• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Method for determining the state of a road and the state of a tire comprising the steps in which: a measurement of a sound signal produced by the tire rolling on the road during a given time frame is recorded; the spectral power density of the sound signal is determined over a given frequency interval, the frequency interval is segmented into a plurality of frequency bands, and a frequency representative of the average sound power is associated with each frequency band; measured in said frequency band, the representative data from a measurement forming the variables of a vector associated with said measurement, - a state of the road and the tire is determined using a discriminant analysis of the data based on on a learning basis The representative data forming the variables of the vector associated with a measurement are obtained by making the ratio between the average sound power measured in a frequency band and the total sound power measured over the entire frequency range.
    公开号:FR3015036A1
    申请号:FR1362879
    申请日:2013-12-18
    公开日:2015-06-19
    发明作者:Antoine Paturle；Jerome Antoni
    申请人:Michelin Recherche et Technique SA Switzerland ；Compagnie Generale des Etablissements Michelin SCA；Michelin Recherche et Technique SA France；
    IPC主号:
    专利说明:

    [0001] [1] The invention relates to a method of detecting the state of the road and the tire fitted to a vehicle traveling on this road, from the noise generated by the tire when it comes into contact with the ground. [2] It is useful to know at all times the state of the road or the tire, to interact with the driver or the driver assistance systems, so as to inform them in real time of the evolution of the rolling conditions, and more generally, the possible modification of the tire adhesion conditions and the handling of the vehicle. [3] These methods are therefore intended to highlight the evolution of weather conditions such as the transition between driving conditions on dry, wet, wet, snow-covered ground. They are based on the observation that the frequency and the sound intensity generated by the tire vary when the state of the road changes, and analyze sound recordings made by microphones placed near the tire and the road. [004] Starting from the recording of a frequency spectrum of sound powers, the meteorological condition of the road is determined using judiciously chosen ratios or by comparing this spectrum with pre-recorded data. [005] However, it is observed that the accuracy and reliability of all these methods rely largely on the introduction of additional parameters such as knowledge of the speed of travel, the temperature or the vehicle load, the degree the most complex methods of recording the visual state of the road ahead of the vehicle. It is therefore necessary to combine information from several sensors, which is not without impact on the cost of implementing the proposed devices. [006] The invention aims to provide a robust solution to the problem of determining the meteorological state of the soil. In a complementary manner, the method proposed by the invention also makes it possible, more unexpectedly, to determine, on the sole basis of a sound recording, other parameters such as the type of coating, the degree of wear of the tire or still the type of sculpture used. [007] The sound recordings are made using a microphone judiciously placed on the vehicle. The spectral density of the sound power is distributed over a given frequency interval. This spectrum varies according to a set of conditions such as the weather conditions, the state of the road, the degree of wear of the tire, the type of tire tread, and to a lesser extent, the pressure. inflating, charging etc. One of the major modalities likely to modify this spectrum, all conditions remaining equal, is the speed of the vehicle at the time when the measurement is made. [8] It has been demonstrated that this variation can be practically neutralized if, instead of considering the power spectrum resulting from the measurement, this spectrum is recalculated by "normalizing" the measured data with the aid of the sound power. total recorded during the measurement time over the entire frequency interval. This is to erase the speed effect and make the measurement substantially invariant to this modality. [9] The object of the invention is to take advantage of this observation. The method for determining the state of a road and the state of a tire mounted on a vehicle traveling on said road according to the invention therefore comprises the steps in which: - a measurement is recorded of an audible signal produced by the tire rolling on a road surface during a given time frame, - a spectral power density of the sound signal is determined over a given frequency interval, - the frequency interval is divided into a frequency interval. a plurality of frequency bands of predetermined widths and associating with each frequency band a piece of data representing a mean sound power measured in said frequency band, the representative data from a measurement forming variables of an associated vector at said measurement, a state of the road and of the tire corresponding to the vector associated with the measurement carried out is determined using a discriminant analysis. a data base based on a learning base formed of a set of vectors associated with measurements previously recorded and made, according to the same steps as above, under known rolling conditions according to modalities each representing a given state of the road and tire. The method is characterized in that the representative data forming the variables of a vector associated with a measurement are obtained by making the ratio between the average sound power measured in a frequency band and the total measured sound power. over the entire frequency range. In this way, it is no longer necessary to take account of the speed, to obtain, from a single acoustic measurement and without it being necessary to introduce additional parameters to interpret the measurement, reliable information on the condition of the road and, as will also be seen, on the condition of the tire. The method according to the invention may also comprise in isolation or in combination the following characteristics: the total measured sound power is equal to the sum of the average sound powers of all the frequency bands of the frequency interval under consideration. - Frequency bands are determined by splitting the frequency range by one-third octave. - The time frame of a measurement is less than or equal to 0.5 seconds, and preferably less than or equal to 0.25 seconds.
    [0002] The frequency range is 0Hz to 20KHz. The frequency range is 200Hz to 20KHz. A weather mode class, formed by different weather conditions of the road, includes a dry state, a wet state and a wet state. - A class of "pavement state" modalities, formed by different states of road pavement, includes a closed state, a medium state, and an open state. - A class of "wear" modalities, formed by different states of wear of the tire, includes a new state, a half-worn state and a worn state. - A class of modalities "sculpture", formed by different types of tire sculptures, includes a sculpture type summer and a sculpture type winter. The discriminant analysis of the data foresees the stages in which: from the learning base, a reduced discriminant space is determined in which zones formed by each modality or combination of modalities are identified, where the associated vector is transformed; to a measurement in said reduced discriminant space and, depending on the location of said vector, the measurement is associated with a probability according to each of the modalities or combinations of modalities, where the most probable modality is determined according to each of the classes of -4 - modalities. A measure is associated with the measurement according to the "pavement condition", "wear" or "sculpture" modality, after having previously determined that the measurement was made on a dry road. A probability is associated with the measure according to each combination of modalities containing this modality, and this measure is assigned the modality of the class with the highest probability. - A diagnosis of the state of the tire is made according to the "wear" mode or the "sculpture" mode by combining the results of measurements made at different time intervals. - The sound signal generated by the tire is measured by means of a microphone placed in the front part of a wheel arch situated at the rear of the vehicle. The invention will be better understood on reading the accompanying figures, which are provided by way of examples and are not limiting in nature, in which: - Figure 1 shows a vehicle equipped with a measuring device and analysis of the sound power of a tire. Figure 2 shows a non-"normalized" sound power spectrum for measurements made at different speeds.
    [0003] Figure 3 shows the same power spectrum after "normalization". Figure 4 shows the average power spectra normalized for different weather conditions of the road. FIG. 5 shows a distribution of measurements in a reduced two-dimensional discriminant space according to the meteorological condition of the road. FIG. 6 represents a functional diagram of the steps for implementing the method according to the invention. The vehicle C traveling on a floor G, shown schematically in Figure 1, comprises front and rear wheel arches in which are housed the wheels equipped with tires T. When the vehicle C moves, the tire T generates a noise whose amplitude and frequency depend on multiple factors. This sound pressure is in fact the superposition of noises of various origins such as the noises generated by the contacting of the breads of the sculpture with the ground G, by the air movements between the carving elements, by the particles of water raised by the tire, or -5- by the air flows related to the speed of the vehicle. Listening to these noises is also superimposed on noise related to the vehicle environment, such as engine noise. All these noises are also dependent on the speed of the vehicle. A listening means, such as a microphone 1 is installed in a wheel arch to listen to rolling noise as close as possible to the place where they are generated. Of course, the usual precautions are taken to protect the microphone from external aggressions such as splashing water, mud or gravel. For this purpose, the microphone is preferably installed at the front of the wheel arch. Ideally, it can be considered that the installation of a microphone in each of the wheel arches is the best way to capture all the rolling noise generated by the tires. However, to determine the condition of the road (meteorological condition and porosity of the coating), only one microphone is sufficient. In the latter case, it is better to isolate it from aerodynamic noise and the engine. A good choice seems to be to install the microphone at the front of one of the wheel arches of the rear axle. It is also possible to place the microphone in the rear bumper or in the front bumper. The vehicle also comprises a computer 2, connected to the microphone, and configured to perform the operations for formatting and analyzing, as will be described in detail later, the raw information from the microphone, and to estimate the state of the ground or the tire according to a measurement of the sound power detected by the microphone. Means of storing the information are associated with the computer. These means make it possible to store in memory the data relating to a learning plan concerning measurements made under known rolling conditions and according to modalities describing different states of the road or of the tire. Finally, the information concerning the state of the road or the tire can be transmitted to display means or to driving assistance systems 3, or on a remote server. Here, by modality, is meant a set of conditions related to the state of the ground or of the tire capable of appreciably varying the measurement of the sound pressure. As has been mentioned, the number of parameters having a potential impact on the noise of the tire can be significant. However, it appears that certain parameters have a low or second order influence on the nature of the noise generated by the tire. This may be the case for example of the internal pressure of the tire or the load of the tire. Surprisingly, it appears that the weather condition of the road seems to be a parameter of the first order. Its impact on the noise of the tire is very important and above all independent of all the other parameters such as the state of the road surface, the state of wear of the tire or the type of tire tread. These other parameters are also likely to a lesser extent to vary the rolling noise as far as it is known to discern their own acoustic signatures. The meteorological condition of the road forms a first class of modalities, called "weather" class, in which a dry road is differentiated from a wet road, characterized by a height of water flush with the natural roughness of the road surface. the road, or a wet road for which the height of water exceeds the level of natural roughness of the road surface. Real-time knowledge of the changing weather conditions of the road is of paramount importance for adapting, for example, driver assistance systems. Also distinguished in a second class of modalities, said class "state of the coating", different states of the pavement of the road. A coating is called a closed coating when it takes on a smooth appearance and without roughness, such as a bitumen that has been squeezed after having undergone severe heat. A pavement will be considered open, when the roughness is important like that of a worn pavement or that of a country road repaired quickly by means of a superficial rendering carried out by projecting pebbles on bitumen. A medium coating describes all the coatings in an intermediate state between the two preceding states and more particularly qualifies the new coatings. It is assumed here that the porosity of the coating influences the permeability or the sound reflection of the noise generated by the tire. Indeed, the phenomenon of pumping air trapped between the ground and the sculpture of the tire, as well as the phenomenon of amplification of the noise by the air wedge formed by the curvature of the tire and the ground, are all the more pronounced as the pavement of the road is closed. Real-time knowledge of the condition of a road can be useful if, for example, this information is returned by a large number of vehicles or a dedicated fleet of vehicles, to a centralized tracking and monitoring system. maintenance of the road network. Regarding the state of the tire, it is possible to recognize the state of wear distinguishing, in a third class of modalities, said class "wear", the new state, the worn state, and a state intermediate considered here as the state of the tire mid-wear. Information about the evolution of the wear characteristic over time is also important, especially if it is linked to the weather information of the road. Indeed, it is known that a vehicle equipped with worn tires that rolls on a wet coating is more likely to lose its adhesion ("aquaplaning") than if it had new tires. Finally, the method according to the invention is capable of discerning a fourth class of modalities, said class "sculpture" and relating to the type of tire sculpture, distinguishing whether it is a sculpture type summer of a winter type sculpture. These two types of tires are distinguished essentially by treads having different carvings, strongly notched and laminated in the case of winter sculptures, more directional and less notched in the case of summer sculptures, as well as by the nature of the materials forming the tread, softer in the case of winter tires, and harder in the case of summer tires. These characteristics are not without influence on the behavior and handling of the vehicle, and can constitute useful information for adapting the driving system, particularly in regions where tire mounts alternate between summer and winter periods. winter periods. The method according to the invention makes it possible to highlight each of the modalities of these different classes in an isolated manner as is the case more particularly for the weather characteristic or in a combined manner for the other characteristics. [0032] Figure 2 is a spectral representation of the sound power recorded by the microphone during a time frame. By time frame is meant the time interval, usually short, during which a recording is made on the basis of which are established the data used as a basis for a measurement. This time frame is less than or equal to 0.5 seconds or ideally less than or equal to 0.25 seconds. This spectral representation represents the received sound power (in dB) as a function of the frequency, over a given frequency interval, typically here, the audible frequency interval, between 0 Hz and 20 KHz. More specifically, the spectral representation of FIG. 2 is obtained by breaking down the frequency interval into frequency bands of predetermined widths, and by assigning to each frequency band a characteristic value equal to the average power measured in this band of frequencies. frequency. A division of the frequency range into one-third octave bands seems to be the most appropriate. Thus, each point of each of the curves of FIG. 2 represents an average sound power for a given frequency band and measured during a time frame under rolling conditions in which, all other things being equal, only the speed is varied. (typically from 30kmh to 110kmh). It is then observed that the curves representing the spectral powers are shifted relative to each other, and that the total sound power dissipated increases as a function of speed. However, the general shape of the curves remains similar. This observation is reproduced when one or more modalities of the other classes are changed and the curves obtained are compared by varying only the speed parameter. It then determines the total sound power over the entire frequency interval, which can be compared to the area between the curve and the abscissa axis and, for each frequency band, the power is divided average observed in this frequency band during a given time frame, by the total power recorded during this time frame over the entire frequency interval. It amounts in a way to "normalize" the measurement. It is then observed in FIG. 3 that the curves obtained previously overlap substantially and have very similar profiles, in particular in the highest and most representative frequency bands of the sound phenomena described above. This "normalization" makes it possible to neutralize the effect related to the speed without, however, significantly modifying the analysis capacity that it is possible to produce from a sound recording during a frame. determined time. This advantage can be decisive when one does not wish to connect the computer 2 to vehicle speed evaluation means, and that one wishes to obtain information on the state of the road or of the vehicle autonomously. For the sake of simplification and speed in the execution of calculations, it can be considered that the total power is equal to the sum of the average powers in each of the frequency bands of the frequency interval considered. Each of the points of the curve of Figure 3 is a representative value of the average sound power in a given frequency band. All these points can then constitute a vector in a vector space comprising as many dimensions as frequency bands. In the example serving as a support for the present description, a vector comprising 21 dimensions is obtained by considering a frequency interval segmented by one-third octave and included in the frequency range located between 200 Hz and 20 kHz. The choice of the frequency interval can also be adapted according to whether it is desired to completely eliminate the noise generated by the motor whose maximum amplitude is between 50Hz and 60Hz, in which case a frequency interval will be considered. for example between 200Hz and 20KHz, or if it is desired to keep the relevant portion of information contained in the frequency range below 200Hz, in which case the spectrum will be taken into account over the entire range between 0Hz and 20KHz. The recording of the sound power during a time frame can be done from a sampling at high frequency (around 40 kHz) of the sound signal. The implementation of the invention comprises a preliminary learning phase, during which a large number of measurements are made by varying in a known manner the modalities described above, and describing the meteorological state, the condition of the road, the state of wear or the type of tire tread. To each of these measurements, a vector obtained is assigned under the conditions described above. This provides a vehicle-specific learning base. Methods of analysis and statistical processing of the data are known per se and are not the subject of this invention. The linear discriminant analysis method that has been used has yielded reliable and robust results. A first step of this method is to determine the main factorial axes that reduce the number of dimensions to the number just needed to describe the vectors assigned to each of the measurements along orthogonal axes. The passage of the vector space whose number of dimensions is equal to the number of frequency bands, typically equal to 21 dimensions, in the reduced discriminant space is done using a linear transformation. A second step then consists, using the discriminant analysis proper to look for, in this discriminative space reduces the areas in which are located the measurements obtained during the learning phase according to a given single modality or according to a combination of terms. By combination of terms is meant here a representative state of a given measurement performed according to a modality chosen in each of the classes. By way of example, a measurement made in the "wet" state, on a "closed" road with a "summer" and "worn" tire represents the combination of "wet-closed-summer-worn" mode. The number of modalities combined is therefore equal to the product of the number of modalities of each class. Then, in this reduced discriminant space, the center of gravity of the zone in which the points representing a modality or a combination of modalities are located, as well as a confidence interval representative of the dispersion of the points of a same area relative to this center of gravity. In the example used as a support for the present description, the reduction in the number of dimensions between the discriminating starting space and the reduced discriminating space makes it possible to go from 21 dimensions to approximately 15 dimensions. This small reduction makes it possible to observe that it is the taking into account of the total form of the spectrum which is characteristic of the expression of the different modalities. And to suggest that the taking into account of the powers of a reduced number of frequency bands does not make it possible to bring out a particular modality concerning the road or the tire, with the exception of the modalities related to the meteorological state of the road. FIG. 4 represents the spectral distribution of the "normalized" sound power, in frequency bands of 1/3 octave for three weather conditions of the road, all the modalities of the other classes being equal elsewhere. FIG. 5 shows in a two-dimensional space the distribution of the measurements according to one of the "dry", "wet" and "wet" modes of the "weather" class of the road. A first observation shows that the measurements made on dry soil are not overlapping with the measurements made on wet or wet soil. A second observation makes it possible to conclude that it is possible to determine the weather condition of the road independently of the modalities of the other classes with good robustness. The ellipses surrounding each of the point clouds are placed at one, two and three standard deviations, and make it possible to evaluate the dispersion of the measurements around the center of gravity, and especially to assess the recovery rate of an area. compared to another which is representative of the risk of misallocating to another modality of a measure carried out according to a different modality. From these data, it is also possible to determine the probability of membership of a new measurement to one of the three modalities of the "weather" class of the road by evaluating the distance from this point to the center of gravity. each of these terms. Table 1 gives the probabilities of classification of the weather condition of the road according to one of the three modes "dry", "wet", "wet". Pj / i J = Dry J = Wet J = Wet i = Dry 1 0 0 i = Wet 0 0.91 0.09 i = Wet 0 0.03 0.97 Table 1 [0058] As can be seen , probabilities high enough to conclude that the vehicle is traveling on a "dry", "wet" or "wet" road. And only the recognition of the "wet" modality can be wrongly attributed to the "wet" modality in 9% of cases. Similarly, it would be possible to determine the condition of the road surface, with less robustness than the weather condition of the road, without it being necessary to know beforehand the condition of the tire. However, it will be preferred to perform this analysis when the road is dry. This observation suggests that certain acoustic phenomena related to porosity and soil reflection are independent of the nature of the tire. On the other hand, by doing similar analyzes, it is found that the zones sheltering the vectors relating to the modalities related to the state of the tire (wear or sculpture) are relatively dispersed and interpenetrate strongly (strong dispersion around the center of the tire). gravity, and low distance from the centers of gravity) which does not allow to conclude to a precise modality without a high risk of determination wrongly, especially when the state of the road is "wet" or "wet". Also, to ensure good robustness, the method provides for a first analysis of the weather condition of the road and, when it is observed that the vehicle is traveling on a "dry" ground, to proceed to a second an analysis which makes it possible to discern the conditions relating to road pavement, to the state of wear of the tire and to the type of tire tread pattern. For more robustness, it then seems preferable to make the discriminant analysis based on the combined modalities of the three classes. In the reduced discriminative space, the clouds of points representative of the vectors and measurements made according to a given combination of modalities selected in each of the three classes of modality "state of the coating", "wear" and "sculpture" are located. The modalities related to the sculpture of the tire are denoted "A" for a tire "winter" and "P" for a tire "summer", the conditions of the state of wear are noted "N" for a tire "new", "M" for a tire "half-worn" and "U" for a "worn" tire, and finally, the conditions of the state of the coating are noted "f" for the modality "closed", " m "for the modality" medium "and" o "for the modality" open ". The combined modalities are then noted respectively: ANf, ANm, ANo, AMf, AMm, AMo, AUf, AUm, AUo, PNf, PNm, PNo, PMf, PMm, PMo, PUf, PUm, PUo. Table 2 gives the probabilities obtained from the results of measurements contained in the learning base, for each of the 18 combinations of modalities. The dispersion of the measurements, observed for the modalities alone, is then much weaker for the combined modalities and makes it possible to proceed to a classification in a more efficient manner. Table 2 [0065] The overall probability of detection of one of the modality combinations is of the order of 0.96. The next step consists in recognizing, for a new measurement given, the modality of each of the classes "state of the coating", "wear" and "sculpture" in which the measurement has been carried out. Table 3 makes it possible to determine the probabilities of detecting the modality of one of the three classes according to the combinations of modalities. This table 3 indicates that, if a measurement is assigned to the class "AUf" (Winter, worn closed coating), one can have a good confidence in the determination of the sculpture (1), the state of wear (U) and the state of the coating (1). Relatively poorer confidence is obtained in the "AUo" class (Winter, Worn, Open) for which the prediction on the type of pneumatic sculpture is less good (0.91). cami AUo = pz pNo king = m ... m ... m ... 0 m. 0 mm .. mm 001 mmmm mm 003 - MMMMMMM MMMM M MII M M BM 0.97 MM MM. MM 001 - .MM 0.91 MMMM MMM ° --.--- M +, 97 MM. MMMMMEM.1.M ° IMMIMMIII 0.01 - MMMMI EIMMIM MMMMMIMM IIMMM 1.11M 0 06 -.- .. MM 0.91 MM 0.1 IMM. 1.1.1M AMo PM m PMo PUo 001 0.96 001 0.1 0 02 0.96 0 03 0.03 0.9 Ezzi AN ° cm - 0.99 μm MM 0.96 0 01 In MIM 0.03 MMMM MM .. " MMMM I. 0.01 IMM MIIMM 0.01 MM ---- 0.03 0.95 0.9 0.02 YEAR ANM ANO AMF AMM AMO AM AUMA AUM PNF PM PN PMF PM PM PM AMM 0.01 0.95 0.01 0.02 PUE PUm PUo -14- Probability of class detection Class Sculpture Wear Coating found ANf 1 0.95 0.98 ANm 1 0.99 1 ANo 0.96 1 1 AMf 0.98 0.95 1 AMm 0.97 0.96 1 AMo 0,99 0,96 0,99 AUf 1 1 1 AUm 0,98 0,98 1 AUo 0,91 1 1 PNf 0,98 0,95 1 PNm 0.99 0,98 1 PNo 1 1 1 PMf 1 0.97 1 PMm 0.97 0.98 1 PMo 1 1 1 PUf 1 0.97 1 PUm 0.94 0.97 1 PU 0.87 0.97 1 Table 3 [0069] From the base d In learning, the areas in which the combined modalities are located in the reduced discriminant space are located, as well as their center of gravity and their dispersion, typically in the case of the present description, in the reduced discriminative space, the 18 zones of the 18 combined modalities considered [0070] Then, from the location of the vector associated with each new measurement and transformed in the reduced vector space, for each modality of a class, a probability is determined for each of the modality combinations containing this modality. and we attribute to this measure the modality of the class with the highest probability. Thus, if the weather class found is "dry", the 21 variables of the vector resulting from the measurement allow, based on the discriminant analysis based on the learning base to determine a probability of belonging to one of the combined methods, according to the class "state of the coating", the class "wear" or the class "sculpture", is typically, in the case serving as a support for the present description, the -15- probability of belonging to a of the combined modalities ANf, ANm, ANo, AMf, AMm, AMo, AUf, AUm, AUo, PNf, PNm, PNo, PMf, PMm, PMo, PUf, PUm, PUo. This probability is calculated for example by evaluating a distance from the center of gravity of the combined class considered. The probability of membership of the measurement to one of the modalities of a particular class, other than the weather class, is then carried out using a second probability calculation called model "sculpture + wear + coating". on dry soil "as follows. The probability of the state condition of the "closed" coating is deduced from the relation p (coating = "closed") = p ("closed") = p (ANf) + p (AMf) + p (AUf) + p (PNf) + p (PMf) + p (PUf) [0074] Similarly, we deduce: p (coating = "medium") = p ("medium") = p (ANm) + p (AMm) + p (AUm) + p (PNm) + p (PMm) + p (PUm), and p (coating = "open") = p ("open") = p (ANo) + p (AMo) + p (AUo) + p (PNo) + p (PMo) + p (PUo). We then look for the one of the three probabilities which is maximum and which gives the modality of the state of the detected coating and the associated probability: p (coating) = max [ ("closed"), p ("medium") , p ("open")]. Similarly, p (sculpture = max [ ("Winter"), p ("Summer")] with: p (sculpture = "Winter") = p ("Winter") = p (ANf) + p (ANm) + p (ANo) + p (AMf) + p (AMm) + p (AMo) + p (AUf) + p (AUm) + p (AUo) and p (sculpture = "Summer") = p ( "Summer") = p (PNf) + p (PNm) + p (PNo) + p (PMf) + p (PMm) + p (PMo) + p (PUf) + p (PUm) + p (PUo). Finally, the wear is given by p (wear = max [ ("nine"), p ("half-worn"), p ("worn")] with: p (wear = "new") = p ("nine") = p (ANf) + p (ANm) + p (ANo) + p (PNf) + p (PNm) + p (PNo), p (wear = "half-worn) = p ( half-worn ") = p (AMf) + p (AMm) + p (AMo) + p (PMf) + p (PMm) + p (PMo) and, p (wear =" worn ") = p (" worn ") = p (AUf) + p (AUm) + p (AUo) + p (PUf) + p (PUm) + p (PUo) [0078] The probability of assignment to a modality resulting from a given measure is then confronted with a threshold determined to decide the validity of the result found and its transmission to a display or driver assistance system, for example, all the detections of which the probability of class is not at least equal to 0.75 are rejected. And if this probability is between 0.95 and 0.75, the result from the measurement must be confirmed by one or more of the following measures. It will be observed here that, unlike the weather condition or road pavement condition that can change abruptly and that require rapid decision-making, the evolution of wear or the type tire tread are much more stable factors over time, typically on time scales corresponding to distances of 100 kilometers or even 1000 kilometers.
    [0004] However, as the detection of these pneumatic parameters depends on the state of the road, we arrive at the paradox that we must be able to detect them almost as quickly as the state of the road. We can greatly reduce the probability of a determination wrongly on these two criteria by combining the observations obtained with several consecutive measurements, before deciding the actual state of wear or type of sculpture of tire mounted on the vehicle. FIG. 6 gives the sequence of the operations implemented in the method according to the invention. After performing the measurement of the frequency spectrum of the sound power over a given time frame, typically for 1/4 of a second, the spectrum is divided into bands of known widths, typically in 1/3 octaves, on a given frequency range, typically from 200Hz to 20KHz. The sampling of the measurement, during the time frame of 0.25 seconds, is made at a frequency of 44100 KHz. With the aid of a linear transformation, the standardized vector resulting from the measurement is passed from the 21-dimensional space in a reduced discriminant space. The meteorological condition of the road is then directly determined and, if the road is in the "dry" mode, the probabilities of association of the measure with each of the combined modalities are determined and the type of sculpture is successively deduced therefrom. of the tire, the tire wear and the state of the road surface apply the model sculpture + wear + coating on dry ground. The cumulative results relating to the state of the tire can then be determined with good robustness by cumulating the results of the measurements made at regular time intervals, so as to transmit a diagnosis of the state of the tire. and condition of the road to a display means or a driver assistance system or to a road condition monitoring unit. The invention therefore makes it possible to determine in a robust manner and on the basis of a single measurement of sound intensity carried out during a short time frame, the weather condition of the road as well as, more unexpectedly, the the condition of the road or tire liner without knowing the speed of the vehicle. Embodiments of the invention serving as a basis for the present description are not limiting, and may be the subject of implementation variants, in particular in the choice of methods of numerical analysis of the data, as far as they make it possible to obtain the technical effects as described and claimed.
    权利要求:
    Claims (15)
    [0001]
    REVENDICATIONS1. A method of determining the condition of a road (G) and the condition of a tire (T) mounted on a vehicle (C) traveling on said road, comprising the steps of: - recording a measuring a sound signal produced by the tire rolling on a road surface during a given time frame, - determining a spectral power density of the sound signal over a given frequency interval, - segmenting the frequency interval into a plurality of frequency bands of predetermined widths, and each frequency band is associated with a piece of data representing a mean sound power measured in said frequency band, the representative data resulting from a measurement forming variables of a vector associated with said measurement, a state of the road and of the tire corresponding to the vector associated with the measurement carried out is determined by means of a discriminant analysis of the data Ndée on a learning base formed of a set of vectors associated with measurements previously recorded and performed, according to the same steps as above, under known rolling conditions in terms each representing a given state of the road and of the tire, characterized in that the representative data forming the variables of the vector associated with a measurement are obtained by making the ratio between the average sound power measured in a frequency band and a total sound power measured over the entire frequency interval.
    [0002]
    2. The method of claim 1, wherein the total measured sound power is equal to the sum of the average sound power of all the frequency bands of the frequency interval considered.
    [0003]
    The method of claim 1 or claim 2, wherein the frequency bands are determined by cutting the frequency range by one-third octave.
    [0004]
    4. Method according to one of claims 1 to 3, wherein the time frame of a measurement is less than or equal to 0.5 seconds, and preferably less than or equal to 0.25 seconds.
    [0005]
    5. Method according to one of claims 1 to 4, wherein the frequency interval is between 0Hz and 20KHz
    [0006]
    6. Method according to one of claims 1 to 4, wherein the frequency interval is between 200Hz and 20KHz.
    [0007]
    7. Method according to one of claims 1 to 6, wherein a class of "weather" modalities, formed by different weather conditions of the road, comprises a dry state, a wet state and a wet state.
    [0008]
    8. Method according to one of claims 1 to 7, wherein a class of conditions "state of the coating", formed by different states of the road surface, comprises a closed state (f), a medium state (m) and an open state (o).
    [0009]
    9. Method according to one of claims1 to 8, wherein a class of modalities "wear", formed by different states of wear of the tire, comprises a new state (N), a half-worn state (M) and a worn state (U).
    [0010]
    10. Method according to one of claims1 to 9, wherein a class of modalities "sculpture", formed by different types of tire sculptures, includes a sculpture type summer (P) and a sculpture type winter (A).
    [0011]
    11. Method according to one of claims 1 to 10, wherein the discriminant analysis of the data provides the steps in which: - from the learning base, a reduced discriminating space in which areas are identified is determined; formed by each modality or combination of modalities, the vector associated with a measurement is transformed into said reduced discriminant space and, depending on the location of said vector, the measurement is associated with a probability according to each of the modalities (dry, wet, wet , closed, open, medium, worn, half-worn, new, summer, winter) or combinations of modalities (ANf, ANm, ANo, AMf, AMm, AMo, AUf, AUm, AUo, PNf, PNm, PNo, PMf , PMm, PMo, PUf, PUm, PUo), the most probable modality is determined according to each of the classes of modalities ("weather", "state of the coating", "wear", "sculpture").
    [0012]
    12. Method according to claim 11, in which the measurement is associated with a modality according to the "state of the coating", "wear" or "sculpture" mode, after having previously determined that the measurement has been carried out on a dry road. 20-
    [0013]
    13. Method according to claim 12, in which: for each of the modalities of a class, the measure is associated with a probability according to each of the combinations of modalities containing this modality, the class is assigned the modality of the class having the highest probability.
    [0014]
    14. The method of claim 13, wherein a diagnosis of the state of the tire according to the modality "wear" or the modality "sculpture" by performing the results of measurements made at different time intervals.
    [0015]
    15. Method according to one of claims 1 to 14, wherein the sound signal generated by the tire (T) is measured by means of a microphone (1) placed in the front part of a wheel arch located at the rear of the vehicle (C).
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3084418B1|2018-10-10|Acoustic detection method for the state of a tire and the state of a street
    FR3036354A1|2016-11-25|METHOD FOR DETERMINING A RUNNING LIMIT SPEED
    EP2189353A2|2010-05-26|Method and device to detect impacts on specific zones on a vehicle, especially a railway vehicle
    WO2017103474A1|2017-06-22|Method for mapping data relating to road conditions
    EP3655770A1|2020-05-27|Method for detecting road and tyre conditions
    EP3458325B1|2021-04-07|Method for proposing a driving speed
    FR3067288B1|2019-06-28|METHOD FOR MONITORING AND / OR MONITORING THE USE OF A TIRE
    US20210255296A1|2021-08-19|Method and device for detecting a noise generated in a receive signal of an ultrasonic sensor
    EP3601001B1|2021-05-05|Method and system for real-time estimation of road conditions and vehicle behavior
    Saeed et al.2020|Comparison of Pattern Recognition Techniques for Classification of the Acoustics of Loose Gravel
    EP3068636A1|2016-09-21|Neural-network-based decision support method in a system for monitoring motor vehicle tyre pressure
    WO2022029382A1|2022-02-10|Method for estimating the state of wear of a tyre
    FR3051422A1|2017-11-24|METHOD FOR DETERMINING A ADHESIVE MARGIN FOR A VEHICLE IN A RUNNING SITUATION
    EP3068637B1|2017-08-30|Method for decision-making assistance by an empirical and matrix-based approach in a system for monitoring the pressure of the tyres of a motor vehicle
    WO2019122719A1|2019-06-27|Method and system for constructing a database of tyre tread acoustic signatures on a driving ground
    WO2021019148A1|2021-02-04|Method for estimating the water level on a roadway when a tyre is running
    CA3120047A1|2020-05-22|Method for determining the firmness of a ground
    EP3880497A1|2021-09-22|Method for determining the slip of a tyre
    WO2015071556A1|2015-05-21|Fuzzy-logic-based decision support method in a system for monitoring motor vehicle tyre pressure
    同族专利:
    公开号 | 公开日
    CN105829883A|2016-08-03|
    US20160349219A1|2016-12-01|
    EP3084418A1|2016-10-26|
    JP2017505430A|2017-02-16|
    FR3015036B1|2016-01-22|
    US10365248B2|2019-07-30|
    EP3084418B1|2018-10-10|
    WO2015092253A1|2015-06-25|
    CN105829883B|2019-10-08|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5586028A|1993-12-07|1996-12-17|Honda Giken Kogyo Kabushiki Kaisha|Road surface condition-detecting system and anti-lock brake system employing same|
    JP2008143508A|2006-11-14|2008-06-26|Yuki Center|Road surface status determining method and its device|
    EP2573594A1|2010-05-19|2013-03-27|Kabushiki Kaisha Bridgestone|Method for estimating condition of road surface|WO2018172464A1|2017-03-24|2018-09-27|Chazal Guillaume|Method and system for real-time estimation of road conditions and vehicle behavior|
    WO2018224790A1|2017-06-08|2018-12-13|Compagnie Generale Des Etablissements Michelin|Method for checking and/or monitoring the use of a tyre|
    WO2020099785A1|2018-11-14|2020-05-22|Compagnie Generale Des Etablissements Michelin|Method for determining the firmness of a ground|
    US11125556B2|2017-12-21|2021-09-21|Bce Inc.|Method and system for monitoring and assessing road conditions|
    FR3113128A1|2020-08-03|2022-02-04|Compagnie Generale Des Etablissements Michelin|PROCEDURE FOR ESTIMATING THE STATE OF WEAR OF A TIRE|JPH06174543A|1992-12-03|1994-06-24|Toyota Motor Corp|Detecting apparatus for road surface condition|
    US5436612A|1994-03-30|1995-07-25|Aduddell; Richard N.|Audible vehicle monitoring apparatus|
    JPH08138187A|1994-11-02|1996-05-31|Honda Motor Co Ltd|Road surface state detecting device|
    JP3473158B2|1995-03-22|2003-12-02|住友電気工業株式会社|Road surface condition detection device|
    JPH08298613A|1995-04-26|1996-11-12|Mitsubishi Electric Corp|Road surface state detecting device|
    JP4629246B2|2001-02-13|2011-02-09|株式会社ブリヂストン|Tire running state detection method, tire running state detection device, road surface state estimation method, and road surface state estimation device|
    JP2002243535A|2001-02-20|2002-08-28|Omron Corp|Road surface condition detecting device|
    JP4046059B2|2002-11-08|2008-02-13|株式会社豊田中央研究所|Road surface condition estimation device|
    DE10259979A1|2002-12-19|2004-07-15|Daimlerchrysler Ag|A method for determining a road condition while driving a Kraffahrzeugs|
    US20050076987A1|2003-10-09|2005-04-14|O'brien George Phillips|Acoustic signal monitoring system for a tire|
    JP4717716B2|2006-05-25|2011-07-06|株式会社ブリヂストン|Road surface state estimating apparatus and method|
    US20080018441A1|2006-07-19|2008-01-24|John Robert Orrell|Tire failure detection|
    US8290662B2|2008-04-25|2012-10-16|Ford Global Technologies, Llc|System and method for tire cornering power estimation and monitoring|
    CN101275900A|2008-05-08|2008-10-01|江汉大学|Method for recognizing road surface types based on vehicle wheel vibration|
    FR2937902B1|2008-11-06|2011-12-09|Michelin Rech Tech|PNEUMATIC WITH SOUND WITNESSES|
    FR2940190B1|2008-12-23|2012-05-18|Michelin Soc Tech|ALERT METHOD FOR WEARING A PNEUMATIC WITH A SILL|
    FR2943276B1|2009-03-19|2013-05-17|Michelin Soc Tech|METHOD FOR MONITORING THE CONDITION OF A TIRE|
    FR2953164B1|2009-12-02|2012-01-06|Michelin Soc Tech|METHOD OF DETECTING THE WEAR OF A TIRE|
    FR2954225B1|2009-12-18|2012-05-11|Michelin Soc Tech|PNEUMATIC HAVING MULTINIVEAL WEAR WITNESSES|
    FR2954224B1|2009-12-18|2013-05-10|Michelin Soc Tech|METHOD OF UNIVERSALLY DETECTING THE WEAR THRESHOLD OF A TIRE|
    JP5218445B2|2010-02-12|2013-06-26|トヨタ自動車株式会社|Tire condition judging device|
    FR2976521B1|2011-06-15|2016-09-09|Soc De Tech Michelin|METHOD OF UNIVERSALLY DETECTING THE WEAR THRESHOLD OF A TIRE|
    FR2976522B1|2011-06-15|2014-05-09|Michelin Soc Tech|PNEUMATIC HAVING MULTINIVEAL WEAR WITNESSES|
    FR2976520B1|2011-06-15|2014-05-09|Michelin Soc Tech|PNEUMATIC COMPRISING MONOBARETTE SOUNDS OF WEAR|
    FR2976852B1|2011-06-23|2016-12-23|Soc De Tech Michelin|PNEUMATIC WITH SOUND CHANNELS|
    JP5758759B2|2011-09-20|2015-08-05|トヨタ自動車株式会社|Vehicle road surface determination device and driving support device|
    FR2981009B1|2011-10-06|2013-12-20|Michelin Soc Tech|IMPROVED METHOD OF DETECTING THE WEAR OF A TIRE|
    FR2999997B1|2012-12-21|2015-02-06|Michelin & Cie|VEHICLE COMPRISING MEANS FOR DETECTING NOISE GENERATED BY A TIRE|FR3036354A1|2015-05-20|2016-11-25|Michelin & Cie|METHOD FOR DETERMINING A RUNNING LIMIT SPEED|
    JP2017020961A|2015-07-14|2017-01-26|住友ゴム工業株式会社|Tire noise display method, and noise performance evaluation method using the same|
    FR3045815A1|2015-12-17|2017-06-23|Michelin & Cie|METHOD FOR CARTOGRAPHIC REPRESENTATION OF DATA CONCERNING THE STATUS OF A ROAD|
    FR3052106B1|2016-06-01|2018-09-28|Ldl Technology|METHOD AND DEVICE FOR MANAGING A VEHICLE LOAD|
    CN106240250A|2016-08-17|2016-12-21|北京中聚高科科技有限公司|A kind of tire wear supervising device based on SO 2 sensor and method|
    US20200158692A1|2017-07-17|2020-05-21|Compagnie Generale Des Etablissements Michelin|Method for detecting road and tire conditions|
    JP2019164107A|2018-03-20|2019-09-26|本田技研工業株式会社|Abnormal sound determination device and determination method|
    CN108777067B|2018-06-07|2021-04-02|郑州云海信息技术有限公司|Road health degree monitoring method and system|
    KR102062271B1|2018-06-28|2020-01-03|만도헬라일렉트로닉스|Apparatus for estimating tire status and control method thereof|
    CN109017166B|2018-08-16|2020-09-08|杭州容大智造科技有限公司|Device for detecting tire pressure by using sound|
    CN108973543B|2018-08-16|2021-05-04|杭州容大智造科技有限公司|Equipment for detecting tire pressure by using electricity consumption|
    WO2020230323A1|2019-05-16|2020-11-19|日本電信電話株式会社|Abnormality detection device, method, system, and program|
    WO2021019148A1|2019-08-01|2021-02-04|Compagnie Generale Des Etablissements Michelin|Method for estimating the water level on a roadway when a tyre is running|
    DE102020116507A1|2020-06-23|2021-12-23|Dr. Ing. H.C. F. Porsche Aktiengesellschaft|Procedure for determining a target variable|
    法律状态:
    2015-12-21| PLFP| Fee payment|Year of fee payment: 3 |
    2016-12-22| PLFP| Fee payment|Year of fee payment: 4 |
    2017-12-21| PLFP| Fee payment|Year of fee payment: 5 |
    2019-09-27| ST| Notification of lapse|Effective date: 20190906 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1362879A|FR3015036B1|2013-12-18|2013-12-18|METHOD OF ACOUSTICALLY DETECTING THE CONDITION OF ROAD AND TIRE|FR1362879A| FR3015036B1|2013-12-18|2013-12-18|METHOD OF ACOUSTICALLY DETECTING THE CONDITION OF ROAD AND TIRE|
    US15/103,987| US10365248B2|2013-12-18|2014-12-15|Method for acoustic detection of the condition of the road and the tire|
    EP14827846.8A| EP3084418B1|2013-12-18|2014-12-15|Acoustic detection method for the state of a tire and the state of a street|
    JP2016541391A| JP2017505430A|2013-12-18|2014-12-15|Method for acoustic detection of road and tire conditions|
    CN201480068249.4A| CN105829883B|2013-12-18|2014-12-15|The method for carrying out sonic detection for the state to road and tire|
    PCT/FR2014/053352| WO2015092253A1|2013-12-18|2014-12-15|Method for acoustic detection of the condition of the road and the tyre|
    [返回顶部]