EP1811479B1 - Magnetic traffic control system - Google Patents

Abstract

The device has a set of sensors (C power x i) designed to be arranged along a direction (2). Another set of sensors (C power y j) is arranged along another direction (4), that intersects the former direction at a point at which common sensor (C power xy o) is placed. A microcomputer calculates a relation between a time signature of a vehicle e.g. car, passing above the common sensor and a spatial profile resulting from measurements made by the former sensors. The microcomputer is used to extract the length and the width of the vehicle by thresholding a spatial representation.

Description

TECHNICAL FIELD AND PRIOR ART

The invention relates to a method and a device for classification of vehicles from their electromagnetic signature.

It collects road data, for example to count and / or classify motor vehicles during their journeys on a roadway.

• l'identification et la classification par type de véhicule: un exemple caractéristique est la classification au péage autoroutier pour paiement automatique. Le système utilisé actuellement sur les autoroutes est basé sur l'association de plusieurs types de capteurs :

• un interrupteur magnétique, formé de deux boucles de courant, qui permet de détecter la présence d'un véhicule,
• un capteur de type piézoélectrique, disposé à la surface de la chaussée, qui permet de détecter le passage des essieux d'un véhicule pour les compter,
• un capteur optique qui forme un rideau placé transversalement à la route: lorsque le véhicule le traverse, il fournit une estimation de sa hauteur.

The invention therefore relates to the field of study and control of road traffic, whose applications are broad. As an example:

• identification and classification by type of vehicle: a typical example is the motorway toll classification for automatic payment. The system currently used on motorways is based on the combination of several types of sensors:

• a magnetic switch, formed by two current loops, which makes it possible to detect the presence of a vehicle,
• a piezoelectric type sensor, disposed on the surface of the roadway, which makes it possible to detect the passage of the axles of a vehicle for counting them,
• an optical sensor that forms a curtain placed transversely to the road: when the vehicle passes through it, it provides an estimate of its height.

The main disadvantages of this device are its cost, a lack of robustness (especially with respect to climatic conditions), a difficult maintenance (due in particular to the wear of current loops) and an average classification error rate.

The invention also allows the identification of particular vehicles, for example for the regulation of car traffic, the monitoring of the use of a road, the optimization of traffic, the monitoring of a vehicle on a limited road area ( pedestrian zone) or in a car park, the allocation to the identified vehicle of a particular service (private parking space, subscription to a gas station, etc.).

Another application of the invention is the authentication of particular vehicles: for example it may be a vehicle equipped with a remote identification system (RFID badge for example) which is validated via the reading and the authentication of the magnetic signature of the vehicle (this signature being obtained by a device or a method according to the invention).

There are systems based on magnetoresistances or networks of current loops. But, they are either expensive or poorly performing in terms of vehicle recognition.

Magnetic traffic control systems rely on the use of the magnetic signature of a vehicle. An automobile is a magnetic mass that modifies the field lines because the field Magnetic tends to take the path of greater magnetic permeability. In addition, an automobile may include ferrous materials that alter the direction and intensity of the magnetic field. The vehicle is generally represented by a set of magnetic dipoles, which add to the earth's magnetic field quiescent (that is to say at rest temporarily) and which create a magnetic anomaly that can be measured by magnetic sensors.

These signals are then used in a detection / classification system whose objective may be to count the vehicles or identify them. Each class can be characterized by a number of parameters, of which the most commonly used are the number of axles, the inter-axle distances, the length of the vehicle, the distances between the roadway and the sills and / or between the axles.

One of the difficulties of classification systems is the time / space correspondence. Indeed, signatures are acquired by magnetic sensors over time. They are therefore dependent on the speed of the vehicle: they can be compressed if the vehicle accelerates, dilated if it brakes, or even if it stops, as shown on the diagram. figure 1 , on which the curves I and II represent, respectively, the temporal deformations of the signature of a vehicle passing quickly on one sensor, and slowly, with stop, on another. Conversely, the spatial magnetic signature of the vehicle is constant.

We are therefore looking for a method to pass temporal signatures in the spatial domain, regardless of the speed and trajectory of the vehicle.

The patent FR-2811789 describes a vehicle classification system for detecting the electromagnetic signature from a single current loop. This signature is scanned, sequenced, and dated. The speed of a vehicle can also be calculated, looking for the moment when the pace of the signature stops following an exponential law.

This calculation is not precise enough and does not allow to check if the vehicle stopped on the sensor. The measured characteristics are restricted to amplitudes of the signal in its temporal representation and in its frequency representation.

The patent US 5331276 discloses a velocity measurement system comprising two biaxial FluxGate magnetometers, separated by a known distance and oriented precisely with respect to each other. The speed of the vehicle traveling in the vicinity of the system is calculated by forming the ratio between the time derivative of the measured field (given by the time derivative of a B signal of one of the magnetometers) and the spatial derivative of the measured signals (calculated by the instantaneous difference of the two B signals measured on the two magnetometers). So that the spatial difference of the fields of the two sensors is approximately equal to the spatial gradient, the spacing between the two sensors must be neither too short, not too big (it must be at most equal to 1/10 of the distance to the nearest point of passage of the vehicle). This constraint limits the use of this device to specific trajectories and vehicles, with a little equivalent magnetic moment.

Several vehicle classification systems propose to determine the speed by exploiting the time difference between two signatures measured by sensors placed at known distances. But for the temporal offset of the signatures to give a good estimate of its speed, it must be constant on the basis of calculation (inter-sensor distance). However, in normal road traffic conditions, vehicles rarely follow a uniform movement, especially near motorway tolls, for example.

The patent EP 0770978 discloses such a multi-sensor vehicle detection system disposed in a floor or a ceiling, placed in tubes arranged transversely to the path of the vehicle. The distance between two adjacent sensors in a tube is less than or substantially equal to the normal width of a tire, so as to detect twin wheels of vehicles. By placing two detector tubes parallel to each other, transverse to the longitudinal direction of the roadway, and separated by a known distance, it is possible to identify the instants of detection of a vehicle and to calculate the time taken by the vehicle to go from one device to another. The patent US4509131 offers to use a correlation to perform a comparable calculation, the device being placed on the vehicle and exploiting the magnetic signatures of the ground.

The patent EP 0841647A1 describes a multi-point measuring device arranged transversely to the road. It allows to map the vehicle, in time and space. A data count reduction calculation is used to extract from the map a set of characteristic values of each vehicle, regardless of its size or the number of axles. This device is used to identify each vehicle for the purpose of monitoring road traffic. This is not a classification system. In addition, this method, while establishing a temporal / spatial relationship, does not make it possible to obtain an image of the object.

The U.S. Patent 5,392,034 describes a multi-sensor vehicle identification system.

There is therefore the problem of finding a method and a device for obtaining such a spatial image of the magnetic signature of the vehicle.

STATEMENT OF THE INVENTION

According to the invention, a multi-sensor device is used and spatial and temporal information is exploited to extract the characteristics of the magnetic signatures of the vehicles.

• au moins un premier ensemble de capteurs (C^x _i), destinés à être disposés le long d'au moins une première direction,
• au moins un deuxième ensemble de capteurs (C^y _j), destinés à être disposés selon au moins une deuxième direction, qui coupe la première en un point auquel est disposé un capteur commun (C^xy ₀), appartenant au premier et au deuxième ensemble,
• des moyens de calcul, pour calculer une relation entre la signature temporelle S_o (t) d'un véhicule passant au-dessus du capteur commun et un profil spatial S_o(x) résultant des mesures effectuées par les capteurs du premier ensemble de capteurs.

The invention firstly relates to a device for measuring magnetic signatures of vehicles, comprising:

• at least a first set of sensors (C ^x _i ), intended to be arranged along at least one first direction,
• at least a second set of sensors (C ^y _j ), intended to be arranged in at least a second direction, which intersects the first at a point at which is disposed a common sensor (C ^xy ₀ ) belonging to the first and second sets ,
• calculating means, for calculating a relationship between the time signature S _o (t) of a vehicle passing over the common sensor and a spatial profile S _o (x) resulting from the measurements made by the sensors of the first set of sensors .

According to the invention, first magnetic measurements are used in the direction of displacement to obtain a law between the time and the position of the sensors, then this law is applied to another series of measurements made in at least one other direction. The temporal notion disappears and we obtain a spatial image of the object, but which is not a photo of the object at a time t since, in a way, the time has been "stretched" on the first sensors.

At least one second direction may be perpendicular to the first direction. A device according to the invention may further comprise a third set of sensors intended to be arranged in at least a third direction, which intersects the first at a point at which is disposed a common sensor belonging to the first and third sets.

The calculation means can furthermore make it possible to calculate the speed of the vehicle.

A device according to the invention may comprise a plurality of first sets of sensors and a plurality of second sets of sensors forming a 2D matrix of sensors, the matrix being able to be hollow.

According to a variant, the device according to the invention may comprise a first set of sensors, at least one second set of sensors, and at least one 2D array of sensors arranged on at least one of the sides of the first set.

At least one field or field gradient sensor, 1D, or 2D or 3D may be arranged in the vertical direction, or be offset.

The calculation means may make it possible to form a spatial representation of the signature of the vehicles, and / or to extract from said spatial representation of the vehicle identification parameters, for example, by thresholding said spatial representation, the length and / or the width of the vehicle, or, by detecting the intensity maxima, the number of axles of the vehicle, and / or calculating the energy of the signature and / or at least part of its Fourier coefficients and / or the angle traveled by the magnetic field vector (using, in addition, a triaxial field sensor), and / or the drift of the signature P (X, Y) along X and / or a gradient map and / or a vertical gradient of the field and the ratio of this gradient to the field.

These parameters can be used in a classification algorithm.

The invention also relates to a method for recognizing the magnetic signature of a moving object comprising the implementation of a device according to the invention, as described above.

According to the invention, first magnetic measurements are used in the direction of displacement to obtain a law between the time and the position of the sensors, then this law is applied to another series of measurements made in at least one other direction. The temporal notion disappears and we obtain a spatial image of the object, but which is not a photo of the object at a time t since, in a way, the time has been "stretched" on the first sensors.

BRIEF DESCRIPTION OF THE DRAWINGS

• The figure 1 represents a simulated example of the temporal deformity of a signature of a vehicle passing rapidly on one sensor, and slowly, with stop, on another,
• the figure 2 represents a device, according to the invention, in "T", with two lines,
• the figure 3 represents a "morphing", making it possible to connect a temporal function S ₀ (t) and a spatial function S ₀ (x),
• the Figures 4A - 4I represent images for 3 components, before and after transformation of "morphing" type,
• the Figures 5A - 5C represent variants of devices according to the invention,
• the figure 6 illustrates an embodiment of a bimatric device,
• the figure 7 illustrates an embodiment of a device with several "T".

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

A device according to the invention and different variants and their implementation will first be described.

The data processing is then described.

A first embodiment of the invention implements a multi-sensor device.

Spatial and temporal information is used to extract the characteristics of the magnetic signatures of vehicles. Each sensor is an element capable of measuring one or more components of the local magnetic field or the local magnetic gradient (such as FluxGate magnetometers for example).

These sensors are distributed, as illustrated on the figure 2 on at least one line 2 oriented parallel to the running direction (X direction, C ^x _i sensors) and on at least one line 4 oriented differently (Y direction, C ^Y _i sensors), these lines comprising at least one common sensor C ^xy ₀ .

We then have temporal and spatial information, which can be linked by a technique called "morphing", such as for example described in the article of CSMyers et al. "The comparative study of several dynamic time-warping algorithms for connected-word recognition", The Bell System Technical Journal, Vol. 60, No7, 1981 . It is thus possible to build the 2D spatial photo of the vehicle signature.

The provision described above covers a set of cases, some of which are illustrated by way of example in the following paragraphs.

According to a first exemplary embodiment, called a basic device, the sensors are available for example in the form of a "T" (case of the figure 2 ).

• N_x (>1) capteurs C^X _i sur une seule ligne 2,
• N_y (>1) capteurs C^Y _i sur une seule ligne 4.

In a first version, the number of sensors is reduced: it is limited to two lines, unlike the general case where we can have more than two lines. We dispose :

• N _x (> 1) C ^X _i sensors on a single line 2,
• N _y (> 1) C ^Y _i sensors on a single line 4.

On the figure 2 Y is perpendicular to X, thus transverse to the rolling direction.

These two lines 2, 4 have at least one sensor C ^XY ₀ in common at their intersection. This can be located anywhere along lines 2 and 4. For example, the figure 2 locates the sensor C ^XY ₀ at the beginning of the line 2 and in the center of the line 4, but other arrangements are possible, the line 4 may for example be located between the ends of the line 2 (see position 4 'on the figure 2 ) with a sensor C ^XY _{0 '} in common between the lines 2 and 4'.

In addition, the sensors can be uniformly distributed on each line, or arranged with a variable pitch. In particular, on line 4, it is interesting to concentrate the density of sensors in the areas where, statistically, the wheels of the vehicles can pass, in order to have in particular the signatures of the axles, important elements in the automotive classification. It is this particular case that is represented on the figure 2 .

At each moment, the measurements from the sensors C ^x _i arranged along the line 2 provide a spatial profile S _o (x), or section along X, of the signature of the vehicle.

A pre-processing, of the thresholding type, for example, makes it possible to detect the beginning and the end of the useful magnetic signature.

Each spatial profile is, in whole or in part, comparable to the time measurement S _o (t), resulting from the sensor C ^xy ₀ when the vehicle passes above the intersection of the lines 2, 4. The main difference comes from the temporal deformation of the spatial signature related to the speed of the vehicle. Minor dissimilarities can also appear locally along the magnetic signature, since S _o (t) is a developed of the local signature (in C ^xy ₀ ) of the vehicle, whereas S _o (x) is a snapshot. Overall, the signal S _o (t) can be seen as a compressed version of the signal S _o (x) (if the vehicle accelerates), dilated (if it brakes), constant (if it stops), or even returned (if the vehicle backs up), and possibly deformed in pieces.

One can use a "morphing" technique (as for example the "Direct Time Warping" algorithm used in speech processing, see bibliographic reference given above, article by CS Myers et al.) To determine the relationship L (t - x) between these two signals S _o (x) and S _o (t).

• un éloignement plus ou moins fort par rapport au point voisin (accélération ou freinage),
• une répétition pendant un certain temps (arrêt),
• un éloignement en sens opposé (recul).

A "morphing" algorithm seeks point-to-point correspondence between two forms, as shown in figure 3 , on which the curves I 'and II' respectively represent S _o (x) and S _o (t). The algorithm makes it possible to find a point of the spatial signature S _o (x) having undergone:

• a distance more or less strong compared to the next point (acceleration or braking),
• a repetition for a while (stop),
• a distance in the opposite direction (recoil).

The technique of "morphing" applies well to this problem because the set of magnetic dipoles that form a vehicle follows the same kinetics.

This is a technique to progressively move from one signal to another, in the most continuous way possible. Such a technique is for example described in the C.S.Myers document already cited above.

Moreover, the relation L (t - x) is also characteristic of the velocity profile of the vehicle when it passes over the sensor C ^xy ₀ . At the end of the "morphing" step, we obtain the relation giving x as a function of t, x = f (t). The speed results from the integration of this function.

Then, the data from the sensors C ^y _i are used.

In the course of time, these measurements form an image I (t, Y) distributed according to time and on line 4. It is possible to apply the relation L (tx), determined previously, at each column i of I (t, Y), that is to say at each time signal coming from the sensors C ^y _i .

• en figures 4A - 4C : les images I(t, Y) pour les 3 composantes B_x, B_y, B_z du champ ;
• en figures 4D - 4F : les coupes centrales illustrant S₀(t) après « morphing » (en trait fin) à S₀(x) (en trait gras) (pour chaque composante B_x, B_y, B_z).
• en figures 4G - 4I : les images P(X,Y) spatiales issues des capteurs C^y _i. (Là encore : pour chaque composante B_x, B_y, B_z).

This gives a photo P (X, Y) of the vehicle signature, from a single line of sensors. Thus are represented:

• in Figures 4A - 4C : the images I (t, Y) for the 3 components B _x , B _y , B _z of the field;
• in Figures 4D - 4F : the central sections illustrating S ₀ (t) after "morphing" (in fine lines) to S ₀ (x) (in bold lines) (for each component B _x , B _y , B _z ).
• in Figures 4G - 4I spatial P (X, Y) images from C ^y _i sensors. (Again: for each component B _x , B _y , B _z ).

P (X, Y) is a time course of the signature, placed in space, without having to determine the speed of the vehicle or without making any hypothesis on its trajectory.

Its acquisition is therefore independent of the driving speed and the trajectory of the vehicle.

• la mesure temporelle So(t) par un capteur magnétique Co (le capteur C^XY0) placé sur la trajectoire de l'objet,
• la mesure S(x) à un instant ta par des premiers capteurs magnétiques C^x _i alignés avec Co selon la direction 2 de déplacement selon X,
• la comparaison point à point du signal temporel So(t) et du signal spatial S(x),
• l'élaboration d'une relation t(x) entre les temps t et les sites ou les positions x le long de la direction de déplacement,
• les mesures Sy(t) au cours du temps par des seconds capteurs magnétiques C^y _i alignés avec Co selon une direction Y (direction 4 ou 4') différente de la direction de déplacement 2,
• la transformation par la relation t(x) des mesures Sy(t) en un signal spatial Sy(x), image magnétique spatiale de l'objet.

According to the invention, a method for recognizing the magnetic signature of a moving object comprises:

• the temporal measurement So (t) by a magnetic sensor Co (the sensor C ^XY 0) placed on the trajectory of the object,
• the measurement S (x) at an instant ta by first magnetic sensors C ^x _i aligned with Co in the direction of displacement 2 along X,
• the point-to-point comparison of the time signal So (t) and the spatial signal S (x),
• the development of a relation t (x) between the times t and the sites or the positions x along the direction of displacement,
• the measurements Sy (t) over time by second magnetic sensors C ^y _i aligned with Co in a direction Y (direction 4 or 4 ') different from the direction of movement 2,
• the transformation by the relation t (x) of the measurements Sy (t) into a spatial signal Sy (x), spatial magnetic image of the object.

According to a second exemplary embodiment, the basic device can take other forms, and the invention described above can be applied in different configurations of the sensors.

In the various configurations, it is sought to have a line 2 arranged in the direction of the vehicle running (X direction), and a common sensor with another line 8 of sensors, along which will be applied the "morphing".

• figure 5A : système à lignes de capteurs 2, 40 formant un « V » ;
• figure 5B : système à lignes de capteurs 2, 42 formant un demi « T » transverse;
• figure 5C : système à plusieurs lignes « Y » 42, 44, 46, chacune faisant un angle avec la ligne 2 ; Comme pour le cas de la figure 2, la technique de « morphing » peut être appliquée pour chaque ligne « Y » 42, 44, 46 à partir de la signature S(x) et de chaque signature temporelle issue du capteur commun entre chaque ligne « Y » et la ligne des capteurs formant S(x).

The Figures 5A - 5C present several examples of geometry according to this embodiment:

• Figure 5A : sensor line system 2, 40 forming a "V";
• Figure 5B : sensor line system 2, 42 forming a transverse half "T";
• Figure 5C : multi-line "Y" system 42, 44, 46, each at an angle to line 2; As for the case of figure 2 , the "morphing" technique can be applied for each "Y" line 42, 44, 46 from the signature S (x) and from each time signature resulting from the sensor common between each line "Y" and the line of sensors forming S (x).

According to a third embodiment ( figure 6 ), a "hollow" matrix device is formed: n lines 2, 2 ₁ , 2 ₂ , 2 ₃ ..., 2 _n are arranged parallel to one another in the direction of movement of the vehicles, whereas m lines 4, 4 ₁ , 4 ₂ , 4 ₃ ..., 4 _m are arranged in the direction Y, parallel to each other. These m lines could be arranged other than perpendicular to the X axis. A sensor is disposed at each intersection 2 _i - 4 _j .

The low cost and compactness of the sensors used to gather a greater amount of information is used: the device forms a 2D matrix, or a carpet, of sensors which are distributed under the roadway, uniformly or otherwise.

This matrix is "hollow" in some places: it lacks sensors or their density is not satisfactory for the precision required by the application. We then use the principle described above ("morphing") to complete the missing data.

In the matrix, two lines as described above are chosen to form a two-line system at the intersection of which a sensor is located, and the morphing technique is used to reconstruct the missing data in the chosen zone. This can be repeated at several places in the matrix.

At every moment, the measurements made by all the sensors form a spatial photo P (X, Y) of the signature of the vehicle, completed in some places by the technique according to the present invention.

As before, the acquisition of this photo is independent of the running speed and the trajectory of the vehicle.

A fourth embodiment is a system with several basic devices ( figure 7 ).

One adds to one of the basic devices (on the figure 7 a "T" device) described above in connection with FIGS. 5A-5C, a small sensor array 300 (denoted M _ij ) placed on one side (or both sides) of the "T", and occupying a length l _x .

Thus, it is possible locally to obtain a snapshot of part of the vehicle signature. In particular, if l _x is approximately 3 m, this matrix provides the spatial unwinding of one or more sets axle + wheel + tire of a car or a truck.

In addition, the matrix M _ij also makes it possible to capture the signatures of small vehicles that could provide a very weak signal on the line 2 of sensors. This can happen in particular when a motorcycle circulates in a toll channel by tightening well on one side to make the transaction.

At every moment, the 2D photo from the sensors of the matrix can locate the bike. Preprocessing provides the beginning and end of the useful signature.

It is thus possible to determine which line of sensors M _i of the matrix coincides best with the running axis of the motorcycle.

This line can then be used with the sensor line 4 to form a new "T" device, as explained above, of dimension and positioning more adapted to this vehicle. By a method of "morphing" identical to that already presented above, it is then possible to recover the photo P (X, Y) of the magnetic spatial signature of the motorcycle.

According to a fifth exemplary embodiment, at least one sensor (field or gradient field, 1D, 2D or 3D) in the vertical direction is added to one of the devices described above. This system makes it possible to measure, at a distance D _z , one or more components of the field (or gradient) in at least one plane different from that of one of the devices described above. This information may be relevant for having vehicle height data.

A sixth exemplary embodiment is a device with a remote reference.

In this version, the device described above is added to the reference measurement means (field or field gradient 1D, or 2D or 3D) remote. This means that these means are located far enough from the measurement zone not to be sensitive to the passage of the vehicle. This reference measurement makes it possible to improve measurement accuracy by subtracting geomagnetic and surrounding noise (industrial noise, streetcar, electrical network, etc.)

During the installation of the device, the sensors can for example be grouped into lines, which are seen as branches of the tree system that manages the acquisition and storage of data.

A line comprises 1 or more nodes, each comprising a mono, bi or tri-axis sensor and the associated electronics (filtering, amplification, digitization, multiplexing). Each node is linked on a digital high speed information exchange bus (USB for example).

A central system 50 ( figure 6 ), for example a microcomputer specially programmed for this purpose, for example offset at the edge of the roadway, manages the multiplexing, the timing of acquisitions, and the storage of data. It also embeds means or the processing system which carries out the exploitation of the measurements (pretreatment, morphing, extraction of the parameters, classification).

Physically, the lines may be in the form of tubes buried under the roadway or bars inserted into grooves on the surface of a road surface. This mechanism has the advantage of greater ease of implementation of the classification device and less maintenance compared to current loops (which undergo "hard" deformation of the road and the incessant passages of vehicles). If a sensor proves defective, the line is removed from the ground, and the sensor easily replaced. The central system 50 is not modified. Similarly, all or part of the lines can be used, depending on the needs of the classification system, without having to intervene on the roadway.

The exploitation of the data will now be described.

All the devices and methods described above make it possible to capture the spatial 2D photo P (X, Y) of the vehicle. In the case where several components of the field or of gradient are recorded, one obtains as many images as of components.

At first, the parameters identifying the vehicle, or its type, are extracted from the photo. This provides the image of the distribution of the characteristic dipoles of the signature.

For example, by thresholding, the spatial dimensions of the signature in the Y and X direction provide the width and length of the vehicle, regardless of its speed, whether it is running, stopped, or even in reverse.

Detection of intensity maxima provides the number of axles, as well as their 2D positioning and relative spacing.

The exploitation of the spectral content of the image gives the energy of the signature and its main Fourier coefficients.

If we have three photos, from tri-axis field sensors, it is also possible to calculate the angle traveled by the vector total magnetic field of the vehicle B = B _x + B _y + B _z . This one is characteristic of the soft or disturbed character of the signature, and can be indicative of the height between vehicle and ground.

With a device according to the invention, the data obtained in the X direction can be strongly oversampled with no additional installation costs related to the sensors and the associated electronics, since they come from a time acquisition. We can then easily approximate the drift of P (X, Y) along X by calculating the difference P (X _i , Y) - P (X _i-1 , Y). This gives a gradients map whose operation can better locate the vehicle axles.

By extending a map of gradients (measured or calculated), we can obtain the vertical gradient and calculate, via the ratio of the vertical gradient on the field, an indication of the distance separating the magnetic sources that characterize the vehicle from the sensors. is a magnitude related to the height of the vehicle.

In a second step, these parameters are used in a classification algorithm. One solution is based on the implementation of learning-restitution type neural network, for example.

A device 50, such as a microcomputer, is programmed to implement one of the methods described above, from the measurements delivered by the sensors.

Claims (19)

1. Device for measuring the magnetic signatures of vehicles including:

   - at least a first set of sensors (C^x _i), designed to be arranged along at least a first direction (2),

   - at least a second set of sensors (C^y _j), designed to be arranged along at least a second direction (4), that intersects the first direction at a point at which a common sensor (C^xy ₀) is placed, belonging to the first and the second set,

   - calculation means (50) to calculate a relation between the time signature S_o(t) of a vehicle passing above the common sensor and a spatial profile S_o(x) resulting from measurements made by the sensors in the first set of sensors.
2. Device set forth in claim 1, at least one second direction being perpendicular to the first direction.
3. Device set forth in claim 1 or 2, including a third set of sensors designed to be arranged along at least a third direction, that intersects the first direction at a point at which a common sensor (C^xy ₁) is placed, belonging to the first and third set.
4. Device set forth in any one of claims 1 to 3, the calculation means being also used to calculate the vehicle speed.
5. Device set forth in any one of claims 1 to 4, including a plurality of first sets of sensors and a plurality of second sets of sensors forming a 2D matrix of sensors.
6. Device set forth in claim 5, the matrix being hollow.
7. Device set forth in any one of claims 1 to 4, including a first set of sensors, at least one second set of sensors, and at least one 2D matrix of sensors arranged on at least one of the sides of the first set.
8. Device set forth in any one of claims 1 to 7, also including at least one 1D, or 2D or 3D field sensor or field gradient sensor along the vertical direction.
9. Device set forth in any one of claims 1 to 8, including at least one 1D, or 2D or 3D offset field sensor or field gradient sensor.
10. Device set forth in any one of claims 1 to 9, the calculation means being used to form a spatial representation of the signature of vehicles.
11. Device set forth in claim 10, the calculation means being used to extract vehicle identification parameters from said spatial representation.
12. Device set forth in claim 11, the calculation means being used to extract the length and/or the width of the vehicle by thresholding said spatial representation.
13. Device set forth in claim 11 or 12, the calculation means being used to extract the number of vehicle axles by detection of intensity maxima.
14. Device set forth in any one of claims 11 to 13, the calculation means being used to calculate the energy of the signature and/or at least part of its Fourier coefficients.
15. Device set forth in any one of claims 11 to 14, also including a triaxial field sensor, the calculation means being used to calculate the angle crossed by the magnetic field vector.
16. Device set forth in any one of claims 11 to 15, the calculation means being used to calculate the derivative of the signature P(X,Y) along X.
17. Device set forth in claim 16, the calculation means being used to calculate a map of gradients.
18. Device set forth in claim 16 or 17, the calculation means being used to calculate a vertical gradient of the field and the ratio of this gradient to the field.
19. Device set forth in any one of claims 11 to 18, said parameters being used in a classification algorithm.

Images