A method and system for dashboard lighting control based on risk prediction and human-machine trust.

Abstract

This invention discloses a dashboard lighting control method and system based on risk prediction and human-machine trust. The method includes the following steps: collecting and extracting driver behavior information, including the driver's eye, hand, and foot status information; collecting road traffic environment images to obtain driving environment and traffic event information; using the driver's behavior information as input to a pre-constructed trust assessment model, outputting the driver's trust assessment value for the autonomous driving system at the current moment; assessing the current autonomous driving risk based on the driving environment and traffic event information to obtain the autonomous driving system risk assessment result; calculating the fit between the trust value and the autonomous driving system risk based on the driver's trust assessment value and the autonomous driving risk assessment result at the current moment, determining the threshold range of the fit, and controlling different colored lights accordingly based on different threshold ranges. This invention can optimize the coordination of human-machine co-driving and improve traffic safety.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving, and in particular to a method and system for controlling dashboard lighting based on risk prediction and human-machine trust. Background Technology

[0002] Human-machine interaction (HMI) systems, as the medium for information exchange between drivers and vehicles, provide drivers with crucial information such as vehicle status, environmental perception, and operational intentions during driving. This information forms the foundation for drivers' risk perception and driving behavior decisions, directly impacting driving safety. Currently, HMI is evolving towards multimodal approaches, offering richer information content and transmitting information through more sensory channels and combinations. However, this also brings negative consequences such as increased cognitive load and driver distraction. Therefore, HMI significantly affects driving safety, necessitating exploration of ways to optimize HMI systems to enhance driver trust in co-driving vehicles.

[0003] Currently, the information prompts regarding autonomous driving on human-computer interaction interfaces are relatively simple, failing to enable drivers to quickly and accurately understand the risk response capabilities of autonomous driving systems. This can lead to over- or under-trust, thereby affecting safe driving. On the other hand, most intelligent connected vehicles currently use LCD instrument panels to display autonomous driving information, requiring drivers to actively identify the information. This method is not eye-catching enough and lacks intuitiveness, which will further increase the cognitive load on drivers. Summary of the Invention

[0004] The main objective of this invention is to provide a dashboard lighting control method and system based on risk prediction and human-machine trust, which enables intuitive and dynamic autonomous driving information prompts, adapts to the dynamic changes in the driver's attention, trust, and ability to predict risk events during driving, significantly reduces the cognitive load of the driver, optimizes the coordination of human-machine co-driving, and thus improves traffic safety.

[0005] The technical solution adopted in this invention is:

[0006] A method for controlling dashboard lighting based on risk prediction and human-machine trust is provided, including the following steps:

[0007] Collect and extract driver behavior information, including the driver's eye, hand, and foot status information;

[0008] Collect images of the road traffic environment and obtain information on the driving environment and traffic incidents;

[0009] The driver's behavioral information is input into a pre-built trust assessment model, and after processing by the trust assessment model, the driver's trust assessment value of the autonomous driving system at the current moment is output.

[0010] The risks of autonomous driving are assessed based on information about the driving environment and traffic incidents, resulting in risk assessment results for the autonomous driving system, including risk assessment coefficients for the technical capabilities of the autonomous driving system, risk assessment coefficients for the driving environment, and risk assessment coefficients for traffic incidents.

[0011] The degree of fit between trust and autonomous driving system risk is calculated based on the driver's trust assessment value and autonomous driving risk assessment results at the current moment, and different colored lights are controlled according to the threshold range of the degree of fit.

[0012] Following the above technical solution, the trust assessment model calculates each risk assessment coefficient using the Kalman filter algorithm.

[0013] Following the above technical solution, the trust level assessment value is specifically calculated based on the probability of the driver's hands being ready to take over, the probability of the feet being ready to take over, and the probability of the eyes being ready to take over.

[0014] Following the above technical solution, the driver's hand-ready-to-take-over state is specifically as follows: the grip force sensor installed on the steering wheel detects whether the driver is touching the steering wheel with one or both hands, resting on the steering wheel, or lightly touching the lower edge of the steering wheel. Assuming that during autonomous driving, when the driver is in a trusted state, their hands do not touch the steering wheel. Only when the driver is in a distrusted state and touches the steering wheel will the grip force sensor detect the corresponding contact signal. If the force detected by the grip force sensor is less than a certain preset threshold, it is considered that the driver has not taken over the steering wheel, but is in a state where the driver's hands are ready to take over.

[0015] Following the above technical solution, the driver's foot is ready to take over in the following specific way: by using pressure sensors installed on the accelerator and brake pedals, the driver's toes are lightly placed on the brake pedal. Assuming that during unmanned driving, when the driver is in a trusted state, their feet do not touch the accelerator or brake pedals. Only when the driver's feet touch the accelerator or brake pedals will the pressure sensor detect the corresponding contact signal. If the pressure sensor detects that the applied pressure is less than a certain preset threshold, it is considered that the driver has not taken over the accelerator or brake pedals, but is in the driver's foot ready to take over state.

[0016] Following the above technical solution, the driver's eye gaze points collected by the eye tracker are specifically divided into the dashboard, the road ahead, the vehicles interacting on the right, and other areas. The state in which the eye gaze point is in the dashboard area is the eye ready to take over.

[0017] According to the above technical solution, if it is determined from the driver's behavior information that the driver has not paid attention to the driving environment or the instrument panel for a certain period of time, it is determined that the driver is in a distracted state and the red light is controlled to flash continuously; if it is determined from the risk assessment results of the autonomous driving system that there is a dangerous event in the driving environment or the vehicle is in a dangerous driving environment, the red light is controlled to flash continuously.

[0018] Following the above technical solution, if the compatibility value is in the first range, the driver is in an abnormally over-trusted state, and the red light will be turned on.

[0019] If the fit value is in the second range, the driver is in a state of significant over-trust, and the orange light will be on.

[0020] If the fit value is in the third range, the driver is in a state of general over-trust and controls the yellow light to be on.

[0021] If the fit value is in the fourth range, the driver is in a state of slightly over-trust and controls the green light to be on.

[0022] If the fit value is in the fifth range, the driver is in a state of slight over-trust and controls the blue light to illuminate.

[0023] This invention also provides an instrument panel lighting control system based on risk prediction and human-machine trust, comprising:

[0024] The driver behavior information collection module is used to collect and extract driver behavior information, including the driver's eye, hand, and foot status information;

[0025] The traffic environment information collection module is used to collect images of the road traffic environment and obtain information on the driving environment and traffic events.

[0026] The trust assessment module is used to input the driver's behavioral information into a pre-built trust assessment model, and output the driver's trust assessment value of the autonomous driving system at the current moment after processing by the trust assessment model.

[0027] The risk assessment module is used to assess the current risks of autonomous driving based on driving environment and traffic event information, and obtain the risk assessment results of the autonomous driving system, including the risk assessment coefficient of the autonomous driving system's technical capabilities, the risk assessment coefficient of the driving environment, and the risk assessment coefficient of the traffic event.

[0028] The fit calculation module is used to calculate the fit between the trust level and the risk of the autonomous driving system based on the driver's trust level assessment value and the autonomous driving risk assessment result at the current moment.

[0029] The lighting control module is used to control different colored lights according to the threshold range of the compatibility.

[0030] The present invention also provides a computer storage medium storing a computer program that can be executed by a processor, the computer program executing the dashboard lighting control method based on risk prediction and human-machine trust as described in the above technical solution.

[0031] The beneficial effects of this invention are as follows: By collecting driver behavior information parameters and environmental parameters around the vehicle, and comprehensively considering the driver's driving psychology and risk factors in the traffic environment, this invention calculates the driver's trust level assessment value for the autonomous driving system and the risk assessment result of the autonomous driving system. In this way, it calculates the fit between the trust level and the risk of the autonomous driving system. By judging the different threshold ranges of the fit level, it controls the corresponding light prompts of different colors, thereby providing drivers with intuitive and efficient information prompts and improving the safety of autonomous driving.

[0032] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a flowchart of the dashboard lighting control method based on risk prediction and human-machine trust in an embodiment of the present invention;

[0035] Figure 2 This is an overall structural diagram of the dashboard lighting control system according to an embodiment of the present invention;

[0036] Figure 3 These are three views of the dashboard lighting control system according to an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram illustrating the use of the dashboard lighting control system according to an embodiment of the present invention;

[0038] Figure 5 This is a schematic diagram of the dashboard lighting control system based on risk prediction and human-machine trust in an embodiment of the present invention;

[0039] Reference numerals: 1. LCD instrument panel; 2. Reflector; 3. Light guide groove; 4. Flexible light guide; 5. Light shield; 6. RGB lamp head; 7. Circuit board. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0041] It should be noted that the illustrations provided in the embodiments of the present invention are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0042] In this invention, it should also be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" are used only for descriptive and distinguishing purposes and should not be construed as indicating or implying relative importance.

[0043] Furthermore, it should be noted that the features of the various embodiments of the present invention can be combined or integrated in whole or in part, and as those skilled in the art will understand, they can interact and operate in different ways. Each embodiment can be implemented independently of each other or in association with one another.

[0044] Example 1

[0045] like Figure 1 As shown, the dashboard lighting control method based on risk assessment and human-machine trust in this embodiment of the invention mainly includes the following steps:

[0046] S1. Collect and extract driver behavior information, including driver's eye, hand, and foot status information;

[0047] S2. Collect images of the road traffic environment and obtain information on the driving environment and traffic events;

[0048] S3. Input the driver's behavior information into the pre-built trust assessment model, and output the driver's trust assessment value of the autonomous driving system at the current moment after processing by the trust assessment model.

[0049] S4. Assess the current risks of autonomous driving based on driving environment and traffic event information to obtain the risk assessment results of the autonomous driving system, including the risk assessment coefficient of the autonomous driving system's technical capabilities, the risk assessment coefficient of the driving environment, and the risk assessment coefficient of the traffic event.

[0050] S5. Calculate the fit between trust level and autonomous driving system risk based on the driver's trust level assessment value and autonomous driving risk assessment results at the current moment, and control the light prompts of different colors according to the threshold range of the fit level.

[0051] The trust assessment model can calculate various risk assessment coefficients using the Kalman filter algorithm.

[0052] Trust ratings can be calculated based on the probability that a driver is ready to take over using their hands, feet, and eyes.

[0053] The driver's hand-ready-to-take-over state is specifically defined as follows: the grip force sensor installed on the steering wheel detects whether the driver is touching, resting on, or lightly touching the lower edge of the steering wheel with one or both hands. Assuming that during autonomous driving, when the driver is in a trusted state, their hands do not touch the steering wheel. Only when the driver is in a distrusted state and touches the steering wheel will the grip force sensor detect the corresponding contact signal. If the force detected by the grip force sensor is less than a certain preset threshold, it is considered that the driver has not taken over the steering wheel, but is in a state where the driver's hands are ready to take over.

[0054] The driver's foot ready-to-take-over state is specifically defined as follows: pressure sensors installed on the accelerator and brake pedals detect the driver's toes lightly touching the brake pedal. Assuming that during autonomous driving, when the driver is in a trusted state, their feet do not touch the accelerator or brake pedals. Only when the driver's feet touch the accelerator or brake pedals will the pressure sensors detect the corresponding contact signal. If the pressure detected by the pressure sensors is less than a certain preset threshold, it is considered that the driver has not taken over the accelerator or brake pedals, but is in the driver's foot ready-to-take-over state.

[0055] Furthermore, the driver's eye gaze points collected by the eye tracker are specifically divided into the dashboard, the road ahead, the vehicles interacting on the right, and other areas. The state where the eye gaze point is in the dashboard area is the eye ready to take over.

[0056] Specifically, if the driver's behavior information indicates that the driver has not paid attention to the driving environment or the dashboard for a certain period of time, it is determined that the driver is in a distracted state, and the red light is controlled to flash continuously.

[0057] If the risk assessment results of the autonomous driving system indicate that there is a dangerous event in the driving environment or that the vehicle is in a dangerous driving environment, the red light will flash continuously.

[0058] In step S5, if the fit value is in the first range, the driver is in an abnormally over-trusted state and controls the red light to turn on.

[0059] If the fit value is in the second range, the driver is in a state of significant over-trust, and the orange light will be on.

[0060] If the fit value is in the third range, the driver is in a state of general over-trust and controls the yellow light to be on.

[0061] If the fit value is in the fourth range, the driver is in a state of slightly over-trust and controls the green light to be on.

[0062] If the fit value is in the fifth range, the driver is in a state of slight over-trust and controls the blue light to illuminate.

[0063] This embodiment collects driver behavior information parameters and environmental parameters around the vehicle, and comprehensively considers the driver's driving psychology and risk factors in the traffic environment to calculate the driver's trust level assessment value of the autonomous driving system and the risk assessment result of the autonomous driving system. It then calculates the fit between trust level and autonomous driving system risk, and controls different colored lights to indicate different threshold ranges of fit level. This achieves intuitive and dynamic autonomous driving information prompts, adapting to the dynamic changes in the driver's attention, trust level, and ability to predict risk events during driving. It significantly reduces the cognitive load of the driver, optimizes the coordination of human-machine co-driving, and thus improves traffic safety.

[0064] Example 2

[0065] This embodiment is based on Embodiment 1 and provides a specific calculation example.

[0066] The specific calculation process for the driver's trust level assessment is as follows:

[0067] Establish an objective trust equation:

[0068] T(t k+1 )=f(t k ,T(t k ),L(t k ),M(t k ),N(t k ))

[0069] Wherein, T(t) k+1 ) for t k+1 The level of trust at any given moment is the true value to be estimated, T(t).k ) for t k Trust level at any given moment, L(t) k ),M(t k ), N(t) k () represents the Boolean value of the human-computer interaction scheme. For a given time t... k In a moving intelligent connected vehicle, if the human-machine interaction solution simultaneously displays the currently encountered risk events and corresponding countermeasures on the dashboard, then L(t) k )=1,M(t k )=0,N(t) k If the human-computer interaction scheme is to only display the currently encountered risk event on the dashboard, then L(t) = 0; k )=0,M(t k ) = 1, N(t) k If the human-computer interaction scheme is to only display how the system prepares to deal with risk events on the dashboard, then L(t) = 0; k )=0,M(t k )=0,N(t) k ) = 1. Therefore, for f, we have f:

[0070] [t0, t f ]×[T min ,T max ]×{0,1} 3 →[T min ,T max ]

[0071] The observed variable and the degree of confidence satisfy the following equation:

[0072]

[0073] Where h(t) k ) for t k The probability of the driver's hands being ready to take over at all times is determined by a grip force sensor installed on the steering wheel, which detects whether the driver is touching, resting on, or lightly touching the lower edge of the steering wheel with one or both hands. Assuming that during autonomous driving, when the driver is in a state of trust, their hands are typically placed on their lap and not touching the steering wheel. The grip force sensor will only detect a contact signal when the driver is not in a state of trust and touches the steering wheel. When the driver does not apply sufficient grip force, although the grip force sensor can detect the contact, the force is too small to affect the rotation of the steering wheel. That is, if the grip force sensor detects a force F... h (t k () less than a certain preset threshold F threshold If the driver has already taken over the steering wheel, it is not considered that the driver is in a state of preparing to take over.

[0074] f(t k ) for t k The probability of the driver's foot being ready to take over is determined by pressure sensors installed on the accelerator and brake pedals. These sensors detect when the driver's toes are lightly resting on the pedals. Assuming that during autonomous driving, the driver is in a state of trust, their feet typically do not touch the accelerator or brake pedals, but are placed elsewhere, such as under the pedals, under the seat, or near the door panel. The pressure sensors only detect contact when the driver's foot touches the accelerator or brake pedal. Even when the driver's foot does not apply sufficient pressure, the pressure sensors can detect contact, but because the applied pressure is too small to change the opening of the brake or accelerator pedal, the contact signal will not be detected. f (t k () less than a certain preset threshold P threshold If the driver has already taken over the accelerator or brake pedal, it is not considered that the driver is in a state of readiness to take over.

[0075] The driver's eye-tracking fixation points, collected by the eye tracker, are divided into the dashboard, the road ahead, the vehicles interacting on the right, and other areas, e(t) k ) for t k The probability that the driver's eye gaze is constantly focused on the dashboard area.

[0076] Therefore, for have

[0077]

[0078] Furthermore, assume f and The function is linear, time-invariant, and contains a random term based on individual driver differences. Furthermore, the observed variables are assumed to follow a Gaussian distribution and be independent of events and mutual observations. Considering the discreteness of the observed variables and trust levels, the LT1 system state-space equation is used to represent the driver's trust level:

[0079]

[0080] Where A, B, and H are coefficient matrices to be determined, and have w(t k The noise is process noise with covariance Q, and follows a Gaussian distribution: w(t) k )~N(0,Q); v(t) k The noise is process noise with covariance R, and follows a Gaussian distribution: v(t) k )~N(0,R).

[0081] Furthermore, based on the Kalman filter solution algorithm, the confidence level and covariance are initialized:

[0082]

[0083] in, This refers to the initial level of trust, where C is the coefficient matrix to be determined, and C = [c 11 c 21 c 31 ] T h(t0) refers to the frequency at which the driver's hand movements indicate a readiness to take over at the initial moment; f(t0) refers to the frequency at which the driver's foot movements indicate a readiness to take over at the initial moment; e(t0) refers to the frequency at which the driver's eye movements focus on the instrument panel at the initial moment. refer to The covariance. Calculate the Kalman gain:

[0084]

[0085] Where K refers to the Kalman gain; Refers to ∑T(t) k The covariance of ) ; ∑v refers to the standard deviation of the observation noise.

[0086] Further, calculate the observation error:

[0087]

[0088] in, t k The frequency with which the driver's hand gestures indicate a readiness to take over control; t k The driver's foot movements at all times indicate the frequency of preparation for taking over; t k The frequency at which the driver's eye focus is on the dashboard at all times; t k The prior estimate at time t is the estimated subject confidence level.

[0089]

[0090] Where u refers to the error between the estimated observation and the actual observation;

[0091] Furthermore, the posterior state estimate is calculated as follows:

[0092]

[0093] Wherein, T(t) k ) refers to t k The posterior estimate of the post-estimation, i.e., the optimal estimate of the subject's trust level; ∑T(t k) refers to T(t) k The covariance of ) is used to update the prior estimate for the next time step:

[0094]

[0095] in, t k+1 Prior trust level at a given moment, i.e., the estimated trust level of the driver at the current moment; refer to covariance; σ w The standard deviation of noise in the estimation process.

[0096] Furthermore, based on the driver's trust level T(t) at the current moment k+i The risk assessment module calculates various risk assessment coefficients, including system technical capability R1, driving environment R2, and traffic incident R3. The following calculation steps are used to determine the driver's trust level and the fit between the risk assessment and the driver's trust level, M(T(t)). k+1 ),R):

[0097]

[0098] Where R refers to the system's multidimensional risk assessment vector, which extends the trust level into a three-dimensional risk assessment vector:

[0099]

[0100] Construct the covariance matrix S:

[0101]

[0102] in The variance of the three risk factors, σ 12 σ 13 σ 23 The covariance refers to the variance between different risk factors, reflecting their correlation. Substituting the confidence vector T′, the system's risk assessment vector R, and the covariance matrix S into the following formula:

[0103]

[0104] Furthermore, based on the fit value, the RGB color space and flashing period of the lamp head are controlled. One control decision scheme for dashboard lighting is as follows:

[0105] If there are dangerous events in the driving environment, including but not limited to slope collapse, traffic accident, sudden braking of the vehicle in front, speeding, or overtaking from the side, the control light will turn red and flash continuously.

[0106] When a smart connected vehicle is in a dangerous driving environment, including but not limited to heavy snow, fog, heavy rain, or driving on a mountain road, the lights will be controlled to be red and flashing continuously.

[0107] The driver is in a distracted state; the driver does not pay attention to the driving environment or the instrument panel for more than 20 minutes, i.e., T distract If the time exceeds 20 minutes, the lamp head will be controlled to turn red and flash continuously;

[0108] The driver is in an abnormally overconfident state, specifically τ1≤M(T(t) k+1 If R) < τ0 and the trust level is greater than the risk assessment, and there is no dangerous event or dangerous environment, then the light head will be controlled to red. A trust level greater than the risk assessment means the driver's trust level T(t) at the current moment. k+1 The value of trust level T(t) is higher than that of the risk assessment vector R in the overall assessment. Specifically, the trust level T(t) k+1 The sum of the weighted sums of the risk assessment vectors R must be greater than T(t). k+1 )>α1·R1+α2·R2+α2·R2, where α1, α2, and α3 are the weighting coefficients for each dimension;

[0109] The driver is in a state of significant over-trust, specifically τ2≤M(T(t) k+1 If R) < τ1 and the trust level is greater than the risk assessment, and there are no dangerous events or dangerous environments, then control the light head to orange;

[0110] The driver is in a state of general over-trust, specifically τ3≤M(T(t) k+1 If R) < τ2 and the trust level is greater than the risk assessment, and there are no dangerous events or dangerous environments, then control the light bulb to yellow;

[0111] The driver is in a state of slightly excessive trust, specifically τ4≤M(T(t) k+1 If R) < τ3 and the trust level is greater than the risk assessment, and there are no dangerous events or dangerous environments, then control the light head to be green;

[0112] The driver is in a state of slight over-trust, specifically τ5≤M(T(t) k+1 If R) < τ4 and the trust level is greater than the risk assessment, and there are no dangerous events or dangerous environments, then control the light head to be blue;

[0113] In a safe driving environment, the distance d between the vehicle on the left front and the vehicle in front is... left-front ≥230m, distance d from the vehicle on the right front rig h t-front ≥230m, distance d from the vehicle on the left rear left-rear ≥230m, distance d from the vehicle on the right rear rig h t-rear ≥230m, distance d from the vehicle directly in frontfront ≥200m, distance d from the vehicle directly behind rear If the distance is ≥200m, the lamp head will be controlled to be blue and illuminated in a breathing mode at a specific frequency;

[0114] Among them, τ0, τ1, τ2, τ3, τ4, and τ5 are preset thresholds, and satisfy the following size relationship: τ0>τ1>τ2>τ3>τ4>τ5, that is, each threshold decreases sequentially.

[0115] Example 3

[0116] This embodiment describes a dashboard lighting control system based on risk prediction and human-machine trust, used to implement the method embodiment. Figure 5 As shown, it includes:

[0117] The driver behavior information collection module is used to collect and extract driver behavior information, including the driver's eye, hand, and foot status information;

[0118] The traffic environment information collection module is used to collect images of the road traffic environment and obtain information on the driving environment and traffic events.

[0119] The trust assessment module is used to input the driver's behavioral information into a pre-built trust assessment model, and output the driver's trust assessment value of the autonomous driving system at the current moment after processing by the trust assessment model.

[0120] The risk assessment module is used to assess the current risks of autonomous driving based on driving environment and traffic event information, and obtain the risk assessment results of the autonomous driving system, including the risk assessment coefficient of the autonomous driving system's technical capabilities, the risk assessment coefficient of the driving environment, and the risk assessment coefficient of the traffic event.

[0121] The fit calculation module is used to calculate the fit between the trust level and the risk of the autonomous driving system based on the driver's trust level assessment value and the autonomous driving risk assessment result at the current moment.

[0122] The lighting control module is used to control different colored lights according to the threshold range of the compatibility.

[0123] Specifically, the trust assessment module is connected to the driver behavior information collection module to process real-time received eye, hand, and foot status data to obtain the driver's trust level; the traffic environment information collection module is used to collect road traffic environment information; the risk assessment module is connected to the traffic environment information collection module to process captured road condition images, identify road driving environment safety risk sources, and assess the risk response capability of the autonomous driving system; the fit calculation module is connected to the trust assessment module and the risk assessment module to determine the degree of fit between the trust level and the risk of the autonomous driving system based on the results of the trust and risk assessments; and the lighting control module adjusts the RGB color space and flashing period of the headlights based on the output results of the fit calculation module.

[0124] The road traffic environment information includes vehicles, pedestrians, road obstacles, weather conditions, etc.

[0125] The risk assessment module includes a fusion algorithm to analyze images acquired by cameras and different types of radar sensors, enabling environmental perception and risk assessment.

[0126] Furthermore, several acquisition modules collect data from the eye tracker, the grip force sensor installed on the steering wheel, and the pressure sensor installed on the accelerator and brake pedals, and extract eye movement trajectory features and hand and foot status information.

[0127] Furthermore, the trust assessment module uses the information extracted by the driver behavior information collection module as input variables, the trust model based on the Kalman filter algorithm as the assessment method, and the objective trust level as the output result.

[0128] The lighting control module is primarily used to control the display on the LCD instrument panel, such as... Figure 2-4 As shown, the LCD instrument panel 1 mainly includes an LCD screen, a reflector 2 at the bottom of the LCD screen, a light guide groove 3 at the bottom of the LCD screen, a flexible light guide 4 inside the light guide groove 3, and a light shield 5 and an RGB lamp head 6 connected to both ends of the flexible light guide 4 respectively. The RGB lamp head 6 is connected to a circuit board 7 located at the bottom of the screen through a circuit.

[0129] Furthermore, the light shield 5 is used to prevent unnecessary light leakage, thereby improving light utilization efficiency and reducing glare. The RGB lamp head 6 is used to provide an adjustable color light source. The RGB lamp head 6 is connected to a circuit board 7 mounted at the bottom of the LCD screen via circuitry. The circuit board 7 provides power and control signals to the RGB lamp head 6. The program for the adaptation calculation module and the lighting control module is programmed onto the circuit board 7.

[0130] Furthermore, the surface of the reflector 2 is curved and made of a highly reflective material to maximize light reflection and improve illumination uniformity. The reflector 2 and the flexible light guide 4 are installed inward relative to the outer contour of the LCD screen at equal intervals.

[0131] The aforementioned dashboard lighting control system is installed in vehicles with autonomous driving capabilities, while also possessing the characteristics and functions of non-autonomous vehicles, such as the steering and braking systems required to complete driving tasks. Autonomous vehicles can operate autonomously in all road environments without driver intervention by controlling actuators such as the accelerator, brake pedal, and steering wheel. Driver intervention is required in emergencies, but the driver can also proactively take over in non-emergency situations. The dashboard lighting control system collects driver behavior information parameters and environmental parameters surrounding the vehicle, comprehensively considering the driver's psychology and risk factors in the traffic environment to provide the driver with intuitive and efficient information prompts.

[0132] The specific operating steps of the dashboard lighting control system are as follows:

[0133] (1) Data collected by the eye tracker, grip force sensor, and pressure sensor in the driver behavior information collection module are used to extract the driver's eye, hand, and foot status information, respectively, h(t k ) = 0.20, f(t) k ) = 0.135, e(t) k ) = 0.07;

[0134] (2) The camera and radar sensor in the traffic environment information collection module collect images of the road traffic environment and obtain information on the driving environment and traffic events.

[0135] (3) The trust assessment module uses the information extracted by the driver behavior information collection module as input variables, executes the driver's trust assessment model for the autonomous driving system, and outputs the driver's trust assessment value at the current moment, including:

[0136] First, obtain the Boolean value of the human-computer interaction scheme provided by the system: L(t) k ) = 1, M(t) k ) = 0, N(t) k ) = 0 indicates that the current risk event and corresponding response measures are displayed on the dashboard at the same time;

[0137] Furthermore, we select model parameters, assuming the confidence level T(t) at the previous time step. k=0.65; the influence weight of the previous trust level on the current trust level A = [0.8]; the influence coefficient matrix of sound, visual, and tactile cues on trust level B = [0.2, 0.1, 0.05]; the influence coefficient matrix of hand, foot, and eye movement frequencies on trust level C = [0.3, 0.2, 0.1]; process noise w(t) k )~N(0,Q), covariance Q=0.01, assume w(t) k ) = -0.02; Process noise v(t) k )~N(0,R), covariance R=0.01, assume v(t)~N(0,R), covariance R=0.01, assuming ... k ) = 0.005, according to the coefficient matrix C:

[0138]

[0139] Furthermore, a new level of trust is calculated based on the human-computer interaction and the trust level from the previous moment:

[0140]

[0141] Calculations show that the driver's trust level at the current moment is 0.7.

[0142] (4) Using the built-in algorithm of the risk assessment module, extract information on the technical capabilities of the autonomous driving system, the driving environment, and traffic events respectively. Use the built-in processing rules to assess the current autonomous driving risks. Assume that the risk assessment coefficient of the system technical capabilities is R1 = 0.8, the risk assessment coefficient of the driving environment is R2 = 0.5, and the risk assessment coefficient of the traffic events is R3 = 0.6.

[0143] (5) The fit calculation module takes the driver's trust assessment value at the current moment, the risk assessment coefficient R1 of the system's technical capabilities, the risk assessment coefficient R2 of the driving environment, and the risk assessment coefficient R3 of the traffic incident calculated by the trust assessment module as input variables, executes the fit calculation model, and outputs the fit between the trust level and the risk of the autonomous driving system:

[0144] First, the covariance matrix S is calculated as follows:

[0145]

[0146] The inverse matrix S of the covariance is calculated. -1 Then, by substituting these values ​​into the formula, the fit between trust level and the risk of the autonomous driving system can be calculated:

[0147]

[0148] (6) The lighting control module will adapt M(T(t) k+1The information extracted by the driver behavior information collection module and the traffic environment information collection module is used as input variables to execute the dashboard lighting control decision scheme and adjust the RGB color space and flashing period of the lamp heads:

[0149] First, select thresholds τ0 = 1, τ1 = 0.7, τ2 = 0.5, τ3 = 0.3, τ4 = 0.1, and τ5 = 0. At this point, the fitness M(T(t) is... k+1 The value of R lies between τ3 and τ2;

[0150] Furthermore, it determines whether a dangerous driving environment or situation exists. If so, it controls the RGB light head to turn red and flash continuously; if not, it further determines the driver's distraction time T. distract If the time exceeds 20 minutes, control the RGB LED head to turn red and flash continuously; if not, further determine the distance d of the vehicle to the left front. left-front The distance d between the vehicle on the right front right-front The distance d of the vehicle to the left rear left-rear The distance d between the vehicle on the right rear right-rear Is it greater than or equal to 230m, and what is the distance d of the vehicle directly in front? front The distance d of the vehicle directly behind rear If the distance is greater than or equal to 200m, control the RGB LED head to be blue and blink at a breathing frequency; otherwise, perform an adaptation check (M(T(t)). k+1 The comparison of ), R) with each threshold controls the RGB lamp head to output the corresponding color.

[0151] In this embodiment, there are three vehicles traveling at locations: 50m to the right front, 155m directly behind, and 300m directly in front. The driving environment is safe, no dangerous driving incidents have occurred, and τ3≤M(T(t) k+1 If R) < τ2, the driver is determined to be in a state of general over-trust, so the RGB light head is controlled to be yellow and constantly lit.

[0152] The above specific case implementation processes demonstrate that the present invention has the following technical advantages and beneficial effects:

[0153] (1) A mathematical model is used to quantify and measure the driver’s trust in the human-machine co-driving car, and the objective trust level of the driver can be continuously estimated.

[0154] (2) By using an adaptation calculation function, the multi-dimensional influencing factors of driver trust, system technical capability, driving environment and traffic events are comprehensively considered, and the adaptation of trust and risk assessment is quantified.

[0155] (3) By using a traffic light control scheme, drivers can be guided to adjust their trust in the autonomous driving system in order to achieve optimal human-machine collaboration and improve driving safety.

[0156] (4) By implementing a signal light installation scheme, drivers are provided with intuitive driving information prompts, reducing their cognitive load.

[0157] Example 4

[0158] This application also provides a computer-readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, server, App application store, etc., which stores a computer program. When the program is executed by a processor, it implements a corresponding function. This embodiment describes a dashboard lighting control method based on risk assessment and human-machine trust, implemented when the computer-readable storage medium is executed by a processor.

[0159] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0160] The order of the steps in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0161] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for controlling dashboard lighting based on risk prediction and human-machine trust, characterized in that, Includes the following steps: Collect and extract driver behavior information, including the driver's eye, hand, and foot status information; Collect images of the road traffic environment and obtain information on the driving environment and traffic incidents; The driver's behavioral information is input into a pre-built trust assessment model, and after processing by the trust assessment model, the driver's trust assessment value of the autonomous driving system at the current moment is output. The risks of autonomous driving are assessed based on information about the driving environment and traffic incidents, resulting in risk assessment results for the autonomous driving system, including risk assessment coefficients for the technical capabilities of the autonomous driving system, risk assessment coefficients for the driving environment, and risk assessment coefficients for traffic incidents. The degree of fit between trust and autonomous driving system risk is calculated based on the driver's trust assessment value and autonomous driving risk assessment results at the current moment, and different colored lights are controlled according to the threshold range of the degree of fit.

2. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 1, characterized in that, The trust assessment model calculates each risk assessment coefficient using the Kalman filter algorithm.

3. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 1, characterized in that, The trust rating is calculated based on the probability that the driver is ready to take over with their hands, feet, and eyes.

4. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 3, characterized in that, The driver's hand-ready-to-take-over state is specifically defined as follows: the grip force sensor installed on the steering wheel detects whether the driver is touching the steering wheel with one or both hands, resting on the steering wheel, or lightly touching the lower edge of the steering wheel. Assuming that during autonomous driving, when the driver is in a trusted state, their hands do not touch the steering wheel. Only when the driver is in a distrusted state and touches the steering wheel will the grip force sensor detect the corresponding contact signal. If the force detected by the grip force sensor is less than a certain preset threshold, it is considered that the driver has not taken over the steering wheel, but is in a state where the driver's hands are ready to take over.

5. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 3, characterized in that, The driver's foot ready-to-take-over state is specifically defined as follows: pressure sensors installed on the accelerator and brake pedals detect the driver's toes lightly touching the brake pedal. Assuming that during autonomous driving, when the driver is in a trusted state, their feet do not touch the accelerator or brake pedals. Only when the driver's feet touch the accelerator or brake pedals will the pressure sensors detect the corresponding contact signal. If the pressure detected by the pressure sensors is less than a certain preset threshold, it is considered that the driver has not taken over the accelerator or brake pedals, but is in the driver's foot ready-to-take-over state.

6. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 3, characterized in that, Specifically, the driver's eye gaze points collected by the eye tracker are divided into the dashboard, the road ahead, the vehicles interacting on the right, and other areas. The state where the eye gaze point is in the dashboard area is the eye ready to take over.

7. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 1, characterized in that, If the driver's behavior information indicates that the driver has not paid attention to the driving environment or the dashboard for a certain period of time, the system will determine that the driver is in a distracted state and control the red light to flash continuously. If the risk assessment results of the autonomous driving system indicate that there is a dangerous event in the driving environment or that the vehicle is in a dangerous driving environment, the system will control the red light to flash continuously.

8. The dashboard lighting control method based on risk prediction and human-machine trust as described in claim 1, characterized in that, If the compatibility value is in the first range, the driver is in an abnormally over-trusted state and controls the red light to be on. If the fit value is in the second range, the driver is in a state of significant over-trust, and the orange light will be on. If the fit value is in the third range, the driver is in a state of general over-trust and controls the yellow light to be on. If the fit value is in the fourth range, the driver is in a state of slightly over-trust and controls the green light to be on. If the fit value is in the fifth range, the driver is in a state of slight over-trust and controls the blue light to illuminate.

9. A dashboard lighting control system based on risk prediction and human-machine trust, characterized in that, include: The driver behavior information collection module is used to collect and extract driver behavior information, including the driver's eye, hand, and foot status information; The traffic environment information collection module is used to collect images of the road traffic environment and obtain information on the driving environment and traffic events. The trust assessment module is used to input the driver's behavioral information into a pre-built trust assessment model, and output the driver's trust assessment value of the autonomous driving system at the current moment after processing by the trust assessment model. The risk assessment module is used to assess the current risks of autonomous driving based on driving environment and traffic event information, and obtain the risk assessment results of the autonomous driving system, including the risk assessment coefficient of the autonomous driving system's technical capabilities, the risk assessment coefficient of the driving environment, and the risk assessment coefficient of the traffic event. The fit calculation module is used to calculate the fit between the trust level and the risk of the autonomous driving system based on the driver's trust level assessment value and the autonomous driving risk assessment result at the current moment. The lighting control module is used to control different colored lights according to the threshold range of the compatibility.

10. A computer storage medium, characterized in that, It contains a computer program that can be executed by a processor, which performs the dashboard lighting control method based on risk assessment and human-machine trust as described in any one of claims 1-8.

Images