System for taking over driver by non-driving related task in man-machine co-driving mode and use method thereof

Abstract

The invention discloses a system for taking over a driver by a non-driving related task in a man-machine co-driving mode. The takeover system comprises a basic theory platform, a takeover test platform, a takeover behavior characteristic analysis platform and a takeover visual characteristic analysis platform. Wherein the basic theory platform is connected with the takeover test platform, the takeover test platform is connected with the takeover behavior characteristic analysis platform, the takeover behavior characteristic analysis platform is connected with the takeover visual characteristic analysis platform, and the basic theory platform provides a man-machine co-driving mode basic theory. According to the invention, a virtual simulation test platform and scene carrying equipment are utilized, a driver takeover and obstacle avoidance scene under a highway is built, takeover request time and mode are designed, an eye tracker and a driving simulator are utilized to record videos and data of a vehicle takeover process of a driver under a man-machine co-driving mode, and the acquired data are subjected to synchronous processing, so that the driver takeover and obstacle avoidance effect is improved. And data support is provided for analyzing the influence of non-driving related tasks on driver takeover characteristics.

Description

Technical Field

[0001] The present invention relates to the technical field of human-machine co-driving mode, and specifically to a system for taking over non-driving related tasks from a driver in the human-machine co-driving mode, and also to a method for using the system for taking over non-driving related tasks from a driver in the human-machine co-driving mode. Background Art

[0002] Nowadays, with the development and use of various driving assistance systems, driving safety, comfort and efficiency are gradually improving. It is not difficult to see that the continuous improvement of the level of automation has become an obvious trend in the development of the automotive industry. According to the different levels of vehicle intelligence and automation development, the development of intelligent vehicles is divided into six levels: no automation, driving assistance, partial automation, conditional automation, high automation and full automation. Under conditional automation or higher automation levels, the driver does not need to supervise the vehicle and the surrounding traffic environment during the entire driving process. However, due to the limitations of many conditions such as current technology and ethics, there are more cases of low automation levels, that is, partial automation or lower levels of automation. This vehicle driving mode is called human-machine co-driving mode. In this mode, the driver needs to monitor the automation system and traffic environment. When encountering system failures or complex traffic environments, it is necessary to control the vehicle in time and perform safe obstacle avoidance operations, that is, take over the vehicle.

[0003] Since the vehicle frees the driver from physical and mental work during automated driving, this will cause the driver to perform various actions that are not related to driving operations, that is, non-driving related tasks. The appearance of this task will cause the driver to lose concentration. When the automatic system cannot handle situations such as obstacles on the road, the end of the automatic driving authorization area, or a system failure, the driver is distracted in the human-machine co-driving mode and cannot monitor the surrounding environment in real time. He loses the situational awareness of the surrounding environment, which makes it very easy to cause traffic accidents due to untimely takeover. In summary, although vehicle automation has the expected positive effects, it is constrained and restricted by technology, ethics, and laws, and the human-machine co-driving mode will exist for a long time.

[0004] However, the traditional human-machine co-driving mode has the following shortcomings: As a type of intelligent driving mode, the human-machine co-driving mode will exist for a long time. When encountering system failures or complex traffic environments, the driver needs to take over control of the vehicle in a timely manner. When the driver is in non-driving related tasks, the driver's perception and reaction ability to the traffic environment will be reduced to a certain extent, and even cause traffic accidents. Summary of the invention

[0005] The first purpose of the present invention is to provide a system for taking over non-driving related tasks from the driver in the human-machine co-driving mode, so as to solve the problem that in the human-machine co-driving mode, when the driver takes over the control of the vehicle in the state of non-driving related tasks, the perception and reaction ability of the traffic environment will be reduced.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a system for taking over the driver for non-driving related tasks in a human-machine co-driving mode, including a basic theory platform, a takeover test platform, a takeover behavior characteristic analysis platform, and a takeover visual characteristic analysis platform; The basic theory platform provides basic theories for the human-machine co-driving model, explains the basic concepts involved in the human-machine co-driving model and takeover, and explains the existing non-driving related task types and contents; The takeover test platform designed four miniature non-driving related tasks and a group of no-task tasks as a control group, constructed a takeover obstacle avoidance scenario on a highway, designed the takeover request method and time, and carried out a driver takeover obstacle avoidance test under human-machine co-driving; The takeover behavior characteristics analysis platform deeply analyzes the impact of different non-driving related tasks on the driver's recovery of situational awareness and the issuance of takeover actions during takeover; The takeover visual characteristics analysis platform analyzes various indicators of the driver's gaze behavior, scanning behavior, and blinking behavior when taking over the autonomous driving vehicle, and understands the driver's focus area and visual change patterns during the takeover and obstacle avoidance process, and supplements the impact of non-driving related tasks in the takeover process on the capture and processing of information during takeover.

[0007] The basic theory platform includes a human-machine co-driving theory module and a non-driving theory module. The human-machine co-driving theory module includes a human-machine co-driving sub-module and a vehicle takeover sub-module. The human-machine co-driving theory module limits the degree of automation to partially automated driving and below, among which: the human-machine co-driving submodule partially or completely replaces the driver's visual behavior function according to the six levels of intelligent vehicle development; the vehicle takeover submodule changes from the person responsible for the driving task to the regulator of the road. When a specific situation that the automatic system cannot handle occurs, the driver regains control of the vehicle; The non-driving theory module focuses on non-driving related tasks that controllably occupy the driver's specific senses and brain load, including sending messages, playing games, watching videos, listening to music, writing emails and reading.

[0008] The takeover test platform includes test equipment module, test design module and data processing module; Among them: the test equipment module records the vehicle's driving data during the driver's takeover process; the test design module uses visual behavior tracking equipment to record the driver's dynamic visual data during the takeover process; and the data processing module analyzes the driver's takeover characteristics.

[0009] The test equipment module includes a driving simulator and an eye tracker, the test design module includes a simulation scenario design submodule, a non-driving related task design submodule and a takeover request design submodule, and the data processing module includes a data preprocessing submodule and a data analysis submodule; Among them: the driving simulator combines pure digital simulation with field tests, and is equipped with basic components equipped with real vehicles, including a driver's seat, steering wheel, tactile feedback device, accelerator pedal, brake pedal, gear shift device and instrument panel; the eye tracker is used to record the driver's visual characteristics when capturing and processing visual information; the simulation scenario design submodule simulates a highway with 6 lanes in both directions, a lane width of 3.5m and a total length of 12.5km; the non-driving related task design submodule uses the traditional driving sub-task selection method to set up a questionnaire to select the non-driving related tasks that are more likely to appear in the human-machine co-driving mode according to their occurrence The frequency and degree of danger are subjectively investigated; in the takeover request design submodule, the vehicle speed before takeover is 100km / h or 27.8m / s, and obstacles are set at 200 meters and 140 meters ahead of the radar detection distance. The time to maintain the vehicle speed until reaching the obstacle is 7s or 5s, specifically 200÷27.8≈7.1s and 140÷27.5≈5s, that is, the takeover request time is taken as the variable; the data preprocessing submodule performs data import, data interception and data grouping respectively, and the data analysis submodule performs descriptive statistical analysis, t-test analysis and correlation analysis respectively.

[0010] The takeover behavior characteristics analysis platform includes a takeover time module and a takeover quality module; Among them: the takeover time module performs takeover time analysis on the preprocessed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker; the takeover quality module performs takeover quality analysis on the preprocessed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker.

[0011] The takeover time module includes a takeover reaction time submodule, a takeover braking time submodule and a takeover steering time submodule, and the takeover quality module includes a vehicle longitudinal motion submodule and a vehicle lateral motion submodule; Among them: the takeover reaction time submodule studies the impact of different non-driving related tasks and different takeover request times on the driver's recovery of takeover consciousness and the length of time it takes to put his right hand back on the steering wheel; the takeover braking time submodule studies the impact of different non-driving related tasks and different takeover request times on the speed of the driver's deceleration action; the takeover steering time submodule studies the impact of different non-driving related tasks and different takeover request times on the speed of the driver's steering action; the vehicle longitudinal motion submodule studies the driver's ability to take over the speed control system under different conditions, and then analyzes and compares the longitudinal speeds under different conditions; the vehicle lateral motion submodule conducts a descriptive analysis of the steering wheel turning rate, vehicle lateral offset distance and lateral acceleration.

[0012] The visual characteristic analysis platform includes a gaze behavior module, a scan behavior module, and a blink behavior module; Among them: the gaze behavior module analyzes the difficulty of the driver in extracting traffic environment information from the distribution of gaze points and the effective gaze duration; the scanning behavior module studies the impact of takeover behavior in different states on the amount of traffic environment information extracted by the driver from the perspective of scanning amplitude and scanning speed; the blinking behavior module studies the driver's nervousness in taking over the vehicle in different states from the perspective of blinking frequency and blinking duration.

[0013] The second purpose of the present invention is to provide a method for using a system for taking over the driver's non-driving related tasks in a human-machine co-driving mode, so as to solve the problem that in the human-machine co-driving mode, when the driver takes over the control of the vehicle in a non-driving related task state, the driver's perception and reaction ability to the traffic environment will be reduced.

[0014] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a method for using a system for taking over the driver for non-driving related tasks in a human-machine co-driving mode, specifically: before the test, fill in basic information, calibrate and decalibrate an eye tracker, sign a written consent form and undergo adaptability training; during the test, after starting the vehicle, the vehicle maintains an automatic driving state, and after the driver's hands and feet leave the vehicle's driving system, the driver must carefully complete non-driving related tasks, and upon hearing a takeover request, resume control of the vehicle as soon as possible, and return to the middle lane after completing the obstacle avoidance operation; after the test, the driver adjusts the state and prepares to enter the next set of tests.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Use the virtual simulation test platform and scene-mounted equipment to build a driver takeover and obstacle avoidance scenario on the highway, and design the takeover request time and method. Use an eye tracker and a driving simulator to record the video and data of the driver taking over the vehicle in the human-machine co-driving mode, and synchronize the acquired data to provide data support for analyzing the impact of non-driving related tasks on the driver's takeover characteristics; 2. Starting from the takeover time, the time after the driver hears the takeover voice request is studied from the time level, mainly from the takeover reaction time, takeover braking time, and takeover steering time. Descriptive analysis and difference analysis are performed on the three indicators respectively, and the difference in the change trend and distribution characteristics of each indicator under different non-driving related tasks is compared. At the same time, the correlation between the three time indicators is compared; 3. Introduce takeover quality indicators, starting from the perspective of vehicle lateral movement and vehicle longitudinal movement during the takeover process, and analyze the differences in takeover quality during the takeover process under non-driving related tasks by comparing these two objective indicators; 4. Starting from the three types of eye movement behaviors, namely gaze, saccade and blink, we selected three types of representative indicators as the basis for analyzing the visual characteristics of the driver's takeover of non-driving related tasks in the human-machine co-driving mode. We conducted descriptive analysis and difference analysis on each indicator, compared the changing trends and distribution characteristics of each indicator under different non-driving related tasks, and compared the correlation between the five visual characteristic indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the architecture of the takeover system of the present invention; Figure 2 This is a schematic diagram of the architecture of the basic theoretical platform of the present invention; Figure 3 It is a schematic diagram of the architecture of the takeover test platform of the present invention; Figure 4 It is a schematic diagram of the architecture of the takeover behavior characteristic analysis platform of the present invention; Figure 5 It is a schematic diagram of the architecture of the visual characteristic analysis platform of the present invention; Figure 6 A model diagram of the driver taking over process in the human-machine co-driving mode of the present invention; Figure 7 The components and connection diagram of the driving simulator hardware of the present invention; Figure 8 It is a flow chart of the test implementation of the present invention; Figures 9 to 17 It is the nozzle quality diagram of the present invention; Figures 18 to 22 It is the distribution diagram of the fixation points of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0018] Example 1 like Figure 1As shown, a takeover system for a driver in a non-driving related task in a human-machine co-driving mode of the present invention includes a takeover system. The takeover system includes a basic theory platform, a takeover test platform, a takeover behavior characteristic analysis platform, and a takeover visual characteristic analysis platform. Among them, the basic theory platform is connected to the takeover test platform, the takeover test platform is connected to the takeover behavior characteristic analysis platform, and the takeover behavior characteristic analysis platform is connected to the takeover visual characteristic analysis platform.

[0019] The basic theory platform provides the basic theory of human-machine co-driving mode, explains the basic concepts involved in human-machine co-driving mode and takeover, and explains the existing non-driving related task types and contents. The takeover test platform designed four miniature non-driving related tasks and a group of no tasks as the control group, established a takeover obstacle avoidance scenario on the highway, designed the takeover request method and time, and carried out the driver takeover obstacle avoidance test under human-machine co-driving. The takeover behavior characteristics analysis platform deeply analyzes the impact of different tasks on the driver's recovery of situational awareness and taking over actions during takeover; the takeover visual characteristics analysis platform analyzes the various indicators of the driver's gaze behavior, scanning behavior, and blinking behavior when taking over the autonomous driving vehicle, and understands the driver's focus area and visual change rules in this takeover obstacle avoidance process, supplementing the impact of non-driving related tasks on the capture and processing of information during takeover.

[0020] Example 2 Different from Example 1, in Example 2 of the present invention, a system for taking over a driver's non-driving related tasks in a human-machine co-driving mode is as follows: Figure 2 As shown: The basic theory platform includes a human-machine co-driving theory module and a non-driving theory module. The human-machine co-driving theory module includes a human-machine co-driving sub-module and a vehicle takeover sub-module. The human-machine co-driving theory module limits the degree of automation to partially automated driving and below; the human-machine co-driving sub-module divides the development of intelligent vehicles into six levels, partially or completely replacing the driver's visual behavior function. The vehicle takeover sub-module changes from the person responsible for the driving task to the regulator of the road. When there are problems that the automatic system cannot handle, such as: obstacles on the road, the end of the automatic driving authorization area, or a system failure, the driver can regain control of the vehicle. The non-driving theory module controls the driver's specific senses and brain load, such as sending messages, playing games, watching videos, listening to music, writing emails, and reading.

[0021] Example 3 Different from Example 2, in Example 3 of the present invention, a system for taking over a driver's non-driving related tasks in a human-machine co-driving mode is as follows: Figure 3 As shown: The takeover test platform includes a test equipment module, a test design module and a data processing module.

[0022] In terms of function, the test equipment module records the vehicle's driving data during the driver's takeover process; the test design module uses visual behavior tracking equipment to record the driver's dynamic visual data during the takeover process; and the data processing module analyzes the driver's takeover characteristics.

[0023] In terms of content, the test equipment module includes a driving simulator and an eye tracker; the test design module includes a simulation scenario design sub-module, a non-driving related task design sub-module and a takeover request design sub-module; the data processing module includes a data preprocessing sub-module and a data analysis sub-module.

[0024] The driving simulator combines pure digital simulation with field tests to build basic components equipped with real vehicles, such as driving seats, steering wheels, tactile feedback devices, accelerator pedals, brake pedals, gear shifters and instrument panels. Eye trackers are used to record the visual characteristics of drivers when capturing and processing visual information. The simulation scenario design submodule simulates a highway with 6 lanes in both directions, 3.5m wide and 12.5km long. The non-driving-related task design submodule uses the traditional driving task selection method to set up a questionnaire and conduct a subjective survey on non-driving-related tasks that appear more frequently in the human-machine co-driving mode according to the two indicators of frequency and degree of danger. The speed of the car in the takeover request design submodule before taking over is 100km / h or 27.8m / s. Obstacles are set at 200 meters and 140 meters ahead of the radar detection distance. The time to maintain the speed until reaching the obstacle is specifically 200÷27.8≈7.1s and 140÷27.5≈5s, that is, the takeover request time is used as a variable. The data preprocessing submodules are used to import data, extract data and group data. The data analysis submodules are used to perform descriptive statistics analysis, t-test analysis and correlation analysis.

[0025] Example 4 Different from Example 3, in Example 4 of the present invention, a system for taking over a driver's non-driving related tasks in a human-machine co-driving mode is as follows: Figure 4 As shown: The takeover behavior characteristics analysis platform includes a takeover time module and a takeover quality module; The takeover time module performs takeover time analysis on the pre-processed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker, and the takeover quality module performs takeover quality analysis on the pre-processed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker. The takeover time module includes a takeover reaction time submodule, a takeover braking time submodule, and a takeover steering time submodule, and the takeover quality module includes a vehicle longitudinal motion submodule and a vehicle lateral motion submodule; The takeover reaction time submodule studies the effects of different non-driving related tasks and different takeover request times on the time it takes for the driver to regain consciousness and put his right hand back on the steering wheel; the takeover braking time submodule studies the effects of different non-driving related tasks and different takeover request times on the speed at which the driver decelerates; the takeover steering time submodule studies the effects of different non-driving related tasks and different takeover request times on the speed at which the driver turns; the vehicle longitudinal motion submodule studies the driver's ability to take over control speed under different conditions, and analyzes and compares the longitudinal speeds under different conditions; the vehicle lateral motion submodule performs descriptive analysis on the steering wheel turning rate, vehicle lateral offset distance and lateral acceleration.

[0026] Example 5 Different from Example 4, in Example 5 of the present invention, a system for taking over a driver's non-driving related tasks in a human-machine co-driving mode is as follows: Figure 5 As shown: The visual characteristic analysis platform includes a gaze behavior module, a scan behavior module, and a blink behavior module; The gaze behavior module studies the difficulty of the driver in extracting traffic environment information from the aspects of gaze point distribution and effective gaze duration. The scanning behavior module studies the impact of taking over in different states on the amount of traffic environment information extracted by the driver from the aspects of scanning amplitude and scanning speed. The blinking behavior module studies the driver's tension in taking over the vehicle in different states from the aspects of blinking frequency and blinking duration.

[0027] Example 6 In the present invention, in conjunction with the specification Figure 6 First, after receiving the takeover request, the driver begins to restore situational awareness, which is mainly manifested by turning the line of sight from non-driving related tasks to the road, and then putting hands on the steering wheel and feet on the pedals to prepare for action. Studies have shown that after receiving the takeover request, the driver's reaction time for the first time to focus on the center of the road is about 0.7s-1s, and the reaction time for the first contact with the steering wheel is about 1.2s-1.8s. Then, the driver begins to capture and process traffic environment information. Finally, the driver chooses one or two of braking, acceleration, and steering as the takeover action according to the traffic environment, and takes over the control of the vehicle until the vehicle returns to a stable state; In conjunction with the instruction manual Figure 7The simulator has the basic components of a real car, such as a driver's seat, steering wheel, tactile feedback device, accelerator pedal, brake pedal, gear shift device and instrument panel, etc. It can complete the design of simulation test scenarios under different needs, such as driving assistance system functions, driver behavior characteristics, vehicle dynamics, unmanned control, etc. With the help of this platform, the driver takes over the obstacle avoidance test in the human-machine co-driving mode. The device can collect steering wheel angle, pedal opening and closing degree, lateral speed, longitudinal speed, and real-time vehicle coordinate data when the vehicle is running; The system uses the Smarteyepro8.0 non-contact eye tracker to record the visual behavior data of the driver when taking over the vehicle, providing data support for the subsequent exploration of visual behavior characteristics. The Smarteyepro8.0 eye tracker includes a scene camera and four Basler GigE cameras. An infrared flash is also set next to the camera, which illuminates the driver's face after power is turned on, so as to capture the position of the driver's head and eyes from all angles and improve the recognition effect. At the same time, the scene camera uses the video function to record the whole process of the driver putting his right hand on the steering wheel to take over after hearing the voice requesting to take over; In conjunction with the instruction manual Figure 8 , a method for using a system for taking over a driver for non-driving related tasks in a human-machine co-driving mode of the present invention is specifically as follows: before the test, fill in basic information, calibration and calibration of an eye tracker, informed consent form and adaptive training; during the test, after starting the vehicle, the vehicle maintains an automatic driving state, and after the driver's hands and feet leave the vehicle, the driver needs to conscientiously complete non-driving related tasks, and upon hearing a takeover request, resume control of the vehicle as soon as possible, and return to the middle lane after completing the obstacle avoidance operation; after the test, the driver adjusts the state and prepares to enter the next group of sub-tests; Table 1 shows the descriptive analysis of takeover reaction time, including the driver’s pre-takeover status for five tasks and two takeover request times (TTC5 seconds and TTC7 seconds).

[0028] Table 1

[0029] When the takeover request time (TTC) is shorter, the driver needs to react faster because the obstacle is closer to the vehicle and the driver feels more urgent, so the takeover reaction time is shorter. Different non-driving related tasks have different effects on the driver's takeover reaction time. The text task leads to the longest takeover reaction time and the largest fluctuation, while the hands-free task and no task have shorter reaction times and less fluctuation.

[0030] Table 2 shows the descriptive analysis of the takeover time. When TTC is 7 seconds, the driver has more time to react, so the takeover time is longer. Compared with TTC of 5 seconds, the vehicle is farther away from the obstacle, and the driver's braking reaction is not as urgent. Under different non-driving related tasks, the text task leads to the longest takeover time and the largest fluctuation, while the no-task state has the fastest reaction and the smallest fluctuation. The takeover time of other tasks is similar.

[0031] Table 2

[0032] Table 3 shows the descriptive analysis of the takeover time. When TTC is 5 seconds, the driver's takeover time is shorter, and when TTC is 7 seconds, the takeover time is longer. Different non-driving related tasks have different effects on the takeover time. Under text and voice tasks, the driver's takeover time is longer and fluctuates more; under no task state, the takeover time is the shortest; the takeover time under handheld and hands-free tasks is between text and voice tasks and no task.

[0033] Table 3

[0034] In conjunction with the instruction manual Figures 9 to 17 Describe the quality of the pipe. Fig. 9 is a longitudinal speed analysis diagram, and Table 4 is a descriptive analysis of the longitudinal speed: Compared with the 7s TTC, the longitudinal speed of the 5s TTC is smaller. The smaller the takeover request time, the more the driver has to reduce the speed to compensate for the time to think about the obstacle avoidance operation, and the deceleration compensation phenomenon of taking over the obstacle avoidance is more obvious. For these five different non-driving related tasks, the average speed difference is not large, and there is no obvious difference in the impact on the longitudinal speed.

[0035] Table 4

[0036] Fig.10 and Fig.11 The curves of instantaneous speed change under two takeover request times are shown. Compared with TTC of 5 seconds, the curve of instantaneous speed change when TTC is 7 seconds is slower, that is, the longitudinal acceleration is smaller. Under the same TTC mode, the curves of text and voice tasks are steeper, while the curves of handheld, hands-free, and no tasks are smoother. Therefore, it can be seen that the driver has better longitudinal control ability of the vehicle under TTC7s, and the longitudinal control ability of text and voice tasks is poor.

[0037] Fig.12 The steering wheel rotation rate is 0.5°. Fig.13The steering wheel turning rate is 2.5°. Table 5 shows the descriptive analysis of the steering wheel turning rate: When the steering wheel turning rate is 0.5°, the mean and standard deviation of the two TTCs are quite different, while at 2.5°, the difference is very small. In non-driving related tasks, the steering wheel turning rate is larger in the text and voice task states, and the steering wheel turning rate is the smallest in the no-task state.

[0038] Table 5

[0039] Fig.14 is the lateral offset distance of the vehicle. Table 6 shows the descriptive analysis of the lateral offset distance of the vehicle (lateral offset distance of the vehicle - LP, standard deviation of the lateral offset distance of the vehicle - SDLP): Because the reserved takeover time is longer when TTC is 7S, the driver can better control the vehicle, so the lateral distance of the vehicle is smaller when TTC is 7S. Different non-driving related tasks have different effects on the lateral distance of the vehicle. The lateral offset distance and fluctuation are larger under text and voice tasks, smaller under no task, and the lateral offset distance of the vehicle under handheld and hands-free tasks is between text and voice tasks and no task.

[0040]

[0041] Fig.15 and Fig.16 The figure is a line graph of the vehicle's lateral position offset. It can be seen that the overall trend is to increase first and then decrease. After hearing the takeover request, the driver regains situational awareness and weakens the control of the vehicle's lateral position. When the driver regains awareness, he also regains control of the vehicle. The lateral control ability of the vehicle is the weakest in the text task state. The hands-free, handheld, and no-task states have little effect on the vehicle's lateral control ability. When the TTC is 7s, the driver has ample time to take over the vehicle, and the curve is relatively flat.

[0042] Fig.17 Table 7 shows the descriptive analysis of lateral acceleration: when TTC is 7S, the driver has more time to react, has better control over the lateral motion of the vehicle, has a smaller mean lateral acceleration, and has better control over the lateral motion of the vehicle. Different non-driving related tasks have different effects on the lateral acceleration of the vehicle. The lateral acceleration of the vehicle is larger under the text task, and smaller when holding and without tasks.

[0043] Table 7

[0044] In conjunction with the instruction manual Figures 18 to 22, the distribution of gaze points of two drivers during the driver takeover process under different non-driving related tasks is selected. The distribution of gaze points under different non-driving related tasks shows that the road front area and the left rearview mirror area are the main observation areas of the driver when taking over. Among them, the gaze points focusing on the road front area (horizontal [0°, 10°], vertical [0°, 5°]) are the most concentrated and the largest in number, followed by the left rearview mirror area (horizontal [-50°, -40°], vertical [-20°, -10°]).

Claims

1. A system for taking over non-driving related tasks from the driver in the human-machine co-driving mode, characterized by: Including basic theory platform, takeover test platform, takeover behavior characteristic analysis platform and takeover visual characteristic analysis platform; The basic theory platform provides basic theories for the human-machine co-driving model, explains the basic concepts involved in the human-machine co-driving model and takeover, and explains the existing non-driving related task types and contents; The takeover test platform designed four miniature non-driving related tasks and a group of no-task tasks as a control group, constructed a takeover obstacle avoidance scenario on a highway, designed the takeover request method and time, and carried out a driver takeover obstacle avoidance test under human-machine co-driving; The takeover behavior characteristics analysis platform deeply analyzes the impact of different non-driving related tasks on the driver's recovery of situational awareness and the issuance of takeover actions during takeover; The takeover visual characteristics analysis platform analyzes various indicators of the driver's gaze behavior, scanning behavior, and blinking behavior when taking over the autonomous driving vehicle, and understands the driver's focus area and visual change patterns during the takeover and obstacle avoidance process, and supplements the impact of non-driving related tasks in the takeover process on the capture and processing of information during takeover.

2. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 1 is characterized in that: The basic theory platform includes a human-machine co-driving theory module and a non-driving theory module, and the human-machine co-driving theory module includes a human-machine co-driving submodule and a vehicle takeover submodule; The human-machine co-driving theory module limits the degree of automation to partially automated driving and below, among which: the human-machine co-driving submodule divides the development of intelligent vehicles into six levels to partially or completely replace the driver's visual behavior function; The vehicle takeover submodule changes from the person responsible for the driving task to the regulator of the road. When a specific situation that the automatic system cannot handle occurs, the driver regains control of the vehicle; The non-driving theory module focuses on non-driving related tasks that controllably occupy the driver's specific senses and brain load, including sending messages, playing games, watching videos, listening to music, writing emails and reading.

3. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 1, characterized in that: The takeover test platform includes a test equipment module, a test design module and a data processing module; Among them: the test equipment module records the vehicle's driving data during the driver's takeover process; the test design module uses visual behavior tracking equipment to record the driver's dynamic visual data during the takeover process; and the data processing module analyzes the driver's takeover characteristics.

4. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 3 is characterized by: The test equipment module includes a driving simulator and an eye tracker, the test design module includes a simulation scenario design submodule, a non-driving related task design submodule and a takeover request design submodule, and the data processing module includes a data preprocessing submodule and a data analysis submodule; Among them: the driving simulator combines pure digital simulation with field tests, and is equipped with basic components equipped with real vehicles, including a driver's seat, steering wheel, tactile feedback device, accelerator pedal, brake pedal, gear shift device and instrument panel; the eye tracker is used to record the driver's visual characteristics when capturing and processing visual information; the simulation scenario design submodule simulates a highway with 6 lanes in both directions, a lane width of 3.5m and a total length of 12.5km; the non-driving related task design submodule uses the traditional driving sub-task selection method to set up a questionnaire to select the non-driving related tasks that are more likely to appear in the human-machine co-driving mode according to their occurrence The frequency and degree of danger are subjectively investigated; in the takeover request design submodule, the vehicle speed before takeover is 100km / h or 27.8m / s, and obstacles are set at 200 meters and 140 meters ahead of the radar detection distance. The time to maintain the vehicle speed until reaching the obstacle is 7s or 5s, specifically 200÷27.8≈7.1s and 140÷27.5≈5s, that is, the takeover request time is taken as the variable; the data preprocessing submodule performs data import, data interception and data grouping respectively, and the data analysis submodule performs descriptive statistical analysis, t-test analysis and correlation analysis respectively.

5. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 1, characterized in that: The takeover behavior characteristic analysis platform includes a takeover time module and a takeover quality module; Among them: the takeover time module performs takeover time analysis on the preprocessed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker; the takeover quality module performs takeover quality analysis on the preprocessed vehicle motion data recorded by the driving simulator and the eye movement video recorded by the eye tracker.

6. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 5, characterized in that: The takeover time module includes a takeover reaction time submodule, a takeover braking time submodule and a takeover steering time submodule, and the takeover quality module includes a vehicle longitudinal motion submodule and a vehicle lateral motion submodule; Among them: the takeover reaction time submodule studies the impact of different non-driving related tasks and different takeover request times on the driver's recovery of takeover consciousness and the length of time it takes to put his right hand back on the steering wheel; the takeover braking time submodule studies the impact of different non-driving related tasks and different takeover request times on the speed of the driver's deceleration action; the takeover steering time submodule studies the impact of different non-driving related tasks and different takeover request times on the speed of the driver's steering action; the vehicle longitudinal motion submodule studies the driver's ability to take over the speed control system under different conditions, and then analyzes and compares the longitudinal speeds under different conditions; the vehicle lateral motion submodule conducts a descriptive analysis of the steering wheel turning rate, vehicle lateral offset distance and lateral acceleration.

7. The system for taking over the driver's non-driving related tasks in the human-machine co-driving mode according to claim 1, characterized in that: The takeover visual characteristics analysis platform includes a gaze behavior module, a scan behavior module and a blink behavior module; Among them: the gaze behavior module analyzes the difficulty of the driver in extracting traffic environment information from the distribution of gaze points and the effective gaze duration; the scanning behavior module studies the impact of takeover behavior in different states on the amount of traffic environment information extracted by the driver from the perspective of scanning amplitude and scanning speed; the blinking behavior module studies the driver's nervousness in taking over the vehicle in different states from the perspective of blinking frequency and blinking duration.

8. The method for using the system for taking over the driver for non-driving related tasks in the human-machine co-driving mode according to any one of claims 1 to 7, characterized in that: Specifically: before the test, fill in basic information, calibrate and decalibrate the eye tracker, sign an informed consent form and undergo adaptive training; during the test, after starting the vehicle, the vehicle remains in automatic driving mode. After the driver's hands and feet leave the vehicle's driving system, the driver needs to carefully complete non-driving related tasks. When hearing a request to take over, the driver must resume control of the vehicle as soon as possible and return to the middle lane after completing the obstacle avoidance operation; after the test, the driver adjusts his status and prepares to enter the next set of tests.

Images