An experimental device and method based on driver dissatisfaction threshold collection

Abstract

The present invention discloses an experimental device and method based on the collection of driver dissatisfaction thresholds, the device comprising a fixed cockpit, a circular screen, a projector, a speaker, a host computer, a slave computer, a steering wheel angle sensor and a brake pedal pressure sensor; the method comprises the following steps: preventing the driver from having a learning effect; calculating the following vehicle time; performing a normal distribution test on the calculated following vehicle time; and calculating the driver's dissatisfaction. The existing experimental method for collecting driver dissatisfaction thresholds has an imperfect experimental process and ignores the learning effect of the driving simulator, resulting in the inability to calculate the effective following vehicle time from the experimental data, and thus the inability to calculate the driver's dissatisfaction threshold. The present invention sets a random interference vehicle in front of the main vehicle, and the interference vehicle will accelerate in a short time after the main vehicle encounters it, so as to avoid the learning effect as much as possible and ensure that the data calculates the effective following vehicle time.

Description

Technical Field

[0001] The present invention relates to a lane-changing decision technology for an intelligent vehicle, and in particular to an experimental device and method based on driver dissatisfaction threshold collection. Background Art

[0002] Lane changing behavior requires the driver to comprehensively consider the surrounding environment information to make the right decision, so as to avoid traffic accidents during the lane changing process. In view of such safety issues, a reasonable lane changing decision model is designed to ensure the lane changing safety of the vehicle.

[0003] At present, lane-changing decision models are mainly divided into decision models based on machine learning and speed factors. The decision model based on machine learning uses data sets to train different networks to predict the driver's lane-changing intention. This model requires a large amount of data for training, has poor interpretability, and cannot obtain a high accuracy rate. The decision model based on speed factors takes into account the driver's dissatisfaction, that is, the driver's dissatisfaction with the speed, and solves the problem that the lane-changing decision model frequently generates and disappears the intention to change lanes due to environmental changes. When the driver's dissatisfaction accumulates to the set threshold, the intention to change lanes will be generated. Since the model considers historical data, there is an accumulation process, thereby avoiding the phenomenon of lane-changing intention oscillation. However, the driver's dissatisfaction threshold is not set according to the actual data collection. The existing method of collecting the dissatisfaction threshold is not perfect for the driving simulator learning effect and experimental process. Therefore, a reasonable experimental method for collecting the driver's dissatisfaction threshold is needed. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention aims to provide an experimental device and method for collecting driver dissatisfaction threshold values ​​which can improve the learning effect of a driving simulator, ensure the effectiveness of the experiment and improve the experimental process.

[0005] In order to achieve the above object, the technical solution of the present invention is as follows:

[0006] An experimental device based on driver dissatisfaction threshold collection includes a fixed cockpit, a ring screen, a projector, a sound box, a host computer, a slave computer, a steering wheel angle sensor and a brake pedal pressure sensor, wherein the fixed cockpit is provided with an instrument panel, a steering wheel and a brake pedal; the steering wheel angle sensor is installed on the steering wheel, the brake pedal pressure sensor is installed on the brake pedal, the steering wheel angle sensor and the brake pedal pressure sensor are respectively connected to the slave computer through a data line, the slave computer is connected to the host computer through a data line, and the host computer is respectively connected to the projector, the sound box and the instrument panel through the data line;

[0007] The fixed cockpit simulates a real vehicle environment, and the instrument panel in the fixed cockpit provides the driver with current vehicle speed information;

[0008] The annular screen is installed in front of the fixed cockpit and is a 120° annular screen, which provides the driver with the front information of the vehicle in the experimental environment;

[0009] The projector is installed above the fixed cockpit and projects road information onto the annular screen;

[0010] The host computer is installed outside the fixed cockpit and is a high-performance computer equipped with a corresponding display. A software platform is installed and run in the host computer, and the software platform is a control system based on MATLAB / Simulink and Carsim. The software platform calculates vehicle state parameters in real time, transmits the vehicle state parameters to the speaker and the instrument panel in the fixed cockpit, and provides the driver with road information ahead on a circular screen through a projector.

[0011] The lower computer is installed outside the fixed cockpit and is a semi-physical simulation working platform to complete the deployment and data collection of the entire experiment;

[0012] The speaker is installed outside the fixed cockpit to provide the driver with the running sound of the vehicle in the software platform;

[0013] The steering wheel angle sensor collects a steering wheel angle signal;

[0014] The brake pedal pressure sensor collects a driver's brake pedal pressure signal.

[0015] An experimental method based on driver dissatisfaction threshold collection is carried out using an experimental device based on driver dissatisfaction threshold collection, and includes the following steps:

[0016] A. Preventing drivers from experiencing learning effects

[0017] After practicing and pre-experimenting in the driving simulator, the driver will conduct the formal experiment. After the driver completes the lane change operation in the formal experiment, the experiment is terminated; in the formal experiment, data is collected in real time according to the designed sampling frequency, the driver's steering wheel angle is collected through the steering wheel angle sensor, and the driver's brake pressure is collected through the brake pedal pressure sensor; and the steering wheel angle and brake pedal pressure of each formal experiment process are saved, and the experimental data are saved as .mat on the host computer. Due to the certain gap between the driving simulator and the real car, the driver will ignore the safety in actual driving after becoming familiar with the designed working conditions, so that the collected data does not conform to the actual driving environment, that is, a learning effect occurs. In order to prevent the driver from having a learning effect after becoming familiar with the environment, the formal experiment will prevent the driver from changing lanes as soon as he encounters a vehicle, resulting in the inability to calculate the required following time for the lane change data collected by the driver, that is, the collected data is invalid, and the driver's dissatisfaction threshold cannot be calculated. Therefore, a random interference vehicle is set in front of the main vehicle, and the interference vehicle will accelerate in a short time after the main vehicle encounters it, ensuring that the following time required for the experiment is calculated for each experiment, and then the effective driver dissatisfaction threshold is calculated.

[0018] B. Calculate the time of following the vehicle

[0019] The following time t is calculated based on the collected experimental data. For the calculated following time t, the time when the driver starts to operate the brake pedal is defined as t a The time when the driver starts the lane change operation is t b , which is the moment when the steering wheel angle begins to increase. The following time t is calculated as follows:

[0020] t=t b -t a

[0021] C. Perform a normal distribution test on the calculated following time

[0022] The Shapiro-Wilk test method is used to test the normal distribution of the calculated following time. First, the null hypothesis H 0 , that is, the driver following time t of the sample size n always conforms to the normal distribution. The following time is arranged from small to large into an order statistic, and the statistic W is calculated as follows:

[0023]

[0024] Where, t k is the following time of the kth driver, is the sample average, and the calculation formula is:

[0025]

[0026] In the formula, when n is an even number, l = n / 2, and when n is an odd number, l = (n-1) / 2. Coefficient a k,n is the coefficient a related to the specific values ​​of n and k, which passes the Shapiro-Wilk test k,n Found in the table.

[0027] The criterion for this statistic is that when W ≤ Z α If Z is negative, the null hypothesis is rejected; otherwise, the null hypothesis is accepted. α It is the alpha quantile related to the sample size n and the significance level α. Its value is obtained from the alpha quantile table of the Shapiro-Wilk test based on α and n. The significance level α is the probability of making an error in the estimated following time t within a certain interval, which is taken as 0.01-0.05.

[0028] At the same time, the calculation formula for the standard deviation S of the following time t is as follows:

[0029]

[0030] If the standard deviation S is between 0.6 and 1.1, it means that there are no outliers in the collected data and the collected data is valid.

[0031] D. Calculate driver dissatisfaction

[0032] According to the driver dissatisfaction formula, when the driver dissatisfaction is greater than the driver dissatisfaction threshold S thr When , the driver has the intention to change lanes, and the driver dissatisfaction threshold is calculated according to the following formula using the following time t:

[0033]

[0034]

[0035] Where S(k) is the driver dissatisfaction at the current moment; S(k-1) is the driver dissatisfaction at the previous moment; V des is the target vehicle speed; V F is the speed of the vehicle in front; the numerical value of c is used to characterize the type of vehicle in front, the c value of a small car is 100, and the c value of a large car is 200; T is the sampling time.

[0036] Compared with the prior art, the present invention has at least the following beneficial effects:

[0037] The existing experimental method for collecting driver dissatisfaction thresholds has an imperfect experimental process and ignores the learning effect of the driving simulator, resulting in the inability to calculate the effective following time from the experimental data, and thus the driver dissatisfaction threshold. The present invention sets a random interference vehicle in front of the main vehicle, and the interference vehicle will accelerate in a short time after the main vehicle encounters it, which can avoid the learning effect as much as possible and ensure that the data can calculate the effective following time. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the composition of the experimental device of the present invention.

[0039] Figure 2 is a flow chart of the experimental method of the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below in conjunction with the accompanying drawings. An experimental device based on driver dissatisfaction threshold collection is shown in FIG. Figure 1 As shown, using Figure 1 The method of conducting the experiment with the experimental device shown is as follows Figure 2 As shown, in order to achieve the effect of preventing the driver from having a learning effect as described in step A, the following process needs to be performed during the experiment:

[0041] A1. Experimental scenarios with different speed differences and target vehicle speeds are designed in the software platform. The designed experimental scenarios meet the requirements of my country's roads.

[0042] A2. Practice in driving simulator. Drivers are not familiar with the operation of driving simulators. At the same time, there are differences in driving experience between driving simulators and real cars, so each driver must practice in driving simulators before the formal experiment. First, the driver must be familiar with the steering wheel, accelerator pedal and brake pedal of the driving simulator in the established scene. The driver can skillfully control the acceleration and deceleration of the vehicle and skillfully complete the complete lane change behavior to ensure that the driver complies with daily driving habits. The driver must practice for at least 20 minutes.

[0043] A3. Pre-experiment. Be familiar with the entire experimental process and experimental purpose, ensure the effectiveness of the formal experiment, and eliminate contingency as much as possible in the formal experiment. The driver keeps driving in the lane, the driving speed is within the required speed range, and the speed of the vehicle in front is the set fixed speed. When encountering the vehicle in front, the driver brakes and slows down to ensure a safe distance, and then chooses when to change lanes according to the driver's personal wishes.

[0044] A4. Formal experiment. The static scene in the formal experiment is the same as that in the pre-experiment to prevent the static scene from affecting the driver. In the dynamic scene, in order to prevent the driver from changing lanes when encountering a vehicle after becoming familiar with the environment, which makes it impossible to collect the driver's lane change data, an interference vehicle is set in front of the main vehicle. The interference vehicle will accelerate in a short time after the main vehicle encounters it. The driver drives the vehicle in the lane at the initial moment, and accelerates to the target speed and maintains the target speed as required. Lane change is not allowed before encountering the front environment vehicle. After encountering the front and rear environment vehicles, the driver brakes and slows down to ensure a safe distance, and then changes lanes according to the driver's personal wishes. After the driver completes the lane change operation, the experiment ends and saves each steering wheel angle, main vehicle speed, main vehicle lateral displacement and brake pedal pressure. The experimental data is saved as .mat on the host computer. Each driver can collect multiple lane change data with different speed gaps or different speeds according to the needs of the experiment. However, in order to avoid the influence of the experimental working condition sequence on the experimental data, the data is not collected in a fixed order. Each driver randomly selects the working condition sequence to conduct this experiment. Since the driver's lane-changing intention is affected by environmental and driver psychological factors, and the driver has a strong learning effect in the driving simulator experiment, the driver can conduct at most three experiments for each working condition.

[0045] The present invention is not limited to this embodiment, and any equivalent concepts or changes within the technical scope disclosed by the present invention are included in the protection scope of the present invention.

Claims

1. An experimental method based on driver dissatisfaction threshold collection, using an experimental device based on driver dissatisfaction threshold collection, characterized in that: The following steps are involved: A. Preventing drivers from experiencing learning effects After practicing and conducting preliminary experiments on the driving simulator, the driver will undergo a formal experiment. After the driver completes the lane change operation in the formal experiment, the experiment is terminated. In the formal experiment, data is collected in real time according to the designed sampling frequency. The driver's steering wheel angle is collected through the steering wheel angle sensor, and the driver's brake pressure is collected through the brake pedal pressure sensor. The steering wheel angle and brake pedal pressure of each formal experiment process are saved, and the experimental data are saved as .mat on the host computer. Since there is a certain gap between the driving simulator and the real car, the driver will ignore the safety in actual driving after becoming familiar with the designed working conditions, so that the collected data does not conform to the actual driving environment, that is, a learning effect occurs. In order to prevent the driver from changing lanes as soon as they encounter a vehicle after becoming familiar with the environment, the formal experiment prevents the driver from experiencing a learning effect after becoming familiar with the environment. This causes the collected lane change data to be unable to calculate the required following time, i.e. the collected data is invalid, and thus the driver dissatisfaction threshold cannot be calculated. Therefore, a random interference vehicle is set in front of the main vehicle. The interference vehicle will accelerate within a short period of time after the main vehicle encounters it, ensuring that the required following time is calculated for each experiment, and then the effective driver dissatisfaction threshold is calculated. B. Calculate the time of following the vehicle Calculate the following time based on the collected experimental data t , the calculated following time t , the moment when the driver starts to operate the brake pedal is The driver starts the lane change operation at , that is, the moment when the steering wheel angle begins to increase; following time t The calculation formula is as follows: C. Perform a normal distribution test on the calculated following time The Shapiro-Wilk test method is used to test the normal distribution of the calculated following time. First, the null hypothesis H0 is defined, that is, the sample size collected n Driver following time t It always conforms to the normal distribution. The following time is arranged from small to large into an order statistic, and the statistic is calculated as follows: W : In the formula, For the k The following time of each driver, is the sample average, and the calculation formula is: In the formula, when n Even number l = n / 2, when n When odd l =( n -1) / 2; coefficient is with n and k The coefficients of the Shapiro-Wilk test for specific values ​​of Check in the table; The criterion for this statistic is that when If , then reject the null hypothesis, otherwise accept the null hypothesis; is related to the sample size n and significance level Related Quantile, whose value is based on , n In the Shapiro-Wilk test Quantile table is available in the table; significance level Estimated following time t The probability of making an error in a certain interval is taken as 0.01-0.05; Simultaneous following time t Standard Deviation S The calculation formula is as follows: If the standard deviation S Between 0.6 and 1.1, it means that there are no abnormal values ​​in the collected data and the collected data is valid; D. Calculate driver dissatisfaction According to the driver dissatisfaction formula, when the driver dissatisfaction is greater than the driver dissatisfaction threshold When the driver has the intention to change lanes, he uses the following time t , the driver dissatisfaction threshold is calculated according to the following formula: In the formula, is the driver’s dissatisfaction at the current moment; is the driver's dissatisfaction at the last moment; is the target vehicle speed; is the speed of the preceding vehicle; c The numerical value of c The value is 100, and the c The value is 200; T is the sampling time.

2. The experimental method based on driver dissatisfaction threshold collection according to claim 1 is characterized by: The experimental device based on driver dissatisfaction threshold collection includes a fixed cockpit, a ring screen, a projector, a speaker, a host computer, a slave computer, a steering wheel angle sensor and a brake pedal pressure sensor, wherein the fixed cockpit is provided with an instrument panel, a steering wheel and a brake pedal; the steering wheel angle sensor is installed on the steering wheel, the brake pedal pressure sensor is installed on the brake pedal, the steering wheel angle sensor and the brake pedal pressure sensor are respectively connected to the slave computer through a data line, the slave computer is connected to the host computer through a data line, and the host computer is respectively connected to the projector, the speaker and the instrument panel through a data line; The fixed cockpit simulates a real vehicle environment, and the instrument panel in the fixed cockpit provides the driver with current vehicle speed information; The annular screen is installed in front of the fixed cockpit and is a 120° annular screen, which provides the driver with the front information of the vehicle in the experimental environment; The projector is installed above the fixed cockpit and projects road information onto the annular screen; The host computer is installed outside the fixed cockpit and is a high-performance computer with a corresponding display. A software platform is installed and operated in the host computer, and the software platform is a control system based on MATLAB / Simulink and Carsim. The software platform calculates vehicle state parameters in real time, transmits the vehicle state parameters to the speaker and the instrument panel in the fixed cockpit, and provides the driver with road information ahead on a circular screen through a projector. The lower computer is installed outside the fixed cockpit and is a semi-physical simulation working platform to complete the deployment and data collection of the entire experiment; The speaker is installed outside the fixed cockpit to provide the driver with the running sound of the vehicle in the software platform; The steering wheel angle sensor collects a steering wheel angle signal; The brake pedal pressure sensor collects a driver's brake pedal pressure signal.

Images