Hierarchical braking method and system based on passenger state recognition and medium

Abstract

The invention relates to a layered braking method and system based on passenger state recognition and a medium, and relates to the technical field of vehicle active safety. The method comprises the steps that information of an obstacle in front of a vehicle is obtained, and the time of collision TTC is calculated; identifying the type, posture and seat position of a passenger in the vehicle, and generating passenger state information; selecting a corresponding braking strategy group from a plurality of preset braking strategy groups according to the passenger state information, wherein different braking strategy groups have different TTC trigger thresholds and braking intensities; and executing corresponding braking control according to the selected braking strategy group and the current TTC value. According to the system, the body impact on different types of passengers can be reduced while the braking safety is guaranteed, and the personalized adaptability and the overall safety protection effect of the braking system are improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle active safety technology, and in particular to a layered braking method, system and medium based on occupant status recognition. Background Technology

[0002] With the rapid development of automotive intelligent technology, Automatic Emergency Braking (AEB) systems have become an important component of modern vehicle active safety technology. Traditional AEB systems primarily use sensors such as forward-facing radar and cameras to monitor obstacles in front of the vehicle. When the system determines there is a collision risk, it automatically activates the braking system to avoid or mitigate a collision. Meanwhile, Occupant Monitoring System (OMS) technology is also constantly evolving, capable of identifying the type, posture, and seatbelt wearing status of occupants through sensors such as cameras. These technological advancements provide a technical foundation for improving vehicle safety performance.

[0003] However, existing AEB systems have certain limitations in terms of braking strategies. Traditional AEB systems typically employ a uniform braking strategy, failing to adequately consider the individual differences and real-time conditions of different occupants. For occupants with lower physical resilience, such as the elderly, pregnant women, and children, uniform aggressive braking may produce excessive deceleration, adversely affecting their health. Furthermore, when occupants are in non-standard postures, such as leaning forward, sideways, or not wearing seatbelts correctly, traditional braking strategies may not provide optimal protection. These issues indicate that existing technologies still have room for improvement in terms of personalized braking strategies and occupant adaptability. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a layered braking method, system and medium based on occupant status recognition, which can reduce the physical impact on different types of occupants while ensuring braking safety, and improve the personalized adaptability and overall safety protection effect of the braking system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] In a first aspect, the present invention provides a layered braking method based on occupant state recognition, which adopts the following technical solution: Acquire information about obstacles in front of the vehicle and calculate the time to collision (TTC). Identify the type, posture, and seat position of occupants in the vehicle and generate occupant status information; Based on the occupant status information, a corresponding braking strategy group is selected from multiple preset braking strategy groups, wherein different braking strategy groups have different TTC trigger thresholds and braking intensities; and Based on the selected braking strategy group and the current TTC value, the corresponding braking control is executed.

[0007] Furthermore, in the aforementioned hierarchical braking method based on occupant state recognition, the identification of the type, posture, and seat position of the occupants includes: The vehicle uses an in-vehicle camera to identify passenger types, which include adults, children, the elderly, and pregnant women. Occupant posture is monitored via in-vehicle sensors, including sitting upright, sideways, leaning forward, leaning back, sleeping, and alighting from the vehicle; and Determine the seating positions of the occupants, including the driver's seat, the front passenger seat, the left rear seat, and the right rear seat.

[0008] Furthermore, in the above-mentioned hierarchical braking method based on occupant state recognition, the plurality of braking strategy groups include: The reminder strategy group is applicable to scenarios where children in the back seat are not in their seats, any occupant in any seat is not wearing a seat belt, or a pregnant woman in the front seat; The comfort braking system is suitable for any rear passenger, whether they are elderly, sitting upright, or sleeping; and The standard braking strategy group is suitable for scenarios where the front-seat passengers are sitting upright or in extremely dangerous conditions.

[0009] Furthermore, in the above-mentioned hierarchical braking method based on occupant status recognition, the braking intensity of the alert strategy group is lower than that of the comfort braking group, and the braking intensity of the comfort braking group is lower than that of the normal braking strategy group.

[0010] Furthermore, in the above-mentioned hierarchical braking method based on occupant state recognition, the execution of the corresponding braking control includes multiple stages: During the voice reminder phase, it is triggered when TTC is less than or equal to a preset first threshold. The forward collision warning phase is triggered when TTC is less than or equal to a preset second threshold. The pre-braking phase is triggered when TTC is less than or equal to a preset third threshold; and During the full braking phase, it is triggered when TTC is less than or equal to the preset fourth threshold.

[0011] Furthermore, in the above-mentioned hierarchical braking method based on occupant status recognition, different braking strategy groups correspond to different stage trigger thresholds, and the stage trigger thresholds of the alert strategy group are greater than the corresponding stage trigger thresholds of the comfort braking group and the normal braking strategy group.

[0012] Furthermore, the aforementioned hierarchical braking method based on occupant state recognition also includes: During the execution of braking control, the driver's braking intervention behavior is monitored in real time; and When the driver's active braking is detected, the automatic braking control is aborted and the braking control is handed over to the driver.

[0013] Furthermore, in the aforementioned hierarchical braking method based on occupant status recognition, the driver's braking intervention behavior includes the brake pedal being pressed and the braking deceleration exceeding a preset threshold.

[0014] Furthermore, in the aforementioned hierarchical braking method based on occupant status recognition, the voice reminder stage includes playing customized voice warnings according to the recognized occupant status. The customized voice warnings can provide corresponding reminder content based on different occupant types and posture states.

[0015] Secondly, the present invention provides a layered braking system based on occupant status recognition, which adopts the following technical solution: The obstacle detection module is used to acquire information about obstacles in front of the vehicle and calculate the time to collision (TTC). The occupant status recognition module is used to identify the type, posture, and seat position of occupants in the vehicle and generate occupant status information; A braking strategy selection module is used to select a corresponding braking strategy group from a plurality of preset braking strategy groups based on the occupant status information, wherein different braking strategy groups have different TTC trigger thresholds and braking intensities; and The braking control module is used to execute corresponding braking control based on the selected braking strategy group and the current TTC value.

[0016] Thirdly, the present invention provides a readable storage medium, which adopts the following technical solution: A readable storage medium storing computer instructions that, when executed by a processor, implement the hierarchical braking method as described in any one of the first aspects above.

[0017] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: The tiered braking method based on occupant status recognition of this invention identifies the type, posture, and seating position of occupants in real time and selects the appropriate braking strategy group according to the occupant status information, thereby achieving personalized braking control. This method can provide appropriate braking intensity for occupant groups with different physical tolerance levels (such as the elderly, pregnant women, and children), avoiding the adverse effects of excessive deceleration on the occupant's body that may result from a uniform braking strategy. Simultaneously, by employing different TTC trigger thresholds and tiered braking control, this method can improve the adaptability of the braking system to different occupant states while ensuring braking safety, thereby optimizing occupant comfort and overall safety protection. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a structural block diagram of a layered braking system based on occupant status recognition according to the present invention.

[0020] Figure 2 This is a flowchart of a hierarchical braking method based on occupant status recognition according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application.

[0022] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments of this application. Furthermore, the descriptions of each embodiment in the following embodiments have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0023] The method steps described in this embodiment of the invention can be executed in the order described in the specific implementation, or the execution order of each step can be adjusted according to actual needs, provided that the technical problem can be solved. These are not listed one by one here.

[0024] The present invention will be further described in detail below with reference to the accompanying drawings.

[0025] Reference Figure 1 This invention discloses a layered braking system based on occupant status recognition.

[0026] The layered braking system collects information through the forward perception module 1 and the occupant detection module 2. After being processed by the strategy filtering module 3, the corresponding control operations are executed by the braking execution module 4 and the HMI interaction module 5.

[0027] The forward-looking perception module 1 is configured to acquire obstacle information ahead of the vehicle and calculate the time-to-collision (TTC). In some embodiments, the forward-looking perception module 1 includes forward-facing radar and camera sensors for continuous monitoring of the road environment ahead. The forward-looking perception module 1 calculates the distance and relative speed between the vehicle and the target object ahead in real time, and calculates the TTC based on these parameters. The forward-looking perception module 1 transmits the acquired obstacle information and the calculated TTC data to the strategy filtering module 3.

[0028] The in-vehicle occupant detection module 2 is configured to identify the type, posture, and seat position of in-vehicle occupants, generating occupant status information. In some embodiments, the in-vehicle occupant detection module 2 includes an OMS camera located on the roof, a seat pressure sensor, and an infrared sensor. The in-vehicle occupant detection module 2 identifies occupant types using image recognition technology, including adults, children, the elderly, and pregnant women. The in-vehicle occupant detection module 2 monitors the dynamic posture of occupants, including sitting upright, leaning to the side, leaning forward, leaning back, sleeping, and various dismounted states. The in-vehicle occupant detection module 2 determines the seat position of the occupant, including the driver's position, the front passenger position, the left rear seat position, and the right rear seat position. The in-vehicle occupant detection module 2 sends the identified occupant status information to the strategy filtering module 3 in the form of a state vector.

[0029] The strategy selection module 3 is configured to select a corresponding braking strategy group from multiple preset braking strategy groups based on occupant status information. Different braking strategy groups have different TTC trigger thresholds and braking intensities. In some implementations, the strategy selection module 3 incorporates a strategy query matrix that defines the mapping relationship between different combinations of occupant status and seat position and specific braking strategy groups. The strategy selection module 3 receives and integrates TTC data from the forward-looking perception module 1 and occupant status information from the in-vehicle occupant detection module 2. Based on preset decision logic, the strategy selection module 3 selects a braking strategy group suitable for the current scenario from the alert strategy group, comfort braking group, and normal braking strategy group. Based on the selected braking strategy group and the real-time TTC value, the strategy selection module 3 issues corresponding control commands to the braking execution module 4 and the HMI interaction module 5.

[0030] The braking execution module 4 is configured to perform corresponding braking control based on the selected braking strategy group and the current TTC value. The braking execution module 4 provides smooth and precise braking output control, achieving a smooth and precise output from gentle pre-braking to customized full braking. In some embodiments, the braking execution module 4 precisely controls the braking system to generate the target deceleration at the corresponding TTC threshold point according to the instructions of the strategy selection module 3. The braking execution module 4 performs multi-stage braking control, including a gentle pre-braking stage, a comfort braking stage, and a full braking stage, each stage corresponding to a different deceleration control range. Through precise deceleration control, the braking execution module 4 avoids secondary injuries to occupants caused by excessive braking.

[0031] HMI interaction module 5 is configured to provide multimodal human-machine interaction warning functions based on instructions from strategy filtering module 3. In some implementations, HMI interaction module 5 generates and plays customized voice warnings at set early TTC threshold points, with the warning content related to the occupant's status. HMI interaction module 5 is equipped with multi-zone, multi-color optical prompting devices, including instrument panel icons, HUD head-up display, and ambient lighting for specific seats. Based on the identified high-risk seats, HMI interaction module 5 flashes a dynamic icon for that seat on the corresponding area's display screen or changes the ambient lighting in that area to a warning color, providing the driver with intuitive and accurate risk location information.

[0032] Based on the layered braking system described in the above embodiments, this invention also provides a layered braking method based on occupant status recognition.

[0033] Reference Figure 1 A layered braking method based on occupant status recognition enables personalized protection for vehicle occupants. This method begins with forward risk detection, proceeds through occupant status recognition and strategy selection, and ultimately executes customized braking control.

[0034] Specifically, in step S1, the forward-looking perception module 1 acquires information about obstacles in front of the vehicle and calculates the time to collision (TTC). The forward-looking radar and camera in the forward-looking perception module 1 work together to continuously scan the road environment ahead to detect potential collision targets. The forward-looking radar provides accurate distance measurements and relative speed data, while the camera provides visual recognition and classification information of the target object. Step S1 calculates the distance and relative speed between the vehicle and the obstacle in front in real time by fusing radar and camera data, and calculates the TTC based on the current relative motion state.

[0035] In step S2, the in-vehicle occupant detection module 2 identifies the type, posture, and seat position of the in-vehicle occupants, generating occupant status information. Step S2 uses an in-vehicle camera to perform image recognition on the occupants, determining that the occupant type includes adults, children, the elderly, and pregnant women. Step S2 monitors the dynamic posture of the occupants using seat pressure sensors and infrared sensors, including sitting upright, leaning to the side, leaning forward, leaning back, sleeping, and various dismounted states. Step S2 determines the specific seat position of each occupant, including the driver's position, the front passenger position, the left rear seat position, and the right rear seat position, and combines this information to form a complete occupant status vector.

[0036] In step S3, the strategy filtering module 3 selects a corresponding braking strategy group from a set of preset braking strategy groups based on occupant status information. Different braking strategy groups have different TTC trigger thresholds and braking intensities. Step S3 receives TTC data from step S1 and occupant status information from step S2, and performs decision analysis using a built-in strategy query matrix. Step S3 maps specific occupant statuses and seat positions to corresponding braking strategy groups, including alert strategy groups, comfort braking groups, and normal braking strategy groups. Each braking strategy group corresponds to a different TTC trigger threshold sequence and braking intensity range to meet the safety protection needs of different occupant groups.

[0037] In step S4, the braking execution module 4 and the HMI interaction module 5 execute corresponding braking control based on the selected braking strategy group and the current TTC value. Step S4 triggers control actions at different stages according to a preset TTC threshold sequence, based on the braking strategy group selected in step S3. Step S4 includes multiple progressive stages such as voice prompts, forward collision warnings, pre-braking, and full braking, each stage corresponding to a specific TTC trigger point and control intensity. Step S4 achieves a complete control process from early warning to final braking protection through precise timing control and deceleration management.

[0038] In some implementations, the process of identifying occupant types using in-vehicle cameras is achieved through image processing algorithms and machine learning models. The in-vehicle camera captures visual information such as facial features, body contours, and body proportions of the occupants, and analyzes and classifies these features using a pre-trained deep learning network. Occupant type identification is based on multiple physiological characteristic parameters, including facial contour maturity, body size proportions, and sitting posture. Adult identification is based on standard facial maturity features and body proportion parameters; child identification is based on relatively small body size and specific infantile facial features; elderly identification combines facial wrinkle texture, hair color changes, and body posture features; and pregnant woman identification is based on specific changes in the shape of the abdominal contour and sitting posture adjustment patterns.

[0039] In-vehicle sensors monitor occupant posture through the coordinated operation of multiple sensing technologies. Seat pressure sensors are distributed across multiple areas, including the seat cushion, backrest, and headrest, determining the occupant's posture by detecting pressure distribution patterns in different areas. A straight-sitting posture corresponds to a uniform pressure distribution on the seat cushion and backrest; a side-sitting posture shows increased pressure on one side of the seat cushion and an asymmetrical pressure distribution on the backrest; a forward-leaning posture shows increased pressure at the front of the seat cushion and decreased pressure on the backrest; and a backward-leaning posture shows increased pressure on the backrest and headrest. A sleeping posture is identified by the continuous pressure of the head on the headrest or side and the relative stillness of the body.

[0040] Infrared sensors supplement posture monitoring by detecting human body thermal radiation signals. They can detect heat distribution patterns across different parts of the occupant's body, making them particularly useful for identifying disengagement. When an occupant leaves their standard sitting position, the thermal signal distribution detected by the infrared sensors changes significantly, including the thermal signal from bending forward to pick up an item, changes in vertical thermal distribution when standing, and lateral shifts in thermal signal during sideways movement. Data fusion analysis between the infrared sensors and seat pressure sensors provides an accurate assessment of the occupant's posture.

[0041] Seat location is determined through the division of the vehicle's interior space into zones and spatial positioning by sensors. The driver's position is identified using sensor signals from the steering wheel area and seat position codes, while the front passenger position is determined based on the activation status of sensors on the front passenger seat. The left and right rear seats are distinguished using independent sensor groups in the rear seat area, with each seat corresponding to an independent set of pressure sensors and infrared detection areas. The in-vehicle camera uses a visual positioning algorithm, combined with the physical coordinates of the seats, to accurately determine the seat assignment for each detected occupant.

[0042] The roof-mounted configuration of the OMS cameras provides optimal field of view coverage for occupant status detection. The roof-mounted cameras acquire top-down images of all seats inside the vehicle, avoiding obstructions from seat backs and other interior trim. The OMS cameras employ a wide-angle lens design, with a single camera's field of view covering the entire interior space, including all seating areas in both the front and rear. In some implementations, multiple OMS cameras are distributed at different locations on the roof, improving the accuracy and reliability of occupant detection through multi-angle image fusion.

[0043] The seat pressure sensors are configured in a distributed layout, with independent pressure detection units installed at key pressure points on each seat. Pressure sensors in the seat cushion area are distributed on the left and right sides and in the front and back areas, forming a four-quadrant pressure detection grid. Pressure sensors in the backrest area are arranged according to the contact areas of the human spine and shoulders, including independent detection points in the lumbar support area, mid-back, and shoulder areas. The headrest area is equipped with dedicated pressure sensors to detect the contact status and pressure changes of the head. Each pressure sensor provides real-time pressure values ​​and trends, which are converted into occupant posture information through data processing algorithms.

[0044] Referring to Table 1, multiple braking strategy groups are categorized and designed based on occupant safety protection needs and physical tolerance. Strategy selection module 3 uses a strategy query matrix to achieve precise mapping between occupant status and seat position combinations to specific braking strategy groups. This strategy query matrix takes different occupant types, posture states, and seat position combinations as input parameters and outputs the corresponding braking strategy group selection results.

[0045] Table 1

[0046] The alert strategy group is applicable to scenarios where rear-seat children are unseated, any occupant is not wearing a seatbelt, or a pregnant woman is in the front seat. The core logic of the alert strategy group is primarily alerting and assisting, with braking as a secondary measure. The goal is to awaken the driver's intervention and create takeover time through very gentle braking. When the strategy query matrix detects a rear-seat child in an unseated state, including playing, standing, or bending over (non-standard sitting postures), the system automatically selects the alert strategy group. When any occupant is not wearing a seatbelt correctly, regardless of occupant type or seat position, the strategy query matrix uniformly maps to the alert strategy group. When a pregnant woman is detected in the front seat, considering the special needs for abdominal protection, the strategy query matrix categorizes this scenario into the alert strategy group.

[0047] The Comfort Braking system is suitable for any rear passenger, whether they are elderly, sitting upright, or sleeping. Its primary goal is smooth deceleration and injury avoidance, maximizing passenger comfort and safety and preventing secondary injuries by sacrificing some collision avoidance probability. When the strategy query matrix identifies any type of rear passenger—including children, the elderly, or adults—in a normal sitting position, the system selects the Comfort Braking system to suit the characteristics of the rear seat safety restraint system. When an elderly passenger is detected sitting upright, the strategy query matrix maps this scenario to the Comfort Braking system, taking into account their bone density and physical endurance. When any passenger is sleeping, the strategy query matrix selects the Comfort Braking system to avoid sudden braking impact due to muscle relaxation and reduced reaction time.

[0048] The standard braking strategy group is suitable for scenarios where a standard adult is sitting upright in the front seat or in extremely dangerous situations. Prioritizing collision avoidance, the standard braking strategy group employs an active braking strategy similar to a traditional AEB system. When the strategy query matrix detects a standard adult sitting upright in the front seat, the system selects the standard braking strategy group to fully utilize the adult's physical capacity and the robust safety restraint system in the front seat. In extremely dangerous situations, when the TTC value under any strategy group deteriorates drastically and a collision is unavoidable, the strategy query matrix will override the original selection and forcibly switch to the standard braking strategy group to maximize collision avoidance effectiveness.

[0049] The braking intensity of the alert strategy group is lower than that of the comfort braking group, and the braking intensity of the comfort braking group is lower than that of the normal braking strategy group. This tiered braking intensity design is based on the physical tolerance and safety protection needs of different occupant groups. The maximum deceleration of the alert strategy group is limited to below 0.5g to avoid excessive inertial impact on vulnerable occupants. The maximum deceleration of the comfort braking group is controlled within the range of 0.6g, maintaining occupant comfort while ensuring a certain braking effect. The maximum deceleration of the normal braking strategy group reaches 1.0g, fully utilizing the vehicle's braking performance to achieve the best collision avoidance effect.

[0050] The mapping relationship of the strategy query matrix is ​​achieved through preset decision rules and priority ranking. When multiple occupants are present in the vehicle simultaneously, the strategy query matrix selects strategies according to the priority of safety protection. The safety protection priority for children and pregnant women is higher than that for adults, and the protection priority for rear-seat occupants is higher in terms of braking intensity selection than that for front-seat occupants. When different types of occupants are present in the vehicle at the same time, the strategy query matrix selects the strategy group with the highest protection level to ensure the safety of the most vulnerable occupant.

[0051] The implementation of child protection principles is reflected in the special handling rules of the strategy query matrix for child-related scenarios. When a child is detected to be out of position, the system disables the full braking phase to prevent the enormous inertial force from throwing the child against hard surfaces inside the vehicle. The child protection principles stipulate that for scenarios involving a child leaving the rear seat, the maximum deceleration of braking control should not exceed 0.5g, and the braking process uses gradual loading to avoid sudden changes in deceleration. Once the child returns to a normal sitting position or fastens their seatbelt, the strategy query matrix dynamically adjusts the strategy selection, allowing the use of higher-intensity braking strategy groups.

[0052] Referring to Table 2, the execution of the corresponding braking control involves a progressive control process with multiple stages, each stage corresponding to a specific TTC trigger threshold and control action. The multi-stage design of the braking control achieves a complete safety management process from early warning to final braking protection through a time-series hierarchical triggering mechanism. Different braking strategy groups employ differentiated TTC trigger threshold settings at each stage to adapt to the safety protection needs and physical tolerance of different occupant groups.

[0053] Table 2

[0054] The voice alert phase, as the first stage of the braking control process, is triggered when the Time To Collision (TTC) is less than or equal to a preset first threshold. The voice alert phase of the alert strategy group activates when the TTC is less than or equal to 3.5 seconds, providing the earliest possible risk warning to vulnerable occupants. The voice alert phase plays customized voice warning content through the HMI interaction module 5, with the warning information personalized based on the identified occupant's status. When a child in the rear seat is detected leaving their seat, the voice alert is "Please note, children in the rear seat, please sit properly." When a front passenger is detected leaning forward, the voice alert is "Please note the road ahead." The goal of the voice alert phase is to awaken the driver's attention while the collision risk is still far off and prompt occupants to adjust to a safe posture.

[0055] The Forward Collision Warning (FCW) phase triggers when the Time To Chance (TTC) is less than or equal to a preset second threshold, providing the driver with enhanced visual and audible warning signals. The FCW phase activates in the Alert Strategy group when the TTC is less than or equal to 2.8 seconds, in the Comfort Braking group when the TTC is less than or equal to 2.2 seconds, and in the Normal Braking Strategy group when the TTC is less than or equal to 2.0 seconds. The FCW phase communicates the impending emergency braking to the driver through a combination of flashing instrument panel icons, a head-up display (HUD), and an audible alarm. The FCW phase also activates ambient lighting warnings in specific seating areas; when a risk is detected for an occupant in a particular seat, the ambient lighting in that area turns amber as a warning.

[0056] The pre-braking phase is triggered when the TTC is less than or equal to the preset third-order value, providing tactile warning feedback through slight braking force. The pre-braking phase for the alert strategy group initiates when the TTC is less than or equal to 2.2 seconds, applying a slight deceleration of approximately 0.15g as a tactile warning. The pre-braking phase for the comfort braking group initiates when the TTC is less than or equal to 1.7 seconds, applying a deceleration of approximately 0.3g. The tactile warning during the pre-braking phase conveys the signal that the braking system has intervened to the driver through the slight deceleration felt on the brake pedal and vehicle body. The deceleration control during the pre-braking phase employs a progressive loading method to avoid sudden braking shocks that could cause discomfort to occupants.

[0057] The comfort braking phase, as a moderate-intensity braking control phase, provides a balanced braking effect between the pre-braking and full-force braking phases. The comfort braking phase for the alert strategy group triggers when the time-to-catch (TTC) is 1.5 seconds or less, applying approximately 0.5g of deceleration. The comfort braking phase for the comfort braking group triggers when the TTC is 1.2 seconds or less, applying approximately 0.6g of deceleration. The comfort braking phase, with a moderate deceleration level of approximately 0.3g, ensures adequate braking while avoiding excessive inertial shock to the occupants. The deceleration control curve for the comfort braking phase features a smooth upward design, gradually increasing the deceleration from the slight level of the pre-braking phase to the target value, ensuring the occupants' physical adaptation process.

[0058] The full braking phase is triggered when the TTC is less than or equal to the preset fourth threshold, providing maximum braking force to avoid or mitigate a collision. The full braking phase for the Comfort Braking group initiates when the TTC is less than or equal to 0.7 seconds, applying a maximum deceleration of approximately 1.0g. The full braking phase for the Normal Braking Strategy group initiates when the TTC is less than or equal to 1.2 seconds, also applying a maximum deceleration of approximately 1.0g. The Alert Strategy group, based on occupant protection principles, disables the full braking phase in specific scenarios, especially when a child is detected to be out of position. The deceleration output during the full braking phase reaches the performance limit of the vehicle's braking system, achieving collision avoidance or a significant reduction in collision severity through maximized braking force.

[0059] Different braking strategy groups correspond to different stage trigger thresholds, reflecting differentiated safety protection strategies based on occupant status. The trigger thresholds for each stage of the alert strategy group are higher than those for the comfort braking group and the normal braking strategy group, providing earlier intervention opportunities and longer reaction time for vulnerable occupants. The voice alert stage trigger threshold for the alert strategy group is 3.5 seconds, while the comfort braking group and the normal braking strategy group do not have independent voice alert stages. The forward collision warning stage trigger threshold for the alert strategy group is 2.8 seconds, higher than the 2.2 seconds for the comfort braking group and the 2.0 seconds for the normal braking strategy group.

[0060] The alert strategy group has a pre-braking phase trigger threshold of 2.2 seconds, higher than the comfort braking group's 1.7 seconds. The normal braking strategy group does not have a separate pre-braking phase and directly enters full braking. The alert strategy group has a comfort braking phase trigger threshold of 1.5 seconds, higher than the comfort braking group's 1.2 seconds. The normal braking strategy group does not have a comfort braking phase. This differentiated threshold setting ensures that vulnerable occupants receive more warning time and a gentler braking process, while standard adult occupants receive more aggressive collision avoidance protection.

[0061] Tactile warning feedback is achieved through a slight deceleration during the pre-braking phase, providing the driver with a direct sense of braking system intervention. The slight deceleration of approximately 0.15g produces tactile feedback sufficient for the driver to perceive without causing significant physical impact on occupants. The deceleration level of the tactile warning feedback is precisely calibrated to ensure consistent warning effectiveness across various road conditions and vehicle load states. Tactile warning feedback works in conjunction with visual and auditory warning signals, enhancing the driver's risk perception and reaction speed through a multi-sensory approach.

[0062] The technical effectiveness of differentiated threshold settings lies in the precise protection of different occupant groups and the optimized balance of braking performance. The higher trigger threshold of the alert strategy group provides vulnerable occupants with ample time to adjust their posture, reducing the risk of secondary injury due to sudden braking. The medium trigger threshold of the comfort braking group maintains occupant comfort while ensuring braking effectiveness, adapting to the physical tolerance of rear-seat occupants and elderly occupants. The lower trigger threshold of the normal braking strategy group maximizes the use of available braking time and distance, providing optimal collision avoidance for standard adult occupants. Through precise control of the time series, differentiated threshold settings achieve a complete safety protection process from personalized warnings to customized braking.

[0063] In some implementations, the tiered braking method based on occupant status recognition also includes real-time monitoring of the driver's braking intervention behavior during braking control. Driver braking intervention monitoring is achieved through the coordinated operation of a brake pedal sensor, a brake pressure sensor, and a vehicle deceleration sensor. The brake pedal sensor detects changes in the brake pedal position and the force applied, the brake pressure sensor monitors changes in hydraulic pressure within the braking system, and the vehicle deceleration sensor measures the actual deceleration of the vehicle. The driver braking intervention monitoring system operates continuously throughout all stages of automatic braking control, including the voice alert stage, forward collision warning stage, pre-braking stage, comfort braking stage, and full braking stage.

[0064] The driver's braking intervention includes criteria for determining whether the brake pedal is depressed and whether the braking deceleration exceeds a preset threshold. Detection of brake pedal depression is achieved through changes in the position signal of the brake pedal sensor. When the brake pedal position sensor detects that the pedal has moved downwards beyond a preset displacement threshold, the system recognizes this as the driver initiating braking. The determination of whether braking deceleration exceeds the preset threshold is achieved through real-time data analysis from the vehicle's deceleration sensor, with the preset threshold set at a deceleration level of 0.2g or 0.3g. When the deceleration value detected by the vehicle's deceleration sensor exceeds the preset threshold, the system confirms that the driver's active braking action has produced an actual braking effect.

[0065] The system employs a dual-confirmation mechanism to determine braking intervention. Only when both conditions are met simultaneously—the brake pedal being depressed and the braking deceleration exceeding a preset threshold—is the system confirmed as valid driver intervention. This dual-confirmation mechanism avoids misjudgments caused by slight brake pedal contact or normal vehicle deceleration fluctuations. The preset threshold for braking deceleration is dynamically adjusted according to different braking control phases. A lower threshold of 0.2g is used in the pre-braking phase to improve detection sensitivity, while a higher threshold of 0.3g is used in the comfort braking and full braking phases to avoid confusion with the deceleration of automatic braking.

[0066] When the system detects driver-initiated braking, it aborts automatic braking control and transfers braking control to the driver. This transfer is achieved through a control switching mechanism within the braking system. The automatic braking control module immediately ceases outputting control commands to the brake actuators, and control of the braking system is completely transferred to the driver's brake pedal input. The response time for this handover is controlled within 100 milliseconds to ensure the driver's braking intentions take effect promptly. A smooth transition control strategy is employed during the handover, with the automatic braking force released gradually to avoid affecting vehicle stability and driver comfort due to sudden changes in braking force.

[0067] After the handover of braking control is complete, the system continues to monitor the collision risk status and the driver's braking effectiveness. When the driver's braking effectively reduces the collision risk, the TTC value increases, or the collision threat is eliminated, the system exits automatic braking mode and resumes normal driver assistance mode. When the driver's braking is insufficient to avoid a collision, and the TTC value continues to decrease and reaches the emergency threshold, the system provides additional braking assistance on top of the driver's braking, enhancing the overall braking effect through superimposed braking force. Dynamic management of braking control ensures that the driver always maintains ultimate control over the vehicle's braking, while providing safety protection functions of the automatic braking system when necessary.

[0068] In some implementations, the voice reminder stage includes playing a customized voice warning based on the identified occupant status, which can provide corresponding reminder content according to different occupant types and posture states.

[0069] The customized voice warnings are generated based on the identified occupant status, using personalized content and multimodal warning coordination. When the occupant detection module 2 detects a child in the rear seat leaving their seat, the HMI interaction module 5 plays a customized voice warning, "Please note, children in the rear seat, please sit properly," while simultaneously activating the amber ambient light warning in the corresponding rear seat area. The dynamic icons for the rear seats on the instrument panel display provide a visual warning in a rapidly flashing mode, and the HUD head-up display projects seat position information. When an elderly occupant is detected in a forward-leaning posture, the system plays a gentle voice reminder, "Please sit properly, elderly person," and the ambient light in the corresponding seat area slowly pulses to amber, while the instrument panel seat icons use a slow flashing animation. When a pregnant occupant is detected, the customized voice warning is "Please pay attention to abdominal protection," accompanied by soft amber area lighting and gentle icon prompt animation. Different occupant types and postures correspond to different voice content, optical warning intensity, and animation rhythm, achieving precise and personalized safety reminders.

[0070] This invention also discloses a readable storage medium.

[0071] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the layered braking method described in any of the above embodiments. The computer-readable storage medium may include any entity or device capable of carrying a computer program, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc. The computer program includes computer program code. The computer program code may be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable storage medium may include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc.

[0072] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0073] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0074] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A hierarchical braking method based on occupant state recognition, characterized by, The method comprises: obtaining information of an obstacle in front of the vehicle and calculating a time to collision (TTC); identifying the type, posture and seat position of an occupant in the vehicle to generate occupant state information; selecting a corresponding braking strategy group from a plurality of preset braking strategy groups according to the occupant state information, wherein different braking strategy groups have different TTC triggering thresholds and braking intensities; and executing corresponding braking control according to the selected braking strategy group and the current TTC value.

2. The hierarchical braking method based on occupant state recognition according to claim 1, characterized in that, The identification of the type, posture and seat position of the occupant in the vehicle comprises: identifying the type of the occupant through an in-vehicle camera, wherein the type of the occupant includes adult, child, old person and pregnant woman; monitoring the posture of the occupant through an in-vehicle sensor, wherein the posture of the occupant includes upright sitting, sideways sitting, forward leaning, backward leaning, sleeping and off-seat state; and determining the seat position of the occupant, wherein the seat position includes driver position, front passenger position, rear left position and rear right position.

3. The hierarchical braking method based on occupant state recognition according to claim 1, characterized in that, The plurality of braking strategy groups comprises: a reminding strategy group suitable for the scene of rear child off-seat, occupant in any seat not wearing safety belt or front pregnant woman; a comfort braking group suitable for the scene of any occupant in the rear, old person sitting upright or occupant in sleeping posture; and a general braking strategy group suitable for the scene of standard adult sitting upright in the front or extreme dangerous working condition.

4. The hierarchical braking method based on occupant state recognition according to claim 3, characterized in that, The braking intensity of the reminding strategy group is lower than that of the comfort braking group, and the braking intensity of the comfort braking group is lower than that of the general braking strategy group.

5. The hierarchical braking method based on occupant state recognition according to claim 3, characterized in that, The execution of corresponding braking control comprises multiple stages: a voice reminding stage triggered when TTC is less than or equal to a preset first threshold value; a forward collision warning stage triggered when TTC is less than or equal to a preset second threshold value; a pre-braking stage triggered when TTC is less than or equal to a preset third threshold value; and a full braking stage triggered when TTC is less than or equal to a preset fourth threshold value. Different braking strategy groups correspond to different stage triggering thresholds, and the stage triggering thresholds of the reminding strategy group are greater than the corresponding stage triggering thresholds of the comfort braking group and the general braking strategy group.

6. The hierarchical braking method based on occupant state recognition according to claim 5, characterized in that, The method further comprises:

7. The occupant status recognition based hierarchical braking method of claim 1, wherein, monitoring the braking intervention behavior of the driver in real time during the execution of the braking control; and suspending the automatic braking control and handing over the braking control to the driver when the driver's active braking is detected. The braking intervention behavior of the driver includes that the brake pedal is depressed and the braking deceleration exceeds a preset threshold value. The voice reminding stage includes playing customized voice warnings according to the identified occupant state, and the customized voice warnings can provide corresponding reminding contents according to different occupant types and posture states.

8. The hierarchical braking method based on occupant state recognition according to claim 7, characterized in that, The method comprises:

9. The hierarchical braking method based on occupant state recognition according to claim 5, wherein, a forward perception module for obtaining information of an obstacle in front of the vehicle and calculating a time to collision (TTC); 10. A hierarchical brake system based on occupant state recognition, characterized by, an in-vehicle occupant detection module for identifying the type, posture and seat position of an occupant in the vehicle to generate occupant state information; a strategy screening module for selecting a corresponding braking strategy group from a plurality of preset braking strategy groups according to the occupant state information, wherein different braking strategy groups have different TTC triggering thresholds and braking intensities; and a braking execution module for executing corresponding braking control according to the selected braking strategy group and the current TTC value. ​ ​ 11. A readable storage medium, characterized by, The readable storage medium stores computer instructions, and the computer instructions are executed by the processor to implement the layered braking method in any one of claims 1-9.

Images