CN207611485U - An Embedded Road Vehicle Type Recognition System - Google Patents

Abstract

The utility model is related to field of road, more particularly to a kind of flush type road vehicle type identifying system.The utility model provides a kind of flush type road vehicle type identifying system, including pavement structure ontology, vibrating sensing optical fiber is equipped in the pavement structure ontology, the vibrating sensing optical fiber includes one or more vibrating sensing Fibre Optical Sensor sections, it is connected by vibrating sensing optical fiber changeover portion between each vibrating sensing Fibre Optical Sensor section, further include vibration optical fiber analytical equipment, the vibration optical fiber analytical equipment is connected with vibrating sensing optical fiber by fiber pigtail.The utility model passes through DOVS（Distributed optical fiber vibration sensory perceptual system）For the needs of vehicle cab recognition technology, a kind of novel, accurate, the road vehicle type identification method based on vibration perception, device and its system that can identify on a large scale are provided.

Description

An Embedded Road Vehicle Type Recognition System

technical field

The utility model relates to the field of road engineering, in particular to an embedded road vehicle type identification system.

Background technique

The automatic detection and identification of road vehicles is an important part of the intelligent transportation system. In the stage of traffic planning, the identification of vehicle types for surveyed traffic flow can provide a more reliable basis for traffic statistics, and is also conducive to the formulation of more scientific and reasonable traffic planning; in the field of road traffic monitoring and control, toll stations, parking lots, etc. Vehicle type identification has a large number of application requirements; in addition, vehicle type identification can also provide evidence and help for the handling of traffic incidents and vehicle tracking.

In order to build an efficient and convenient vehicle identification system, it is necessary to ensure high identification accuracy, durability and ease of installation. At present, for automatic vehicle recognition technology, there are mainly three types of methods: image-based recognition, sound feature-based recognition and embedded sensor-based recognition. Image recognition uses ordinary cameras or infrared cameras to collect and recognize vehicle outlines, which can accurately determine the type of vehicle, but the effect is not good at night and in low-visibility conditions; sound feature recognition uses microphones and other equipment for data collection and recognition, but it is greatly affected by environmental noise and cannot be applied to vehicle identification under multi-lane and high-flow conditions; the method based on embedded sensors is the current mainstream vehicle identification method, including ground induction coils, piezoelectric sensors, Geomagnetic sensors and acceleration sensors. Among these sensors, the ground induction coil is relatively cheap, but the recognition accuracy is limited. The other types of sensors have defects such as inconvenient networking, limited life, and high price, and their reliability is not good.

Therefore, in order to facilitate real-time traffic management and control, a road vehicle type identification system and method with higher reliability and more convenient networking are needed.

Utility model content

In view of the above-mentioned shortcomings of the prior art, the purpose of this utility model is to provide an embedded road vehicle type identification system and method for solving the problems in the prior art.

In order to achieve the above purpose and other related purposes, the first aspect of the utility model provides an embedded road vehicle type identification system, including a road structure body, a vibration sensing optical fiber is arranged in the road structure body, and the vibration sensor The optical fiber includes one or more vibration sensing optical fiber sensing sections, each vibration sensing optical fiber sensing section is connected by a vibration sensing optical fiber transition section, and also includes a vibration optical fiber analysis device, the vibration optical fiber analysis device is connected with the vibration sensing The optical fibers are connected by optical fiber pigtails.

In some embodiments of the present utility model, the vibration sensing optical fiber is a single-mode optical fiber.

In some embodiments of the present utility model, the vibration sensing optical fiber is a metal armored optical fiber.

In some embodiments of the present utility model, the diameter of the vibration sensing optical fiber is 2-5 mm.

In some embodiments of the present utility model, the long-term allowable tensile force of the vibration sensing optical fiber is ≥600N, and the short-term allowable tensile force is ≥1500N.

In some embodiments of the present invention, the attenuation of the vibration sensing optical fiber is ≤0.2db/Km.

In some embodiments of the present utility model, the vibration sensing optical fiber includes a plurality of vibration sensing optical fiber sensing sections, and the distance between each vibration sensing optical fiber sensing section is ≥0.2m and ≤0.5m.

In some embodiments of the present invention, the sensing section of the vibration sensing optical fiber is spiral, and the long axis of the spiral is distributed vertically. The diameter of the sensing optical fiber sensing section is 250-350mm, and the length of the optical fiber in each vibration sensing optical fiber sensing section is ≥4m and ≤6m.

In some embodiments of the present utility model, multiple rows of vibration sensing optical fiber sensing sections are arranged in the road surface structure body.

In some embodiments of the present utility model, the vibration sensing optical fiber sensing segments are evenly distributed in the width direction of the lane, and the coverage rate is 2-3 per meter.

In some embodiments of the present utility model, 2 to 3 rows of vibration-sensing optical fiber sensing sections are arranged in the road surface structure body according to the length direction of the lane, and the distance between each row of vibration-sensing optical fiber sensing sections is 3-3. 10m.

In some embodiments of the present utility model, the pavement structure body is a combination of one or more of cement concrete pavement, asphalt concrete pavement or composite pavement structure;

In some embodiments of the present utility model, multiple rows of vibration-sensing optical fiber sensing sections are arranged in the road surface structure body, and at least part of the vibration-sensing optical fiber sensing sections are connected in series in sequence.

The second aspect of the present invention provides a road vehicle type identification method, using the embedded road vehicle type identification system The embedded road vehicle type identification system includes the following steps:

1) Using the embedded vehicle type identification system, the vibration data of the road surface structure when the vehicle passes through the vibration optical fiber analysis device is collected;

2) Carry out feature extraction to each vibration data group;

3) According to the characteristic information of the above-mentioned vehicles, classify the driving vehicles.

In some implementations of the present utility model, in the step 2), the features are selected from one or more combinations of moving speed, axle type, wheel type, wheelbase and vibration spectrum distribution characteristics of the traveling vehicle.

In some embodiments of the present utility model, in the step 3), the feature information used to classify the driving vehicle is the shaft type and/or the vibration spectrum distribution characteristics.

In some embodiments of the present invention, in the step 2), the method for feature extraction of each vibration data group specifically includes the following steps:

a) Compile the original vibration data observed at the measuring points in the same cross-section into the same vibration data group, and perform empirical mode decomposition (EMD) to obtain multiple vibration components after EMD processing;

b) Superimpose vibration components of a specific order to obtain the superimposed vibration curve fr(t) (x-axis is time, y-axis is vibration intensity):

Among them, IMF _i is the vibration component of the i-th order after EMD treatment, and n and m are the lowest and highest orders of specific orders.

c) Calculate the sum of the squares of the vibration data per unit time (τ) as the short-term energy within this period, and thus obtain the time-history curve E(T) of the short-term vibration energy (the x-axis is time, and the y-axis is the short-term vibration energy time energy):

The unit time τ is the length of the short-term energy analysis time frame, and the initial value can be set to 0.5s;

d) Set the vehicle vibration energy judgment threshold, and the part of the short-term energy time-history curve used to determine whether the vehicle passes through that meets the threshold requirements is regarded as the vibration period [T ₁ , T ₂ ] generated by the excitation of the driving vehicle. Then intercept the vibration curve fr'(t) in this period from the superimposed vibration curve fr(t);

e) According to the time difference between the excitation and vibration of the vehicle at the adjacent measurement section, the moving speed of the vehicle is obtained as:

Among them, L is the distance between adjacent sections; Δt is the time difference;

f) According to the minimum wheelbase of the driving vehicle (that is, S in the publicity below), calculate and determine the short-term energy analysis time frame length for axle type feature recognition as:

Among them, v represents the moving speed of the vehicle; S is the coupling wheelbase, usually 1m;

g) Normalize the intercepted vibration curve fr'(t) to obtain the normalized vibration curve fn(t):

According to the short-term energy analysis time frame length used for axial feature recognition, the short-term energy time-history curve E'(T) of the normalized vibration curve fn(t) is calculated. Set the axle type recognition threshold and determine the number of curve peaks that meet the threshold requirements, judge the axle type of the driving vehicle based on the number of peaks, and then determine the wheelbase based on the distance between the peaks;

h) With reference to steps a)-g), calculate the intercepted short-term energy distribution curves in each vibration sensing optical fiber sensing section when the front and rear axles of the vehicle pass through the same measurement section. The sensing segments are sorted from left to right according to the driving direction of the vehicle, and the ordinal numbers of the vibrating optical fiber sensing segments are defined accordingly, and the maximum short-term energy values when passing by different axes are calculated respectively, which are denoted as E(i,j), where i is the ordinal number of the axle, and j is the ordinal number of the vibrating optical fiber sensing section;

i) Based on the front axle, calculate the ratio of the maximum short-term energy value of the vehicle rear axle to the front axle in each vibration optical fiber sensing segment, which is recorded as Er(i), where i=1,2,3...n,n For the number of sensing segments on a single measurement section, calculate the standard deviation of Er(i) to measure the vibration difference caused by the front and rear axles in the same section, and set the wheel type judgment threshold. If the standard deviation exceeds this threshold, it means that the axle type is Double wheel group, and vice versa for single wheel group;

j) Obtain the frequency spectrum distribution of the intercepted vibration curve fr'(t) by means of time-frequency analysis, and obtain the distribution characteristic S(f,t) of the amplitude-frequency characteristic of the signal in time (the x-axis is time, the y-axis is frequency, The z-axis is the vibration amplitude), and the amplitude-frequency distribution data is superimposed on the time direction (x-axis), and the superposition result is used as the amplitude-frequency curve S(f) of the vehicle vibration response (the x-axis is the frequency, and the y-axis is the vibration amplitude );

k) Intercept the frequency band sensitive to the vehicle vibration response as the characteristic frequency band, and calculate the weighted frequency f _w in this frequency band to characterize the spectrum distribution characteristics of the vehicle vibration response. The calculation method is:

Among them, f _h and f _l are the upper and lower cut-off frequencies of the characteristic frequency band respectively.

In some embodiments of the present utility model, in the road vehicle type identification method, prior parameter training for vehicle type identification is also performed on the driving vehicle;

In some embodiments of the present utility model, in the road vehicle type identification method, the vehicle type identification probability is also determined for the driving vehicle.

In some embodiments of the present invention, during the prior parameter training of the vehicle type identification for the driving vehicle, the acquired moving speed, wheelbase, and vibration spectrum distribution characteristics of the driving vehicle are used for parameter training through the support vector machine.

In some embodiments of the present invention, in the process of identifying the probability of vehicle type identification of the driving vehicle, the driving vehicle is divided into "small and medium-sized passenger cars", "large trucks", " According to the three categories of large buses or small and medium-sized trucks, the moving speed, wheelbase, and vibration spectrum characteristics of the driving vehicles are extracted, and the joint probability discrimination method is used to distinguish "large buses or small and medium-sized trucks" into "double-axle buses" and "twin-axle trucks".

Description of drawings

Fig. 1 shows the front view of the embedded vibration sensing optical fiber of the utility model.

Fig. 2 shows a top view of the embedded vibration sensing optical fiber of the utility model.

Fig. 3 is a schematic diagram showing the structure of the embedded road vehicle type identification system of the present invention.

Fig. 4 shows the flow chart of the utility model embedded vehicle identification method.

Fig. 5 shows a flow chart for extracting the moving speed of a driving vehicle in the embedded vehicle type identification method of the present invention.

Fig. 6 shows a flowchart for extracting the axle type and wheelbase of the driving vehicle in the embedded vehicle type identification method of the present invention.

Fig. 7 shows the flow chart of extracting the wheel type of the embedded vehicle type identification method of the present invention.

Fig. 8 shows a flow chart of extracting vibration spectrum characteristics of the embedded vehicle type identification of the utility model.

Fig. 9 is a schematic diagram showing the decomposition of EMD in the embodiment.

Fig. 10 shows the short-term energy distribution curves of a two-axle truck and a four-axle truck in Example 1.

FIG. 11 is a schematic diagram showing the layout of the embedded vibration measurement device in Embodiment 2.

Component designation description

1 Vibration sensing fiber

11 Vibration sensing fiber optic sensing section

12 Vibration sensing optical fiber transition section

2 Vibration fiber optic analysis device

3 Optical fiber lead-out wires

Detailed ways

The implementation of the present utility model is illustrated by specific specific examples below, and those skilled in the art can easily understand other advantages and effects of the present utility model from the content disclosed in this specification.

See Figures 1 through 9. It should be noted that the structures, proportions, sizes, etc. shown in the drawings attached to this specification are only used to match the content disclosed in the specification, for those who are familiar with this technology to understand and read, and are not used to limit the implementation of the utility model Therefore, it has no technical substantive meaning. Any modification of structure, change of proportional relationship or adjustment of size shall still fall within the scope of The technical content disclosed by the utility model must be within the scope covered. At the same time, terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description and are not used to limit this specification. The practicable range of the utility model, and the change or adjustment of its relative relationship, without any substantial change in the technical content, shall also be regarded as the practicable scope of the utility model.

As shown in Figures 1-3, the utility model provides an embedded road vehicle type identification system, which includes a road surface structure body, and a vibration sensing optical fiber 1 is arranged in the road surface structure body, and the vibration sensing optical fiber includes One or more vibration sensing optical fiber sensing sections 11, each vibration sensing optical fiber sensing section 11 is connected by a vibration sensing optical fiber transition section 12, and also includes a vibration optical fiber analysis device 2, and the vibration optical fiber analysis device 2 and Vibration sensing optical fibers 1 are connected through optical fiber lead-out lines 3 .

In the embedded road vehicle type identification system provided by the utility model, the road surface structure body 1 can be one or more combinations of cement concrete pavement, asphalt concrete pavement or composite pavement structure, etc. The sensing optical fiber 1 is laid in the road structure body 1 so as to collect structural vibration information caused by vehicles passing the road.

In the embedded road vehicle type identification system provided by the utility model, those skilled in the art can select a suitable type of optical fiber for collecting the vibration data of the road surface structure when the vehicle passes by, for example, the vibration sensing optical fiber 1 can be a single Mode fiber (Single Mode Fiber), the single mode fiber usually refers to the optical fiber that can only transmit the light of one mode, the vibration sensing fiber 1 can be a metal armored fiber, the diameter of the vibration sensing fiber 1 (including the jacket) can be 2 to 5mm, the long-term allowable tensile force of the vibration sensing optical fiber 1 usually needs to be ≥ 600N, and the short-term allowable tensile force usually needs to be ≥ 1500N (the short-term and long-term tensile performance tests of optical fibers are based on YD/T 769-2003 "Optical Cables for Core Networks - Outdoor Optical Cables for Central Tube Communications"), so that it has a certain tensile strength to ensure that the optical fiber will not be damaged or broken when it is stretched. The vibration sensing optical fiber 1 Attenuation ≤0.2db/Km to ensure the smooth operation of DOVS.

In the embedded road vehicle type identification system provided by the utility model, the vibration sensing optical fiber 1 usually includes a plurality of vibration sensing optical fiber sensing sections 11, and the distance between each vibration sensing optical fiber sensing section 11 is usually ≥0.2m and ≤0.5m, the 11 sections of the vibration sensing fiber optic sensor are usually helix, the long axis of the helix can be distributed vertically (usually relative to the road surface), the vibration sensing fiber optic sensor The height of the long-axis direction of the sensing section 11 can be 5-30mm, the diameter of the vibration-sensing optical fiber sensing section 11 can be 250-350mm, and the length of the optical fiber in each vibration-sensing optical fiber sensing section 11 is usually ≥ 4m and ≤ 6m .

In the embedded road vehicle type identification system provided by the utility model, multiple rows of vibration-sensing optical fiber sensing sections 11 are arranged in the road surface structure body, and the vibration-sensing optical fiber sensing sections 11 can Evenly distributed, the coverage rate is 2 to 3 per meter. When measuring, multiple rows of vibration sensing optical fiber sensing sections 11 can be arranged in the road surface structure body according to the length direction of the lane. More specifically, 2 to 3 There are three rows of vibration-sensing optical fiber sensing sections 11, the distance between each row of vibration-sensing optical fiber sensing sections 11 can be 3-10m, and at least some or all of the vibration-sensing optical fiber sensing sections 11 are connected in series in sequence. Those skilled in the art can adjust the parameters of the vibration-sensing optical fiber transition section 12 as required, and the parameters of the vibration-sensing optical fiber transition section 12 can be basically the same as the vibration-sensing optical fiber sensing section 11 on the whole. Keep a certain length of the vibration sensing optical fiber transition section 12, so as to ensure that the vibration sensing has sufficient spatial resolution, for example, in an embodiment of the present utility model, the vibration sensing between each vibration sensing optical fiber sensing section 11 The optical fiber transition section 12 may be ≥0.2 meters.

In the embedded road vehicle type identification system provided by the utility model, the vibration optical fiber analysis device 2 can be selected from various distributed optical fiber vibration demodulators in the field, which can detect the vibration signal along the connected optical fiber, For example, it can be the FT630-02 fiber optic vibration sensor analyzer produced by Shanghai Baian Sensing Technology Co., Ltd.

The utility model also provides a road vehicle type identification method, using the above-mentioned embedded road vehicle type identification system, when measuring, the road vehicle type identification system can be arranged in the road structure, when the vehicle passes by, Collect the vibration data of the optical fiber through the vibration fiber demodulation device, and save it as the original vibration data, as shown in Figure 4, which may specifically include the following steps:

1) Using the embedded vehicle identification system as described above, the vibration data of the road surface structure when the vehicle passes through the vibration optical fiber analysis device is used to obtain the original vibration data. Vibration data, and save as the original vibration data;

2) feature extraction is carried out to each vibration data group, and described feature can comprise one or more in the speed of movement of driving vehicle, axle type, wheel type, wheelbase and vibration spectrum distribution;

3) According to the characteristic information of the above-mentioned vehicles, the characteristic information used to classify the driving vehicles is the axis type and/or the vibration spectrum distribution characteristics. In one embodiment of the present invention, the driving vehicles can be classified, and the driving vehicles There are three types of "small and medium passenger cars", "large trucks" and "big buses or small and medium trucks". The judgment method can be, for example: first judge the axle type of the vehicle. The root axle, the type is "small and medium-sized passenger car" or "big bus or small and medium-sized truck"; if there are 3 or more peaks, the type is "large truck". If the wheelbase is smaller than the threshold, it is a "small and medium-sized passenger car", otherwise it is "large bus or small and medium-sized truck";

4) Carry out vehicle type recognition prior parameter training on driving vehicles, more specifically, it may be to carry out vehicle type recognition prior parameter training on "buses or small and medium-sized trucks" driving vehicles, during the process of performing vehicle type recognition prior parameter training on driving vehicles, The obtained moving speed, number of axles, wheelbase, and vibration spectrum distribution characteristics of the driving vehicle can be used for parameter training through the support vector machine;

5) Carry out vehicle type identification probability discrimination for driving vehicles, more specifically, it can be used to determine the vehicle type identification probability of "buses or small and medium-sized trucks". During the process of vehicle type identification probability discrimination, the movement The speed, wheelbase, and vibration spectrum features are judged according to the joint probability discrimination method.

In the road vehicle type identification method provided by the utility model, in the step 2), the method for feature extraction of each vibration data group specifically includes the following steps:

a) Compile the original vibration data observed at the measuring points in the same cross-section into the same vibration data group, and perform empirical mode decomposition (EMD) to obtain multiple vibration components after EMD processing;

b) Superimpose the vibration components of a specific order (according to the measured vibration data to adjust the interval range and quantity of the order (frequency interval), for example, it can be 2-9 or 3-10) vibration components, and obtain the superimposed vibration curve fr( t) (x-axis is time, y-axis is vibration intensity):

Among them, IMF _i is the vibration component of the i-th order after EMD treatment, and n and m are the lowest and highest orders of specific orders.

c) Calculate the sum of the squares of the vibration data per unit time (τ) as the short-term energy within this period, and thus obtain the time-history curve E(T) of the short-term vibration energy (the x-axis is time, and the y-axis is the short-term vibration energy time energy):

The unit time τ is the length of the short-term energy analysis time frame, and the initial value can be set to 0.5s.

d) Set the vehicle vibration energy judgment threshold, which is used to judge whether the part of the short-term energy time-history curve that the vehicle passes through meets the threshold requirement as the vibration period [T ₁ , T ₂ ] generated by the excitation of the driving vehicle, and then from The superimposed vibration curve fr(t) intercepts the vibration curve fr'(t) within this period;

e) According to the time difference between the excitation and vibration of the vehicle at the adjacent measurement section, the moving speed of the vehicle is obtained as:

Among them, L is the distance between adjacent sections; Δt is the time difference.

f) According to the minimum wheelbase of the driving vehicle (that is, S in the publicity below), calculate and determine the short-term energy analysis time frame length for axle type feature recognition as:

Among them, v represents the moving speed of the vehicle; S is the coupling wheelbase, usually 1m;

g) Normalize the intercepted vibration curve fr'(t) to obtain the normalized vibration curve fn(t):

According to the short-term energy analysis time frame length used for axial feature recognition, the short-term energy time-history curve E'(T) of the normalized vibration curve fn(t) is calculated. Set the axle type recognition threshold and determine the number of curve peaks that meet the threshold requirements, judge the axle type of the driving vehicle based on the number of peaks, and then determine the wheelbase based on the distance between the peaks;

h) According to the above steps (steps a-g), calculate the intercepted short-term energy distribution curves in each vibration sensing optical fiber sensing section when the front and rear axles of the vehicle pass through the same measurement section, and in a single measurement section, the vibration sensing optical fiber sensor The sensing segments are sorted from left to right according to the driving direction of the vehicle, and the ordinal numbers of the vibrating optical fiber sensing segments are defined accordingly, and the maximum short-term energy values when passing by different axes are calculated respectively, which are denoted as E(i,j), where i is the ordinal number of the axle, and j is the ordinal number of the vibrating optical fiber sensing section;

i) Based on the front axle, calculate the ratio of the maximum short-term energy value of the vehicle rear axle to the front axle in each vibration optical fiber sensing section (refer to the above E(i,j)), denoted as Er(i), where i= 1,2,3...n, n is the number of sensing segments on a single measurement section, calculate the standard deviation of Er(i) to measure the vibration difference caused by the front and rear axles in the same section, set the wheel type judgment threshold, standard deviation If the threshold is exceeded, it means that the shaft type is a double-wheel set, otherwise it is a single-wheel set.

j) Obtain the spectral distribution of the intercepted vibration curve fr'(t) by means of time-frequency analysis, which can be wavelet decomposition, short-time Fourier transform or S-transform, and finally obtain the amplitude-frequency characteristics of the signal in time The distribution characteristics of S(f,t) (the x-axis is time, the y-axis is frequency, and the z-axis is vibration amplitude), the amplitude-frequency distribution data is superimposed on the time direction (x-axis), and the superposition result is used as the vehicle vibration response The amplitude-frequency curve S(f) (x-axis is frequency, y-axis is vibration amplitude).

k) Intercept the frequency band sensitive to the vehicle vibration response as the characteristic frequency band, and calculate the weighted frequency f _w in this frequency band to characterize the spectrum distribution characteristics of the vehicle vibration response. The calculation method is:

Among them, f _h and f _l are the upper and lower cut-off frequencies of the characteristic frequency band respectively.

Aiming at the shortcomings of the existing vehicle identification methods, the utility model provides a new, accurate, and widely identifiable road vehicle based on vibration perception through DOVS (Distributed Optical Vibration Sensing System) to meet the needs of vehicle identification technology. Type identification method, device and system thereof.

Example 1

The auxiliary road of the Zhouhai section of the Pudong Outer Ring Road is a cement concrete pavement, and the traffic composition of this section is complex, including passenger cars, buses and various types of trucks.

The DOVS-based embedded vehicle type identification system is used to detect and identify the type of driving vehicle, and the system is installed in a prefabricated construction method. Firstly, the embedded vehicle identification device is constructed according to the winding method shown in Fig. 1-Fig. 2. The information of the identification device used in the embodiment is as follows: a 400m long vibration sensing optical fiber is used to construct the identification device, and the vibration optical fiber sensing section adopts Single-mode optical fiber, the sheath is metal armored, the total diameter (including the sheath) is 3mm. The optical fiber sensing section is wound in a circular shape, with a winding diameter of 300mm, 4 turns each time, and a length of about 3.8m.

Then build an embedded vehicle identification system through the layout shown in Figure 3. In the embodiment, the transition section of the vibration optical fiber is a single-mode optical fiber with a diameter of 3mm, and the length of each transition section is the same, which is 0.2m. The multiple vibration measurement devices After the optical fiber is connected in series through the transition section, it is fixed on the steel mesh in the prefabricated concrete paving panel by cable ties. A total of 24 vibration measuring devices are arranged on the whole board, among which there are 3 rows arranged along the length direction of the board, each row has 8 devices, and the distance between each row is about 2.2m. After the optical fiber coil is bound to the steel mesh, it is poured with concrete into the prefabricated paving panel, and the position of the optical fiber is about 7cm away from the top surface of the prefabricated panel. After the system is buried in the assembly plate, it is installed in the existing road surface by means of hoisting. After the system is built with the assembly board, it is exported and connected to the vibration analysis equipment through the optical fiber lead-out line. The type and parameters of the optical fiber lead-out line are the same as those of the sensing section and the transition section, and the length is 25m. The vibration analysis equipment adopts the FT630-02 fiber optic vibration sensor analyzer produced by Shanghai Bayan Sensing Technology Co., Ltd.

According to the buried vehicle identification system, the vibration signals of the road surface were collected for more than 5 hours, including the vibration signals generated by 217 moving vehicles. After obtaining the original vibration signal, the embedded vehicle type identification method (as shown in Figure 4-8) is used to identify different vehicle types, and the whole process is realized by MATLAB software. [l1]

After the original vibration signal is captured, the vehicle model is identified according to the steps shown in Figure 4. In step 2, according to the captured vibration signal of this road section, after the original vibration signal is decomposed by EMD, the 2nd to 9th order components are superimposed to obtain the corresponding time history curve (Fig. 9), namely:

Among them, IMF _i is the i-th order vibration component after EMD treatment.

After obtaining the superimposed vibration time-history curve, calculate the sum of the squares of the vibration data per unit time (τ) as the short-term energy within this period, and thus obtain the time-history curve E(T) of the vibration short-term energy (x-axis is time, and the y-axis is vibration short-term energy):

The unit time τ is the length of the short-term energy analysis time frame, and the value is 0.5s.

Based on experience, the vehicle vibration judgment threshold is set to be twice the average value of the short-term energy in the no-car state, and the time period when the short-term energy exceeds the vehicle vibration judgment threshold is determined as the time period when the vehicle is driving, so as to intercept the vibration curve of the vehicle passing time period fr'(t).

When the front axle of the vehicle passes through two adjacent measurement sections, the difference of the moment corresponding to the peak value of the short-term energy curve is used as the time difference of the vehicle passing through the two measurement sections, and the driving speed of the vehicle is calculated accordingly. The buried sections of the first and third rows of optical fibers.

The length of the short-term energy analysis time frame of each vehicle is calculated according to the driving speed. In this embodiment, the speed of the measured vehicle is mainly distributed in the range of 8-25m/s, and the corresponding short-term energy analysis time frame length is distributed in the range of 0.01s-0.06s.

The short-term energy corresponding to the time history curve is calculated by using the calculated time frame length, and the short-term energy distribution curve of the vibration data in the time domain is obtained. Use the findpeaks function in MATLAB to extract the peaks, set the axle type recognition threshold to 3 (threshold after normalization processing), and determine the number of axles of the driving vehicle accordingly. According to the number of axles, a category of "large trucks" can be distinguished. Figure 10 shows the short-term energy distribution curves of a two-axle truck and a four-axle truck.

Calculate the time difference between adjacent peaks of a single vehicle and multiply it by the speed of the vehicle to calculate the wheelbase of the vehicle. Set the wheelbase judgment threshold to 3.4m, that is, a vehicle with a wheelbase less than 3.4m can be judged as a "small and medium-sized passenger car", and vice versa as a "large bus or small and medium-sized truck".

The classification of buses or small and medium-sized trucks is carried out by the prior parameter training of vehicle type identification and the probability discrimination method of vehicle type identification. The prior parameter training of car model recognition is carried out by support vector machine method, and the extracted training parameters are spectral distribution features, driving speed and front and rear axle distance. First, short-time Fourier transform is used for the intercepted vibration curve:

Wherein γ(t) is a window function, a rectangular window is selected in this embodiment, and the window width is 0.256s. S(t,f) is a function of amplitude versus time and frequency (time on the x-axis, frequency on the y-axis, and vibration amplitude on the z-axis). Then S(t, f) is superimposed on the time direction (x-axis), and the superposition result is used as the amplitude-frequency curve S(f) of the vehicle vibration response (x-axis is frequency, y-axis is vibration amplitude).

Taking 0-40Hz as the characteristic frequency band of the driving vehicle in this embodiment, the weighted frequency _fw in this frequency band is calculated to characterize the spectrum distribution characteristics of the vehicle vibration response. The calculation method is:

After obtaining the spectrum distribution characteristics of each driving vehicle, select the weighted frequency, vehicle speed, and wheelbase of 60 driving vehicles for support vector machine training, and construct a support vector machine network. The kernel function in the training process of the support vector machine adopts the radial basis function, and its formula is as follows:

Among them, x' is the value of the support vector, x is the sample value to be classified, σ is the width of the kernel function, and the value is 0.5 in the vector machine network. In this embodiment, the training and verification of the support vector machine are realized by using the LIBSVM toolbox in the MATLAB software. Randomly select 60 pieces of driving vehicle data as the training set, and the rest of the data as the verification set, and use the vector machine network generated by training the training set for classification verification.

According to the above steps, the vehicles are classified, and the classification results are shown in Table 1. It can be seen that the embedded vehicle type identification device, system and method provided by this aspect can effectively identify different vehicle types, avoid the disadvantages of high cost and poor reliability of traditional identification methods, and realize real-time identification of vehicle types on cement concrete roads. identify.

Table 1 Vehicle type recognition results

Example 2

At the T-shaped intersection of Huajing Road-Taipei East Road, Pudong New District, Shanghai, the pavement structure is a concrete structure, and the traffic composition includes passenger cars, buses, and minivans.

The DOVS-based embedded vehicle type identification system is used to detect and identify the types of vehicles in the intersection, and the system is installed in a prefabricated construction method. First, build the embedded vehicle identification device according to the winding method shown in Figure 1. The information of the identification device used is as follows: a 1200m long vibration sensing optical fiber is used to construct the identification device, the vibration optical fiber sensing section adopts single-mode optical fiber, and the sheath adopts Metal armored, with an overall diameter (including sheath) of 3mm. The optical fiber sensing section is wound in a circular shape, with a winding diameter of 300mm, 4 turns each time, and a length of about 4m.

Then build an embedded vehicle identification system through the layout shown in Figure 3. The transition section of the vibrating optical fiber in each paving panel is a single-mode optical fiber with a diameter of 3mm. The length of each transition section is the same, both being 0.2m. After the device is connected in series through the optical fiber in the transition section, it is fixed on the steel mesh by cable ties. There are 21 identification devices arranged in 7 boards, and 3 rows are arranged along the length direction of the board, with 7 devices in each row, that is, 3×7 layout; 2 boards are arranged with 7 pieces (1×7); there is 1 board arranged The quantity is 14 (2×7); there is 1 board and the quantity is 36 (6×6). The vibrating optical fiber is led out from each board through the optical fiber lead-out line and placed in the fiber box reserved on the edge of the board. The type and parameters of the optical fiber lead-out line are the same as those of the sensing section and the transition section, and the length is 25m. After all the prefabricated panels are assembled on site, the vibrating optical fibers in the optical fiber boxes between adjacent paving panels are fused and connected to the vibration analysis equipment through the optical fiber lead-out lines. Form the embedded vehicle type identification system (as shown in Figure 11) of 13 boards as a whole. The vibration analysis equipment adopts the FT630-02 fiber optic vibration sensor analyzer produced by Shanghai Bayan Sensing Technology Co., Ltd.

Finally, the embedded vehicle type identification method (as shown in Figure 4-8) is used to measure and identify the types of 40 vehicles. Due to the small amount of data, only based on the axle type parameters, small and medium-sized buses, buses or small and medium-sized Trucks and large trucks are identified, and the data processing method and flow in the identification process are the same as those in Embodiment 1. The overall recognition accuracy of the three types of vehicles is 95%.

To sum up, the utility model effectively overcomes various shortcomings in the prior art and has high industrial application value.

The above-mentioned embodiments only illustrate the principles and effects of the present utility model, but are not intended to limit the present utility model. Anyone familiar with this technology can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed in the utility model should still be covered by the claims of the utility model.

Claims (10)

1. a kind of flush type road vehicle type identifying system, which is characterized in that including pavement structure ontology, the pavement structure Vibrating sensing optical fiber (1) is equipped in ontology, the vibrating sensing optical fiber includes one or more vibrating sensing Fibre Optical Sensor sections (11), it is connected by vibrating sensing optical fiber changeover portion (12) between each vibrating sensing Fibre Optical Sensor section (11), further includes vibration light Fine analytical equipment (2), the vibration optical fiber analytical equipment (2) are connected with vibrating sensing optical fiber (1) by fiber pigtail (3) It connects.

2. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the vibrating sensing optical fiber (1) it is single mode optical fiber, the vibrating sensing optical fiber (1) is metal armouring optical fiber.

3. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the vibrating sensing optical fiber (1) a diameter of 2~5mm, long-term permission drawing force >=600N of the vibrating sensing optical fiber (1), in short term allow drawing force >= 1500N, decaying≤0.2db/Km of the vibrating sensing optical fiber (1).

4. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the vibrating sensing optical fiber (1) include multiple vibrating sensing Fibre Optical Sensor sections (11), spacing >=0.2m between each vibrating sensing Fibre Optical Sensor section (11) and ≤0.5m。

5. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the vibrating sensing optical fiber It is spiral shape to sense section (11), and spiral long axis is distributed vertically, the height of vibrating sensing Fibre Optical Sensor section (11) long axis direction For 5~20mm, a diameter of 250-350mm of vibrating sensing Fibre Optical Sensor section (11), each vibrating sensing Fibre Optical Sensor section (11) Length >=the 4m and≤6m of middle optical fiber.

6. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the pavement structure ontology In be equipped with multiple rows of vibrating sensing Fibre Optical Sensor section (11).

7. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the vibrating sensing optical fiber Sensing section (11) is uniformly distributed in lane width direction, and coverage rate is 2~3/meter.

8. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the pavement structure ontology In by the length direction in track be laid with 2~3 row's vibrating sensing Fibre Optical Sensor sections (11), it is each to arrange vibrating sensing Fibre Optical Sensor section (11) spacing between is 3~10m.

9. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the pavement structure ontology For one or more combinations in cement concrete paving, bituminous (bitumen)concrete pavement or compound pavement structure.

10. flush type road vehicle type identifying system as described in claim 1, which is characterized in that the pavement structure sheet Multiple rows of vibrating sensing Fibre Optical Sensor section (11) is equipped in body, at least part of vibrating sensing Fibre Optical Sensor section (11) is sequentially connected in series.