070] It will be noted here that if several sensors are used to determine the coefficients characterizing a trackside line, their uncertainties will be associated with each other in order to obtain the covariance matrix of a detected line.

[0071] At this stage, the computer 11 assigns an identifier to each of the traffic lanes. VO, VI detected.

[0072] This identifier is for example an integer i, 0 for the leftmost lane, 1 for the lane immediately to the right of the latter...

[0073] It will be noted in this regard that before the computer assigns a new integer for a new track, it can consult its memory and search whether among the tracks detected during the previous iteration, there was one whose track edges already carried the same identifiers, in which case the same identifier will be reused.

[0074] The calculator also assigns to each detected track edge line 31, 32, 33 a type characterizing it (for example: continuous line, broken lines, safety barrier, shoulder, etc.).

[0075] Finally, during step E0, the computer 11 acquires data relating to the posture and / or the kinematics of the motor vehicle 10 on the road 30. This is kinematic data which is preferentially obtained using only the sensors on board the motor vehicle 10.

[0076] Here, these data include values ​​of: - yaw rate of the motor vehicle, expressed in rad / s, - speed V of the motor vehicle relative to the ground, expressed in m / s, - longitudinal acceleration ax of the motor vehicle along the X axis, expressed in m / s 2 - lateral acceleration ay of the motor vehicle along the Y axis, expressed in m / s2, - yaw acceleration of the motor vehicle, expressed in rad / s2, - angle af of the front wheel of the motor vehicle relative to the X axis, expressed in rad, - angle ar of the rear wheel of the motor vehicle relative to the X axis, expressed in rad, and - instantaneous curvature c' of the trajectory of the motor vehicle (this curvature being defined here by the inverse of the radius of curvature, expressed in m1).

[0077] It will be noted here that the instantaneous curvature c' can be defined here as:

[0078] [Math.4] c' = tan(af / D)

[0079] With D the distance between the front and rear axles.

[0080] At this stage, it will be noted that only instantaneous data (measured or calculated at time step T) are used here.

[0081] Alternatively, one could further use previously acquired data to smooth the data used.

[0082] As shown by the dotted arrow in [Fig.3], if all of these data could not be obtained, the method proceeds directly to step S5.

[0083] Otherwise, the method continues in a step S2 of determining a characteristic point of the motor vehicle 10, hereinafter called “critical point Pc” and whose coordinates are noted (xc, yc).

[0084] This critical point Pc is preferably located on or contained in the envelope of the motor vehicle (defined in particular by its bodywork, its windows, etc.). However, as a variant, it could be a point arranged, for safety reasons, at a reduced distance from this envelope, outside of it.

[0085] This critical point Pc could have an invariable and predefined position. In this case, its position would simply be read in the computer's memory.

[0086] However, as a variant, it may have a variable position on the vehicle, this position being for example different depending on the autonomous piloting function which will use the result of the present method (obstacle avoidance function, function of maintaining the center of the traffic lane, etc.).

[0087] For example, the critical point could be located at a corner of the vehicle, in particular if the objective is to detect a change of lane as early as possible by determining the moment when the critical point crosses the intermediate edge line 32. In this case, it will be the front point of the vehicle located on the same side as the steering wheel.

[0088] According to another example, this could be the center of gravity of the motor vehicle 10. The position of this center of gravity could be estimated once and be invariable, or could be calculated as a function, for example, of the number of passengers in the vehicle 10. This critical point centered on the center of gravity will be useful in particular for carrying out emergency braking such that the vehicle 10 remains generally on its traffic lane. This critical point centered on the center of gravity will also be useful when it is desired to display on the interior screen of the vehicle the lane on which the vehicle is mainly located.

[0089] Alternatively, one could define not a single critical point, but a set of several critical points Pc, typically the four corners of the motor vehicle. Indeed, with these four corners, the detection of a change of traffic lane will always be carried out very early, regardless of the side towards which the vehicle 10 is heading.

[0090] The third step S3 then consists, for the computer 11, in determining the channel in which this critical point Pc is located.

[0091] In this case, the critical point Pc is assigned to the traffic lane V0, V1 for which it is located to the right of its left edge line and to the left of its right edge line.

[0092] To carry out this check, the critical point Pc is projected onto the clothoids of the different track edge lines. Then the calculator proceeds successively, track by track track, by determining whether the orthogonal projections Pc* of the critical point Pc on the two edge lines bordering this track have ordinates that are one positive and the other negative.

[0093] More precisely here, the selected traffic lane V0, V1 is the one for which: - the scalar product of the vector Pc^c and the vector n (namely the vector normal to the clothoid at the level of the orthogonal projection Pc*) for the clothoid representative of one of the two edge lines 31, 32, 33 delimiting this traffic lane V0, VI, multiplied by: - the scalar product of the vector pc ^Pc and the vector » (namely the vector normal to the clothoid at the level of the orthogonal projection Pc*) for the clothoid representative of the other of the two edge lines 31, 32, 33 delimiting this traffic lane V0, VI, is negative.

[0094] In practice, the computer performs this calculation successively for each traffic lane, starting with the leftmost lane. As soon as the product of the scalar products is negative, the critical point Pc is assigned to the corresponding traffic lane. This traffic lane is then said to be “pre-selected”.

[0095] It will be noted that in the variant where a set of several critical points Pc is considered, we will search here in which traffic lanes these critical points are located. It is thus possible for several traffic lanes to be preselected.

[0096] If this step S3 has made it possible to assign the critical point Pc to a traffic lane, the method continues in a step S4.

[0097] Otherwise, an intermediate step S3' is provided.

[0098] This intermediate step S3' consists of extending the clothoids defining the track edge lines before implementing the process of step S3 again.

[0099] Indeed, as each clothoid has a finite length L, it may happen that the orthogonal projection Pc* of the critical point Pc is located not on the clothoid, but on the extension of the latter, upstream of the original M of the clothoid or downstream of the curvilinear abscissa point L of the clothoid.

[0100] The idea is therefore, if no orthogonal projection Pc* has been found on a clothoid, to extend the latter by a limited length (this length is less than 5 meters and can typically be equal to 3 meters). This length is limited since it is difficult to extend a clothoid reliably.

[0101] Here, before carrying out this extension, the computer carries out a check. This check consists of verifying, for each track, whether the coefficients Wo and c0 of the clothoids of the lines bordering this track are greater or not than predetermined threshold values. If they are less than these thresholds, the clothoids are extended backwards over a length of 3 meters. Otherwise, they are only extended over a reduced length less than strictly 3 meters.

[0102] As shown in [Fig.3], if this step S3' does not allow a solution to be reached, no traffic lane is selected and the method ends (step S5).

[0103] Otherwise, the method continues in a step S4.

[0104] Before describing this step S4, we can explain how the orthogonal projection of the critical point Pc onto a clothoid is carried out.

[0105] Here, the projection is carried out on the basis of the Newton Raphson method.

[0106] This method makes it possible to find a precise approximation of a zero (i.e. a root) of a real function f of a real variable t.

[0107] To introduce this method, we can define the set C(t) as being the set of values ​​of the coefficients of the clothoid considered at the level of the curvilinear abscissa point l=t.

[0108] We can also define the real variable t* which is such that the set C(t*) is the orthogonal projection Pc* of the critical point Pc on the clothoid defined by this set C. This real variable t* therefore constitutes the curvilinear abscissa of the orthogonal projection Pc* of the critical point Pc on the clothoid.

[0109] We can finally define the function f in the following form:

[0110] [Math.5] f(t) =<Pc-C(t\C'(t)>

[0111] In this equation, the brackets mean that we are producing the scalar product of two vectors. The first vector is formed by the difference between the coordinates (xc, yc) of the critical point and the Cartesian coordinates (x(t), y(t)) of a point located along the clothoid, at the curvilinear abscissa t. The second vector is formed by the derivative with respect to the curvilinear abscissa t of the Cartesian coordinates (x, y) of said point.

[0112] We can also write this equation in the form:

[0113] [Math.6] + (y^t) ) *sin

[0114] In this equation, the term Wo>proj is equal to the heading angle of the clothoid at the orthogonal projection Pc*.

[0115] Therefore, we can also write:

[0116] [Math.7] 2 / (,)=<P-C(t), C'u) > -||c"(f)||

[0117] As well as:

[0118] [Math. 8] / (0 =<P-C(t), cav(\p(t)) ) * ^(t) > - 1

[0119] The objective is then to find the zero of the function f.

[0120] In practice, to find this zero, we consider as input the coefficients of the clothoid considered, as well as the coordinates of the critical point Pc to be projected.

[0121] Then, the calculator selects a starting point on the clothoid from which the orthogonal projection Pc* will be sought. Indeed, this method consists of testing different starting points then, for each starting point, attempting to converge the algorithm towards a solution constituting the orthogonal projection Pc*. This attempt to converge the algorithm is carried out a finite number of iterations (for example about twenty times). Thus, as long as the method does not converge quickly, we start again with other starting points.

[0122] For example, the starting points successively used could be the coordinate points: - (x(L / 2),y(L / 2)), - (x(0),y(0)), - (x(L),y(L)).

[0123] In other words, if we consider that s0 is the curvilinear abscissa of the starting point on the clothoid, then the aforementioned points correspond to: - s0 = L / 2, - s0 = 0, - So = L.

[0124] We can then formulate the search algorithm for the orthogonal projection Pc* in the following manner.

[0125] As long as the number n of iterations remains below a predetermined threshold Nmax and the solution t* has not been found, the following process is repeated in a loop.

[0126] This method involves determining the terms of the clothoid at the curvilinear abscissa point s considered (during the first loop s=s0), so as to obtain: ■ MojProj (x(s), y(s)), - ^O.proj = W, " C0,proj = C(S).

[0127] Therefore, the value of the function f(s) can be obtained, as well as that of its derivative f'(s), using the equations Math.6 and math.8.

[0128] If the absolute value of this derivative is less than a predetermined threshold e very close to zero (of the order of 0.01 meters), then we consider that the projected is the correct one, which can be written t*=s.

[0129] Otherwise, we start again at a point offset from the point considered with abscissa s, this offset point having an abscissa s+ds, with:

[0130] [Math.9]

[0131] For this, the first step is to check if the new point is on the clothoid (the latter having a finite length L). In other words, the calculator checks that the point with coordinates (x(s+ds),y(s+ds)) belongs to the set C considered.

[0132] Then, if ds is strictly positive, the value ds is assigned the minimum of the values ​​of Ls and ds, while if ds is strictly negative, the value ds is assigned the maximum of the values ​​-s and ds. Thus, in the case where the point s was outside the clothoid, the value 0 is assigned to the point s if s was negative or L if s was positive.

[0133] Then, if the number of iterations n has not reached the maximum threshold Nmax, the value s+ds is assigned to the abscissa s, and the counter n is incremented.

[0134] To decide whether the solution is preserved, the process calculates the value of the function f(s) and checks whether it is sufficiently close to 0 (given e). This is the case when the point Pc is very close to the clothoid, with a precision of e.

[0135] If this is not the case, the calculator calculates the cosine at the point of the solution which must be close enough to zero for the solution to be preserved:

[0136] [Math. 10] cosine = ..............................................y ^(x^-a^s) )2+(y;)-y (.<))“

[0137] This calculation is carried out when the critical point Pc has an orthogonal projection on the clothoid.

[0138] Thus, if the orthogonal projection Pc* of the critical point Pc on the clothoid exists, then the calculator determines the Euclidean distance between the critical point Pc and its orthogonal projection Pc*, as well as the scalar product<Pc*Pc,n> .

[0139] The same steps are applied for the two trackside lines considered.

[0140] After the two track edge lines have been tested, it is considered that the critical point Pc belongs to the tested traffic lane if the Euclidean distances between the points Pc* and Pc are non-zero and if the product of the two scalar products<Pc*.Pc, n> calculated gives a negative result.

[0141] At step S4, it is considered that a traffic lane could have been preselected. The idea then consists of verifying that this traffic lane can be used by the motor vehicle 10.

[0142] Indeed, it may happen that non-existent or unusable traffic lanes are detected. Typically, the computer may have wrongly detected a traffic lane between a road marking line and a safety barrier, or a cycle lane.

[0143] Then, during this step, the calculator compares the width of the motor vehicle 10 (which is known and stored in its memory) with the width of the preselected traffic lane.

[0144] The width of the traffic lane can be obtained by determining the lateral gap between the two projections of the critical point onto the edge lines of the preselected traffic lane. Alternatively, it could be obtained in another way (e.g. by reading this data into mapping software).

[0145] If the width of the vehicle is greater than that of the traffic lane, this lane is not selected and the method ends (step S5).

[0146] Otherwise, the method continues in a step S6.

[0147] It will be noted that in the variant where a set of several critical points Pc is considered, if it has been detected that the vehicle is in several traffic lanes simultaneously (a main lane in which the most important critical point is located and a secondary lane), the following three cases are possible: - the widths of the two traffic lanes are less than the width of the vehicle, in which case the process ends (step S5), - the width of the main traffic lane is greater than the width of the vehicle, in which case the process continues (“step S6) considering this main traffic lane, - the width of the main traffic lane is less than the width of the vehicle but the width of the secondary traffic lane is greater than the width of the vehicle, in which case the method continues (“step S6) considering this secondary traffic lane.

[0148] Step S6 then simply consists of changing the preselected traffic lane to “selected”.

[0149] At this stage, during a step S7, the computer acquires in its memory the identifier of the traffic lane selected at the previous time step T1.

[0150] It is thus able to determine whether the motor vehicle has remained on its lane, has changed lanes to the left, or has changed lanes to the right, by simply comparing the identifiers of the lanes selected at the current time step T and at the previous time step T1.

[0151] In the aforementioned variant where several critical points are considered and where a main path and a secondary path have been detected, this detection makes it possible to confirm the result of this step S7.

[0152] It may happen that a traffic lane selected at the previous iteration (T1) no longer exists at the current iteration (T). This may be due to a particular road configuration (merging of two traffic lanes into one, separation of one lane into two, change of road shape to enter a roundabout, etc.) or to a failure of the lane detection by the sensors.

[0153] This type of event will be perceived using the unique identifiers of the traffic lanes (which will disappear between times T1 and T).

[0154] Thus, the computer will be able to detect this event, so that in this eventuality, the computer will not issue a valid lane change signal.

[0155] During a step S8, when a traffic lane has been selected, the calculator calculates the time required for the vehicle to cross one of the two edge lines which border this traffic lane and / or the position of the zone in which the vehicle will cross this edge line.

[0156] These two pieces of information are in fact useful for executing an automated piloting function of the vehicle 10.

[0157] In practice, this step is implemented for the two edge lines which border the selected traffic lane, except if one or the other of these two lines borders the road (in particular if it is a safety barrier).

[0158] To calculate the time ø required for the vehicle to cross the edge lines, it is necessary to generate a trajectory T0 of the vehicle using at least one model (see [Fig.2]).

[0159] Different models can be used for this purpose.

[0160] Typically, a constant curvature and longitudinal acceleration model (CCLA model) may be used, but alternatively, a constant curvature and longitudinal velocity model (CCALV model) could be used. Other models could also be used.

[0161] The determination of the trajectory is carried out over a predetermined prediction duration, of a few seconds at most (here 5 seconds).

[0162] Once the trajectory T0 is obtained, the computer determines the position of the possible intersection between the trajectory T0 and the edge line(s) bordering the selected traffic lane.

[0163] Here again, this calculation is carried out using the Newton-Raphson method in two dimensions.

[0164] Note that when no intersection is detected, the solution ôt will be formed by infinite time.

[0165] Otherwise, the solution ôt will be formed by the time required to arrive at this intersection.

[0166] Here we can describe in more detail how the intersection between the trajectory T0 of the vehicle 10 and each clothoid considered is sought.

[0167] At this stage, the calculator knows the set C of parameters defining the clothoid illustrating the edge line considered (the two lines being considered one after the other).

[0168] The calculator also knows the data relating to the dynamics of the vehicle at time T, namely: - its position xx0, yyo, - its heading 0W, - its speed v0, - its acceleration a0, - the instantaneous curvature c of its trajectory (see equation Math.4), and - the AT validity period of the model.

[0169] It is therefore able to calculate the trajectory of the motor vehicle, for example in Cartesian form.

[0170] Then, the calculator determines, if it exists, the root of the expression:

[0171] [Math. 11] MX XX] = MV y »V')]

[0172] This expression means that we are looking for the curvilinear abscissa s of the clothoid considered and the instant t' where the trajectory of the vehicle intersects this clothoid.

[0173] In this expression, s and t' are therefore respectively included in the intervals [0; L] and [0; AT],

[0174] In this expression also, [x(s);y(s)] are the Cartesian coordinates of the clothoid considered, and [xx(t');yy(t')] are the Cartesian coordinates of the trajectory considered, here of CCLA type.

[0175] We can then first define the concepts and principles allowing Newton's method to be applied to the determination of this intersection.

[0176] We can thus introduce a new function f:R2 -» R2.

[0177] Given a starting point u0, we can write the following sequence:

[0178] [Math. 12] u(n + 1) = i^n) - inv(J(u(H^ * fMn))

[0179] In this equation J is the Jacobian of the function f.

[0180] Here, the function f is defined as follows:

[0181] [Math. 13] t') = [ / ; fy] = [a0 - - J^')]

[0182] The terms x(s) and y(s) are defined by the equation Math.3.

[0183] The terms xx(t') and yy(t') are defined using the model considered, as follows:

[0184] [Math. 14] xx(t) = xxO yy(t) = yyO 4(cos(0(t')) - COs(0o))

[0185] In these equations, we used the term 0 which can be defined by:

[0186] [Math. 15] 0(0 = 0O+ c(v0.f + a^.t' / 2)

[0187] Where:

[0188] [Math. 16] v(0 = v0 + a^'

[0189] Since we consider the acceleration constant.

[0190] The Jacobian is given by the expression:

[0191] [Math. 17] J (s, t') = [««( tp(s) ), - v(f )£os(3(t')), sm(y{s) ), -v(t

[0192] rp(s) is defined in equation Math.2.

[0193] Thus, we can write:

[0194] [Math. 18] | J(5,t') | = v ( f ) sin (y{s) - 0(t '))

[0195] As well as:

[0196] [Math. 19] = 1 / |J|(-), v(t').cos(3(t')), -sin(y^s)\

[0197] It can be noted that to avoid the pair (s,t') leaving its interval [0;L]x[0;T], we limit the displacement to the largest displacement which keeps this pair of values ​​in its interval.

[0198] It should also be noted that there may be more than one solution, and that convergence to a solution is not guaranteed.

[0199] In practice, the steps for determining the intersection are as follows.

[0200] As long as the number of iterations m already carried out is less than a maximum threshold Mmax, the calculator determines the function f(s,t'), first of all for an initial value of curvilinear abscissa s and on the basis of the acquired values ​​of the vehicle parameters.

[0201] It then checks whether the sum fx2+fy2 is less than a predetermined threshold, close to zero.

[0202] If this is the case, the process stops since the solution found is satisfactory. Indeed, the solution is located in the immediate vicinity of the intersection point.

[0203] Otherwise, it calculates the Jacobian J(s,t').

[0204] If the absolute value of this Jacobian is less than a threshold close to zero, it is considered that no solution has been found and the process is interrupted.

[0205] Otherwise, the calculator calculates the product inv(J(s,t')).f(s,t').

[0206] It should be noted here:

[0207] [Math.20] (cls, dt)= inv(J{s, t'}

[0208] It is then necessary to define constraints on the displacement (ds, dt) to prevent the couple (s,t') from leaving the interval, since Newton's method does not provide for this.

[0209] For this, the idea is to calculate the positive value 1 closest to the displacement up to 1 in this direction, that is to say find the maximum value of 1 included in the interval [0;l] which is such that the sum s + l.ds is in the interval [0; L] and the sum t' + l.dt is in the interval [0; T].

[0210] The displacement (ds, dt) is then updated.

[0211] If the sum ds2+dt2 is too small (for example less than a threshold of 0.0001), no solution is found and the loop stops.

[0212] If the number m of iterations has not yet reached the maximum threshold Mmax, the values ​​of s and t' are updated, as well as those of the sums s+ds and t'+dt associated with the intervals [0; L] and [0; T].

[0213] If the number m of iterations has reached the maximum threshold Mmax, the iterations stop.

[0214] As soon as the solution has already been detected, the appropriate values ​​on the clothoid are indicated, so that it is possible to deduce the time required for the vehicle to cross the edge line.

[0215] Thus, this process makes it possible to determine both the time required for the vehicle to cross one of the two edge lines which border the selected traffic lane and the position of the zone in which the vehicle will cross this edge line.

[0216] Finally, during a step S9, the selected traffic lane is stored in the memory of the computer, with its identifier and the equations of its left and right lines. On the basis of this memory, at each iteration, the computer will be able, at step S6, to obtain information relating to a change of lane.

[0217] In step S10, it is planned to develop different data for use by several “clients”, i.e. for use in different automated piloting functions of the vehicle 10.

[0218] Thus, an indicator is developed to indicate whether a route has been selected.

[0219] In addition, the calculator determines the identifier of the selected traffic lane.

[0220] Data is also generated to indicate whether a lane change occurs, and to which side.

[0221] Finally, all of this data is grouped into a new memory record, along with all other calculated data (including the estimated position of crossing an edge line and the time required to get there).

[0222] At this stage, we can describe step S5 cited above.

[0223] As explained, this step corresponds to the case where no traffic lane has been selected, for example if the pre-selected lane was too narrow or if the data relating to the lane edge lines were not available, or if the vehicle is placed next to the detected lanes.

[0224] In these different eventualities, step S5 interrupts part of the process, but step S7 remains implemented to develop an indicator informing that no lane change could not be detected due to a problem. This indicator is then stored in step S10.

[0225] The stored data is then either directly transmitted to the clients who need it (automatic braking system, system for centering the vehicle in its lane, obstacle avoidance system, system for displaying a representation of the vehicle in its lane, etc.), or transmitted over the vehicle network so that the clients can read the information necessary for their proper functioning.

[0226] The present invention is in no way limited to the embodiment described and shown, but those skilled in the art will be able to provide any variant in accordance with the invention.

[0227] Typically, in the embodiment described above, the computer seeks to characterize the posture of the vehicle that takes it onto the road. Of course, it will be able to use this same method to characterize the posture of each mobile object located in the environment of the vehicle 10.

[0228] Furthermore, the method described above may be applied in fields other than the automotive field, for example in the field of robotics. Typically, it may be applied to robots designed to move around in a warehouse, on multi-lane traffic routes.

Claims

Claims

1. Method for characterizing a posture of a mobile object moving on a traffic axis comprising several traffic lanes (V0, VI), the method comprising steps implemented by a computer (11): - acquiring data characterizing edge lines (31, 32, 33) delimiting each traffic lane (V0, VI), - projecting a critical point (Pc) attached to said mobile object onto curves representative of said edge lines (31, 32, 33), said representative curves being deduced from the acquired data, and - selecting the traffic lane (V0, VI) on which the critical point (Pc) is located, taking into account said projections, the selected traffic lane (V0, VI) being that for which the product of two scalar products is negative, each scalar product being calculated for one of the two curves representative of the two edge lines (31, 32, 33) delimiting this traffic lane (V0, V1) between,on the one hand, the vector formed by the critical point (Pc) and the projection of the critical point (Pc) on said curve and a vector normal to said curve at the level of the projection of the critical point (Pc).,

2. A characterization method according to claim 1, wherein the projection and selection steps are repeated with at least one further critical point (Pc) attached to said moving object.

3. A characterization method according to claim 2, wherein if two traffic lanes (V0, VI) are selected, it is determined that the moving object changes traffic lane (V0, VI).

4. Characterization method according to one of claims 1 to 3, in which, the acquisition, projection and selection steps being carried out at a current instant, it is provided to acquire which traffic lane (V0, VI) was selected at an instant preceding the current instant, and to determine whether the mobile object has changed traffic lane (V0, VI) by comparing the traffic lane (V0, VI) selected at the previous instant with the traffic lane (V0, VI) selected at the current instant.

5. Characterization method according to one of claims 1 to 4, in which it is provided to determine the width of the selected traffic lane (V0, VI), for example taking into account the projections of the critical point (Pc), and to validate the selection of the traffic lane (V0, VI). selected only if said width is greater than a predetermined width of the moving object.

6. A characterization method according to one of claims 1 to 5, wherein for the representative curves of said edge lines (31, 32, 33) having limited lengths, if the projection of the critical point (Pc) on one of the representative curves does not exist, it is provided to extend said representative curve.

7. A characterization method according to one of claims 1 to 6, wherein it is provided to determine an estimated time before the moving object cuts at least one of the edge lines (31, 32, 33) bounding the preselected traffic lane.

8. A characterization method according to claim 7, wherein said time is estimated by modeling a future trajectory of the moving object and by determining the intersection between the future trajectory and the edge line (31, 32, 33), preferably using a Newton - Raphson method.

9. Method for piloting a motor vehicle (10) traveling on a road (30) comprising several traffic lanes (V0, VI), the method comprising: - an operation of characterizing the posture of the motor vehicle (10) by means of a characterization method according to one of claims 1 to 8, and - an operation of automatically piloting the motor vehicle (10) as a function of the selected traffic lane, for example to maintain or center the motor vehicle (10) in the selected traffic lane.

10. Motor vehicle (10) comprising data acquisition means characterizing edge lines (31, 32, 33) delimiting each traffic lane (V0, VI) of a road (30) taken by the motor vehicle (10), characterized in that it further comprises a computer (11) programmed to implement a characterization method according to one of claims 1 to 8 and / or a piloting method according to claim 9.

Images