EP0841647A1 - Procédé et dispositif de surveillance du traffic - Google Patents

Procédé et dispositif de surveillance du traffic Download PDF

Info

Publication number
EP0841647A1
EP0841647A1 EP97118758A EP97118758A EP0841647A1 EP 0841647 A1 EP0841647 A1 EP 0841647A1 EP 97118758 A EP97118758 A EP 97118758A EP 97118758 A EP97118758 A EP 97118758A EP 0841647 A1 EP0841647 A1 EP 0841647A1
Authority
EP
European Patent Office
Prior art keywords
magnetic field
vehicle
sensor signals
field sensors
dimensional
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP97118758A
Other languages
German (de)
English (en)
Inventor
Klaus Wasserberg
Alfred Lehmann
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Deutsches Zentrum fuer Luft und Raumfahrt eV
Original Assignee
Deutsches Zentrum fuer Luft und Raumfahrt eV
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Deutsches Zentrum fuer Luft und Raumfahrt eV filed Critical Deutsches Zentrum fuer Luft und Raumfahrt eV
Publication of EP0841647A1 publication Critical patent/EP0841647A1/fr
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/042—Detecting movement of traffic to be counted or controlled using inductive or magnetic detectors
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/015—Detecting movement of traffic to be counted or controlled with provision for distinguishing between two or more types of vehicles, e.g. between motor-cars and cycles

Definitions

  • the invention relates to a method and a device for traffic monitoring with the or it is possible is to record and recognize individual vehicles.
  • Induction loops embedded in the pavement are known, for example to control traffic lights be used. Individual vehicles can not detected and recognized with this device can, at best, by evaluating the induction signals the vehicles certain categories, e.g. Vehicle types. Besides, they are Induction loops both in procurement as well expensive to install.
  • the induction loops are in one or several rows in a row arranged to the roadway.
  • the sensitivity of the induction loops is the actual sensor loop by several surrounding auxiliary loops increased.
  • the Sensor loops are connected to an evaluation device which is the phase angle between voltage and Measures current in the sensor loop.
  • optical detection of vehicles In addition to magnetic detection, there are also devices known for the optical detection of vehicles. Such systems are used, for example, in traffic management systems used on highways. With these optical Sensors is an identification of a vehicle not possible. It is only recognized whether a Vehicle has passed.
  • DE 35 21 655 A1 describes a device for detecting described by vehicle traffic with magnetic field detectors.
  • the magnetic field detectors measure interference a given magnetic field by a vehicle, so that the presence and / or movement of a vehicle can be determined. With this facility however different vehicle types can be distinguished.
  • the magnetic Characteristics of a passing vehicle several magnetic field sensors distributed across the road in each case on several spatially separated Places measured selectively. This is done in several successive times. From the multitude of a sensor vector is generated, which characterizes the vehicle and therefore it makes it recognizable. The process works so precisely that even differentiated between vehicles of the same model can be. Because only the magnetic characteristic of the vehicle is not measured It is possible to draw conclusions about the owner of the vehicle.
  • the device for traffic monitoring has a cross to the direction of traffic arranged row of magnetic field sensors, whose signals are processed in a computer unit.
  • the analog sensor signals go digital converted and summarized in one measured value, several of these that have arisen at successive times Measured values are then processed into a one-dimensional one Combined measured value sequence.
  • This processed Measured value sequence is with a transformation matrix to a feature vector identifying the respective vehicle transformed. Due to this processing The sensor signals make the vehicles recognizable increases and the scope of the vehicle identifying Record is reduced, reducing processing speed the device is increased.
  • the Magnetic field sensors so sensitive that they affect the earth's magnetic field measure and those caused by the vehicle Determine changes in the earth's magnetic field. Thereby high measuring accuracy of the individual measurements guaranteed, so that same vehicle models with identical equipment can be distinguished. Tests of the invention have shown that vehicles just because of the different rolling directions the body panels can be distinguished.
  • this corresponds to one lane, at least 12 magnetic field sensors arranged in a row. This will sufficient measurement resolution is achieved to the vehicles to be able to record and recognize them safely.
  • the Magnetic field sensors with a clock pulse the sensor signals out at the same time. This has the advantage that the sensor signals can be processed in parallel, making them quicker and easier to get a reading can be summarized.
  • the measured values can initially be multidimensional Measurement sequence are summarized, which then to a prepared one-dimensional measurement sequence in vector form is reduced. So one becomes a vehicle characterizing matrix a vector that is still sufficient Information about the characterization of the vehicle contains.
  • the vector can be easier and faster can be processed as a matrix.
  • the transformation matrix preferably runs out a covariance matrix that is in a learning phase follow from the prepared one-dimensional measured value several vehicles is formed.
  • a covariance matrix is one known from image recognition, among other things statistical means with which deviations of a single pattern calculated from the average of all samples can be. With this method, the scope of the feature vector can be reduced since the Vehicles characterized by deviations from the average and not by a complete sentence of measured values. This enables quick processing of the sensor signals.
  • the analog-digital conversion of the analog sensor signals can be done directly in the sensor units, whereby the susceptibility to interference of data transmission from the magnetic field sensors to the computing unit is reduced.
  • FIG. 1 shows a traffic monitoring device 1 with sixteen magnetic field sensors 2 that are in a straight line Row are laid across a lane 3.
  • the sensors 2 are either embedded in the pavement or under a threshold that can be driven over attached to the pavement.
  • the magnetic field sensors 2 measure the magnetic characteristics of a passing vehicle Vehicle 4. They are so sensitive that they measure the earth's magnetic field. While no vehicle 4 located in the measuring range of the magnetic field sensors 2 the sensor signals corresponding to the value of the earth's magnetic field of the magnetic field sensors 2 stored in order as Reference values are available. Is a Vehicle 4 in the measuring range of the magnetic field sensors 2 the measured value is greater than the stored reference value and the difference is used as a sensor signal.
  • the sensors 2 are connected to a via a bus line 5 Computer unit 6 connected.
  • the on a clock signal the Computer unit 6 output by the magnetic field sensors 2
  • Analog sensor signals are in the magnetic field sensors 2 digitally converted and parallel over the Transfer bus line 5 to the computer unit 6.
  • a digital measured value 7 is input to the computer unit 6 (Fig. 2) with a word length of 16 digits (corresponding the 16 sensors), with each digit the associated Amplitude value A indicates.
  • the measured value 7 sets are thus digitally transformed and sensor signals transmitted in parallel from all 16 magnetic field sensors 2 together.
  • a measured value sequence generator 8 they are passed of a vehicle 4 successively generated measured values 7 to a characteristic multidimensional sequence of measured values 9 summarized. It is after each clock pulse a measured value 7 for the previously recorded measured values added until the vehicle 4 the series of Has completely run over magnetic field sensors 2.
  • Fig. 2 is such a measurement sequence 9 in three-dimensional Simplified form or for a very slow moving Vehicle shown. There is usually a multi-dimensional one Measured value sequence 9 from significantly more measured values 7.
  • the points S1 to S16 on the coordinate S directed transverse to the roadway are assigned to the individual magnetic field sensors 2.
  • the coordinate t of time is plotted in the longitudinal direction of the lane.
  • the current measured value 7 is added to the measured values recorded up to the time t n at the time t n + 1 .
  • the index n continuously indicates the number of clock cycles. This creates the grid spacing, which is only determined by the timing, in the direction of the coordinate t.
  • the amplitude A of a grid point is in relation to the measured magnetic field at the respective time t n at the respective magnetic field sensor S.
  • the grid point 11 characterizes a sensor signal which at time t 5 , ie five clock cycles after the vehicle 4 in the measuring range of the sensors 2 arrived, from which sensor S2 was recorded.
  • the strongest amplitudes A are 10 at four places observe. These are through the wheels or the metal rims of the vehicle 4 caused. Because magnetic fields fall exponentially, parts of the vehicle 4, which are closest to the magnetic field sensors 2, stronger detected as more distant parts. This explains the high amplitudes at points 10 in the area of the wheels of the vehicle 4.
  • the discrete quantized samples of the characteristic multidimensional measurement sequence 9 of vehicles can already be considered vehicle descriptive will.
  • the amount of data per vehicle is relatively large. For example, for a 30th km / h fast vehicle of 4 m length 480 measurements per Sensor (at a frequency of 1000 Hz). Used at 16 Sensors result in a total of 7680 feature components, which is not in one for vehicle detection realistic time can be processed. In the extraction stage 8a will show the number of feature components 64 reduced.
  • the multidimensional measurement value sequences 9 additionally contain Information that is not vehicle-specific (Irrelevance), such as that of the wheels of the Characteristic components evoked by vehicles 10.
  • the multidimensional measurement value sequences 9 are prepared in such a way that object-specific properties like instantaneous speed of vehicles, vehicle lengths, Noise and interference components, sensitivity and amplification factors the detectors have different ground clearances, Deflection in the event of uneven road surfaces eliminated are.
  • the multidimensional measurement sequence 9 in Direction of the t-coordinate to a length of 32 support points reduced.
  • This compression corresponds to one Time normalization because of the speed or the length of the vehicle is eliminated.
  • a time normalization factor is calculated from the ratio between the number of recorded Bases of the multidimensional sequence of measured values 9 in the direction of the t-coordinate and the predefined Number (32) of support points defined.
  • This normalization factor determines the number of bases of the original multidimensional measurement sequence leading to a Support point of the time-standardized multidimensional measurement sequence be summarized. By summarizing them Bases of the original measurement sequence a degree of regression is set, the slope converted for the time-standardized sequence of measured values becomes.
  • the feature values are summarized in regions and a prepared one-dimensional measurement value sequence 9a in vector form (FIG. 3) is formed.
  • a heuristic analysis of the multidimensional measurement value sequences 9 shows that there is a statistical link between the feature components. It is mainly caused by the statistical dependence of the input signal. General design principles influence the metal mass distribution on the underside of the vehicle, so that even when viewed ideally, no independent feature components can be obtained. This includes, in particular, the feature components 10 caused by the wheels of the vehicle. For further information compression, the feature components which have been recorded by the sensors which are arranged in the area of the wheels of the vehicle are therefore removed from the multidimensional measurement value sequence 9.
  • the feature components of the sensors S 1 to S 5 and S 12 to S 16 which have detected the vehicle wheels, are suppressed.
  • these feature components can be used to determine the vehicle width and the center distance. They are therefore well suited for differentiating between different types of vehicles, especially for car, truck and motorcycle classification.
  • the remaining feature components of sensors S 6 to S 11 are combined into two groups.
  • the values of the first three sensors (S 6 to S 8 ) at the 32 support points are combined to form an average, so that a sequence of 32 averaged feature components is produced.
  • the feature components of the three sensors S 9 to S 11 are summarized accordingly.
  • the two sequences are attached to each other so that a one-dimensional measurement sequence 9a in the form of a measurement vector is created (FIG. 3).
  • the first 32 values of the one-dimensional measurement value sequence 9a come from the sensors S 6 to S 8 , the last 32 values from the sensors S 9 to S 11 .
  • the number of values that characterize the vehicle can be limited to 64 per vehicle. This number is of a magnitude that is favorable for vehicle detection.
  • the amplitude of the one-dimensional measurement value sequence is standardized in the extraction stage 8a.
  • the amplitude value of a measuring point (X (n) samples ) is divided by the value of the maximum amplitude (X max ) and multiplied by 255 in order to obtain amplitude-standardized values.
  • Fig. 3 along the abscissa are the numbers of the Vehicle characterizing values and along the ordinate the amounts of these values (standardized to 255) specified.
  • the characteristic multi-dimensional measurement sequence an inclined, e.g. overtaking vehicle can also lead to a sequence of measured values prepared with other comparable measured value vector 9a normalized will.
  • the measured value vector 9a is multiplied in a feature vector generator 12 by a transformation matrix M T (FIG. 5).
  • the transformation matrix M T is obtained from a covariance matrix M K , which is shown in FIG. 4.
  • the covariance matrix M K is formed in a learning phase 13 from the measured value vectors 9a of numerous vehicles in a learning part 13.
  • the eigenvalues of the covariance matrix M K are the diagonal elements of the actual transformation matrix M T.
  • the other elements of the transformation matrix M T are zero. With the assumed Gaussian distribution, this means the statistical independence (uncorrelation) of the components.
  • the diagonal components are very large at first (this corresponds to a very high information content) and then decrease very sharply. The most important components are selected by neglecting those components which, according to their intrinsic values, have only a small amount of information.
  • V M M T x X .
  • the feature vector V M obtained from the transformation is shown in FIG. 6.
  • the elements of the feature vector V 11 are plotted along the abscissa, while the ordinate shows the non-normalized values of the elements that were created during the transformation.
  • the transformation matrix M T was applied to the measured value vector, which has 64 elements. Therefore, the feature vector V M consists of 64 elements.
  • the elements with the highest values, ie with the highest distinctive character, are among the first 32 elements of the feature vector V M.
  • the last 32 elements have very small values or the value zero, ie they do not contribute any essential information to distinguish vehicles. Therefore the last 32 elements can be neglected. This results in a feature vector V M with 32 digits which uniquely characterizes a vehicle.
  • the feature vector V M generated in the feature vector generator 12 is stored in a comparator 15 and compared with the feature vectors of earlier vehicles previously stored in the computer 6. If the new feature vector V M is unknown, a new data record for the vehicle 4 is created and saved. If the vehicle 4 can be identified on the basis of its feature vector V M , the location and time of the vehicle 4 are recorded in the data record.
  • FIG. 7 shows the structure of a magnetic field sensor 2 shown.
  • the magnetic field sensor 2 consists of two individual ones Sensor components 2a and 2b in their longitudinal direction are arranged rotated by 90 ° to each other.
  • the sensor components 2a and 2b have preferred directions regarding the sensitivity. Towards her The longitudinal sensitivity is low. To everyone Directional components of the magnetic field with the same sensitivity To be able to detect the sensor components 2a and 2b offset by 90 ° so that the disadvantageous measurement characteristics of the individual sensor components 2a, 2b is balanced. To generate the Sensor signal, the two component signals become geometric added.
  • the sensor signal is converted digitally and via the bus line 5 to the computer unit 6 passed on.
  • the measured value sequence generator 8, the extraction stage 8a, the feature vector generator 12 and the comparator 15 can also in the computer unit 6 by appropriate Software can be realized.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)
EP97118758A 1996-11-12 1997-10-29 Procédé et dispositif de surveillance du traffic Withdrawn EP0841647A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE1996146632 DE19646632C1 (de) 1996-11-12 1996-11-12 Verfahren und Vorrichtung zur Verkehrsüberwachung
DE19646632 1996-11-12

Publications (1)

Publication Number Publication Date
EP0841647A1 true EP0841647A1 (fr) 1998-05-13

Family

ID=7811372

Family Applications (1)

Application Number Title Priority Date Filing Date
EP97118758A Withdrawn EP0841647A1 (fr) 1996-11-12 1997-10-29 Procédé et dispositif de surveillance du traffic

Country Status (2)

Country Link
EP (1) EP0841647A1 (fr)
DE (1) DE19646632C1 (fr)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000049590A1 (fr) * 1999-02-18 2000-08-24 The University Court Of The University Of Edinburgh Detecteur et classificateur de vehicules
FR2790834A1 (fr) * 1999-03-12 2000-09-15 Philippe Gendrier Systeme d'imagerie magnetique par reseau matriciel de magnetometres
EP1139086A3 (fr) * 2000-03-29 2004-01-07 MTU Friedrichshafen GmbH Procédé de détection de ratés d'allumage utilisant la vitesse du vilebrequin
EP1361555A3 (fr) * 2002-05-10 2004-03-03 Siemens Aktiengesellschaft Système pour exploiter un parking et/ou pour enregistrer des vehicules à l'interieur comme à l'exterieur
FR2896070A1 (fr) * 2006-01-11 2007-07-13 Commissariat Energie Atomique Systeme magnetique de controle de trafic
DE102011014855A1 (de) * 2011-03-24 2012-09-27 Thales Defence & Security Systems GmbH Verfahren und Vorrichtung zum Erfassen und Klassifizieren von fahrenden Fahrzeugen
EP3091372A1 (fr) * 2015-05-05 2016-11-09 Centro de Cálculo Igs Software S.L. Système de détection de véhicule

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB201503855D0 (en) 2015-03-06 2015-04-22 Q Free Asa Vehicle detection

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE3521655A1 (de) * 1985-06-18 1987-01-15 Mueller Ind Management Gmbh Einrichtung zum erfassen von fahrzeugverkehr mit magnetfelddetektoren
US5264793A (en) * 1991-04-11 1993-11-23 Hughes Aircraft Company Split array dipole moment detection and localization
WO1995028693A1 (fr) * 1994-04-19 1995-10-26 Honeywell Inc. Detecteur magnetometrique de vehicules

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE4231881A1 (de) * 1992-09-24 1994-03-31 Ant Nachrichtentech Anordnung zum Erfassen von Verkehrsgrößen

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE3521655A1 (de) * 1985-06-18 1987-01-15 Mueller Ind Management Gmbh Einrichtung zum erfassen von fahrzeugverkehr mit magnetfelddetektoren
US5264793A (en) * 1991-04-11 1993-11-23 Hughes Aircraft Company Split array dipole moment detection and localization
WO1995028693A1 (fr) * 1994-04-19 1995-10-26 Honeywell Inc. Detecteur magnetometrique de vehicules

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000049590A1 (fr) * 1999-02-18 2000-08-24 The University Court Of The University Of Edinburgh Detecteur et classificateur de vehicules
FR2790834A1 (fr) * 1999-03-12 2000-09-15 Philippe Gendrier Systeme d'imagerie magnetique par reseau matriciel de magnetometres
EP1139086A3 (fr) * 2000-03-29 2004-01-07 MTU Friedrichshafen GmbH Procédé de détection de ratés d'allumage utilisant la vitesse du vilebrequin
EP1361555A3 (fr) * 2002-05-10 2004-03-03 Siemens Aktiengesellschaft Système pour exploiter un parking et/ou pour enregistrer des vehicules à l'interieur comme à l'exterieur
FR2896070A1 (fr) * 2006-01-11 2007-07-13 Commissariat Energie Atomique Systeme magnetique de controle de trafic
EP1811479A1 (fr) * 2006-01-11 2007-07-25 Commissariat à l'Energie Atomique Système magnétique de contrôle de trafic
JP2007188503A (ja) * 2006-01-11 2007-07-26 Commiss Energ Atom 磁気交通制御システム
US7765056B2 (en) 2006-01-11 2010-07-27 Commissariat A L'energie Atomique Magnetic traffic control system
DE102011014855A1 (de) * 2011-03-24 2012-09-27 Thales Defence & Security Systems GmbH Verfahren und Vorrichtung zum Erfassen und Klassifizieren von fahrenden Fahrzeugen
EP3091372A1 (fr) * 2015-05-05 2016-11-09 Centro de Cálculo Igs Software S.L. Système de détection de véhicule

Also Published As

Publication number Publication date
DE19646632C1 (de) 1998-05-14

Similar Documents

Publication Publication Date Title
EP0560023B1 (fr) Dispositif pour la classification d'une forme, en particulier d'un billet de banque ou d'une pièce de monnaie
DE19620627C2 (de) Objekterfassungsverfahren in einem Fahrzeug
DE69302975T2 (de) Fahrzeugpositionserfassungsanlage auf einer Spur
DE69306954T2 (de) Vorrichtung und verfahren zum erkennen von einem oder mehreren fahrzeugrädern.
EP0915350A2 (fr) Dispositif pour déterminer des données indicatives de l'évolution des voies de circulation
DE102011014855A1 (de) Verfahren und Vorrichtung zum Erfassen und Klassifizieren von fahrenden Fahrzeugen
DE102017131114A1 (de) Verfahren und Vorrichtung zum Bestimmen wenigstens eines Parameters eines Objektes
DE19646632C1 (de) Verfahren und Vorrichtung zur Verkehrsüberwachung
EP2819113B1 (fr) Installation de mesure destinée à l'enregistrement d'une plaque d'immatriculation d'un véhicule lors du passage d'une section de mesure
WO2022135978A1 (fr) Dispositif de détermination de données pour la détermination de la vitesse d'un véhicule, dispositif d'analyse et procédé associé
DE19638069A1 (de) Verfahren und Einrichtung zur Erfassung von Verkehrsdaten von Fahrzeugen
EP2320384A1 (fr) Procédé de reconnaissance de caractéristique et de catégorisation de péage de véhicules
DE69722014T2 (de) Verfahren für die detektion,identifizierung und verfolgen von unsichtbaren gegenständen
DE102019211459A1 (de) Verfahren und Vorrichtung zum Überprüfen einer Kalibrierung von Umfeldsensoren
DE102021100526A1 (de) Verfahren zum Betrieb eines Radarsensors, Radarsensor, Kraftfahrzeug, Computerprogramm und elektronisch lesbarer Datenträger
EP0710927A2 (fr) Méthode de reconnaissance orientée-objet d'objets mouvants
DE3016901C2 (fr)
DE3009679A1 (de) Verfahren zur unterscheidung von fahrzeugen unterschiedlicher bauart im strassenverkehr
DE4328939A1 (de) Verfahren zur Erfassung von streckenbezogenen Verkehrsgrößen
DE69715948T2 (de) Verfahren und Vorrichtung zur Messung der Geschwindigkeit eines Fahrzeuges
DE102005034988B4 (de) Verfahren und Anordnung zur Erfassung von ein Erdmagnetfeld örtlich beeinflussenden Fahrzeugen
EP0903711A2 (fr) Procédé de transmission d'information sur le trafic
DE102019204914A1 (de) Bestimmung einer eindeutigen Relativgeschwindigkeit mittels Modellanpassung
DE19727895A1 (de) Anordnung zur Verkehrserfassung auf einem Streckenabschnitt einer Straße
WO2002056051A1 (fr) Procede pour le traitement de signaux de sortie ou de base d'un dispositif servant a determiner une distance d'un objet

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

AX Request for extension of the european patent

Free format text: AL;LT;LV;RO;SI

AKX Designation fees paid
RBV Designated contracting states (corrected)
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 19981114

REG Reference to a national code

Ref country code: DE

Ref legal event code: 8566