EP0635813A1 - Acoustic evaluation process of road vehicles traffic flow - Google Patents

Abstract

- The invention relates to a method of evaluating the flow of traffic of road vehicles, by an acoustic technique. The method consists: . in defining analysis intervals of predetermined duration (Ti) in the course of which the sound signal is picked up, . in determining, for each of the analysis intervals, on the one hand, the difference between the maximum (Lmax) and minimum (Lmin) levels of the sound signal so as to determine a first acoustic indicator (P1) corresponding to the dynamic range of the signal and, on the other hand, a level (Lp) of the sound signal which is reached during a predetermined fraction of the duration (Ti) of the interval in question, for the purpose of defining a second acoustic indicator (P2) corresponding to the presence of vehicles on the roadway, . and in comparing the acoustic indicators (P1, P2) previously defined with datum values, so as to evaluate the state of the flow of the traffic. <IMAGE>

Description

The present invention relates to the technical field of controlling or evaluating the flow of traffic from road vehicles establishing themselves on a traffic lane in the general sense.

In a conventional manner, a traffic lane of a road or motorway nature comprises a determined number of traffic queues, so that the saturation level is only reached in a very exceptional manner. When for some reason the speed of the flow of vehicles decreases, causing local saturation of the traffic lane, it not only results in an inconvenience for the movement of users, but also an attack on their safety. There therefore appears to be a need, both to detect or prevent accidents and to facilitate traffic conditions for users, to assess traffic in order to be able to control or regulate it. To satisfy this need, it is important to be able to have at a central post, the characteristics of the traffic existing at all times over as long a length as possible of the traffic lane, so as to facilitate operations of prevention, intervention or deviation of the saturated channel.

Numerous systems are known in the state of the art suitable for collecting information along a road axis to be monitored and transmitting this information in real time to a central station, from which it can be taken. intervention and / or prevention measures.

It is thus known practice to install induction loop detectors directly on the roadway at regular intervals along the traffic lane, making it possible to count traffic. Such a technical solution is unsatisfactory mainly because of the high installation, operation and maintenance costs involved in using induction loop detectors installed directly in the roadway.

In an attempt to remedy the drawbacks of this technique, patent application FR-2 675 610 has proposed an installation for evaluating a traffic flow of road vehicles providing for placing at regular intervals, in edge of the lane, microphones which transform the acoustic information emitted by road traffic into an electrical signal intended to be transmitted to a central station. Depending on the noise level obtained in a specific frequency band, the installation makes it possible to determine the state of traffic. It must be considered that this frequency band corresponds to the emission of the pneumatic-road contact. This acoustic emission is a function among other things, of the speed of the vehicles, but also of the nature of the coating placed at the right of the receiving point. Such a solution therefore has the drawback of requiring a calibration of the installation at each detection point. Furthermore, the principle of signal processing recommended by this document only allows to determine if the traffic is fluid or saturated, which prohibits the taking of adequate intervention measures according to the state of saturation of the traffic. In addition, it appears that such a system is not able to prevent periods of disturbance that may occur on the so-called slow lane of the traffic lane, since this technique analyzes the noise emitted by the lane so-called fast.

The present invention therefore aims to remedy the drawbacks set out above by proposing a method designed to ensure almost uninterrupted, even permanent, monitoring of the flow of road traffic and making it possible to detect different traffic conditions, with a view to evaluating at best the intervention measures to be taken.

Another object of the invention is to propose a method for evaluating road traffic over a non-negligible length of the traffic lane, while offering the advantage of having a low cost of installation, operation and maintenance.

• à définir des intervalles d'analyse de durée prédéterminée au cours desquels le signal sonore est recueilli,
• à déterminer pour chacun des intervalles d'analyse, d'une part, la différence entre les niveaux maximum et minimum du signal sonore afin de déterminer un premier indicateur acoustique correspondant à la dynamique du signal et, d'autre part, un niveau du signal sonore qui est atteint pendant une fraction prédéterminée de la durée de l'intervalle considéré, en vue de définir un second indicateur acoustique correspondant à la présence de véhicules sur la voie,
• et à comparer les indicateurs acoustiques précédemment définis à des valeurs de consigne, afin d'évaluer l'état de l'écoulement du trafic.

To achieve these objectives, the evaluation method according to the invention consists of:

• defining analysis intervals of predetermined duration during which the audible signal is collected,
• to determine for each of the analysis intervals, on the one hand, the difference between the maximum and minimum levels of the sound signal in order to determine a first acoustic indicator corresponding to signal dynamics and, on the other hand, a level of the sound signal which is reached during a predetermined fraction of the duration of the interval considered, with a view to defining a second acoustic indicator corresponding to the presence of vehicles on the track,
• and to compare the acoustic indicators previously defined with reference values, in order to assess the state of the traffic flow.

Various other characteristics will emerge from the description given below with reference to the appended drawings which show, by way of nonlimiting examples, embodiments and implementation of the subject of the invention.

Fig. 1 is a schematic plan view of a section of taxiway equipped with an installation implementing the evaluation method according to the invention.

Fig. 2 illustrates the evolution curves of the sound signal in dB, as a function of time in seconds, making it possible to highlight a characteristic of the invention.

Figs. 3A to 3C are curves showing a characteristic of the process according to the invention.

Fig. 4 is a diagram showing the evolution between the two acoustic indicators determined by the method according to the invention.

The method according to the invention is designed to acoustically control or evaluate the flow of traffic from established road vehicles, as illustrated in FIG. 1 , on a traffic lane 1 comprising at least one, and in the example illustrated, three traffic lanes 1₁ to 1₃ . As shown in the figure, queue 1₁ is said to be slow, while queue 1₃ is considered to be fast queue, insofar as the direction of traffic is established from left to right. Of course, it could be envisaged that the method according to the invention could be applied to assess the traffic flow establishing on track 2 in the opposite direction.

The method according to the invention aims to have at the edge of the traffic queue, called the slowest 1₁ , one or more microphones 3 distributed along the section of the traffic lane 1 to be monitored. According to a preferred variant of embodiment, the microphones 3 are placed inside the emergency stop terminals which are generally located at the edge of the traffic lanes.

Each microphone 3 , for example of the electret type, is connected to a processing device 4 which is connected to a central control station 5 via a communication link 6 ensuring for example the power supply to the detectors 3 and the transmission of information from devices 4 .

In the example illustrated, each processing device 4 comprises a stage 7 for amplifying and filtering the signal delivered by the associated microphone 3 . Stage 7 is followed by a circuit 8 for converting the amplified signal into a logarithmic value, in order to obtain a linear scale in decibels. The conversion circuit 8 is connected to a management module 9 managed by a microprocessor which is responsible, in particular, for sampling the signal, for example at the rate of 32 samples per second, and for digitizing them. This module 9 , which performs other functions explained in the description of the process which follows, is connected to a communication module 10 capable of transmitting the information to the central station 5 .

The method according to the invention aims to determine the characteristics of road traffic from the sound waves collected by the microphones 3 . These sound waves originate, in particular, from the powerplant of vehicles and from the acoustic emission due to the contact of tires on the road. It should be noted that due to the sharp decrease in the acoustic level received, as a function of the distance from the microphone to the traffic lanes, the traffic noise from the distant lanes, namely 1₂ , 1₃ in the example illustrated, are masked by those of the nearest queue 1 file.

In order to study the acoustic signal from each microphone 3 , provision is made to collect such a signal S during analysis intervals of predetermined duration T _i . Thus, as shown in FIG. 2 , the sound signal S collected is divided into analysis intervals each having a duration T₁ , T₂, ... T _i , for example equal to 6 minutes. Preferably, the analysis intervals are chosen to be consecutive over time.

For each of the analysis intervals T _i , the method according to the invention aims to determine two acoustic indicators making it possible to assess the state traffic flow. To this end, the method according to the invention aims to define, in each interval T _i , the difference between the maximum level Lmax and the minimum level Lmin of the sound signal S appearing in the corresponding analysis interval T _i . Indeed, as is clear from FIG. 2 , the sound signal received by each microphone comprises a succession of maxima and minima corresponding to the successive passages of vehicles in front of the microphone, at the right of which the acoustic power is maximum. The instantaneous sound power is a function of the distance between the vehicles, the position of the vehicle closest to the microphone and the distance between the traffic axis and the microphone.

The difference between the maximum Lmax and minimum Lmin levels makes it possible to determine a first acoustic indicator P₁ corresponding to the dynamics of the acoustic signal. In the examples illustrated in FIG. 2 , the dynamic indicator P₁ is equal to 3.7 dB in the interval T₁ , to 8.6 dB in the interval T₂ and to 14.2 dB in the interval T _i . The difference between the maximum and minimum values depends only on the distance between the vehicles, as long as the microphones are installed in a fixed manner. It should be considered that the distance between vehicles is an increasing function with the speed of traffic. Indeed, for safety reasons, the speed of traffic determines the distance between vehicles. The dynamics P₁ of the signal therefore constitutes an acoustic indicator making it possible to distinguish either a concentration of vehicles around the microphone, if the value of the dynamics is low, or a relatively fluid circulation, if its value is high. Thus, the indicators P₁ detected during the intervals T₁, T₂, T _i make it possible to observe respectively a strong slowdown, dense traffic or fluid traffic.

The method according to the invention aims to determine a second acoustic indicator P₂ which makes it possible to know whether there is absence or presence of traffic on the monitored lane. Indeed, knowledge of the indicator P₁ alone does not make it possible to know whether there is traffic. An absence of traffic could lead to a low value of the dynamics, which would be interpreted as a strong slowdown, when there is no traffic.

The presence indicator P₂ corresponds to at least one given level L _P of the sound signal which is reached during a predetermined fraction of the duration of the interval considered T _i . This acoustic indicator also makes it possible to assess the state of traffic, insofar as the sound energy emitted increases as a function of speed. The values of the indicator P₂ increase with the speed of the flow of vehicles due to the noise of the engines and the noise of tire-road contact.

The P₂ indicator offers the advantage of overcoming differences in noise levels emitted by vehicles depending on their nature (heavy goods vehicles or passenger cars), while making it possible to assess the presence of vehicles according to a level acoustics reached or exceeded during a given time compared to the analysis time. It should be understood that the indicator P₂ is not expressed as a function of absolute values recorded in a given time space, but is based on the percentage of appearance of one or more noise levels in a time interval predetermined.

The acoustic indicators P₁ and P₂ previously defined are compared with setpoints making it possible to assess the state of traffic flow. Indeed, the comparison with a set level of the acoustic indicator P₁ makes it possible to detect traffic disturbances, while that carried out with the indicator P₂ makes it possible to confirm the presence of a traffic flow. Knowledge of the two acoustic indicators P₁, P₂ is necessary and sufficient to assess the state of saturation or fluidity of a stream of vehicles.

According to a preferred embodiment of the invention, provision is made to determine the acoustic indicators P₁ and P₂ by proceeding, for each analysis interval, to a sampling of the sound signal, then to a classification of the samples obtained according to their noise level. The measured samples are then accumulated to determine their statistical distribution. To this end, provision is made to determine the so-called Henry's straight line of cumulative distribution of the samples, as shown for example in FIG. 3A . Conventionally, this figure is called a histogram of cumulative values, and includes a GAUSS scale along the ordinate axis graduated as a percentage of the duration T _i of the analysis, while the levels in decibels of the acoustic signal appear on the axis. of abscissa. Acoustic indices L₁, L₂ , ..., Li , ..., L₉₉ are thus obtained whose noise level is reached or exceeded during a corresponding fraction of the duration of the analysis. For example, the index L₁ has a noise level which is reached or exceeded during 1% of the duration of analysis, while the index L₉₉ has a noise level which is reached or exceeded during 99% of the duration d 'analysis. The L₁ index therefore corresponds to the index evaluating nuisance noise, while the L₉₉ index evaluates the level of ambient noise on the site.

Figs. 3A to 3C illustrate three families of curves recorded at different times of the day and making it possible to highlight different traffic conditions. Each curve corresponds to an analysis time T _i , for example equal to 6 minutes. It should be noted that the slope of each curve corresponds to the dynamics of the acoustic signal. Also, advantageously, the dynamic indicator P₁ is determined by making the difference between two acoustic indices L _i , namely L₉₉ and L₁ and, preferably, between the indices L₈₄ and L₁₆ . Indeed, it was found that there is a good correlation with the speed for the closest indices surrounding the index L₅₀ . The choice of space L₈₄-L₁₆ makes it possible to increase the precision of the measurements. The average acoustic index P₁ is equal to 16 dB, 4dB and 7 dB respectively for the families of curves illustrated in Figs. 3A to 3C . This method of determining the acoustic indicator P₁ , using a statistical distribution, constitutes an improvement of the method consisting in determining the dynamics of the signal by the difference between the values Lmax and Lmin , due to the very different acoustic powers. issued by the various categories of vehicles.

Preferably, the presence indicator P₂ is obtained by taking an acoustic index substantially close to the index L₅₀ . For this index, it appears, in fact, a good correlation with vehicle speed. In the example illustrated, the average presence index P₂ is equal to 73 dB, 86 dB and 85 dB respectively for the families of curves illustrated in FIGS. 3A to 3C .

It should be noted that the curves illustrated in FIG. 3A correspond to traffic recorded between 2 and 3 am. The statistical distributions are random and the dynamic P₁ (16 dB) is high, which indicates low traffic.

The curves which are shown in Fig. 3B corresponds to a traffic flow recorded between 7 a.m. and 7:30 a.m. The low value of the dynamic index P₁ (4 dB) indicates a saturation of the taxiway.

Fig. 3C , which shows curves corresponding to a traffic recorded between 10 and 11 am, shows that the acoustic indicator P₁ has a value (7 dB) slightly higher than that of the index determined in FIG. 3B . The corresponding circulation is considered dense but more fluid than in the example illustrated in FIG. 3B .

It follows that the saturation of the traffic lane is characterized by a decrease in the value of the dynamic index P₁ and by a change to a minimum value for the presence index P₂ .

In order to distinguish more or less critical states of road traffic, it is planned to analyze the changes in the presence indicator P₂ as a function of the dynamic indicator P₁ . Fig. 4 illustrates, by way of example, the evolution of the presence indicator L₅₀ as a function of the dynamic indicator L₁₆ - L₈₄ . The existing relationship between the presence indicator P₂ and the dynamic P₁ makes it possible to detect critical traffic periods from a single noise measurement. The following description gives an example of a method for determining the set points corresponding to significant traffic conditions.

The analysis of critical traffic conditions consists in analyzing the data recorded during periods of heavy traffic corresponding to a low value of the dynamic indicator P₁ . In the example illustrated, it is chosen to analyze only the data when the dynamic indicator P₁ has a value less than 7 dB (point M₁ ).

The noise samples recorded, during the time when the dynamic indicator P₁ is less than 7 dB, are accumulated in order to determine their distribution. An acoustic index is then chosen, the noise level of which is reached or exceeded during a corresponding fraction of the analysis time.

In a preferred, but not exclusive, manner, the noise level chosen for the determination of the set point corresponds to the noise level which is reached or exceeded during 50% of the duration of analysis. In the example shown, the setpoint value, for such a noise level, is equal to 82.5 dB (point M₂ ).

• une période de blocage lorsque P₁ < 7 dB et P₂ < 82,5 dB,
• une période de ralentissement lorsque P₁ < 7 dB et P₂ > 82,5 dB,
• et une période de trafic fluide pour P₁ > 7 dB.

Thus, it can be determined:

• a blocking period when P₁ <7 dB and P₂ <82.5 dB,
• a period of deceleration when P₁ <7 dB and P₂ > 82.5 dB,
• and a smooth traffic period for P₁ > 7 dB.

Thus, the dynamic indicator P₁ unambiguously allows the detection of a concentration of vehicles near the monitoring point. In this area, strong traffic disruptions, corresponding to a dynamic indicator P₁ <7 dB in the example illustrated, the possibility of detecting different states of the traffic speed is given by the presence indicator P₂ . In the field of the chosen values of the dynamic indicator P₁ , a decrease in the values of the presence indicator P₂ indicates the increase in traffic blocking, while an increase in the level of the indicator P₂ means an increase in the traffic flow.

It should be noted that a refined study of the dynamic indicator P₁ , in its domain of strong disturbances, makes it possible, in conjunction with the presence indicator P₂ , to detect intermediate stages of saturation. Thus, in the example illustrated, it is possible to decompose the period of deceleration into a period of strong deceleration for P₁ between 2 and 5 dB and P₂ > 82.5 dB and into a period of prelude to decelerations for P₁ between 5 and 7 dB and P₂ > 82.5 dB.

Such a principle appears clearly in FIG. 4 on which are placed the set points M₁, M₂ of the indicators P₁ and P₂ , namely 7 and 82.5 dB. The set points M₁, M₂ define the coordinates of a point of rotation of a circle C centered on this point and divided, for example, into sectors numbered from 1 to 8 making it possible to detect different states of saturation of the channel.

• une période de blocage pour P₁ < 7 dB et pour P₂ < 82,5 dB (secteurs 1 et 2),
• une période de fort ralentissement pour P₁ compris entre 2 et 5 dB et P₂ > 82,5 dB (secteur 3),
• une période de prélude aux ralentissements pour P₁ compris entre 5 et 7 dB et P₂ >82,5 dB (secteur 4),
• et une période de trafic fluide pour P₁ > 7 dB (secteurs 5 à 8).

It is thus determined:

• a blocking period for P₁ <7 dB and for P₂ <82.5 dB (sectors 1 and 2),
• a period of strong deceleration for P₁ between 2 and 5 dB and P₂ > 82.5 dB (sector 3),
• a period of prelude to deceleration for P₁ between 5 and 7 dB and P₂ > 82.5 dB (sector 4),
• and a smooth traffic period for P₁ > 7 dB (sectors 5 to 8).

The statistical study of the noise samples thus allows an evaluation of the state of the traffic flow based solely on the acoustic signal emitted by the flow of vehicles.

The invention is not limited to the examples described and shown, since various modifications can be made thereto without departing from its scope.

Claims (8)

1 - Method for acoustically evaluating the flow of traffic from road vehicles, of the type consisting in placing at least one microphone ( 3 ) at the edge of a traffic lane ( 1, 2 ) and in collecting the sound signal ( S ) from the microphone, in order to determine the characteristics of the flow of traffic, characterized in that it consists: - to define analysis intervals of predetermined duration ( T _i ) during which the sound signal is collected, to determine for each of the analysis intervals, on the one hand, the difference between the maximum ( Lmax ) and minimum ( Lmin ) levels of the sound signal in order to determine a first acoustic indicator ( P₁ ) corresponding to the dynamics of the signal and , on the other hand, a level ( L _P ) of the sound signal which is reached during a predetermined fraction of the duration ( T _i ) of the interval considered, with a view to defining a second acoustic indicator ( P₂ ) corresponding to the presence vehicles on the track, - and to compare the acoustic indicators ( P₁ , P₂ ) previously defined with reference values, in order to assess the state of traffic flow. 2 - Method according to claim 1, characterized in that it consists in determining the acoustic indicators ( P₁ , P₂ ) by proceeding, for each analysis interval ( T _i ): - a sampling of the sound signal, - a classification of the samples according to their noise level, - and the accumulation of samples to determine their statistical distribution ( L₁, L₂ ..., L₉₉ ) whose noise level is reached or exceeded during a determined fraction of the analysis interval. 3 - Method according to claim 2, characterized in that it consists in determining the dynamic indicator ( P₁ ) by making the difference between two acoustic indices, namely ( L₉₉-L₁ ) and, preferably, ( L₁₆-L₈₄ ) . 4 - Method according to claim 2, characterized in that it consists of determine the presence indicator ( P₂ ) by taking an acoustic index substantially close to the index ( L₅₀ ). 5 - Method according to preceding claims, characterized in that it consists in analyzing the changes in the presence indicator ( P₂ ) as a function of the dynamic indicator ( P₁ ), in order to determine different phases of traffic flow. 6 - Method according to claim 5, characterized in that it consists in analyzing the changes in the indicators ( P₁ ) and ( P₂ ): - by choosing a low value for the dynamic indicator ( P₁ ) corresponding to periods of heavy traffic, - by accumulating the noise samples recorded during the time when the dynamic indicator ( P₁ ) is less than the determined value, - and by choosing a noise level for the indicator ( P₂ ) which is reached or exceeded during a predetermined fraction of the analysis time. 7 - Method according to claim 6, characterized in that it consists in choosing, for the dynamic indicator ( P₁ ), a value close to 7 dB and, for the presence indicator ( P₂ ), the noise level reached or exceeded for 50% of the analysis time. 8 - Method according to claim 1, characterized in that it consists in collecting the sound signal coming from at least one microphone ( 3 ) placed on the edge of the so-called slowest traffic queue of the lane.

Images