Driver fatigue monitoring method and system based on kss drowsiness scale

Abstract

The application discloses a driver fatigue monitoring method and system based on KSS drowsiness grade, and relates to the technical field of driving safety. The method acquires the KSS drowsiness grade, driving speed and driving time length parameter of a driver in real time, divides different driving modes according to the speed, and dynamically calculates a fatigue weighting coefficient based on different modes. In the fatigue accumulation mode, the coefficient accumulates with the driving time length and the drowsiness grade; in the fatigue recovery mode, the coefficient attenuates. Finally, the comprehensive fatigue evaluation value is obtained by multiplying the weighting coefficient and the KSS grade, and compared with the dynamic threshold to determine whether to alarm. The application effectively solves the false alarm and missed alarm problems caused by the fixed threshold in the prior art, and realizes the precise and adaptive monitoring of the fatigue state of the driver.

Description

Technical Field

[0001] This invention relates to the field of driving safety technology, specifically to a method and system for monitoring driver fatigue based on the KSS drowsiness level. Background Technology

[0002] With the popularization of intelligent vehicle technology, driver fatigue monitoring has become a core element in ensuring road traffic safety. Fatigue driving, a key contributing factor to major traffic accidents, urgently requires high-precision, low-interference real-time monitoring solutions. Current mainstream monitoring technologies mainly rely on three approaches: analysis based on driver facial behavioral characteristics, such as eye closure frequency, yawning, and head posture changes; indirect inference based on vehicle dynamic parameters, including lane departure trajectories and abnormal steering wheel operation; and direct acquisition based on physiological signals, such as electroencephalograms (EEGs) and heart rate variability. Some solutions attempt to integrate multi-source information to improve accuracy, but significant limitations remain in practical applications.

[0003] In existing technologies, the Karolinska Drowsiness Scale (KSS) has been introduced into fatigue monitoring systems as a widely accepted subjective drowsiness assessment tool. However, its application faces fundamental challenges: the system's alarm thresholds are typically set in a fixed manner, failing to adapt to complex and ever-changing driving scenarios. For example, during sustained high-speed driving on highways, driver fatigue accumulates rapidly over time, but existing methods fail to dynamically correlate vehicle speed with driving duration to the KSS level, resulting in rigid threshold settings. When the threshold is too low, the system becomes overly sensitive to non-dangerous states such as brief eye closure or momentary distraction, generating numerous false alarms and severely interfering with the driving experience; when the threshold is too high, it struggles to capture the progressive deterioration of fatigue caused by prolonged medium-to-high-speed driving, leading to the risk of missed alarms.

[0004] The root cause lies in the lack of a dynamic recognition mechanism for driving modes in existing solutions: high-speed or medium-speed driving should be considered a phase of accelerated fatigue accumulation, while low-speed or stationary states may correspond to the fatigue recovery process. This static approach severs the intrinsic connection between the objective driving environment and subjective drowsiness perception, making it impossible to achieve adaptive adjustment of thresholds. Furthermore, traditional methods often employ linear accumulation models when quantifying fatigue risk, neglecting the nonlinear characteristics of fatigue changes at different speeds. For example, a small time increment at high speeds can trigger a significant fatigue spike, while the same time increment has a relatively milder impact at low speeds. These shortcomings cause monitoring results to become disconnected from actual risks, making it difficult to balance the timeliness and reliability of warnings. Summary of the Invention

[0005] In view of this, in order to overcome the shortcomings of traditional methods, the purpose of this invention is to propose a driver fatigue monitoring method and system based on KSS drowsiness level, which can dynamically adjust the fatigue assessment threshold according to the driving mode, effectively combine the objective driving environment and subjective drowsiness perception, avoid false alarms and missed alarms caused by fixed thresholds, and improve the accuracy and reliability of fatigue monitoring.

[0006] To achieve the above objectives, the present invention provides a driver fatigue monitoring method based on the KSS drowsiness level, comprising the following steps:

[0007] Obtain the driver's KSS level, driving speed, and driving duration;

[0008] Based on the driving speed, the current driving mode is determined, and the driving mode includes at least a fatigue accumulation mode and a fatigue recovery mode;

[0009] Based on the driving mode and the KSS level, a fatigue weighting coefficient is dynamically calculated, wherein:

[0010] When in the fatigue accumulation mode, the fatigue weighting coefficient is accumulated based on the driving duration-related parameters and the KSS level;

[0011] When in the fatigue recovery mode, the fatigue weighting coefficient is attenuated;

[0012] Based on the fatigue weighting coefficient and the KSS level, a comprehensive fatigue assessment value is determined;

[0013] The comprehensive fatigue assessment value is compared with a preset threshold, and a fatigue alarm is triggered based on the comparison result.

[0014] As a preferred embodiment of the present invention, the driving mode includes at least a low-speed mode, a medium-speed mode, and a high-speed mode; wherein, the low-speed mode is determined as the fatigue recovery mode, and the medium-speed mode and the high-speed mode are determined as the fatigue accumulation mode.

[0015] As a preferred embodiment of the present invention, the driving mode is determined based on the comparison between the driving speed and preset low-speed thresholds and medium-speed thresholds.

[0016] As a preferred embodiment of the present invention, the step of dynamically calculating the fatigue weighting coefficient includes:

[0017] Based on the driving mode, determine the basic rate of change;

[0018] Under the fatigue accumulation mode, the rate of change of the fatigue weighting coefficient is determined by the base rate of change and the acceleration term determined by the KSS level.

[0019] As a preferred embodiment of the present invention, in high-speed mode, the rate of change is further determined by an acceleration term related to speed.

[0020] In a preferred embodiment of the present invention, the driving time is obtained by processing a continuous sequence of video frames, and the fatigue weighting coefficient is updated based on the frame interval time.

[0021] As a preferred embodiment of the present invention, the comprehensive fatigue assessment value is obtained by multiplying the fatigue weighting coefficient by the KSS level.

[0022] As a preferred embodiment of the present invention, the value of the fatigue weighting coefficient is limited to a preset minimum and maximum value.

[0023] As a preferred embodiment of the present invention, the preset threshold is dynamically determined based on the maximum fatigue level Lmax.

[0024] Corresponding to the method described above, the present invention also provides a driver fatigue monitoring system based on KSS drowsiness levels, comprising:

[0025] The data acquisition module is used to acquire the driver's KSS level, driving speed, and driving duration;

[0026] The mode determination module is used to determine the current driving mode based on the driving speed, wherein the driving mode includes at least a fatigue accumulation mode and a fatigue recovery mode;

[0027] The weighting coefficient calculation module is used to dynamically calculate the fatigue weighting coefficient based on the driving mode and the KSS level;

[0028] The fatigue assessment module is used to determine the comprehensive fatigue assessment value based on the fatigue weighting coefficient and the KSS level.

[0029] The alarm judgment module is used to compare the comprehensive fatigue assessment value with a preset threshold and decide whether to trigger a fatigue alarm based on the comparison result.

[0030] This invention achieves accurate quantification of fatigue status in different driving scenarios by dynamically identifying driving modes and adjusting the fatigue weighting coefficient accordingly. It has the advantages of dynamically adjusting the fatigue assessment threshold according to the driving mode, effectively combining the objective driving environment with subjective drowsiness perception, avoiding false alarms and missed alarms caused by fixed thresholds, and improving the accuracy and reliability of fatigue monitoring. Attached Figure Description

[0031] 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 only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0032] In the diagram:

[0033] Figure 1 This is a flowchart of a driver fatigue monitoring method based on KSS drowsiness levels, according to an embodiment of the present invention.

[0034] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to specific examples and the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0036] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two different entities or different parameters with the same name. Therefore, "first" and "second" are merely for convenience of expression and should not be construed as limiting the embodiments of the present invention. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as other steps or units inherent in a process, method, system, product, or device that includes a series of steps or units.

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] In traditional driver fatigue monitoring systems, the lack of a dynamic fusion mechanism between objective driving scenario parameters and subjective Karolinska drowsiness levels leads to inappropriate alarm threshold settings. Specifically, existing technologies do not perform real-time interactive modeling of driving duration and real-time vehicle speed with KSS levels, making it impossible for threshold settings to reflect the dynamic changes in the rate of fatigue accumulation. Furthermore, the static threshold mechanism lacks adaptability when switching driving scenarios. When the threshold is too low, the system overreacts to momentary drowsiness, causing false alarms; when the threshold is too high, the system lags in responding to progressive fatigue, leading to missed alarms. This problem directly affects the accuracy and reliability of the warning system, making it difficult to achieve a balance between false alarms and missed alarms, thus weakening the system's warning effectiveness in practical applications.

[0039] For example, in continuous driving scenarios on highways, because the system uses a fixed alarm threshold, when the driver's Karolinska drowsiness level (KSS) momentarily rises to level 3 due to brief distraction, the system erroneously triggers a fatigue alarm, causing false alarms. When the driver drives at speeds above 100 km / h for more than two hours, the Karolinska drowsiness level (KSS) gradually rises from level 1 to level 4. If the threshold is set too high to avoid false triggers, the system may fail to identify severe fatigue in time, leading to missed alarms. Furthermore, in this scenario, false alarms distract the driver, potentially causing abnormal operation; missed alarms leave the driver in a high-risk state without warning, and the system cannot effectively avoid potential safety hazards.

[0040] If the aforementioned problems are not addressed, false alarms and missed alarms in fatigue monitoring systems will persist, severely undermining the system's reliability. Drivers may ignore subsequent genuine warning signals due to frequent false alarms, or fail to receive timely risk alerts in the event of a missed alarm, significantly increasing the probability of traffic accidents. Furthermore, insufficient system reliability will hinder the large-scale application of this technology in actual vehicles, impacting the overall ability to ensure road traffic safety.

[0041] Based on this, embodiments of the present invention provide a driver fatigue monitoring method based on the KSS drowsiness level, comprising the following steps:

[0042] Obtain the driver's Karolinska Slump Level (KSS), driving speed, and driving duration.

[0043] Based on driving speed, determine the current driving mode, which includes at least a fatigue accumulation mode and a fatigue recovery mode.

[0044] The fatigue weighting coefficient is dynamically calculated based on the driving mode and KSS level, where:

[0045] When in fatigue accumulation mode, the fatigue weighting coefficient is accumulated based on driving duration-related parameters and KSS level;

[0046] When in fatigue recovery mode, the fatigue weighting coefficient is reduced;

[0047] The comprehensive fatigue assessment value is determined based on the fatigue weighting coefficient and KSS level;

[0048] The comprehensive fatigue assessment value is compared with a preset threshold, and a fatigue alarm is triggered based on the comparison result.

[0049] In practical applications, the KSS level (fatigue level) is a scalar value used to quantify the driver's subjective level of drowsiness. In this invention, this level can be obtained using any known or future driver state monitoring technology, for example, by analyzing driver facial image features (such as eye closure, blinking frequency, and yawning frequency) and mapping them to the output of an algorithm module on a standard KSS scale, or by obtaining the level through driver input. The core protection of this invention lies in the method of dynamically weighting the KSS level after acquisition, while the specific calculation or acquisition method of the KSS level itself is not limited.

[0050] Driving mode refers to the state category classified according to driving conditions, used to distinguish the accumulation or recovery trend of fatigue. It can be determined by road type, such as classifying highway driving as fatigue accumulation mode and urban road driving as fatigue recovery mode; or by ambient light conditions, such as classifying night driving as fatigue accumulation mode and daytime driving as fatigue recovery mode.

[0051] Furthermore, the fatigue weighting coefficient refers to a parameter used to dynamically quantify the degree of historical fatigue accumulation. It can be calculated using a linear accumulation model. In the fatigue accumulation mode, the coefficient increases linearly with driving time; in the fatigue recovery mode, the coefficient decreases linearly with time. For example, a fixed value can be set for each unit of time, or the rate of change can be adjusted according to external environmental parameters. Specifically, the comprehensive fatigue assessment value refers to the final assessment index that integrates the current drowsiness state with historical fatigue accumulation. It can be determined by a weighted summation method, such as adding the fatigue weighting coefficient to the KSS level according to a preset ratio; or by nonlinear fusion through a fuzzy logic system. Thus, this invention dynamically calculates the fatigue weighting coefficient and interacts with objective driving scenario parameters and subjective KSS levels in real time, thereby achieving adaptive generation of the comprehensive fatigue assessment value. This allows the comparison process of alarm thresholds to be dynamically adjusted according to actual driving conditions, effectively avoiding false alarms and missed alarms caused by improper static threshold settings.

[0052] When implementing this driver fatigue monitoring method, the driver's Karolinska drowsiness level (KSS level), driving speed, and driving duration-related parameters are acquired as basic input data for assessing fatigue status. Driving speed is used to determine the current driving mode, which includes a fatigue accumulation mode and a fatigue recovery mode to distinguish the dynamic characteristics of fatigue evolution under different speed conditions. Further, based on the determined driving mode and KSS level, a fatigue weighting coefficient is dynamically calculated: when in the fatigue accumulation mode, this coefficient is accumulated based on driving duration-related parameters and the KSS level, reflecting the nonlinear cumulative effect of fatigue during continuous driving; when in the fatigue recovery mode, the coefficient decays, reflecting the natural relief process of fatigue during low-speed driving. Thus, based on the fatigue weighting coefficient and the KSS level, a comprehensive fatigue assessment value is determined, which integrates both historical fatigue accumulation and current drowsiness status information. Finally, the comprehensive fatigue assessment value is compared with a preset threshold, and a fatigue alarm is triggered based on the comparison result, achieving adaptive adjustment of the warning decision.

[0053] As a specific implementation, during long-distance driving, when the driving speed is at a high level (e.g., the vehicle is continuously driving on a highway), the system determines it to be in fatigue accumulation mode, where the fatigue weighting coefficient accumulates with driving time, and the accumulation amount is affected by the KSS level. When the speed decreases to a lower level (e.g., the vehicle is driving at low speed in urban areas), the system switches to fatigue recovery mode, and the fatigue weighting coefficient gradually decreases. In this scenario, the comprehensive fatigue assessment value is dynamically updated in real time, and an alarm is activated when the assessment value exceeds a threshold. Through this mechanism, objective driving scenario parameters and subjective KSS drowsiness levels are effectively and dynamically integrated, enabling adaptive adjustment of the alarm threshold, thereby reducing false alarms and false negatives and improving the accuracy and reliability of fatigue monitoring.

[0054] In some of the embodiments of the present invention, a driving mode based on driving speed is proposed to distinguish between fatigue accumulation and recovery. However, in its implementation, the specific classification criteria of driving mode are not clearly defined, which makes it impossible for the system to accurately identify the dynamic fatigue scenario based on actual speed changes. For example, when driving at low speed, it may be mistakenly judged as a fatigue accumulation mode and the coefficient may continue to accumulate, or it may be mistakenly judged as a recovery mode and the fatigue may be prematurely decayed at medium and high speeds, resulting in a distortion of the comprehensive fatigue assessment value. This leads to the alarm threshold being out of touch with the actual driving conditions, increasing the risk of false alarms or missed alarms.

[0055] In response, this invention further proposes that the driving modes include at least a low-speed mode, a medium-speed mode, and a high-speed mode; wherein, the low-speed mode is determined as a fatigue recovery mode, and the medium-speed mode and the high-speed mode are determined as fatigue accumulation modes.

[0056] In an exemplary embodiment of the present invention, the system parameter configuration is as follows:

[0057] Camera frame rate: V FPS = 15.0;

[0058] Maximum fatigue level: L max = 4;

[0059] Minimum weighting coefficient: W min = 0;

[0060] Maximum weighting coefficient: W max = L max ×2.0 = 8.0;

[0061] For the corresponding KSS drowsiness level, the step size of the weighting coefficient is: S0 = 1.0;

[0062] For high-speed driving conditions, the step size of the weighting coefficient is: S1 = 0.3;

[0063] Speed ​​threshold, low speed: LowSpeed ​​= 10.0 km / h;

[0064] Speed ​​threshold, medium speed: MidSpeed ​​= 30.0 km / h.

[0065] When the speed is less than or equal to LowSpeed, it is in low-speed mode and in the fatigue recovery phase. When the speed is greater than LowSpeed ​​but less than or equal to MidSpeed, it is in medium-speed driving. When the speed is greater than MidSpeed, it is in high-speed driving.

[0066] It should be noted that the above parameter values, including but not limited to frame rate, time thresholds (such as 3 hours of cumulative, 10 minutes of decay), change step size (S0, S1), and speed threshold, are all example values ​​based on experience and are not the only limitation of this invention. Those skilled in the art will understand that these parameters can be configured and adjusted according to different vehicle models, driving environments, or individual driver differences without departing from the protection scope of the core principles of this invention.

[0067] In practical applications, low-speed mode refers to a driving state where the vehicle's speed is in a lower range. This can be achieved through real-time monitoring using speed sensors or by using road type information from a navigation system. Its purpose is to accurately capture the characteristics of lower driver attention requirements in scenarios such as urban congestion, thus providing a basis for judging the natural fatigue recovery process. Medium-speed mode refers to a driving state where the vehicle's speed is in a moderate range. This can be dynamically divided using vehicle speed signals combined with environmental perception data. Its purpose is to reflect the characteristics of drivers needing to maintain a moderate level of attention in scenarios such as suburban roads, ensuring that the fatigue accumulation assessment matches the actual cognitive load. High-speed mode refers to a driving state where the vehicle's speed is in a higher range. This can be achieved by identifying real-time vehicle speed or external traffic flow data transmitted via the vehicle's CAN bus. The purpose is to strengthen the assessment logic for rapid fatigue accumulation, especially considering the characteristics of drivers needing to maintain a high level of concentration in scenarios such as highways. The fatigue recovery mode can be understood as an operating state where the system determines that the driver's fatigue level may be reduced. It is automatically triggered by the speed classification result, aiming to activate the attenuation mechanism of the fatigue weighting coefficient to avoid overreacting to temporary drowsiness. The fatigue accumulation mode can be understood as an operating state where the system determines that the driver's fatigue level may be aggravated. It is directly determined by the speed classification result, aiming to drive the accumulation process of the fatigue weighting coefficient to accurately reflect the dynamic accumulation law of fatigue with driving time.

[0068] Specifically, the solution of this invention quantifies driving speed into three intervals: low speed, medium speed, and high speed, and establishes a mapping relationship between speed intervals and fatigue dynamic modes, achieving refined identification of driving scenarios. When the system acquires real-time driving speed, it first determines the driving mode category based on the interval the speed value falls into: if the speed is in the low speed interval, it is determined to be a fatigue recovery mode, triggering the attenuation logic of the fatigue weighting coefficient; if the speed is in the medium or high speed interval, it is determined to be a fatigue accumulation mode, triggering the accumulation logic of the fatigue weighting coefficient. This classification mechanism ensures that in low-speed driving scenarios (such as slow-moving urban roads), the system can accurately identify the natural fatigue relief process, avoiding continuous coefficient accumulation due to incorrect identification as an accumulation mode; simultaneously, in medium- and high-speed driving scenarios (such as suburbs or highways), the system can reliably maintain the determination of the fatigue accumulation mode, preventing premature attenuation of the coefficient due to misjudgment of speed fluctuations as a recovery mode. Overall, the rigid division of speed intervals and the direct correlation with mode determination ensure that the fatigue assessment logic remains synchronized with changes in the cognitive load of the driving environment, thus providing a reliable basis for calculating the comprehensive fatigue assessment value.

[0069] As a specific implementation method, the solution of the present invention is implemented as follows: The system acquires real-time vehicle speed signals through an onboard speed sensor. When the detected vehicle speed is lower than the typical speed limit range of urban roads, the system automatically classifies the current driving mode as a low-speed mode and activates the attenuation algorithm corresponding to the fatigue recovery mode. When the vehicle speed is within the normal driving speed range of suburban roads, it is classified as a medium-speed mode and the accumulation algorithm of the fatigue accumulation mode is started. When the vehicle speed is close to the cruise standard of highways, it is classified as a high-speed mode and the accumulation algorithm of the fatigue accumulation mode is also executed. For example, in the scenario where the vehicle enters a congested section and the vehicle speed is consistently below 30 km / h, the system identifies it as a low-speed mode and begins to linearly attenuate the fatigue weighting coefficient. When the vehicle enters a highway and the vehicle speed is stable above 100 km / h, the system identifies it as a high-speed mode and accumulates the coefficient based on the KSS level and driving time.

[0070] Through the above technical solution, the present invention can accurately distinguish fatigue dynamic scenarios based on actual driving speed, effectively avoiding the problem of incorrectly judging fatigue accumulation mode at low speed driving, resulting in continuous accumulation of coefficients, or incorrectly judging fatigue recovery mode at medium and high speed driving, resulting in premature decay of coefficients. This ensures that the comprehensive fatigue assessment value truly reflects the matching relationship between driver fatigue state and driving conditions, and significantly reduces the false alarm rate and missed alarm risk of the fatigue alarm system.

[0071] In some of the embodiments of the present invention, driving modes including low speed mode, medium speed mode and high speed mode are proposed to distinguish fatigue recovery and accumulation. However, in its implementation, there is a lack of specific threshold mechanism to objectively determine the mode to which the current driving speed belongs, which leads to the mode judgment relying on subjective settings or fuzzy rules, which may cause misclassification of driving modes and thus affect the accuracy of dynamic calculation of fatigue weighting coefficient.

[0072] In response, the present invention further proposes to determine the driving mode based on a comparison between the driving speed and preset low-speed and medium-speed thresholds.

[0073] Specifically, the preset low-speed threshold is a speed reference value used to define the starting point of the low-speed mode. It can be set by using a fixed value or by dynamically adjusting it based on historical driving data. Its purpose is to provide clear boundary conditions for the objective classification of driving modes and avoid ambiguity in classification caused by continuous speed changes. The preset medium-speed threshold is a speed reference value used to distinguish between medium-speed mode and high-speed mode. It can be achieved by using empirical values ​​or by adaptively determining it through vehicle type. Its purpose is to accurately capture the differentiated impact of different speed ranges on the fatigue accumulation rate and ensure that the mode classification matches the actual driving scenario.

[0074] Specifically, the solution of this invention performs a hierarchical comparison of the real-time acquired driving speed with preset low-speed and medium-speed thresholds. When the driving speed is below the low-speed threshold, it is reliably identified as a low-speed mode; when the driving speed is between the low-speed and medium-speed thresholds, it is identified as a medium-speed mode; and when the driving speed is above the medium-speed threshold, it is identified as a high-speed mode. This transforms continuous speed parameters into discrete driving mode categories. This threshold comparison-based mechanism ensures the objectivity and repeatability of mode judgment, making the determination of the driving mode strictly dependent on objective speed values ​​rather than empirical judgments. This provides a stable and reliable foundation for the subsequent dynamic calculation of fatigue weighting coefficients and effectively supports the generation process of comprehensive fatigue assessment values.

[0075] As a specific implementation method, the present invention is implemented as follows: During the initialization phase, the system presets a low-speed threshold and a medium-speed threshold. The low-speed threshold corresponds to the lower limit of the typical speed range for low-speed driving scenarios on urban roads, and the medium-speed threshold corresponds to the upper limit of the typical speed range for medium-speed driving scenarios on suburban roads. During operation, the driving speed is acquired in real time and compared step-by-step with the preset thresholds. If the speed is lower than the low-speed threshold, it is determined to be in low-speed mode; if the speed is higher than the low-speed threshold but lower than the medium-speed threshold, it is determined to be in medium-speed mode; if the speed is higher than the medium-speed threshold, it is determined to be in high-speed mode. This implementation method achieves automated classification of driving modes through clearly defined threshold boundary conditions.

[0076] Through the above scheme, the present invention achieves objective and accurate classification of driving modes, reduces mode misjudgment caused by the continuity of speed changes, thereby improving the accuracy of dynamic calculation of fatigue weighting coefficients and ultimately enhancing the reliability of comprehensive fatigue assessment values.

[0077] Specifically, in some embodiments of the present invention, a dynamic calculation of fatigue weighting coefficients is proposed to assess driver fatigue status. However, in its implementation, the calculation method of the rate of change is relatively static and does not fully consider the dynamic impact of driving mode (such as speed) and KSS level on the rate of change. Specifically, in fatigue accumulation mode, the rate of change may be fixed or depend only on driving time, and the acceleration amplitude cannot be dynamically adjusted according to the real-time KSS level. This results in insufficient fatigue accumulation rate when the driver's KSS level is high, making it impossible to provide timely warning of severe fatigue status; or in high-speed driving scenarios, the rate of change does not take speed factors into account, causing a lag in fatigue risk assessment, thereby leading to missed or false alarms.

[0078] In response, this invention further proposes a step for dynamically calculating fatigue weighting coefficients, including:

[0079] Determine the basic rate of change based on driving mode;

[0080] In the fatigue accumulation mode, the rate of change of the fatigue weighted coefficient is determined by the base rate of change and the acceleration term determined by the KSS level.

[0081] In this embodiment, the rate of change parameter includes:

[0082] Fatigue growth factor: R inc = L max / (3.0 × 60 × 60 × VFPS)

[0083] Fatigue attenuation factor: R dec = L max / (1.0 × 10 × 60 × VFPS)

[0084] R inc Based on the basic growth rate of change, that is, after 3 hours, the weighted coefficient reaches L. max At low speeds, the weighting coefficients decrease, with an attenuation factor of R. dec The current configuration is that after a 10-minute rest period, the weighting coefficient decays to W. min .

[0085] Among them, the basic rate of change refers to the baseline rate of change of the fatigue weighted coefficient set according to the driving mode. It can be implemented by a preset discrete threshold interval mapping method, such as dividing the driving speed into multiple intervals and corresponding to different basic rate of change values, so as to adapt to the fatigue accumulation characteristics under different speed environments. The acceleration item determined by the KSS level can be understood as an additional adjustment factor dynamically generated based on the Karolinska drowsiness level. It can be implemented by a mapping method between the KSS level and a preset function relationship, such as converting the KSS level into an acceleration item value through a linear or piecewise function. Its purpose is to make the rate of change dynamically adjusted according to the driver's real-time drowsiness state.

[0086] Specifically, the present invention achieves precise control of the rate of change through a dual dynamic adjustment mechanism of driving mode and KSS level. First, the driving mode is determined based on a comparison between driving speed and a preset threshold, thereby obtaining the corresponding basic rate of change, ensuring that the initial value of the rate of change matches the current driving risk level. Second, in fatigue accumulation mode, the basic rate of change is superimposed with the KSS level acceleration term to form the final rate of change, where the acceleration term increases significantly with the KSS level, thus actively increasing the fatigue accumulation rate when the driver's drowsiness intensifies. This design allows the rate of change to respond simultaneously to dynamic changes in the driving scenario and the driver's subjective drowsiness, avoiding the shortcomings of fixed rates of change that cannot adapt to high KSS level scenarios, and ensuring that the update rate of the fatigue weighting coefficient remains synchronized with the actual risk level.

[0087] As a specific implementation method, the present invention is implemented as follows: The processing unit periodically acquires driving speed and KSS level data. When the system identifies that the driving mode is in fatigue accumulation mode, the basic rate of change is configured as a benchmark value matching the current speed range. Simultaneously, the KSS level is input to the acceleration term calculation module, which generates acceleration term values ​​through a preset nonlinear mapping relationship. For example, when the KSS level reaches a high level, the acceleration term increases exponentially. Specifically, the processing unit can be an ARM Cortex-M7 series microcontroller used to perform the superposition calculation of the basic rate of change and the acceleration term, and apply the result to the real-time update of the fatigue weighting coefficient.

[0088] The fatigue accumulation mechanism in this embodiment is as follows:

[0089] Applicable scenarios: low speed / stop, fatigue recovery;

[0090] In this scenario, the weighting coefficients decay, reaching a minimum of Wmin. The rate of change of weights (C0): C0 = -Rdec, meaning that when the vehicle is parked and the driver is resting, the weighting coefficients decrease.

[0091] Applicable scenarios: Medium-speed driving, where fatigue accumulates slowly;

[0092] In this scenario, the corresponding weighted coefficient calculation consists of two parts: basic growth + drowsiness level acceleration.

[0093] Rate of change of weight (C0): C0 = R inc + fatigueLevel×S0×R inc , where R inc Representing basic growth, fatigueLevel represents the KSS fatigue level, and the fatigue level is accelerated as "fatigueLevel × S0 × R". inc "The higher the level, the more likely it is to become fatigued, accelerate faster, and the faster the weighted coefficient increases."

[0094] Low speed / stop: fatigue recovery, weighted coefficient decay, minimum is W min The rate of change of weight (C0): C0 = -R dec That is, when the driver is resting, the weighting coefficient decreases.

[0095] Through the above technical solution, the present invention can dynamically adjust the rate of change of the fatigue weighting coefficient according to the real-time KSS level, accelerate the fatigue accumulation assessment process when the driver's drowsiness intensifies, effectively solve the problem of early warning delay caused by static setting of the rate of change, and significantly improve the response sensitivity of the comprehensive fatigue assessment value to severe fatigue state and the reliability of alarm triggering.

[0096] In some of the embodiments of the present invention described above, a feature is proposed that the rate of change in the fatigue accumulation mode is determined by the base rate of change and the KSS level. However, in its implementation, the dynamic influence of speed on the fatigue accumulation rate during high-speed driving is not included in the rate of change calculation, which results in the accumulation rate of the fatigue weighting coefficient in high-speed scenarios not being able to adaptively increase with the increase of speed. This may underestimate the risk of rapid fatigue accumulation in high-speed driving, causing early warning delays or missed reports.

[0097] In this regard, the present invention further proposes a step in which, in high-speed mode, the rate of change is further determined by a speed-related acceleration term, including:

[0098] In high-speed mode, the rate of change is further determined by the acceleration term related to speed.

[0099] The fatigue accumulation mechanism in this embodiment is as follows:

[0100] Applicable scenarios: High-speed driving, where fatigue accumulates rapidly;

[0101] In this scenario, the corresponding weighted coefficient calculation consists of three parts: base growth + speed acceleration + drowsiness level acceleration. Weighted change rate (C0): C0 = R inc + R inc ×S1 + fatigueLevel×S0×R inc , where R inc ×S1 represents "speed acceleration".

[0102] In practical applications, speed-related acceleration terms refer to parameters that dynamically modulate the rate of change based on real-time driving speed. These parameters can be implemented using linear functions, nonlinear functions, or piecewise functions based on preset speed ranges. The purpose is to introduce driving speed as a dynamic variable into the rate of change calculation process, enabling the system to distinguish the impact of different speed levels on the fatigue accumulation rate, thereby avoiding the problem of fixed cumulative acceleration rate caused by the lack of speed factors in high-speed scenarios.

[0103] Specifically, the solution of this invention couples the speed-related acceleration term with the acceleration term determined by the base rate of change and the KSS level in high-speed mode to form a dynamically adjusted rate of change output. Since the fatigue accumulation rate is non-linearly positively correlated with speed in high-speed driving scenarios, this solution uses real-time driving speed data to incrementally compensate the rate of change through the speed-related acceleration term, thereby adaptively enhancing the accumulation rate of the fatigue weighting coefficient as speed increases. Given the strong correlation between driving speed and fatigue accumulation risk, this design ensures that the rate of change accurately reflects the rapid fatigue accumulation characteristics unique to high-speed driving. Simultaneously, the accurate identification of driving modes by the mode judgment module avoids introducing unnecessary speed interference in low-speed or recovery modes, achieving a refined match between the driving scenario and the fatigue assessment logic.

[0104] As a specific implementation method, the present invention is implemented as follows: In high-speed mode, the system can use the square function relationship of speed to calculate speed-related acceleration terms. For example, when the driving speed is detected to exceed the preset medium speed threshold, the acceleration terms are dynamically scaled according to the square ratio of speed. The acceleration terms are combined with the basic rate of change and the acceleration terms determined by the KSS level through addition to form the final rate of change. The calculation process of speed-related acceleration terms is executed in real time by the weighted coefficient calculation module without relying on additional physiological signal acquisition equipment.

[0105] Through the above technical solution, the present invention achieves adaptive enhancement of the fatigue weighting coefficient accumulation speed in high-speed driving scenarios, effectively avoiding the problem of underestimation of fatigue accumulation risk due to the lack of speed factors, ensuring that the comprehensive fatigue assessment value can reflect the rapid fatigue deterioration trend in high-speed driving in a timely manner, thereby reducing the probability of early warning delay or missed reporting, and improving the reliability and adaptability of the fatigue monitoring system in complex driving scenarios.

[0106] In some of the embodiments of the present invention described above, driving time-related parameters are proposed for dynamically calculating fatigue weighting coefficients. However, in the implementation process, parameter acquisition relies on coarse timing mechanisms or non-behavioral sensor data, which cannot accurately capture changes in the driver's actual behavior (such as short rests or distracted states), resulting in distorted driving time estimation. Consequently, the dynamic update of the fatigue weighting coefficient lacks real-time performance and accuracy, leading to deviations in the comprehensive fatigue assessment value and causing false alarms or missed alarms in the alarm system.

[0107] In response, the present invention further proposes that the driving duration-related parameters are obtained by processing a continuous video frame sequence. Specifically, by analyzing the facial behavior features of the driver in the video frame sequence, the continuous period during which the driver is in a focused driving state is effectively identified, and this identified continuous period is used as the driving duration-related parameter for updating the fatigue weighting coefficient. The update of the fatigue weighting coefficient is based on the frame interval time of the video frame sequence.

[0108] Specifically, driving duration-related parameters refer to quantitative indicators reflecting the duration of a driver's actual driving behavior. These can be implemented using behavior recognition algorithms based on video frame sequence analysis. For example, by detecting changes in the driver's eye closure or head posture, effective driving periods can be dynamically distinguished from non-driving periods. This aims to avoid the shortcomings of traditional timing methods that cannot respond to dynamic changes in behavior. Processing continuous video frame sequences can be understood as continuously processing the real-time video stream captured by the vehicle camera. This can be achieved using computer vision techniques such as optical flow or convolutional neural network models to extract driver behavior features. The goal is to accurately identify short rests or distracted states. In practical applications, the update of the fatigue weighting coefficient is specifically to dynamically adjust the coefficient value according to time changes. This can be achieved using an incremental calculation mechanism based on frame interval time to ensure that the update rhythm is synchronized with video processing. The goal is to eliminate the response lag caused by fixed time interval updates. Frame interval time refers to the time difference between consecutive video frame captures. This can be achieved using the timestamp difference recorded by the system clock as a precise time increment reference. The goal is to provide a time reference synchronized with video processing.

[0109] In this embodiment, the weighting coefficients are calculated or updated based on the frame interval and the weight change rate:

[0110] W0 = mCurrentWeight + deltaX × C0,

[0111] mCurrentWeight = max(W min ,min(W0,W max ))

[0112] Where mCurrentWeight represents the current weighting coefficient, initialized to W. min deltaX represents the frame interval. The frame interval (deltaX) refers to the difference in the sequence numbers of two consecutive video frames used for update calculations, and this difference represents the time span. For example, if the frame sequence number difference is 6 and the camera frame rate is 15 frames / second, then the corresponding actual time interval is 6 / 15 = 0.4 seconds.

[0113] Specifically, the solution of this invention first processes a continuous sequence of video frames to accurately obtain parameters related to driving duration. These parameters dynamically distinguish between actual driving and resting states based on driver behavior characteristics (such as the duration of eye closure or the frequency of yawning), thereby truly reflecting the fatigue accumulation process. Subsequently, using the frame interval time as a time increment benchmark, the fatigue weighting coefficient is updated in real time during each video frame processing, so that the coefficient calculation is synchronized with changes in driving behavior in real time. For example, when switching driving modes (such as from high-speed fatigue accumulation mode to low-speed recovery mode), the coefficient decay or cumulative acceleration rate is immediately adjusted according to the accurate value of the frame interval time, ensuring that the update process is completely matched with the video processing, effectively avoiding evaluation deviations caused by parameter distortion or update delays.

[0114] As a preferred embodiment, the present invention is implemented as follows: the vehicle-mounted camera captures a video stream of the driver's face at a predetermined frame rate; the video processing unit is specifically an embedded microcontroller that analyzes the eye region features of each frame in real time; when multiple consecutive frames of closed-eye states are detected, it is determined to be a short rest, and the driving pause time is accumulated; the frame interval time is provided by the system clock and used to calculate the time increment, triggering an immediate update of the fatigue weighting coefficient, for example, adjusting the coefficient value immediately when a change in vehicle speed causes a switch in driving mode.

[0115] Through the above solution, the present invention achieves accurate acquisition of driving time parameters and real-time dynamic updating of fatigue weighting coefficients, effectively solves the problem of deviation in comprehensive fatigue assessment values ​​caused by parameter distortion, significantly reduces the false alarm rate and false alarm rate of the alarm system, and improves the reliability and practicality of driver fatigue monitoring.

[0116] In practical applications, some embodiments of the present invention propose a comprehensive fatigue assessment value to integrate the fatigue weighting coefficient and the KSS level to quantify driver fatigue risk. However, if the specific combination method of the two is not clearly defined during its implementation, the assessment results may not be able to dynamically adapt to changes in driving scenarios. For example, relying solely on the KSS level in the fatigue accumulation mode may easily lead to false alarms (such as amplifying temporary drowsiness), or ignoring objective driving parameters in the recovery mode may result in missed alarms (such as the KSS level not decaying in time when driving at low speeds), thus failing to achieve adaptive adjustment of the alarm threshold.

[0117] In response, this invention further proposes that the comprehensive fatigue assessment value is obtained by multiplying the fatigue weighting coefficient by the KSS level.

[0118] Among them, the comprehensive fatigue assessment value refers to a comprehensive indicator used to quantify driver fatigue risk. It can be calculated by multiplying the fatigue weighting coefficient by the Karolinska drowsiness level. Its purpose is to integrate subjective drowsiness level with objective driving parameters to form a dynamic assessment. The fatigue weighting coefficient is a dynamic parameter reflecting the degree of fatigue accumulation or recovery. It can be dynamically adjusted based on the driving mode through software algorithms. For example, it can be accumulated in the fatigue accumulation mode and attenuated in the fatigue recovery mode. Its purpose is to quantify the impact of driving scenarios on fatigue state. The Karolinska drowsiness level refers to the Karolinska drowsiness level, which can be extracted from the driver's facial video through image processing technology, such as based on eye closure analysis. Its purpose is to provide an objective measure of the driver's subjective fatigue state. In practical applications, the multiplication operation refers to the mathematical operation of multiplying two values. It can be implemented by executing multiplication instructions by a general-purpose processor. Its purpose is to establish the numerical combination relationship between the two.

[0119] Specifically, the present invention determines the comprehensive fatigue assessment value by multiplying a fatigue weighting coefficient by the KSS level. In this process, the fatigue weighting coefficient acts as a dynamic modulation factor, its value changing in real time according to the driving mode. When the system is in fatigue accumulation mode, the fatigue weighting coefficient increases with driving time, and the multiplication operation amplifies the impact of the KSS level on the assessment value, enabling the system to respond promptly to real fatigue risks. When the system is in fatigue recovery mode, the fatigue weighting coefficient decays, and the multiplication operation weakens the impact of instantaneous fluctuations in the KSS level on the assessment value, avoiding false alarms caused by temporary drowsiness. Through this mechanism, the comprehensive fatigue assessment value can dynamically adjust its sensitivity according to the driving scenario, ensuring the accuracy of alarm judgments.

[0120] In one specific implementation, the solution of the present invention can be implemented by an on-board electronic control unit. This electronic control unit may include a microcontroller for receiving KSS level data from a driver monitoring system and driving speed data from the vehicle bus. When calculating the comprehensive fatigue assessment value, the microcontroller performs a multiplication operation, multiplying the dynamically calculated fatigue weighting coefficient by the KSS level. For example, when the system detects that the vehicle is in high-speed mode and the driving time is long, the fatigue weighting coefficient is high, and the multiplication result will significantly increase the assessment value; when the vehicle switches to low-speed mode, the fatigue weighting coefficient decreases, and the multiplication result will reduce the sensitivity of the assessment value to fluctuations in the KSS level.

[0121] Through the above scheme, the present invention can dynamically adjust the sensitivity of the comprehensive fatigue assessment value according to the driving scenario, improve the alarm sensitivity in the fatigue accumulation mode to avoid missed alarms, and reduce the sensitivity in the fatigue recovery mode to prevent false alarms, thereby achieving more accurate fatigue warning.

[0122] In some of the embodiments of the present invention described above, a dynamic calculation of fatigue weighting coefficients is proposed to assess driver fatigue status. However, in the implementation process, since the fatigue weighting coefficients continuously accumulate in the fatigue accumulation mode and continuously decay in the fatigue recovery mode, the coefficient values ​​may exceed the reasonable range. For example, the coefficients may be too large after long-term driving, causing false alarms, or the coefficients may be too small during the recovery process, causing missed alarms, thereby affecting the accuracy and reliability of fatigue monitoring.

[0123] In response, the present invention further proposes that the value of the fatigue weighting coefficient is limited to a preset minimum and maximum value.

[0124] Specifically, the fatigue weighting coefficient refers to the coefficient used to adjust the KSS level during dynamic calculation, which can be updated in real time using software algorithms. The preset minimum value refers to the lower threshold of the fatigue weighting coefficient, which can be implemented using system configuration parameters or hardware limiting circuits. The preset maximum value refers to the upper threshold of the fatigue weighting coefficient, which can be implemented using software threshold settings or data verification mechanisms. The setting of the preset minimum and maximum values ​​is based on the dynamic characteristics of the driving scenario, aiming to ensure that the coefficient fluctuates within a reasonable range and avoid interference with the accuracy of the comprehensive fatigue assessment due to extreme values.

[0125] The present invention monitors the fatigue weighting coefficient in real time during dynamic calculation and performs amplitude limiting when a preset boundary is reached, ensuring that the coefficient value remains within a reasonable range. When the coefficient value is lower than a preset minimum value, the system fixes it at that minimum value to prevent excessive attenuation in fatigue recovery mode, which could lead to a loss of sensitivity to fatigue status. When the coefficient value is higher than a preset maximum value, the system fixes it at that maximum value to prevent excessive growth in fatigue accumulation mode, which could lead to misjudgment of temporary drowsiness. This boundary limiting mechanism, combined with the dynamic switching of driving modes, enables the fatigue weighting coefficient to more accurately reflect the actual fatigue evolution process of the driver, without being affected by numerical drift during the calculation process.

[0126] As a specific implementation method, the present invention is implemented as follows: During the system initialization phase, the preset minimum value can be set to a lower threshold, and the preset maximum value can be set to a higher threshold. During real-time monitoring, the system continuously calculates the fatigue weighting coefficient and checks whether it exceeds the preset range each time it is updated. If it is lower than the minimum value, it is adjusted to the minimum value; if it is higher than the maximum value, it is adjusted to the maximum value. This process is performed synchronously with the acquisition of parameters related to driving speed, KSS level, and driving duration, ensuring that the calculation of the comprehensive fatigue assessment value is based on coefficients within the effective range.

[0127] Through the above technical solution, the present invention effectively avoids the problem that the fatigue weighting coefficient may exceed the reasonable range due to continuous accumulation or decay, ensuring the stability and reliability of the comprehensive fatigue assessment value, thereby reducing the occurrence of false alarms and missed alarms, and improving the accuracy and robustness of the driver fatigue monitoring system.

[0128] In practical applications, existing technologies using fixed thresholds for fatigue alarm judgment cannot dynamically adjust according to driving scenarios. This results in thresholds that are too high in high-risk scenarios such as high-speed driving, leading to missed reports of true fatigue status, while thresholds that are too low in scenarios such as low-speed recovery, resulting in a large number of false alarms. Therefore, adaptive optimization of alarm thresholds cannot be achieved.

[0129] In response, this invention further proposes a preset threshold based on the maximum fatigue level L. max It is dynamically determined.

[0130] Specifically, the preset threshold refers to the critical value used to compare with the comprehensive fatigue assessment value to determine whether to trigger a fatigue alarm. It can be based on the maximum fatigue level L. max The numerical values ​​are dynamically calculated to adaptively adjust the alarm sensitivity; maximum fatigue level L max This refers to the highest fatigue level that the system can identify. It can be achieved using a theoretical upper limit determined based on driving modes and historical data, with the aim of providing a dynamic reference benchmark for threshold setting.

[0131] Specifically, the solution of the present invention is to use a preset threshold and the maximum fatigue level L max This correlation allows the threshold to be dynamically adjusted based on the driving scenario. In fatigue accumulation mode, L max This can be reduced, leading to a corresponding lower threshold and increasing the system's sensitivity to fatigue changes; in fatigue recovery mode, L max The threshold can be increased, leading to a corresponding increase in the alarm rate. This mechanism ensures that the threshold always matches the range of the comprehensive fatigue assessment value, thereby improving the accuracy of alarm decisions.

[0132] In this embodiment, an alarm threshold (THD) is set, where THD = 1.2 × L. max Product = mCurrentWeight × fatigueLevel; When Product is greater than or equal to THD, a fatigue alarm is triggered.

[0133] As a specific implementation method, the solution of the present invention is implemented as follows: when the vehicle is detected to be in high-speed mode, L max It can be set to a relatively low value; the preset threshold can be set to L. max A smaller proportion; when the vehicle is in low-speed mode, L maxIt can be set to a relatively high value; the preset threshold can be set to L. max A relatively large proportion. In this way, the threshold can adaptively reflect changes in the driving scenario.

[0134] Through the above technical solution, the alarm threshold can be dynamically adjusted according to the driving scenario. In high-risk scenarios, the threshold is lowered to avoid missed alarms, and in low-risk scenarios, the threshold is raised to reduce false alarms, thereby improving the accuracy and reliability of the fatigue monitoring system.

[0135] See Figure 1 As shown in the figure, the flow of the driver fatigue monitoring method based on KSS drowsiness level provided in this embodiment of the invention is as follows:

[0136] Step S101: System initialization.

[0137] Initialize relevant parameters, including but not limited to: camera frame rate (V FPS ), Maximum fatigue level (L) max ), the minimum value of the weighting coefficient (W) min ) and maximum value (W) max The weight change step size (S0, S1) for each mode, and the speed thresholds used to determine the driving mode (LowSpeed ​​and MidSpeed).

[0138] Step S102: Acquire and process real-time data.

[0139] The system runs in a loop, acquiring the driver's video data for the latest frame, the frame index, and the current driving speed from the vehicle bus. Simultaneously, it calculates the driver's current KSS level by analyzing the current video frame (e.g., using a pre-trained facial behavior recognition model).

[0140] Step S103: Determine the driving mode and calculate the weight change rate (C0).

[0141] Based on the current driving speed obtained in step S102, determine the driving mode to which it belongs:

[0142] If the speed is less than or equal to LowSpeed, it is determined to be in low-speed mode (i.e., fatigue recovery mode), and the weight change rate C0 is set to -R. dec (Attenuation factor);

[0143] If LowSpeed ​​< speed ≤ MidSpeed, it is determined to be a medium-speed mode (i.e., fatigue accumulation mode), and the weight change rate C0 = R is calculated. inc + fatigueLevel × S0 × R inc ;

[0144] If the speed > MidSpeed, it is determined as the high-speed mode (i.e., the fatigue accumulation mode), and the weight change rate C0 = R is calculated inc + R inc × S1 + fatigueLevel × S0 × R inc .

[0145] Step S104: Update the fatigue weighting factor

[0146] Based on the frame interval time (deltaX) of the video frame and the weight change rate (C0) calculated in step S103, update the current weighting factor on the basis of the previous fatigue weighting factor (mCurrentWeight). The calculation formula is: W0 = mCurrentWeight + deltaX × C0. Subsequently, perform a clipping process on W0 to make it satisfy: mCurrentWeight = max(W min , min(W0, W max )).

[0147] Step S105: Calculate the comprehensive fatigue evaluation value

[0148] Multiply the updated fatigue weighting factor (mCurrentWeight) by the current KSS drowsiness level (fatigueLevel) to obtain the comprehensive fatigue evaluation value (Product). That is: Product = mCurrentWeight × fatigueLevel

[0149] Step S106: Fatigue alarm judgment

[0150] Compare the comprehensive fatigue evaluation value (Product) calculated in step S105 with the preset dynamic alarm threshold (THD, for example, THD = 1.2 × Lmax):

[0151] If Product ≥ THD, it is determined that the driver is in a fatigued state and a fatigue alarm is triggered

[0152] If Product < THD, no alarm is triggered

[0153] Step S107: Execute in a loop

[0154] After completing the processing and judgment of the current frame, the process returns to step S102, continues to obtain the next frame of data, and repeats steps S102 to S106 to achieve real-time and continuous monitoring of the driver's fatigue state until the system is shut down

[0155] Through the above process, the present invention achieves the goal of dynamically adjusting the fatigue assessment sensitivity according to the real-time driving mode, effectively avoiding the problems of false alarms and missed alarms caused by fixed thresholds.

[0156] In another embodiment, the present invention also discloses a driver fatigue monitoring system based on KSS drowsiness levels, comprising:

[0157] The data acquisition module is used to acquire the driver's Karolinska drowsiness level (KSS), driving speed, and driving duration.

[0158] The mode determination module is used to determine the current driving mode based on driving speed. The driving mode includes at least a fatigue accumulation mode and a fatigue recovery mode.

[0159] The weighted coefficient calculation module is used to dynamically calculate the fatigue weighted coefficient based on the driving mode and KSS level;

[0160] The fatigue assessment module is used to determine the comprehensive fatigue assessment value based on the fatigue weighting coefficient and KSS level;

[0161] The alarm judgment module is used to compare the comprehensive fatigue assessment value with the preset threshold and decide whether to trigger a fatigue alarm based on the comparison result.

[0162] The core innovation of this embodiment lies in constructing an adaptive fatigue monitoring framework by dynamically weighting and fusing objective driving scenario parameters with the subjective Karolinska Slump Level (KSS). This enables adaptive assessment of fatigue risk and effectively solves the problems of false alarms and missed alarms caused by static threshold settings. Specifically, the mode judgment module distinguishes between fatigue accumulation mode and fatigue recovery mode in real time based on driving speed, while the weighting coefficient calculation module dynamically adjusts the fatigue weighting coefficient according to the driving mode and KSS level. This ensures that the comprehensive fatigue assessment value can truly reflect the impact of changes in the driving scenario on the fatigue accumulation rate. Through this mechanism, the system avoids false alarms caused by momentary drowsiness in scenarios such as continuous driving on highways, while ensuring timely alarm triggering when the KSS level continuously increases, significantly improving the accuracy and reliability of the warning.

[0163] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0164] It should be understood that, as used herein, the singular form "a" is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and / or" refers to any and all possible combinations of one or more of the associatedly listed items. The embodiment numbers disclosed above are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0165] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A method for monitoring driver fatigue based on the KSS drowsiness level, characterized in that, Includes the following steps: Obtain the driver's KSS level, driving speed, and driving duration; Based on the driving speed, the current driving mode is determined, and the driving mode includes at least a fatigue accumulation mode and a fatigue recovery mode; The driving modes include at least a low-speed mode, a medium-speed mode, and a high-speed mode; wherein, the low-speed mode is determined as the fatigue recovery mode, and the medium-speed mode and the high-speed mode are determined as the fatigue accumulation mode. Based on the driving mode and the KSS level, a fatigue weighting coefficient is dynamically calculated, wherein: When in the fatigue accumulation mode, the fatigue weighting coefficient is accumulated based on the driving duration-related parameters and the KSS level; When in the fatigue recovery mode, the fatigue weighting coefficient is attenuated; The steps for dynamically calculating the fatigue weighting coefficients include: Based on the driving mode, determine the basic rate of change; Under the fatigue accumulation mode, the rate of change of the fatigue weighting coefficient is determined by the base rate of change and the acceleration term determined by the KSS level. A comprehensive fatigue assessment value is determined based on the fatigue weighting coefficient and the KSS level; the comprehensive fatigue assessment value is obtained by multiplying the fatigue weighting coefficient by the KSS level. The comprehensive fatigue assessment value is compared with a preset threshold, and a fatigue alarm is triggered based on the comparison result.

2. The method according to claim 1, characterized in that, The driving mode is determined based on a comparison between the driving speed and preset low-speed and medium-speed thresholds.

3. The method according to claim 1, characterized in that, In high-speed mode, the rate of change is further determined by an acceleration term related to speed.

4. The method according to claim 1, characterized in that, The driving duration is obtained by processing a continuous sequence of video frames, and the fatigue weighting coefficient is updated based on the frame interval time.

5. The method according to claim 1, characterized in that, The fatigue weighting coefficient is limited to a preset minimum and maximum value.

6. The method according to claim 1, characterized in that, The preset threshold is dynamically determined based on the maximum fatigue level Lmax.

7. A driver fatigue monitoring system based on KSS drowsiness levels, characterized in that, include: The data acquisition module is used to acquire the driver's KSS level, driving speed, and driving duration; The mode determination module is used to determine the current driving mode based on the driving speed, wherein the driving mode includes at least a fatigue accumulation mode and a fatigue recovery mode; The driving modes include at least a low-speed mode, a medium-speed mode, and a high-speed mode; wherein, the low-speed mode is determined as the fatigue recovery mode, and the medium-speed mode and the high-speed mode are determined as the fatigue accumulation mode. The weighting coefficient calculation module is used to dynamically calculate the fatigue weighting coefficient based on the driving mode and the KSS level. The step of dynamically calculating the fatigue weighting coefficient includes: determining the basic rate of change based on the driving mode; under the fatigue accumulation mode, the rate of change of the fatigue weighting coefficient is jointly determined by the basic rate of change and the acceleration term determined by the KSS level. The fatigue assessment module is used to determine a comprehensive fatigue assessment value based on the fatigue weighting coefficient and the KSS level; the comprehensive fatigue assessment value is obtained by multiplying the fatigue weighting coefficient and the KSS level. The alarm judgment module is used to compare the comprehensive fatigue assessment value with a preset threshold and decide whether to trigger a fatigue alarm based on the comparison result.

Images