A method and system for constructing driving behavior spectrum in tunnel sections

Abstract

The present invention belongs to the technical field of tunnel driving behavior, and provides a method and system for constructing a driving behavior spectrum of a tunnel section. The present invention constructs a driving behavior chain based on physiological data and driving data, and then expands and extends the driving behavior chain in different dimensions to construct a driving behavior spectrum of a tunnel section, thereby realizing research on the driving behavior spectrum of the tunnel. When establishing the driving behavior chain, the driving behavior chain is obtained by connecting the psychophysiological data and the driving data in series using the time dimension, and the data in the driving process and the time relationship between the data are fully considered. On this basis, the driving behavior spectrum obtained by expanding and extending the driving behavior chain in different dimensions can fully and accurately reflect the driving behavior.

Description

Technical Field

[0001] The present invention belongs to the technical field of tunnel driving behavior, and in particular relates to a method and system for constructing a tunnel section driving behavior spectrum. Background Art

[0002] In recent years, as researchers in related fields have deepened their research on tunnel operation safety, highway tunnel safety assurance technology has been updated, and driver behavior detection and recognition technology has also been developed. However, due to the lack of research methods, most of the research on tunnel driver behavior is statistical analysis on single factors or low-dimensional features, which makes it difficult to form a research system. As a result, designers lack practical and reliable basis when formulating highway tunnel design plans and safety management methods, and tunnel operations are difficult to achieve optimal conditions.

[0003] Behavior spectrum is a systematic data recording of the research results after detailed and comprehensive observation and research on the behavior of individual animals or groups. In the field of transportation, driving behavior spectrum can also be established by making as comprehensive and systematic records of driver behavior data in different driving scenarios and extracting corresponding features.

[0004] The inventors found that there is no research on tunnel driving behavior spectrum in the existing research, so that designers lack practical and reliable basis when formulating highway tunnel design plans and safety management measures; and in the process of constructing the existing driving behavior spectrum, the data in the driving process and the relationship between the data are not fully considered, so that the characteristics in the existing driving behavior spectrum cannot fully and accurately reflect the driving behavior. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a method and system for constructing a driving behavior spectrum of a tunnel section. The present invention proposes a driving behavior chain research method for the driving data of drivers in tunnel sections, and constructs a driving behavior chain based on the collected data of the driver's steering wheel, eye movement, heart rate, speed, etc., combined with environmental variable information; the driver's driving behavior under different working conditions is analyzed from the characteristic performance of micro and macro dimensions, and on the basis of the constructed driving behavior chain, the scope and time scale are expanded and extended, and the driving behavior spectrum is constructed with the matrix as the basic unit; the matrix is ​​used as the research form, the behavior spectrum is divided into a number of basic matrix units, the basic matrix units are iteratively calculated by designing a sliding window, and the tunnel driving behavior is qualitatively analyzed with the spectrum radius characteristic parameter.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0007] In a first aspect, the present invention provides a method for constructing a tunnel section driving behavior spectrum, comprising:

[0008] Obtain the psychological and physiological data and driving data of drivers in tunnel sections;

[0009] Preprocessing the cardiophysiological data and the driving data, filtering out invalid data in the cardiophysiological data, and correcting the driving data;

[0010] Extract features from preprocessed cardiophysiological data and driving data;

[0011] Based on the feature extracted data, a driving behavior chain is established according to the time dimension;

[0012] The driving behavior chain is expanded and extended in different dimensions to construct a tunnel section driving behavior spectrum.

[0013] Furthermore, the psychophysiological data includes electroencephalogram signals, eye movement trajectory signals and heart rate signals; and the driving data includes tunnel section environmental variable information, steering wheel angle data and driving speed.

[0014] Furthermore, irrelevant frequency data in the EEG signal are filtered out, and the invalid data in the eye movement trajectory signal is filtered out using the Laida criterion; the driver's cross-sectional speed inside and outside the tunnel is marked by the driver's tunnel driving timestamp to obtain the speed change pattern of the driver inside and outside the tunnel; the steering wheel angle data is baseline corrected.

[0015] Furthermore, wavelet packet decomposition and reconstruction are used to process the EEG signals, and then the expression of the EEG rhythmic signal on the power spectrum is obtained by comparing the indicators of the increase or decrease of the fast wave and slow wave power energy in the time-frequency domain; for the eye movement trajectory signal, the pupil area, gaze characteristics, scanning characteristics and blinking characteristics are extracted; the heart rate is used as a characterization indicator of the driver's mental state, and the heart rate growth rate is selected as a psychological characteristic research indicator; the steering wheel angle entropy is used as the characteristic variable to quantitatively describe the driver's steering wheel angle changes in the time series to quantitatively analyze the driver's lateral handling stability; the collision risk is used to describe the relationship between the front and rear vehicles in time and space when the following behavior occurs in the tunnel group section.

[0016] Furthermore, the driver, vehicle, road condition and environment are regarded as multi-dimensional unified elements, and connected in series through driving routes. The multi-dimensional unified elements are reflected to the driving path in the order of time development, thus constructing a driving behavior chain.

[0017] Furthermore, the Fourier transform algorithm was used to perform frequency domain analysis on the data in the driving behavior chain in tunnel scenarios of different lengths. The longer the tunnel length, the greater the probability that the driver drives at a fixed steering angle and speed, and the driver's participation decreases. As the tunnel length increases, the uncertainty of the driving speed increases, and the driver's ability to maintain speed deteriorates. As the tunnel length increases, the heart rate discreteness increases.

[0018] Furthermore, based on the driving behavior chain, the scope and time scale are expanded and extended, and the driving behavior spectrum is constructed with the square matrix as the basic unit; in the square matrix, the row vector represents the spatiotemporal information or state information of the driver that changes over time in a certain dimension, and the column dimension is the time index of the matrix, reflecting the values ​​of indicators in different dimensions at a certain moment; all the values ​​of the row vector in the driving behavior time series constitute the driving behavior spectrum.

[0019] In a second aspect, the present invention further provides a system for constructing a tunnel section driving behavior spectrum, comprising:

[0020] The data acquisition module is configured to: acquire the psychological and physiological data and driving data of the driver in the tunnel section;

[0021] A preprocessing module is configured to: preprocess the cardiophysiological data and the driving data, filter out invalid data in the cardiophysiological data, and correct the driving data;

[0022] The feature extraction module is configured to: extract features from the pre-processed cardiophysiological data and driving data;

[0023] The driving behavior chain establishment module is configured to: establish a driving behavior chain according to the time dimension based on the feature extracted data;

[0024] The driving behavior spectrum building module is configured to: expand and extend the driving behavior chain in different dimensions to build a tunnel section driving behavior spectrum.

[0025] In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for constructing a tunnel section driving behavior spectrum as described in the first aspect.

[0026] In a fourth aspect, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for constructing a driving behavior spectrum for a tunnel section described in the first aspect are implemented.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] 1. The present invention constructs a driving behavior chain based on the psychophysiological data and driving data, and then expands and extends the driving behavior chain in different dimensions to construct a driving behavior spectrum for tunnel sections, thereby realizing the research on the driving behavior spectrum for tunnels. When establishing the driving behavior chain, the psychophysiological data and driving data are connected in series using the time dimension to obtain the driving behavior chain, which fully considers the data in the driving process and the time relationship between the data. On this basis, the driving behavior spectrum obtained by expanding and extending the driving behavior chain in different dimensions can fully and accurately reflect the driving behavior.

[0029] 2. The present invention constructs a driving behavior chain based on the collected data of the driver's steering wheel, eye movement, heart rate, speed, etc., combined with environmental variable information; analyzes the driver's driving behavior under different working conditions from the characteristic performance of both micro and macro dimensions, and on the basis of the constructed driving behavior chain, expands the scope and time scale, and constructs a driving behavior spectrum with the matrix as the basic unit; using the matrix as the research form, the behavior spectrum is divided into several basic matrix units, and the basic matrix units are iteratively calculated by designing a sliding window, and the tunnel driving behavior is qualitatively analyzed with the spectrum radius characteristic parameter. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0031] Figure 1 This is a flow chart of Embodiment 1 of the present invention;

[0032] Figure 2 The data acquisition test structure of Example 1 of the present invention;

[0033] Figure 3 The driving behavior process unit decomposition of the first embodiment of the present invention;

[0034] Figure 4 It is a visual expression of the driving behavior chain of Example 1 of the present invention;

[0035] Figure 5 Comparison of angular velocity time domain characteristics of different tunnel scenes in Example 1 of the present invention;

[0036] Figure 6 is the average acceleration of different tunnel scenarios in Example 1 of the present invention;

[0037] Figure 7 Comparison of time domain characteristics of speed in different tunnel scenarios in Example 1 of the present invention;

[0038] Figure 8 Comparison of time domain characteristics of heart rate and its growth rate in different tunnel scenarios in Example 1 of the present invention;

[0039] Fig. 9 The driving behavior spectrum of Example 1 of the present invention;

[0040] Fig.10 This is an example of a basic square matrix unit of the behavior spectrum of Embodiment 1 of the present invention;

[0041] Fig.11 This is an example of a behavior spectrum calculation unit according to Embodiment 1 of the present invention;

[0042] Fig.12 Schematic diagram of the behavior spectrum matrix sliding window of Example 1 of the present invention. DETAILED DESCRIPTION

[0043] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0045] Embodiment 1:

[0046] This embodiment provides a method for constructing a tunnel section driving behavior spectrum, including:

[0047] Acquire the psychophysiological data and driving data of the driver in the tunnel section; the psychophysiological data at least includes an electroencephalogram signal, an eye movement trajectory signal and a heart rate signal; the driving data at least includes tunnel section environmental variable information, steering wheel angle data and driving speed;

[0048] Preprocessing the cardiophysiological data and the driving data, filtering out invalid data in the cardiophysiological data, and correcting the driving data;

[0049] Extract features from preprocessed cardiophysiological data and driving data;

[0050] Based on the feature-extracted data, a driving behavior chain is established according to the development pattern in the time dimension;

[0051] The driving behavior chain is expanded and extended in different dimensions to construct a tunnel section driving behavior spectrum.

[0052] This embodiment comprehensively considers the data in the driving process and the relationship between the data, and constructs a driving behavior chain based on physiological data and driving data; then the driving behavior chain is expanded and extended in different dimensions to construct a driving behavior spectrum for a tunnel section, thereby realizing the research on the driving behavior spectrum for tunnels; on the basis of establishing the driving behavior chain, the driving behavior spectrum obtained by expanding and extending the driving behavior chain in different dimensions can comprehensively and accurately reflect the driving behavior.

[0053] The equipment is used to collect the driver's physiological data and driving data for data preprocessing; feature extraction and analysis are performed on driving behavior characteristics; based on the driving process unit decomposition and behavior chain theory, a driving behavior chain is established according to the time dimension development form, and by expanding and extending the driving behavior chain in different dimensions, matrix and vector forms are introduced to establish the tunnel section driving behavior spectrum from macro and micro multi-dimensional indicators. The specific steps of this embodiment are:

[0054] S1. In this embodiment, it is assumed that there are multiple drivers participating in the vehicle driving test in the tunnel. The electroencephalogram, eye tracker and heart rate monitor can be used to collect various psychological and physiological indicators of each driver during the driving process; the CAN-OBD device, computer, steering wheel angle meter and the like can be used to collect the data stream generated by the vehicle during the driving process, and the illuminance meter and camera can be used to collect the environmental variable information such as the tunnel environment illumination and traffic conditions, such as Figure 2 The data acquisition test structure is shown in FIG.

[0055] During the test, the driver can be asked about his mental state based on his changes in state, rated according to the Karolinska Sleepiness Scale, and the time and place can be recorded.

[0056] S2. You can use software such as Matlab or Origin to pre-process the collected data.

[0057] S2.1. Considering that other bio-waves will cover the EEG spectrum range and hinder the recognition of EEG rhythm, it is necessary to filter out irrelevant frequencies. Specifically, the frequency range of brain rhythm waves is usually 0-30Hz, while the frequencies of human electrooculogram, electromyography and electrocardiogram signals can be as high as 100Hz. Therefore, other bio-waves will cover the EEG spectrum range and hinder the recognition of EEG rhythm, and it is necessary to filter out irrelevant frequencies. Specifically, first perform bandpass filtering to eliminate interference from high-frequency bio-electric signals. Then perform independent component analysis on the processed EEG to separate the source signal from the mixed signal, which is used to remove interference signals such as electrooculogram and electromyography remaining in the EEG signal.

[0058] S2.2. Use the Laida criterion to filter out invalid data from the data obtained by the eye tracker. The basic idea of ​​Laida is as follows:

[0059]

[0060] When a set of data observations x i The average value of the group When the difference is greater than 3 times the standard deviation σ, the observation is considered an outlier and should be discarded.

[0061] S2.3. Use the tunnel driving timestamp to mark the driver's cross-sectional speed inside and outside the tunnel, so as to understand the speed change pattern of the driver inside and outside the tunnel.

[0062] S2.4. Perform baseline correction on the steering wheel angle data. If within a preset time, for example, within 4 seconds, the steering wheel angle has the same steering direction and the sum of the absolute values ​​of the steering amplitudes is greater than or equal to a preset angle, which can be set to 20°, it is considered that the angle data contains curve linear correction information. Then find the steering wheel angle data of the curve correction section, calculate the average and subtract it, and obtain the angle data that eliminates the linear influence.

[0063] S3. Extract features and conduct preliminary data analysis on steering wheel steering, eye movement, heart rate and other data.

[0064] Specifically, the steering wheel angle entropy is used as the characteristic variable to quantitatively describe the changes in the driver's steering wheel angle in the time series, and the driver's lateral handling stability is quantitatively analyzed; according to the driver's visual characteristics, the pupil area, gaze characteristics, scanning characteristics and blinking characteristics are extracted; the heart rate is used as a characterization indicator of the driver's mental state, and the heart rate growth rate is selected as a research indicator of psychological characteristics; the electroencephalography (EEG) signal is processed using wavelet packet decomposition and reconstruction, and then the indicators of the increase or decrease of fast and slow wave power energy in the time-frequency domain are compared to achieve an accurate expression of the EEG rhythm signal in the power spectrum; the 85% vehicle speed is used as the reference value for speed safety judgment; the collision risk (TTC) is used to describe the relationship between the front and rear vehicles in time and space when the following behavior occurs in the tunnel group section.

[0065] S4. Based on the micro level of the driving process and the decomposition of the driving behavior process, the driving process is a time series, which takes the driver, vehicle, road condition and environment as multi-dimensional unified elements, that is, the "human-vehicle-road-environment" system elements are taken into consideration in a unified manner, and through the driving route connection, the multi-dimensional unified elements are reflected to the driving path in the order of time development, and the driving behavior chain is constructed.

[0066] like Figure 3As shown in the figure, driving behavior in tunnel sections can be generally divided into three stages: subjective perception, judgment and decision-making, and manipulation. The first stage is subjective perception, where the driver relies on sensory functions such as vision and hearing to obtain individual attributes of the tunnel environment; the second stage is judgment and decision-making, where the brain forms cognition of the obtained multi-dimensional tunnel environment information and further generates strategies; the third stage is the generated strategy that determines how the driver manipulates the vehicle. Therefore, the decomposition of the driving behavior process can be expressed as the evolution of the driver's subjective perception to driving operation, which is gradually refined as the driving scene changes.

[0067] like Figure 4 As shown in Figure 1, according to the development morphological characteristics of driving behavior in a specific travel time dimension, a series of driving behavior event vectors arranged in chronological order are defined as a driving behavior chain CDB. The driving behavior chain CDB is defined as C = [I1, I2, I3, ..., I t ,...,I T ], where I t =[D t1 ,D t2 ,M t ,R t ,W t ] T , in the definition, I1,I2,I3,...,I t ,...,I T It represents a series of driving behaviors of finite length T (T nodes in time discreteness) arranged in chronological order. The multidimensional vector formed by their connection in the time dimension is the driving behavior chain C.

[0068] I t represents the driving event corresponding to the tth node of the driving behavior chain, such as Figure 4 The star in the figure; and through the combined vector [D t1 ,D t2 ,M t ,R t ,W t ] T The combined vector at this moment is one of the constituent elements of the driving behavior chain vector: Element = [D t1 ,D t2 ,M t ,R t ,W t ] T , where D t1 It represents the driver's subjective perception and decision-making process. The subjective perception process is that the driver relies on the senses to obtain the individual attributes of the tunnel environment, and the decision-making process is that the brain further generates strategies based on the cognition of the obtained multi-dimensional environment information of the tunnel; D t2Represents the driver's psychological and physiological state vector, such as eye movement, EEG and heart rate; M t Represents the vehicle handling index vector, such as speed, acceleration and steering wheel angle; R t Represents road scene variables, such as tunnels and open roads; W t Represents climate environment variables, such as sunny, cloudy and rainy days.

[0069] The driving behavior chain reflects the performance and changes of the driver's driving behavior in a limited time dimension in a specific scenario from a micro perspective. Table 1 shows a partial example of the driver's behavior chain in a tunnel.

[0070] Table 1. Partial example of driver behavior chain in tunnel.

[0071]

[0072] S5. Select the turning, speed and heart rate characteristics of the behavior chain in tunnel scenarios of different lengths for time domain analysis.

[0073] S5.1. Use the steering angular velocity feature to compare the driver's time domain features during steering, such as Figure 5 As shown in the figure, in tunnel scenes of different lengths, the change trends of the mean, root mean square value and standard deviation of the steering angular velocity are basically the same, showing a trend of first decreasing and then increasing. This shows that as the length of the tunnel increases, the steering force applied by the driver when driving in the tunnel tends to decrease, so the steering angular velocity shows a trend of decreasing with the increase of the tunnel length.

[0074] S5.2, such as Figure 6 and Figure 7 As shown in the figure, the speed index and acceleration index are used to compare the time domain characteristics of the driver in terms of speed index. In tunnel scenarios of different lengths, the mean values ​​of the driver's positive and negative acceleration do not show differences, so the influence of tunnel length on the driver's acceleration and deceleration behavior is not significant; in terms of speed index, the speed difference between the short tunnel and the other three tunnels is large, and the speed characteristic changes between the other three tunnels tend to be consistent, indicating that when driving in a non-short tunnel, the driver is forced to drive at a slightly lower speed than the short tunnel section due to safety concerns, but from the perspective of standard deviation change characteristics, as the tunnel length increases, the speed discreteness increases in a wave-like manner.

[0075] S5.3. Use the heart rate index and the heart rate growth rate index to compare the driver's time domain characteristics in terms of heart rate, such as Figure 8 As shown in the figure, with the increase of tunnel length, the mean heart rate and the mean heart rate growth rate of the driver show an upward trend, and the heart rate growth rate increases significantly, indicating that with the increase of tunnel length, the driver's heart rate fluctuates more violently, which makes the driver's heart rate more discrete, more prone to psychological tension, and increases driving load.

[0076] S6. Use the Fourier transform algorithm to perform frequency domain analysis on the turning, speed and heart rate characteristics of the behavior chain in tunnel scenarios of different lengths.

[0077] As an important tool for signal analysis, the core connotation of Fourier transform is that any function can be represented by the sum of a set of sine and cosine functions. Based on Fourier transform, the fast Fourier transform algorithm is derived, which significantly reduces the difficulty and the amount of calculation.

[0078] Fourier Transform is an important tool for signal analysis. Its core meaning is that any function can be represented by the sum of a set of sine and cosine functions.

[0079] A composite periodic function is composed of multiple frequencies superimposed on each other, and can be processed by Fourier series expansion. This is because there is a certain simple relationship between the Fourier transform of a finite periodic function and the Fourier series after its period extension. It is known that the function f(t) has a finite period, and its sampling points are {t n |n=0,1,…,N-1}, then the Fourier series F(ω) can be approximately transformed into the following form:

[0080]

[0081] Among them, f d (t n )=f(t n )w(t n ); w(t n ) is the definite integral at t n The weighting coefficient at the point is determined by the distribution of the sampling points.

[0082] If the sampling frequency point set is {f m , m=0,1,…,M-1}, then ω m =2πf m , formula (1) can be transformed into a generalized discrete Fourier transform:

[0083]

[0084] The data collected in the experiment usually has a fixed sampling frequency, so it can be called uniform sampling, then:

[0085]

[0086] Based on the above Fourier transform, the Fast Fourier Transform (FFT) algorithm was derived and has been widely used in spectrum analysis. The principle is as follows:

[0087] The number of sampling points of FFT is determined by the calculation process, usually N = 2k That is, the number of sampling points is an integer power of 2. Let the integer binary conversion of m and n in formula (3) be as follows:

[0088] n=n k-1 2 k-1 +n k-2 2 k-2 +...+n12 1 +n0 (4)

[0089] m=m k-1 2 k-1 +m k-2 2 k-2 +...+m12 1 +m0 (5)

[0090] Then formula (3) is replaced by:

[0091]

[0092] in,

[0093]

[0094] Further decomposition yields:

[0095] f m (n0,n1,...,n k-1 )=F(m k-1 ,m k-2 ,...,m1,m0) (7)

[0096] It can be seen that compared with the discrete Fourier transform, which needs to solve complicated integral approximation estimates and complex function transformation relationships, the FFT algorithm is significantly less difficult and the amount of calculation is significantly reduced. It is more widely used in the field of modern mathematical analysis, especially in large sample spaces or massive experimental data, with higher computing efficiency.

[0097] S6.1. The driver's angular velocity is transformed into the frequency domain through Fourier transform for spectrum analysis. During the driver's transition from a short tunnel to an extra-long tunnel, the amplitude of the high-frequency component shows a slowly increasing trend, while the amplitude of the low-frequency component decreases exponentially. This indicates that the longer the tunnel, the greater the probability that the driver drives at a fixed steering angular velocity, that is, the driver's participation decreases.

[0098] S6.2. Fourier transform is used to transform the driver's speed into the frequency domain for spectrum analysis. During the transition from a short tunnel to an extra-long tunnel, the speed signal frequency generally moves closer to the low-frequency component, indicating that as the tunnel length increases, the uncertainty of the driving speed increases, that is, the driver's ability to maintain speed gradually deteriorates.

[0099] S6.3. Use Fourier transform to transform the driver's heart rate signal data into the frequency domain for spectrum analysis. As the tunnel length increases, the driver's heart rate signal frequency gradually approaches the zero-frequency component, indicating that the more complex the heart rate signal is, the more signal components there are, the greater the change, that is, the greater the discreteness of the heart rate.

[0100] S6.4. In order to more clearly study the characteristics of the driver's driving behavior in a specific scene or area, a driver behavior spectrum theory system is established. The behavior spectrum layer is the target layer. From the target layer, the object layer is constructed with the "human-vehicle-road-environment" system as the research object, and then the "human-vehicle-road-environment" system is expanded to construct the main indicator layer. The object layer is modularized from the macro and micro levels to obtain an intuitive behavior spectrum description. Then, the main indicator layer is deeply mined to construct the sub-indicator layer, and the characteristic expression of each main indicator of the driving system is deeply mined through feature extraction.

[0101] S7. The driving behavior spectrum reflects the basic data system of the "human-vehicle-road-environment" system. It is an expression of multi-dimensional spatiotemporal driving data. By extracting certain characteristics of the driver and vehicle in the spectrum, the corresponding driving behavior can be analyzed.

[0102] S8. Based on the construction of the driving behavior chain, the scope of the research scene and the time scale are expanded, and the driving behavior spectrum is constructed with the square array as the basic unit. In the process of constructing the driving behavior spectrum, the driver's driving behavior under different working conditions is analyzed from the micro and macro characteristics. Among them, the micro level mainly reflects the driver's behavior that is not easy to detect when performing driving tasks, such as changes in heart / brain waves, eye movement indicators, brake pedal / accelerator pedal pedaling force, etc.; the macro level is the change of the vehicle's external motion state and the change of the relative state of the vehicle when driving, such as headway, collision time, and vehicle distance.

[0103] The matrix can record data of different dimensions. According to the relationship between vectors and matrices, a single vector is regarded as a one-dimensional matrix, and the combination of multiple vectors constitutes a multi-dimensional matrix. Therefore, the driving behavior chain with a lower time dimension constitutes a behavior spectrum after multiple superpositions and combinations.

[0104] The driving behavior spectrum uses a matrix as a specific form of expression. The row vector is the information expansion of the object layer and the indicator layer, which can represent the spatiotemporal information or state information of the driver that changes over time in a certain dimension, such as: specific values ​​or category labels such as psychological / physiological indicators, road types, vehicle position and motion status, and weather conditions. The column dimension is the time index of the matrix, which can reflect the values ​​of indicators of different dimensions at a certain moment. Therefore, all the values ​​of the row vector in the driving behavior time series constitute the driving behavior spectrum. An example of the driving behavior spectrum element at the i-th moment is as follows:

[0105] Ethogrami =[Dri i Veh i Sce i Env i ] T

[0106] Among them, Ethogram i It is the one-dimensional vector element of the driver behavior spectrum at the i-th moment; Dri represents the driver sub-index layer vector; Veh represents the vehicle motion state sub-index layer vector; Sce represents the road scene sub-index layer vector; Env represents the operating environment sub-index layer vector.

[0107] Table 2 Simple notation of driving behavior spectrum in a certain period of time.

[0108]

[0109]

[0110] Table 2 reflects the basic composition of the driving behavior spectrum, which comprehensively covers the characteristic indicators of driver behavior and their changes in time series. i Cog stands for Perception Behavior i stands for cognitive behavior; Jud i represents decision-making behavior; Epri represents brain wave power ratio; Hra i stands for heart rate; Hrgr i Represents the heart rate growth rate; Fixd i represents the fixation duration; Fixf i represents the fixation frequency; Sacd i Sacf represents the duration of the saccade; i Represents the scanning frequency; Blid i Represents the duration of a blink; i represents the blinking frequency; v i represents vehicle speed; a i Represents acceleration; Sped i Represents the speed difference; Stewa i Represents the steering wheel angle; Steas i Represents the steering angular velocity; THW i Stands for Time Headway; TTC i Represents the collision time; Slim i Represents the speed limit value; loca i Represents lane position; Roali i Represents the road alignment; Free i stands for Free Flow; Cro i represents crowded flow; Blo iRepresents blocked flow; Illu i Represents illumination; Sun i Represents sunny weather; Clo i Represents cloudy weather; Rain i Represents rainy day.

[0111] S9. Considering that the square matrix is ​​the manifestation of the basic spectrum of the behavior spectrum, some characteristics of the spectrum are extracted from the square matrix to conduct qualitative and quantitative analysis of the driver's driving behavior.

[0112] Considering that the matrix feature is in square form when being solved, the basic square matrix unit and calculation unit of the driving behavior spectrum are extracted, such as Fig.10 As shown in the figure, the basic matrix unit contains all characteristic indicators. At the same time, in order to study driving behavior separately, the matrix unit dimension is reduced while ensuring the richness of indicators. Based on the basic matrix unit, the variables related to driving behavior are extracted to form calculation units, as shown in Fig.11 shown.

[0113] Considering that although the behavior spectrum numerical matrix contains characteristic indicators of multiple dimensions, these characteristic indicators do not belong to the same category in essence and have great differences in attributes, the 14 characteristic indicators are divided into 6 groups according to the source category of the data stream, as shown in Table 3. At the same time, in order to take into account the maximum time resolution and the number of unit matrix eigenvalues ​​in the characteristic calculation, the calculation is performed in units of 3D square matrix units. If the indicator dimension is less than 3 dimensions, the remaining dimensions are padded with zeros. In the process of eigenvalue calculation, a matrix sliding window is first established with each group of 3D square matrix units. Each window represents 3s data on different dimensions of the group. Then, the square matrix unit eigenvalue is iteratively calculated with a step size of 1s, as shown in Table 3. Fig.12 shown.

[0114] Table 3: Grouping of behavioral spectrum eigenvalue matrix.

[0115]

[0116] Taking the spectrum radius as the system steady-state indicator and combining the intrinsic data characteristics of characteristic indicators of different groups, the driving performance of different groups is qualitatively analyzed.

[0117] Embodiment 2:

[0118] This embodiment provides a tunnel section driving behavior spectrum construction system, including:

[0119] The data acquisition module is configured to: acquire the psychological and physiological data and driving data of the driver in the tunnel section;

[0120] A preprocessing module is configured to: preprocess the cardiophysiological data and the driving data, filter out invalid data in the cardiophysiological data, and correct the driving data;

[0121] The feature extraction module is configured to: extract features from the pre-processed cardiophysiological data and driving data;

[0122] The driving behavior chain establishment module is configured to: establish a driving behavior chain according to the development form of the time dimension based on the feature extracted data;

[0123] The driving behavior spectrum building module is configured to: expand and extend the driving behavior chain in different dimensions to build a tunnel section driving behavior spectrum.

[0124] The working method of the system is the same as the method for constructing the tunnel section driving behavior spectrum in Example 1, and will not be repeated here.

[0125] Embodiment 3:

[0126] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method for constructing a tunnel section driving behavior spectrum described in Example 1 are implemented.

[0127] Embodiment 4:

[0128] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the method for constructing a tunnel section driving behavior spectrum described in Example 1 are implemented.

[0129] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A method for constructing a driving behavior spectrum in a tunnel section, characterized in that: include: Obtain the psychological and physiological data and driving data of drivers in tunnel sections; Preprocessing the cardiophysiological data and the driving data, filtering out invalid data in the cardiophysiological data, and correcting the driving data; Extract features from preprocessed cardiophysiological data and driving data; Based on the feature extracted data, a driving behavior chain is established according to the time dimension; The driving behavior chain is expanded and extended in different dimensions to construct a driving behavior spectrum for tunnel sections; the driver, vehicle, road condition and environment are taken as multi-dimensional unified elements; By connecting driving routes in series, multi-dimensional unified elements are reflected to the driving path in the order of time development, thus building a driving behavior chain; The behavior chain turning, speed and heart rate characteristics in tunnel scenarios of different lengths are selected for time domain analysis; Conduct frequency domain analysis on the turning, speed and heart rate characteristics of the behavior chain in tunnels of different lengths; Based on the construction of the driving behavior chain, the scope of the research scenario and the time scale are expanded, and the driving behavior spectrum is constructed with the square matrix as the basic unit. The driving behavior chain with a lower time dimension is superimposed and combined multiple times to form a behavior spectrum. The row vector of the driving behavior spectrum is the information expansion of the object layer and the indicator layer, representing the spatiotemporal information or state information of the driver that changes over time in a certain dimension. The column dimension is the time index of the matrix, which can reflect the values ​​of indicators of different dimensions at a certain moment. All the values ​​of the row vector in the driving behavior time series constitute the driving behavior spectrum; the spectral features are extracted from the square matrix to conduct qualitative and quantitative analysis of the driver's driving behavior.

2. The method for constructing a tunnel section driving behavior spectrum according to claim 1, characterized in that: The psychophysiological data includes electroencephalogram signals, eye movement trajectory signals and heart rate signals; the driving data includes tunnel section environmental variable information, steering wheel angle data and driving speed.

3. The method for constructing a tunnel section driving behavior spectrum according to claim 2, characterized in that: The irrelevant frequency data in the EEG signal is filtered out, and the invalid data in the eye movement trajectory signal is filtered out using the Laida criterion; the driver's cross-sectional speed inside and outside the tunnel is marked by the driver's tunnel driving timestamp to obtain the speed change pattern of the driver inside and outside the tunnel; the steering wheel angle data is baseline corrected.

4. The method for constructing a tunnel section driving behavior spectrum according to claim 2, characterized in that: The EEG signals are processed using wavelet packet decomposition and reconstruction, and then the expression of the EEG rhythm signal on the power spectrum is obtained by comparing the indicators of the increase or decrease of the fast wave and slow wave power energy in the time-frequency domain; the pupil area, gaze characteristics, scanning characteristics and blinking characteristics are extracted for the eye movement trajectory signal; the heart rate is used as a characterization indicator of the driver's mental state, and the heart rate growth rate is selected as a psychological characteristic research indicator; the steering wheel angle entropy is used as a characteristic variable to quantitatively describe the changes in the driver's steering wheel angle in the time series to quantitatively analyze the driver's lateral handling stability; Collision risk is used to describe the temporal and spatial relationship between the front and rear vehicles when following behavior occurs in tunnel sections.

5. The method for constructing a tunnel section driving behavior spectrum according to claim 1, characterized in that: The Fourier transform algorithm is used to perform frequency domain analysis on the data in the driving behavior chain in tunnel scenarios of different lengths; The longer the tunnel, the greater the probability that the driver drives at a fixed steering angle and speed, and the driver's participation decreases; as the tunnel length increases, the uncertainty of driving speed increases, and the driver's ability to maintain speed becomes worse; as the tunnel length increases, the heart rate discreteness increases.

6. A tunnel section driving behavior spectrum construction system, characterized in that: include: The data acquisition module is configured to: acquire the psychological and physiological data and driving data of the driver in the tunnel section; A preprocessing module is configured to: preprocess the cardiophysiological data and the driving data, filter out invalid data in the cardiophysiological data, and correct the driving data; The feature extraction module is configured to: extract features from the pre-processed cardiophysiological data and driving data; The driving behavior chain establishment module is configured to: establish a driving behavior chain according to the time dimension based on the feature extracted data; The driving behavior spectrum building module is configured to: expand and extend the driving behavior chain in different dimensions to build a tunnel section driving behavior spectrum; take the driver, vehicle, road condition and environment as multi-dimensional unified elements; By connecting driving routes in series, multi-dimensional unified elements are reflected to the driving path in the order of time development, thus building a driving behavior chain; The behavior chain turning, speed and heart rate characteristics in tunnel scenarios of different lengths are selected for time domain analysis; Conduct frequency domain analysis on the turning, speed and heart rate characteristics of the behavior chain in tunnels of different lengths; Based on the construction of the driving behavior chain, the scope of the research scenario and the time scale are expanded, and the driving behavior spectrum is constructed with the square matrix as the basic unit. The driving behavior chain with a lower time dimension is superimposed and combined multiple times to form a behavior spectrum. The row vector of the driving behavior spectrum is the information expansion of the object layer and the indicator layer, representing the spatiotemporal information or state information of the driver that changes over time in a certain dimension. The column dimension is the time index of the matrix, which can reflect the values ​​of indicators of different dimensions at a certain moment. All the values ​​of the row vector in the driving behavior time series constitute the driving behavior spectrum; the spectral features are extracted from the square matrix to conduct qualitative and quantitative analysis of the driver's driving behavior.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method for constructing a tunnel section driving behavior spectrum as described in any one of claims 1 to 5 are implemented.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method for constructing a tunnel section driving behavior spectrum as described in any one of claims 1 to 5 are implemented.

Images