Sound-light early warning interaction method considering driving style influence

Abstract

The invention provides a sound-light early warning interaction method considering the influence of a driving style, and the method comprises the steps: employing a K-means clustering recognition algorithm to calibrate the driving style to be a conservative type, a normal type and a radical type based on the electromyographic signal of a driver and a vehicle signal in a simulated driving environment, and carrying out the scoring of the driving safety through combining an entropy weight method and a TOPSIS method. According to the method, the credibility and the acceptability of a driver on an early warning strategy are obtained through bidirectional MANCOVA analysis, parameters of a sound-light early warning strategy are adjusted based on the credibility and the acceptability, finally experimental verification is carried out, and the result shows that the sound-light early warning strategy considering the driving style can effectively improve the driving safety of drivers of different styles.

Description

Technical Field

[0001] The present invention belongs to the field of driving behavior and driving style recognition, and specifically designs an acoustic-optical warning interaction method that takes the influence of driving style into consideration. Background Art

[0002] With the rapid development of the automobile industry, frequent traffic accidents are a serious problem for countries around the world. Existing studies have pointed out that traffic accidents are mainly caused by operational errors caused by human factors. This requires an efficient driving warning system to help drivers avoid traffic accidents. The driver's driving style is highly correlated with driving safety, and drivers with different driving styles have different experiences with the warning system. Most of the current warning strategies cannot meet the requirements of drivers with different styles, resulting in many errors when warning dangerous driving behaviors, which reduces people's trust in the warning system. Therefore, it is necessary to study driving warning strategies that take into account the impact of driving style. Summary of the invention

[0003] In order to solve the above problems, the present invention overcomes the shortcomings of the existing methods and proposes an interactive sound-light warning method that takes driving style into consideration. The method collects the driver's electromyographic signals and driving data through a driving simulation experiment, uses the K-means clustering recognition algorithm to calibrate the driving style into conservative, normal and ordinary types, combines the entropy weight method and the TOPSIS method to conduct a comprehensive safety assessment of the driving behavior of drivers with different styles, and finally proposes corresponding sound-light warning strategies for different driving styles.

[0004] In order to achieve the above technical features, the purpose of the present invention is achieved as follows: a sound-light warning interaction method considering the influence of driving style comprises the following steps: Step 1: Build a driver-in-the-loop system, let the experimenters complete the driving simulation experiment in turn, and collect their electromyography data and driving data; Step 2, preprocessing the electromyographic data and driving data collected in step 1; then slicing the preprocessed electromyographic data and driving data into driving segments according to a set time period to obtain the required slices, and extracting data from the slices to construct feature parameters for subsequent driving style recognition; Step 3, using principal component analysis to reduce the dimension of the selected characteristic parameters, retaining the principal components with characteristic values ​​greater than 1, and using K-means clustering to classify the retained characteristic values, clustering multiple driving characteristic parameters into three driving style types, which are conservative, normal, and aggressive; Step 4: Combine the entropy weight method and TOPSIS method to establish a driving behavior evaluation system to give a comprehensive safety score to the driving behaviors of drivers with different driving styles; Step 5: Propose corresponding sound-light warning strategies for drivers with different styles.

[0005] In step 1, the driver-in-the-loop system includes a real-time simulator and a host PC, wherein the real-time simulator and the host PC each run a simulink model, speedgoat runs a vehicle model, and the host PC runs a driving scenario-related model, and finally the synchronous collection of electromyographic signals and vehicle status information is realized on the real-time simulator; The simulink model in the host PC can receive the driver's steering, throttle and brake control signals. The host PC then inputs the control signals and rolling resistance information into the real-time simulator through UDP communication. The real-time simulator calculates the vehicle status based on this information and sends it to the host PC, synchronously collecting the driver's physiological signals and vehicle signals.

[0006] The driving data includes steering wheel angle, lateral and longitudinal speed, lateral and longitudinal acceleration, impact degree, accelerator pedal signal, and brake pedal signal; The electromyographic data include the surface electromyographic signals of the gastrocnemius muscle and the surface electromyographic signals of the right sternocleidomastoid muscle.

[0007] Extracting data from the slices in step 2 includes: selecting the average and standard deviation of the driver's speed, absolute value of acceleration, absolute value of impact, throttle pedal opening, and throttle pedal opening change rate, as well as the mean value of muscle activation of the gastrocnemius and right sternocleidomastoid muscle and the median value of the electromyographic amplitude to construct characteristic parameters for subsequent driving style identification.

[0008] In the step 3, a safety evaluation index is selected to determine the driver's operating behavior index and the vehicle status information index from both longitudinal and lateral aspects; Longitudinal velocity standard deviation: ; In the formula, V std is the standard deviation of longitudinal velocity; V i is the actual value of longitudinal speed; is the average longitudinal velocity; Standard deviation of longitudinal acceleration: ; In the formula, is the standard deviation of longitudinal acceleration; is the actual value of longitudinal acceleration; is the average longitudinal acceleration; Longitudinal acceleration rate of change: ; In the formula, is the rate of change of longitudinal acceleration, is the actual value of longitudinal acceleration, is the length of the sampling period; Steering wheel angle entropy: ; In the formula, SE is the steering wheel angle entropy, The probability distribution value of each interval is determined according to the frequency of the error value falling in each interval; Steering wheel angle standard deviation: ; In the formula, is the standard deviation of the steering wheel angle; is the actual value of the steering wheel angle; is the average value of the steering wheel angle in one cycle; Standard deviation of lateral acceleration: ; In the formula, is the standard deviation of lateral acceleration; is the actual value of lateral acceleration; is the average lateral acceleration; RMS value of electromyographic signal: ; In the formula, is the root mean square value of the electromyographic signal, is the time period length, is the EMG data collected during this time period; Average frequency of electromyographic signal: ; In the formula, is the average power of the EMG signal, is at frequency The power spectral density at is the frequency, is the maximum frequency of the signal.

[0009] The steps of determining the objective weight of the indicator by using the entropy method in step 4 are as follows: Step 4.1.1, calculate the j Under the indicator, i The feature weight of each object: ; Step 4.1.2, calculate the j Entropy value of the indicator: ; Step 4.1.3, calculate the coefficient of variation of the indicator: ; Step 4.1.4, determine the objective weight: .

[0010] Features m A driver, n Evaluation Metrics C n , X ij It is i The driver's j Item indicator value; The steps of using the TOPSIS method to make multi-attribute decision in step 4 are as follows: Step 4.2.1, establish a decision matrix based on the original data: (13); Step 4.2.2, normalize the data of different indicators to obtain the normalized matrix σ: (14); Step 4.2.3, assign weights to each indicator, and finally obtain the evaluation matrix V: (15); in, , ; Step 4.2.4, get the positive ideal solution and negative ideal solution : , (16); If the metric is a very large metric, then: (17); If the indicator is a tiny indicator, then: (18); Step 4.2.5, calculate the distance between each experimenter and the positive and negative ideal solutions: ; ; Step 4.2.6, calculate the score of each experimenter: ; Score The closer it is to 1, the safer the driving behavior is.

[0011] In step 5, a head-up display device is used to project signal lights of different colors to achieve light warning. As the driver's safety score changes from 1 to 0, the color of the signal light also changes from green, yellow, orange, and red. According to the color vision principle of ergonomics, a red light is used to represent a warning signal; a yellow light and an orange light are used to represent a warning signal; and a green light is used to represent normal operation.

[0012] In step 5, two-way MANCOVA analysis is used to obtain the influence of different styles of drivers on the trust and acceptance of the warning strategy, and the adjustment parameters are introduced: , , , , and determine the optimal values ​​of these parameters using the particle swarm optimization algorithm: Model initial slope and the intercept is a known fixed value, and the two external influencing factors are trust and acceptance For model parameters and To produce an impact, define the following parameter adjustment model: ; ; The particle swarm optimization algorithm is used to minimize the mean square error between the model prediction value and the actual observation value. The optimization objective function is defined as follows: ; in, is the output predicted by the model, It is the actual observed data.

[0013] The specific functional relationship of the adjustment parameters is: Conservative: ; ; Normal type: ; ; Aggressive: ; ; In this strategy, after the parameters are adjusted, if The value is less than 500, the light warning system still lights up green. If the value is greater than 500, the light warning system will still light up red; The value is less than 500, the buzzer frequency of the sound warning system is still 500Hz. If the value is greater than 2000, the buzzer frequency of the sound warning system is still 2000Hz; after adjustment, the threshold for conservative drivers to trigger the sound warning is , the threshold for normal drivers to trigger the sound warning is , the threshold for triggering the sound warning for aggressive drivers is .

[0014] The present invention has the following beneficial effects: 1. The mainstream driving style recognition method only uses a single vehicle control and vehicle operation status signal. The method of the present invention integrates the driver's physiological signals with the vehicle control and operation status, and adds an objective evaluation of the driver's own feelings. Physiological signals can not only serve as a basis for dividing different driving styles, but also reflect the differences in driving behaviors of drivers with different driving styles.

[0015] 2. Driving style is used to characterize the driver's behavioral characteristics of vehicle operation. In actual driving, the driver may show different driving styles due to subjective and objective factors such as road conditions, emotions, time cost, etc. This method first identifies the driving style in different time intervals during the entire driving process, so as to finely judge the driving behavior and accurately find the driving style with the highest driving safety.

[0016] 3. The present invention proposes corresponding sound-light warning strategies for drivers with different driving styles, which greatly improves the trust and acceptance of the warning strategies by drivers with different driving styles. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0018] Figure 1 This is a logic block diagram of the sound-light warning strategy provided by the present invention.

[0019] Figure 2 Signal flow diagram of the driver-in-the-loop system constructed for the present invention.

[0020] Figure 3 (a) (b) show the location of the gastrocnemius muscle and right sternocleidomastoid muscle.

[0021] Figure 4 (a) (b) are schematic diagrams comparing the original electromyographic signal collected by the present invention and the electromyographic signal obtained after preprocessing by the Butterworth filter.

[0022] Figure 5 This is the K-means clustering result diagram.

[0023] Figure 6It is the driving behavior safety score obtained by combining the entropy weight method and TOPSIS method.

[0024] Figure 7 It is the change of safety score before and after warning. DETAILED DESCRIPTION

[0025] The embodiments of the present invention are further described below in conjunction with the accompanying drawings.

[0026] Embodiment 1: An acoustic-optical warning interaction method considering the influence of driving style comprises the following steps: Step 1: Build Figure 2 The driver-in-the-loop system shown in the figure asks the experimenters to complete the driving simulation experiment in turn, and collects the electromyographic data and driving data of these experimenters. The driving data includes steering wheel angle, lateral and longitudinal speed, lateral and longitudinal acceleration, impact degree, accelerator pedal signal, and brake pedal signal. The electromyographic data includes the surface electromyographic signal of the gastrocnemius muscle and the surface electromyographic signal of the right sternocleidomastoid muscle. The gastrocnemius muscle and the right sternocleidomastoid muscle are as follows. Figure 3 shown.

[0027] Step 2: pre-process the electromyographic signal collected in step 1 to remove noise. Then slice the raw data collected in step 1 into 8-second segments to obtain the required driving segments, and select the mean and standard deviation of the driver's speed, absolute value of acceleration, absolute value of impact, accelerator pedal opening, and accelerator pedal opening change rate, as well as the mean value of muscle activation of gastrocnemius and right sternocleidomastoid muscle and the median value of electromyographic amplitude as parameters for driving style classification and identification.

[0028] Step 3: Use principal component analysis (PCA) to reduce the dimension of the selected feature parameters and retain the principal components with eigenvalues ​​greater than 1. Use K-means clustering method to classify the experimental data and divide the experimenters into conservative, normal, and radical types.

[0029] Step 4: Combine the entropy weight method and TOPSIS method to establish a driving behavior evaluation system to give a comprehensive safety score to the driving behaviors of drivers with different styles.

[0030] Step 5: Propose corresponding sound-light warning strategies for drivers with different styles.

[0031] Embodiment 2: In view of the shortcomings of existing driving warning methods, the present invention proposes an acoustic-optical warning interactive method that takes driving style into consideration. The specific implementation method is as follows: 1. Build as Figure 2The driver-in-the-loop system, in which the real-time simulator and the host PC each run a simulink model, speedgoat runs the vehicle model, and the host PC runs the driving scenario-related model. Finally, the synchronous acquisition of electromyographic signals and vehicle status information is realized on the real-time simulation platform. Figure 2 As shown in the figure, the simulink model in the PC can receive the driver's steering, throttle, brake and other control signals. The PC then inputs the control signals and rolling resistance and wind resistance information into the real-time simulation platform through UDP communication. The real-time simulator calculates the vehicle status based on this information and sends it to the host PC. The driver's physiological signals and vehicle signals are collected synchronously.

[0032] 2. Preprocess the collected EMG data and driving data, clean the driving data, remove unreasonable data, and ensure that all parameters are within a reasonable range. Use a Butterworth 6th-order filter to perform a 10-450Hz bandpass filter on the EMG signal, use a 50Hz notch filter to eliminate power frequency interference, and perform median smoothing. The data after preprocessing is as follows Figure 4 The preprocessed data was sliced ​​into 8-second driving segments, and the mean and standard deviation of the driver's speed, absolute value of acceleration, absolute value of impact, accelerator pedal opening, and accelerator pedal opening change rate, as well as the mean and median of the muscle activation of the gastrocnemius and right sternocleidomastoid muscle, were selected to construct feature parameters for subsequent driving style recognition.

[0033] 3. Use principal component analysis (PCA) to reduce the dimension of the constructed characteristic parameters and reduce the complexity of the data. Select the principal components with eigenvalues ​​greater than 1 and discard the principal components with eigenvalues ​​less than 1. Use K-means clustering analysis to cluster multiple driving characteristic parameters into three driving style types, namely conservative, normal and aggressive. The clustering results are as follows: Figure 5 The driving style of each driver is determined by the composition of his driving clips. For example, if the normal driving clips account for the largest proportion of all the driving clips of a driver, then the driver is a normal driver.

[0034] 4. Select safety evaluation indicators and determine the driver's operating behavior indicators and vehicle status information indicators from both longitudinal and lateral aspects; Longitudinal velocity standard deviation: ; In the formula, V std is the standard deviation of longitudinal velocity; V i is the actual value of longitudinal speed; is the average longitudinal velocity; Standard deviation of longitudinal acceleration: ; In the formula, is the standard deviation of longitudinal acceleration; is the actual value of longitudinal acceleration; is the average longitudinal acceleration; Longitudinal acceleration rate of change: ; In the formula, is the rate of change of longitudinal acceleration, is the actual value of longitudinal acceleration, is the length of the sampling period; Steering wheel angle entropy: ; In the formula, SE is the steering wheel angle entropy, The probability distribution value of each interval is determined according to the frequency of the error value falling in each interval; Steering wheel angle standard deviation: ; In the formula, is the standard deviation of the steering wheel angle; is the actual value of the steering wheel angle; is the average value of the steering wheel angle in one cycle; Standard deviation of lateral acceleration: ; In the formula, is the standard deviation of lateral acceleration; is the actual value of lateral acceleration; is the average lateral acceleration; RMS value of electromyographic signal: ; In the formula, is the root mean square value of the electromyographic signal, is the time period length, is the EMG data collected during this time period; Average frequency of electromyographic signal: ; In the formula, is the average power of the EMG signal, is at frequency The power spectral density at is the frequency, is the maximum frequency of the signal.

[0035] 5. Combine the entropy weight method and TOPSIS method to score driving safety.

[0036] The steps to determine the objective weight of indicators using the entropy method are as follows: (1) Calculate the j Under the indicator, i The feature weight of each object: ; (2) Calculate the j Entropy value of the indicator: ; (3) Calculate the coefficient of variation of the indicator: ; (4) Determine the objective weight: .

[0037] Features m A driver, n Evaluation Metrics C n , X ij It is i The driver's j Item indicator value; The steps of using the TOPSIS method to make multi-attribute decision in step 4 are as follows: (1) Establish a decision matrix based on the original data: (13); (2) Normalize the data of different indicators to obtain the normalized matrix σ: (14); (3) Assign weights to each indicator and finally obtain the evaluation matrix V: (15); in, , ; (4) Get the positive ideal solution and negative ideal solution : , (16); If the metric is a very large metric, then: (17); If the indicator is a tiny indicator, then: (18); (5) Calculate the distance between each experimenter and the positive and negative ideal solutions: ; ; (6) Calculate the score of each experimenter: ; Score The closer it is to 1, the safer the driving behavior is.

[0038] 6. Most of the previous vehicle visual warnings displayed warning information on the main console and instrument panel, while the head-up display device (HUD) can optimize the driver's attention allocation, so the present invention selects the head-up display device to project signal lights of different colors to achieve light warning. The driver's safety score changes from 1 to 0, and the color of the signal light will also change from green, yellow, orange, and red. The wavelength of green light is 500-570nm, the wavelength of yellow light is 570-590nm, the wavelength of orange light is 590-620nm, and the wavelength of red light is 620-700nm. According to the color vision principle of ergonomics, red can trigger a strong emotional response and make people highly alert, so the red light represents a warning signal; yellow and orange can remind people to be alert, but they do not necessarily need to stop immediately or take emergency measures, so the yellow light and orange light represent warning signals; green is usually related to concepts such as safety, permission, and normal operation, indicating that you can continue to move forward or everything is normal, so the green light represents normal operation.

[0039] 7. The influence of different driving styles on the trust and acceptance of warning strategies is obtained through two-way MANCOVA analysis. In order to enhance the adaptability of the model and enable it to dynamically respond to the changes in the trust and acceptance of warning methods by drivers of different driving styles counted in the previous section, the adjustment parameter ( , , , ), and determine the optimal values ​​of these parameters using a particle swarm optimization algorithm.

[0040] Model initial slope and the intercept is a known fixed value, and the two external influencing factors are trust and acceptance For model parameters and To produce an impact, define the following parameter adjustment model: ; ; The particle swarm optimization algorithm is used to minimize the mean square error between the model prediction value and the actual observation value. The optimization objective function is defined as follows: ; in, is the output predicted by the model, It is the actual observed data.

[0041] In summary, by introducing adjustment parameters and using the particle swarm optimization algorithm, we can adaptively adjust the parameters of the linear model so that it can show the best fitting effect under different driving style influence conditions. The following is the specific functional relationship of the adjustment parameters.

[0042] Conservative: ; ; Normal type: ; ; Aggressive: ; ; In this strategy, after the parameters are adjusted, if The value is less than 500, the light warning system still lights up green. If the value is greater than 500, the light warning system will still light up red. The value is less than 500, the buzzer frequency of the sound warning system is still 500Hz. If the value is greater than 2000, the buzzer frequency of the sound warning system is still 2000Hz. After adjustment, the threshold for conservative drivers to trigger the sound warning is , , the threshold for normal drivers to trigger the sound warning is , , the threshold for triggering the sound warning for aggressive drivers is , , the early warning strategy logic diagram is as follows Figure 1 The safety scores before and after the warning are shown in Figure 7 shown.

Claims

1. A sound-light warning interaction method considering the influence of driving style, characterized in that: The following steps are involved: Step 1: Build a driver-in-the-loop system, let the experimenters complete the driving simulation experiment in turn, and collect their electromyography data and driving data; Step 2, pre-processing the electromyographic data and driving data collected in step 1; Then, the pre-processed electromyographic data and driving data are sliced ​​into driving segments according to a set time period to obtain the required slices, and data is extracted from the slices to construct feature parameters for subsequent driving style recognition; Step 3, using principal component analysis to reduce the dimension of the selected characteristic parameters, retaining the principal components with characteristic values ​​greater than 1, and using K-means clustering to classify the retained characteristic values, clustering multiple driving characteristic parameters into three driving style types, which are conservative, normal, and aggressive; Step 4: Combine the entropy weight method and TOPSIS method to establish a driving behavior evaluation system to give a comprehensive safety score to the driving behaviors of drivers with different styles; Step 5: Propose corresponding sound-light warning strategies for drivers with different styles.

2. According to claim 1, a sound-light warning interaction method considering the influence of driving style is characterized in that: In step 1, the driver-in-the-loop system includes a real-time simulator and a host PC, wherein the real-time simulator and the host PC each run a simulink model, speedgoat runs a vehicle model, and the host PC runs a driving scenario-related model, and finally the synchronous collection of electromyographic signals and vehicle status information is realized on the real-time simulator; The simulink model in the host PC can receive the driver's steering, throttle and brake control signals. The host PC then inputs the control signals and rolling resistance information into the real-time simulator through UDP communication. The real-time simulator calculates the vehicle status based on this information and sends it to the host PC, synchronously collecting the driver's physiological signals and vehicle signals.

3. According to claim 2, a sound-light warning interaction method considering the influence of driving style is characterized in that: The driving data includes steering wheel angle, lateral and longitudinal speed, lateral and longitudinal acceleration, impact degree, accelerator pedal signal, and brake pedal signal; The electromyographic data include the surface electromyographic signals of the gastrocnemius muscle and the surface electromyographic signals of the right sternocleidomastoid muscle.

4. The sound-light warning interaction method considering the influence of driving style according to claim 3 is characterized in that: Extracting data from the slices in step 2 includes: selecting the average and standard deviation of the driver's speed, absolute value of acceleration, absolute value of impact, throttle pedal opening, and throttle pedal opening change rate, as well as the mean value of muscle activation of the gastrocnemius and right sternocleidomastoid muscle and the median value of the electromyographic amplitude to construct characteristic parameters for subsequent driving style identification.

5. The sound-light warning interaction method considering the influence of driving style according to claim 4, characterized in that: In the step 3, a safety evaluation index is selected to determine the driver's operating behavior index and the vehicle status information index from both longitudinal and lateral aspects; Longitudinal velocity standard deviation: ； In the formula, V std is the standard deviation of longitudinal velocity; V i is the actual value of longitudinal speed; is the average longitudinal velocity; Standard deviation of longitudinal acceleration: ； In the formula, is the standard deviation of longitudinal acceleration; is the actual value of longitudinal acceleration; is the average longitudinal acceleration; Longitudinal acceleration rate of change: ； In the formula, is the rate of change of longitudinal acceleration, is the actual value of longitudinal acceleration, is the length of the sampling period; Steering wheel angle entropy: ； In the formula, SE is the steering wheel angle entropy, The probability distribution value of each interval is determined according to the frequency of the error value falling in each interval; Steering wheel angle standard deviation: ； In the formula, is the standard deviation of the steering wheel angle; is the actual value of the steering wheel angle; is the average value of the steering wheel angle in one cycle; Standard deviation of lateral acceleration: ； In the formula, is the standard deviation of lateral acceleration; is the actual value of lateral acceleration; is the average lateral acceleration; RMS value of electromyographic signal: ； In the formula, is the root mean square value of the electromyographic signal, is the length of the time period, is the EMG data collected during this time period; Average frequency of electromyographic signal: ； In the formula, is the average power of the EMG signal, is at frequency The power spectral density at is the frequency, is the maximum frequency of the signal.

6. The sound-light warning interaction method considering the influence of driving style according to claim 5, characterized in that: The steps of determining the objective weight of the indicator by using the entropy method in step 4 are as follows: Step 4.1.1, calculate the j Under the indicator, i The feature weight of each object: ； Step 4.1.2, calculate the j Entropy value of the indicator: ； Step 4.1.3, calculate the coefficient of variation of the indicator: ； Step 4.1.4, determine the objective weight: 。 7. The sound-light warning interaction method considering the influence of driving style according to claim 6, characterized in that: Features m A driver, n Evaluation Metrics C n , X ij It is i The driver's j Item indicator value; The steps of using the TOPSIS method to make multi-attribute decision in step 4 are as follows: Step 4.2.1, establish a decision matrix based on the original data: （13）； Step 4.2.2, normalize the data of different indicators to obtain the normalized matrix σ: （14）； Step 4.2.3, assign weights to each indicator, and finally obtain the evaluation matrix V: （15）； in, , ; Step 4.2.4, get the positive ideal solution and negative ideal solution : ， （16）； If the metric is a very large metric, then: (17)； If the indicator is a tiny indicator, then: (18)； Step 4.2.5, calculate the distance between each experimenter and the positive and negative ideal solutions: ； ； Step 4.2.6, calculate the score of each experimenter: ； Score The closer it is to 1, the safer the driving behavior is.

8. The sound-light warning interaction method considering the influence of driving style according to claim 7, characterized in that: In step 5, a head-up display device is used to project signal lights of different colors to achieve light warning. As the driver's safety score changes from 1 to 0, the color of the signal light also changes from green, yellow, orange, and red. According to the color vision principle of ergonomics, a red light is used to represent a warning signal; a yellow light and an orange light are used to represent a warning signal; and a green light is used to represent normal operation.

9. The sound-light warning interaction method considering the influence of driving style according to claim 7, characterized in that: In step 5, two-way MANCOVA analysis is used to obtain the influence of different styles of drivers on the trust and acceptance of the warning strategy, and the adjustment parameters are introduced: , , , , and determine the optimal values ​​of these parameters using the particle swarm optimization algorithm: Model initial slope and the intercept is a known fixed value, and the two external influencing factors are trust and acceptance For model parameters and To produce an impact, define the following parameter adjustment model: ； ； The particle swarm optimization algorithm is used to minimize the mean square error between the model prediction value and the actual observation value. The optimization objective function is defined as follows: ； in, is the output predicted by the model, It is the actual observed data.

10. The sound-light warning interaction method considering the influence of driving style according to claim 9, characterized in that: The specific functional relationship of the adjustment parameters is: Conservative: ； ； Normal type: ； ； Aggressive: ； ； In this strategy, after the parameters are adjusted, if The value is less than 500, the light warning system still lights up green. If the value is greater than 500, the light warning system will still light up red; The value is less than 500, the buzzer frequency of the sound warning system is still 500Hz. If the value is greater than 2000, the buzzer frequency of the sound warning system is still 2000Hz; after adjustment, the threshold for conservative drivers to trigger the sound warning is , the threshold for normal drivers to trigger the sound warning is , the threshold for triggering the sound warning for aggressive drivers is .

Images