An AEB (Automatic Emergency Braking) Control Method for Commercial Buses Turning at Intersections, Considering Passenger Comfort

Abstract

This invention belongs to the field of vehicle braking control systems and relates to an AEB control method for commercial buses turning at intersections that considers passenger comfort. The method first assesses potential collision risks and outputs a basic AEB decision. Then, it calculates the response parameters of a representative seat during the turning braking process and determines the comfort state of the most unfavorable representative seat. A direction-sensitive comfort index is constructed. Based on the collision risk results, the comfort state of the most unfavorable representative seat, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, the AEB trigger threshold is corrected. Finally, the current control mode is determined by combining the basic risk state determined based on the trigger threshold before correction and the comfort-corrected risk state determined based on the trigger threshold after correction. Based on the determined control mode, the target braking deceleration is optimized online, and braking control commands are generated. This achieves a synergistic balance of safety, comfort, and stability in AEB control of commercial buses under intersection turning conditions.

Description

Technical Field

[0001] This invention belongs to the field of vehicle braking control systems, and relates to automatic emergency braking (AEB) control for commercial vehicles, specifically to an AEB control method for commercial buses turning at intersections that takes into account passenger comfort. Background Technology

[0002] With the continuous development of my country's comprehensive transportation system, road passenger transport remains a crucial component of ensuring public travel. Within the key operational vehicle supervision system, "two-passenger-one-dangerous-goods" vehicles (passenger buses, bus services, and hazardous goods transport vehicles) are consistently a priority for safety supervision. Among these, commercial passenger transport, including scheduled passenger transport, chartered passenger transport, and tourist passenger transport, directly impacts the quality of daily and tourist travel services for the public. Meanwhile, commercial passenger transport primarily serves passengers, and its operational quality is reflected not only in driving safety but also in passenger comfort. Especially during intersection turns and emergency braking, the vehicle's longitudinal deceleration, lateral acceleration, sudden acceleration changes, and pitch and roll responses can easily cause passengers to experience forward lurching, side swaying, dizziness, fatigue, and motion sickness. Therefore, in the automatic emergency braking control of commercial passenger vehicles, in addition to effectively mitigating collision risks, it is also essential to fully consider passenger comfort needs to improve the overall performance of the vehicle's active safety control.

[0003] Current research on AEB control strategies mainly focuses on target identification and information acquisition, collision risk assessment, braking intervention timing determination, and braking deceleration control. Related methods can be broadly categorized into two types: those based on safe time models and those based on safe distance models. The former often uses TTC, THW, and their improved forms as criteria for determining the degree of danger, while the latter often establishes a safe distance model through vehicle braking process kinematic analysis, driver reaction characteristic modeling, or road adhesion condition correction, and uses this model to achieve warning and automatic braking control. For example, Chinese patent CN116901974A modifies the initial AEB trigger time limit by collecting vehicle body signals and vehicle perception signals, combined with driver attention level and feedback state level. Chinese patent CN120024334A modifies active braking parameters such as function trigger collision time and execution deceleration by establishing a black dot region and combining historical trigger data. Chinese patent CN119018107A proposes an automatic emergency braking function trigger timing correction method considering steering constraints, modifying the AEB trigger conditions by introducing longitudinal collision time, vehicle steering time, and special scenario correction factors. Building upon this foundation, some studies have further optimized AEB control strategies from a comprehensive performance perspective, focusing on issues such as premature system intervention interfering with normal driving and excessive intervention increasing braking impact. These studies employ methods such as graded braking, limiting braking deceleration, constraining the rate of change of deceleration, or optimizing the braking control process to improve the driving experience. In summary, existing AEB control strategy research mainly revolves around collision risk assessment, trigger timing correction, adaptive parameter adjustment, and braking process optimization.

[0004] However, automatic emergency braking not only relates to collision avoidance effectiveness and driving safety, but also directly affects the vehicle's longitudinal deceleration, body posture, and occupant comfort during intervention. Therefore, braking comfort has gradually become an important focus of related research. Braking comfort research mainly focuses on the impact of factors such as brake pitch, acceleration vibration, rate of change of acceleration, and the brake pressure build-up and release process on ride comfort. For example, existing research has focused on the coupling relationship between braking force and vehicle pitch, reducing brake plunge and rebound by optimizing the braking force timing and subsequent control strategies. Other research, based on brake-by-wire systems, uses methods such as target longitudinal acceleration planning and feedforward and feedback tracking control to suppress braking shock and improve braking smoothness. Furthermore, Chinese patent CN121106243A proposes a cornering active comfort braking method that determines the comprehensive desired deceleration by combining information such as road adhesion coefficient, corner radius, and forward and backward safe speeds, and then controls the braking actuator accordingly for smooth braking. Chinese patent CN121157857A proposes an automatic emergency braking threshold adjustment method based on in-cabin multimodal information. This method calculates driver and passenger state coefficients by acquiring visual and audio information from the driver and passengers, and then adaptively adjusts the AEB trigger threshold. Overall, existing research on braking comfort mainly focuses on three aspects: vehicle dynamics optimization, smooth control of the braking process, and occupant state perception.

[0005] In summary, existing comfort studies primarily focus on the driver and front-seat occupants of the vehicle or passenger cars, and the research scenarios mainly involve straight-line following, rear-end collision warning, and longitudinal emergency braking. There is insufficient research on the differences in occupant comfort across different seats in commercial buses during intersection turning scenarios. Meanwhile, while existing AEB control strategy research has conducted considerable studies on target recognition, collision risk assessment, warning threshold setting, and braking intensity control, it still primarily focuses on collision avoidance and longitudinal braking safety as the core objectives. Comfort considerations are mostly limited to post-event evaluation, supplementary analysis, or parameter correction, and occupant comfort indicators have not yet been directly used as constraints for AEB trigger timing correction and control decision optimization. Therefore, in intersection turning scenarios, it is impossible to simultaneously consider safety, comfort, and stability. Summary of the Invention

[0006] In view of the shortcomings and deficiencies of the existing technology, the purpose of this invention is to provide a commercial bus AEB control method for intersection turning that takes into account passenger comfort. By establishing a turning braking dynamic model of the commercial bus and a response model of the representative seats in the vehicle, passenger comfort index is introduced on the basis of collision risk assessment, and the AEB trigger threshold and control mode are adaptively corrected. Furthermore, the target braking demand is optimized, thereby achieving a synergistic balance of safety, comfort and stability in the AEB control of commercial buses under intersection turning conditions.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] An AEB (Automatic Emergency Braking) control method for commercial buses turning at intersections, taking into account passenger comfort, includes the following steps:

[0009] Step S1. Establish a vehicle dynamics model of a commercial bus under the turning condition at an intersection; at the same time, based on the vehicle body reference point, construct a position coordinate model representing the seats and the seat point response mapping relationship;

[0010] Step S2. Obtain information about the vehicle and the target vehicle or obstacle, assess the potential collision risk in the intersection turning scenario, identify the key conflict vehicle, and output the basic risk status and basic AEB decision.

[0011] Step S3. Based on the vehicle dynamics model and the position coordinate model of the representative seats established in Step S1, calculate the response parameters of the representative seats during the cornering and braking process, construct the comprehensive comfort index of each representative seat, and determine the comfort state of the most unfavorable representative seat; at the same time, combined with the key conflict vehicle and its conflict direction information relative to the vehicle obtained in Step S2, construct a direction-sensitive comfort index that considers the coupling relationship between conflict direction, vehicle steering direction, and representative seat spatial position; then, based on the collision risk results, the comfort state of the most unfavorable representative seat, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, correct the AEB trigger threshold; finally, combine the basic risk state determined based on the trigger threshold before correction and the comfort-corrected risk state determined based on the trigger threshold after correction to determine the current control mode;

[0012] Step S4. Based on the control mode determined in step S3, the target braking deceleration is optimized online by comprehensively considering the collision risk level, the most unfavorable representative seat comfort state, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, and braking control commands are generated.

[0013] As a preferred embodiment of the present invention, the vehicle dynamics model in step S1 adopts a nonlinear dynamics model, establishing the vehicle longitudinal motion equation, the vehicle lateral motion equation, the vehicle vertical motion equation, and the... The equations of motion for the unsprung mass of each wheel position are as follows: vertical motion equation for vehicle body pitch, roll, and yaw. The models of tire force, suspension force, and normal load for each wheel position are also established. The tire slip angle model and wheel rotation dynamics equation for each wheel position are also established.

[0014] As a preferred embodiment of the present invention, in step S1, the vehicle body reference point is used. Establish a position coordinate model of six representative seats inside the vehicle based on the reference; wherein, the six representative seats are the front left seat Front right seat Middle left seat Middle right seat Rear left seat and the rear right seat ;

[0015] The seat point response mapping relationship is a mapping relationship from vehicle body movement to representative seat response, used to obtain the first... Acceleration response of each representative seat in the longitudinal, lateral, and vertical directions. and the The rate of change of longitudinal acceleration for each representative seat and lateral acceleration rate of change .

[0016] As a preferred embodiment of the present invention, step S2 includes the following steps:

[0017] Step S21. Based on the vehicle dynamics model established in step S1 and the on-board sensors, obtain the vehicle's operating status information in real time, and directly measure or estimate the target vehicle's operating status information through the environmental perception module.

[0018] Step S22. Based on the two-dimensional minimum jerk turning trajectory prediction model, predict the future smooth turning trajectories of the vehicle and each target vehicle respectively;

[0019] Step S23. Based on the position and attitude changes of each vehicle in the prediction time domain, construct the dynamic safety envelopes of the vehicle and each target vehicle respectively, and determine whether they overlap in the prediction time domain. Then calculate the equivalent collision time between the vehicle and each target vehicle, and take the minimum equivalent collision time as the overall equivalent collision time of the vehicle facing the intersection vehicle-to-vehicle conflict in the current control cycle. The target vehicle that meets the minimum equivalent collision time is designated as the critical conflict vehicle; at the same time, the angle of approach of the critical conflict vehicle relative to the vehicle itself is determined. ;

[0020] Step S24. Based on the overall equivalent collision time Preset warning threshold Level 1 braking threshold and secondary braking threshold The basic risk status of current intersection turning vehicle conflicts. R The system is categorized and the corresponding basic AEB intervention requirements are output; when When, AEB is not triggered; when When, output an audible and visual warning signal; when At this time, it enters the first-stage braking phase and outputs the first-stage braking target deceleration. and use it as the basic target for deceleration; when At this time, it enters the secondary braking stage and outputs the secondary braking target deceleration. And use it as the basic target for deceleration.

[0021] As a preferred embodiment of the present invention, the two-dimensional minimum jerk turning trajectory prediction model in step S2 minimizes the prediction time domain. The vehicle is constructed using the square integral of the third derivative of the inner trajectory as the objective. The minimum jerk performance index is:

[0022] ;

[0023] in, and Representing vehicles Under the fixed coordinate system of the road direction and The jerk of direction; minimized under given initial position boundary, initial acceleration boundary, final position boundary, and final acceleration boundary condition constraints. Obtain the vehicle in the prediction time domain The smooth turning trajectory within the range, i.e., the two-dimensional minimum jerk prediction trajectory. .

[0024] As a preferred embodiment of the present invention, step S3 includes the following steps:

[0025] Step S31. Based on the acceleration response of each representative seat in the longitudinal, lateral, and vertical directions. and the rate of change of longitudinal acceleration and lateral acceleration rate of change , construct the first The system first generates three sub-indices: acceleration impact index, acceleration vibration index, and pitch angle vibration index. Then, these three sub-indices are normalized to obtain normalized acceleration impact index, acceleration vibration index, and pitch angle vibration index. Finally, a weighted fusion method is used to construct the first... A comprehensive comfort index representing each seat;

[0026] Step S32. Based on the comprehensive comfort index of the six representative seats obtained in step S31, determine the most unfavorable representative seat at the current moment and its corresponding comprehensive comfort index, and statistically analyze the average comprehensive comfort index and the dispersion index of comfort distribution in the vehicle; at the same time, combined with the key conflict vehicle and its conflict direction information relative to the vehicle obtained in step S2, construct a direction-sensitive comfort index that considers the coupling relationship between conflict direction, vehicle turning direction and representative seat spatial position.

[0027] Step S33. Construct a risk zoning proxy model for AEB trigger threshold correction, adaptively correct the AEB trigger threshold, and determine the current control mode. The trigger threshold includes a warning threshold, a first-level braking threshold, and a second-level braking threshold. The risk zoning proxy model first distinguishes basic risk states based on the state space. For each risk sub-region, a proxy model corresponding to each risk sub-region is established. The goal is to reduce the discomfort at the most unfavorable representative seat, reduce the imbalance in comfort experience among occupants in different positions within the vehicle, reduce the discomfort of occupants in sensitive areas under the influence of critical collisions, and ensure that the collision avoidance safety margin is still met after correction. Under constraints, the NSGA-II algorithm is used to solve the Pareto optimal solution set to initially determine the trigger threshold correction amount. Then, the objective function values ​​of each candidate solution in the Pareto optimal solution set are normalized, and the risk adaptive coefficient, comfort distribution adaptive coefficient, and direction sensitivity adaptive coefficient are comprehensively considered to calculate the optimal threshold correction amount at the current moment, thereby obtaining the corrected trigger threshold.

[0028] The comfort correction risk status is then determined based on the revised trigger threshold. And based on the basic risk status With comfort correction risk status Determine the control mode under the current control cycle. for:

[0029] ;

[0030] in, This indicates the comfort-priority AEB control mode. This indicates the AEB control mode with safety as the priority.

[0031] As a preferred embodiment of the present invention, in step S3, the first... Acceleration impact index for each representative seat for:

[0032] ;

[0033] in, The weighting coefficient for the rate of change of lateral acceleration;

[0034] No. An acceleration vibration index representing a seat. for:

[0035] ;

[0036] in, and These are the weighting coefficients for longitudinal acceleration and lateral acceleration, respectively.

[0037] No. Pitch angle vibration index of a representative seat for:

[0038] ;

[0039] in, The acceleration is the pitch angle of the vehicle body. This represents the longitudinal position of the seat.

[0040] As a preferred embodiment of the present invention, the orientation-sensitive comprehensive comfort index in step S3... for:

[0041] ;

[0042] ;

[0043] ;

[0044] in, For the first Each representative seat has a direction-sensitive weight. For the first Each representative seat represents an overall comfort index. To prevent positive numbers from appearing in the numerator or denominator that are not positive, For the first The original weights of the seats are sensitive to orientation. The normalized values ​​of the longitudinal and lateral coordinates of each representative seat relative to the vehicle reference point are respectively... , The vehicle's steering direction coefficient is the vehicle's current yaw rate. , Conversely, it is 0.

[0045] As a preferred embodiment of the present invention, the optimal threshold correction amount at the current time in step S3 Determined through the following online adaptive solution selection criteria:

[0046] ;

[0047] Revised warning threshold Level 1 braking threshold Secondary braking threshold for:

[0048] ;

[0049] ;

[0050] ;

[0051] in, , These are the optimal correction values ​​for the warning threshold, the first-level braking threshold, and the second-level braking threshold, respectively. For risk adaptive coefficient, For direction-sensitive adaptive coefficients, For the comfort distribution adaptive coefficient, , , and For the Pareto optimal solution set The result is the normalized result of the four objective function values ​​corresponding to each candidate solution.

[0052] As a preferred embodiment of the present invention, step S4 determines the target braking demand based on the current control mode, when At that time, the basic target deceleration output in step S2 is directly used. As the current control cycle The safety priority objective is deceleration; when At that time, the deceleration is corrected for comfort. For decision variables, in the interval The internal structure includes a candidate deceleration set. Candidate decelerations must satisfy the modified risk threshold constraint that the overall equivalent collision time under the action of the candidate decelerations must be within the required range. The comfort-corrected risk state must not transition to the safety-first AEB control mode. When the candidate deceleration set is empty, let... Conversely, within the candidate deceleration set, with the objectives of reducing discomfort at the most unfavorable representative seat, minimizing the imbalance in comfort experience among occupants in different positions within the vehicle, reducing discomfort in sensitive areas under the influence of critical conflicts, and ensuring sufficient safety margin for the current candidate deceleration, the Pareto nondominated solution set is solved. Taking into account risk adaptive coefficients, comfort distribution adaptive coefficients, and direction sensitivity adaptive coefficients, the optimal comfort-corrected deceleration for the current control cycle is calculated. This optimal comfort-corrected deceleration is then used as the target deceleration and converted into a control variable directly invoked by the execution layer. Finally, it is corrected based on a PID feedback control law to obtain the final braking control signal.

[0053] As a preferred embodiment of the present invention, the current control cycle in step S4 Optimal comfort-corrected deceleration:

[0054] ;

[0055] in, For the current control cycle Next candidate deceleration The corresponding Pareto non-dominated solution set, , , and These are the normalized results of the objective function values ​​for each candidate solution in the Pareto non-dominated solution set. , , The current control cycle The adaptive coefficients for risk, direction sensitivity, and comfort distribution are as follows.

[0056] The advantages and beneficial effects of this invention are as follows:

[0057] 1) To address the problem that existing AEB methods for commercial buses are unable to reflect the differences in dynamic perception of passengers in different seats during cornering braking at intersections, this invention establishes six representative seat position models and seat point response mapping relationships for front left, front right, middle left, middle right, rear left, and rear right. It also constructs acceleration impact, acceleration vibration, and pitch angle vibration indices for each representative seat to further identify the most unfavorable representative seat and characterize the comfort distribution characteristics inside the vehicle. This improves the AEB control's ability to identify differences in the riding experience of passengers in different seats of the bus, making the braking control more consistent with the actual riding characteristics of commercial buses.

[0058] 2) In response to the problem that existing commercial bus AEB systems do not make sufficient use of the overall comfort distribution in the vehicle and have difficulty identifying local high discomfort areas, this invention, in addition to considering the most unfavorable representative seat, further introduces the average comprehensive comfort index and the comfort distribution dispersion index in the vehicle to jointly characterize the overall level and unevenness of the comfort response of each representative seat. This improves the AEB control's overall perception of the comfort status of the group of occupants in the vehicle, so that the control results not only focus on the worst local position, but also take into account the balance of the feelings of occupants in different positions in the vehicle.

[0059] 3) To address the problem that existing commercial bus AEB methods lack utilization of the coupling relationship between the direction of external critical conflict vehicles, vehicle turning state, and the spatial distribution of occupants in intersection turning scenarios, this invention further constructs a direction-sensitive comfort index based on the comfort characterization of occupants in different seats. This index is a three-element coupling of conflict direction, vehicle turning direction, and representative seat spatial position. It is used to correct the AEB trigger threshold and optimize the target deceleration, improving the pertinence and scenario adaptability of comfort constraints for different conflict directions at intersections. This enables commercial bus AEB to more precisely coordinate external conflict risks and in-vehicle passenger comfort during turning collision avoidance.

[0060] 4) In view of the problem that existing AEB systems mostly use fixed trigger thresholds and single braking control methods, making it difficult to dynamically adjust the intervention timing and control strategy according to collision risk and comfort constraints, this invention adaptively corrects each threshold based on the overall equivalent collision time, the comprehensive comfort index of the most unfavorable representative seat, and the dispersion of in-vehicle comfort distribution. This improves the adaptive decision-making ability of the AEB system in the process of intersection turning risk evolution and the rationality of intervention timing.

[0061] 5) To address the issue that existing AEB systems lack flexibility in switching control objectives across different risk stages, making it difficult to balance collision mitigation needs with passenger comfort requirements, this invention constructs a switching mechanism between a comfort-priority AEB control mode and a safety-priority AEB control mode. The control mode is jointly determined based on the basic risk state, the corrected risk state, and the current comfort adjustment space. When the collision risk is within a controllable range and comfort constraints still have room for adjustment, the system adopts the comfort-priority AEB control mode, smoothly adjusting the braking deceleration. When the collision risk increases, or comfort adjustments may affect safe braking requirements, the system switches to the safety-priority AEB control mode to prioritize braking safety. This improves the AEB system's ability to coordinate and allocate safety and comfort at different risk levels, enabling the control strategy to transition reasonably from comfort to safety as the risk level changes. This enhances the flexibility and practicality of AEB control for commercial buses turning at intersections.

[0062] 6) Existing AEB systems, after determining braking intervention requirements, often output target braking demands using preset fixed values ​​or fixed gradations, making it difficult to adaptively adjust according to actual collision risks and comfort constraints. This invention optimizes the target braking deceleration and its rate of change online based on the collision risk state, the most unfavorable representative seat comfort state, and the in-vehicle comfort distribution characteristics. It also generates braking control commands through the inverse model of the braking system and feedback correction control based on deceleration errors. This improves the adaptability of target braking demand generation, as well as the deceleration tracking accuracy and smoothness during braking execution, enabling commercial buses to better balance safety, comfort, and stability during AEB intervention at intersections. Attached Figure Description

[0063] Figure 1 A flowchart of an AEB control method for commercial buses turning at intersections, taking into account passenger comfort, is provided by the present invention.

[0064] Figure 2 This is a nonlinear model diagram of a commercial bus. Detailed Implementation

[0065] To enable those skilled in the art to better understand the technical solutions and advantages of the present invention, the present application will be described in detail below with reference to the accompanying drawings, but this is not intended to limit the scope of protection of the present invention.

[0066] like Figure 1 As shown in the figure, this embodiment provides an AEB control method for commercial buses turning at intersections, taking into account passenger comfort. The method includes the following steps:

[0067] Step S1. Establish a turning braking dynamics model and a representative seat position model for the commercial bus:

[0068] A vehicle dynamics model for commercial buses under intersection turning conditions is established to describe the longitudinal, lateral, vertical, yaw, roll, and pitch motion characteristics of the vehicle. At the same time, based on the vehicle body reference point, position coordinate models and seat point response mapping relationships of six representative seats (front left, front right, middle left, middle right, rear left, and rear right) are constructed to provide a basis for subsequent dynamic response calculation and comfort evaluation of occupants in different seats.

[0069] Step S2. Obtain intersection turning scenario information and assess collision risk:

[0070] The system acquires the vehicle's operating status, intersection road geometry information, and target vehicle or obstacle information. Based on the relative position, relative speed, and future motion relationship between the vehicle and the target, it assesses the potential collision risk in intersection turning scenarios, outputs the risk status and basic AEB decision, and provides a basis for subsequent trigger threshold correction and control decisions.

[0071] Step S3. Calculate the responses of each representative seat occupant and jointly determine the AEB triggering timing and control mode:

[0072] Based on the vehicle dynamics model and the position coordinate model of the representative seats established in step S1, the response parameters of the six representative seats during the turning and braking process are calculated, the occupant comfort evaluation index of each representative seat is constructed, and the comfort state of the most unfavorable representative seat is determined. Further, combined with the key conflict vehicle index and its conflict direction information relative to the vehicle obtained in step S2, a direction-sensitive comfort index considering the coupling relationship between conflict direction, vehicle steering direction, and representative seat spatial position is constructed. Then, based on the collision risk results, the comfort state of the most unfavorable representative seat, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, the AEB trigger threshold is corrected, and the current control mode is jointly determined according to the collision risk level and occupant comfort constraints: comfort-priority control is adopted when the collision avoidance requirements are met, and safety-priority control is switched when the safety braking requirements cannot be met.

[0073] Step S4. Optimize the target braking demand and execute closed-loop correction control:

[0074] Based on the control mode determined in step S3, the target braking deceleration and its rate of change are optimized online by comprehensively considering the collision risk level, the most unfavorable representative seat comfort state, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, and braking control commands are generated. Through the execution layer correction control, AEB braking intervention that balances safety, comfort, and stability is achieved in the intersection turning scenario.

[0075] Furthermore, in this embodiment, step S1 specifically includes the following steps:

[0076] Step S11: Using the sprung mass center of gravity of the commercial bus Establish a vehicle coordinate system with the origin. ,in The axle is along the longitudinal direction of the vehicle's movement. The axle runs along the left side of the vehicle. The axis runs vertically upward. Treating the sprung parts of the vehicle body as a rigid body, we consider its three translational degrees of freedom and three rotational degrees of freedom: the longitudinal, lateral, and vertical motions of the vehicle reference point, as well as the pitch, roll, and yaw motions of the vehicle. Correspondingly, the longitudinal, lateral, and vertical velocities at the vehicle reference point are denoted as... , and ; , and Let represent the first derivatives of the above components with respect to time, i.e., the longitudinal, lateral, and vertical acceleration components of the vehicle reference point in the vehicle coordinate system. Further, let the ... The vertical displacement of the unsprung mass at each wheel position is ,in Indicates the front axle and the rear axle. To indicate the left and right wheel positions, and These represent the vertical velocity and vertical acceleration of the unsprung mass at that wheel position, respectively. The vehicle pitch angle, roll angle, and yaw angle are denoted as... , and Their angular velocities are respectively , and angular acceleration is , and At the same time, vertical degrees of freedom of unsprung mass and rotational degrees of freedom of wheel are set at the four wheel positions, and the coupling effect of suspension, tires and brake input on the vehicle's cornering and braking process is realized.

[0077] To highlight the key dynamic characteristics closely related to AEB control and occupant comfort under intersection turning braking conditions, the model makes the following assumptions: the vehicle body is a rigid body; the suspension only provides vertical spring damping; suspension geometric kinematic effects, tire self-centering torque, air resistance, and rolling resistance are ignored; the road surface on which the vehicle travels is approximately a flat and level road surface; the vehicle pitch angle and roll angle are both small angle changes, therefore a small angle approximation is adopted.

[0078] Let the total mass of the vehicle be... The sprung mass is The distances from the center of mass to the front and rear axles are respectively and The front and rear track widths are respectively and ; car body around , , The moments of inertia of the shafts are respectively , and The height of the vehicle's center of gravity is Since the vehicle's pitch and roll angles are both relatively small, the vertical displacements of the four suspension mounting points relative to the vehicle's reference point are uniformly expressed as:

[0079]

[0080]

[0081] in, These represent the front axle and the rear axle, respectively. Indicates the left and right wheel positions; Corresponding to the left wheel position, Corresponding to the right wheel position; This refers to the vertical displacement of the sprung mass of the vehicle body.

[0082] Based on this, a nonlinear dynamic model of the commercial bus is established. The longitudinal motion equation of the vehicle is:

[0083]

[0084] in, .

[0085] The equation of motion for the vehicle in the lateral direction is:

[0086]

[0087] in, .

[0088] The equation of motion for the vehicle's vertical movement is:

[0089]

[0090] in, .

[0091] No. The equation of motion for the unsprung mass at each wheel position is:

[0092]

[0093] in, This refers to the unsprung mass of the vehicle.

[0094] The equation of motion for vehicle body pitch is:

[0095]

[0096] The equation of motion for vehicle body roll is:

[0097] The equation of motion for the vehicle body yaw is:

[0098]

[0099] Among them, yaw moment Represented as: .

[0100] in, For the front wheel steering angle, and These are the longitudinal acceleration and lateral acceleration at the vehicle reference point, respectively.

[0101] Furthermore, to ensure the closure of the aforementioned vehicle dynamics model, models for tire forces, suspension forces, and normal loads at each wheel position are established. The longitudinal force, lateral force, and vertical force of each wheel position tire are expressed as follows:

[0102]

[0103]

[0104]

[0105] in, For the first The coefficient of friction between each tire and the road surface. For longitudinal slip ratio, Side slip angle, For tire normal load, and The first Vertical displacement and vertical velocity are input to the road surface at each wheel position. and The first The vertical equivalent stiffness and damping coefficient of each wheel position tire. and These represent the longitudinal force model and the lateral force model of the tire, respectively. Preferably, the Magic Formula tire model is used for calculation.

[0106] No. The vertical force of the suspension at each wheel position is expressed as:

[0107]

[0108] in, and The first Individual wheel suspension stiffness and damping coefficient and The first Vertical displacement and vertical velocity at the vehicle suspension mounting points at each wheel position.

[0109] No. The normal load of each wheel position tire is expressed as:

[0110]

[0111] in, For the first Static normal load on each wheel position.

[0112] To describe the tire kinematic characteristics under cornering and braking coupling conditions, tire slip angle models for each wheel position are established. The slip angles of the front and rear wheels are as follows:

[0113]

[0114]

[0115] Furthermore, the first The velocity components of the wheel center along the wheel plane are represented as follows:

[0116]

[0117]

[0118]

[0119]

[0120] Since this invention primarily addresses automatic emergency braking scenarios under braking conditions, the slip ratio expression under braking conditions is preferably used subsequently. The longitudinal slip ratio of each wheel position tire is:

[0121]

[0122] in, The effective rolling radius of the tire. For the first The wheel angular velocity at each wheel position.

[0123] Finally, the equations of motion for wheel rotation are established:

[0124]

[0125] in, For the first Moment of inertia of each wheel position Indicates the first The wheel angular acceleration of each wheel position; For the first The driving or braking torque received by each wheel position.

[0126] Step S12. Based on the commercial bus turning braking dynamics model established in step S11, in order to determine the differences in dynamic perception of occupants at different positions during the cornering braking process at an intersection, the vehicle body reference point is used. Establish a position coordinate model of six representative seats inside the vehicle based on the reference; wherein, the six representative seats are the front left seat Front right seat Middle left seat Middle right seat Rear left seat and the rear right seat .

[0127] The position coordinates of each representative seat relative to the vehicle body reference point are defined as follows:

[0128]

[0129] in, Indicates the first Each represents the longitudinal coordinate of the seat relative to a reference point on the vehicle body, with the front being positive; This represents the horizontal coordinate, with the left side being positive; This represents the vertical coordinate, with the top being positive.

[0130] This leads to the construction of a set of representative seat positions within the vehicle, i.e., a position coordinate model:

[0131]

[0132] The front left and front right seats represent the positions of occupants near the front overhang, the middle left and middle right seats represent the positions of occupants in the middle of the vehicle, and the rear left and rear right seats represent the positions of occupants near the rear overhang. This position coordinate model can simultaneously represent the impact of the differences in front-to-back and left-to-right positions of commercial buses on passenger comfort.

[0133] After obtaining the motion state of the vehicle body reference point in step S11, a mapping relationship between the vehicle body motion and the response of the representative seats is established in order to further obtain the dynamic response at each representative seat.

[0134] Let the absolute linear acceleration vector of the vehicle body reference point O be denoted as... , The transpose symbol represents the symbol, and its components are as follows: , , The angular velocity vector of the vehicle body is The angular acceleration vector of the vehicle body is ;No. The position vectors of each representative seat relative to the vehicle reference point are: Then the first The acceleration vector representing each seat can be expressed as:

[0135]

[0136] Therefore, we can obtain the first... The acceleration responses of each representative seat in the longitudinal, lateral, and vertical directions are denoted as follows:

[0137]

[0138] Furthermore, the first The rates of change of longitudinal acceleration and lateral acceleration for each representative seat are defined as follows:

[0139]

[0140]

[0141] in, and Used to characterize the feeling of braking lurch and braking impact. and Used to characterize the feeling of lateral swaying and discomfort from sudden lateral changes during cornering. Used to characterize the vertical additional vibration response; vehicle body pitch angle acceleration. and roll acceleration These are used to characterize the effects of the vehicle body pitching and roll effects on the comfort of occupants in each representative seat.

[0142] Step S12 establishes a mapping relationship from the vehicle dynamics state to the response states of six representative seats, providing a foundation for the subsequent steps S3 to construct occupant comfort evaluation indicators for each representative seat, identify the most unfavorable representative seat, and jointly determine the AEB triggering timing and control mode.

[0143] Furthermore, in this embodiment, step S2 specifically includes the following steps:

[0144] Step S21. To facilitate subsequent planar trajectory prediction and collision assessment, a fixed coordinate system for the road is established in the ground plane. Let the position of the vehicle body reference point in the fixed road coordinate system be... The first one within the sensing range of the intersection The positions of the target vehicle body reference points in the road fixed coordinate system are: ,in, , This represents the total number of target vehicles within the intersection at the current moment.

[0145] The vehicle's operating status information includes the longitudinal velocity at a reference point on the vehicle body. lateral velocity Longitudinal acceleration lateral acceleration Front wheel steering angle lateral angle and yaw rate ;No. The operational status information of each target vehicle includes its longitudinal velocity at its body reference point. lateral velocity Longitudinal acceleration lateral acceleration lateral angle and yaw rate The vehicle's operating status information can be obtained in real time from the vehicle dynamics model and onboard sensors established in step S11, while the target vehicle's operating status information can be directly measured or estimated by the environmental perception module.

[0146] Construct the set of target vehicle states within the intersection at the current moment as follows And combine this set of states with the current vehicle state. As input for the minimum jerk trajectory prediction in step S22.

[0147] Step S22. Predict the future turning trajectories of this vehicle and each target vehicle based on the minimum jerk theory:

[0148] After obtaining the state information of the vehicle and each target vehicle within the intersection in step S21, a two-dimensional minimum jerk turning trajectory prediction model is established for both the vehicle and each target vehicle. Let the vehicle... In the prediction time domain The predicted trajectory within the plane is:

[0149]

[0150] in, This vehicle, Indicates the first The target vehicle. For vehicles The undetermined planar trajectory function in the future prediction time domain.

[0151] To ensure the predicted trajectory reflects the continuity and smoothness of the vehicle's turning process, the objective is to minimize the square integral of the third derivative of the trajectory in the prediction time domain. The minimum jerk performance index is:

[0152]

[0153] in, and Representing vehicles Under the fixed coordinate system of the road direction and The jerk of direction. Minimize under given boundary constraints. It can obtain the smooth turning trajectory of the vehicle in the future short time domain.

[0154] (1) Initial boundary conditions:

[0155] At the current control moment Location, vehicle The initial position boundary is:

[0156]

[0157] vehicle The initial velocity boundary in the road fixed coordinate system is defined by the longitudinal velocity at its vehicle reference point. lateral velocity and yaw angle To be determined together, namely:

[0158]

[0159]

[0160] Furthermore, in combination with vehicles longitudinal acceleration lateral acceleration and yaw rate The initial acceleration boundary of the road in the fixed coordinate system can be obtained as follows:

[0161]

[0162]

[0163] For this vehicle Its front wheel steering angle Used to describe the current turning input state, and through the vehicle's current... , , , , and These are all reflected in the initial boundary conditions mentioned above.

[0164] (2) Final boundary conditions:

[0165] To avoid overly complex prediction models, in a single prediction time domain Inside, assuming the vehicle If the longitudinal acceleration, lateral acceleration, and yaw rate remain approximately constant, then the vehicle The yaw angle, longitudinal velocity, and lateral velocity at the predicted endpoint are respectively expressed as:

[0166]

[0167]

[0168]

[0169] Accordingly, the predicted longitudinal and lateral accelerations at the endpoint can be approximated as:

[0170]

[0171]

[0172] Based on the current velocity and acceleration state, a second-order extrapolation is performed on the predicted endpoint position to obtain the final position boundary:

[0173]

[0174]

[0175] From the final yaw angle Final longitudinal velocity and final lateral velocity The final velocity boundary can be constructed:

[0176]

[0177]

[0178] Similarly, the final acceleration boundary is expressed as:

[0179]

[0180]

[0181] (3) Solving for the minimum jerk trajectory:

[0182] Under the aforementioned initial and final boundary conditions, the vehicle The planar trajectories in the prediction time domain are represented in fifth-order polynomial form:

[0183]

[0184]

[0185] in, and The coefficients to be determined are: By substituting the initial and final boundary conditions for position, velocity, and acceleration into the aforementioned polynomials and their first and second derivative expressions, the coefficients of each polynomial can be uniquely determined, thereby obtaining the vehicle's... In the prediction time domain Two-dimensional minimum jerk prediction trajectory within .

[0186] Therefore, by solving for the problem for both the vehicle itself and each target vehicle within the intersection, we can obtain the following results:

[0187]

[0188] That is, the future turning trajectories of this vehicle and all target vehicles in the intersection within the predicted time domain, which provides the basis for constructing the dynamic safety envelope and calculating the equivalent collision time in the subsequent step S23.

[0189] Step S23: In step S22, obtain the prediction time domain of the vehicle and each target vehicle within the intersection. After determining the future turning trajectory within the prediction time domain, based on the position and attitude changes of each vehicle in the prediction time domain, the dynamic safety envelopes of the vehicle and each target vehicle are constructed respectively, and it is determined whether they overlap in the prediction time domain, thereby calculating the TTC between the vehicle and each target vehicle.

[0190] Based on the trajectory results obtained in step S22, this vehicle With the The target vehicle at time The vehicle body reference point positions are respectively and The corresponding yaw angles are respectively and .

[0191] A rectangular dynamic safety envelope, centered on each vehicle body reference point, is established to vary with the vehicle's attitude. Let the vehicle... The body length is The width of the vehicle body is The longitudinal safety margin is The lateral safety margin is Then the vehicle The dynamic safety envelope half-length and half-width are defined as follows:

[0192]

[0193] Let any point in the plane The coordinates in the road fixed coordinate system are: Then the point is relative to the vehicle The position vector of the vehicle body reference point can be represented as:

[0194]

[0195] Furthermore, this position vector is expressed through the yaw angle. Transformation to vehicle In the vehicle body coordinate system, we can obtain:

[0196]

[0197] If satisfied Then the point is considered Falling into the vehicle At any moment Within the dynamic safety envelope. Therefore, the vehicle... At any moment The dynamic safety envelope can be represented as:

[0198]

[0199] Based on the above definition, in the prediction time domain Inside, determine the dynamic safety envelope of the vehicle respectively. Dynamic safety envelope of each target vehicle Does overlap occur? If so, at what time? Make Then it is considered that this car is related to the first one. The target vehicle is at potential collision risk at that moment.

[0200] Furthermore, define this vehicle and the first The equivalent collision time between the target vehicles is:

[0201]

[0202] If within the entire prediction time domain, this vehicle and the... If the dynamic safety envelopes of all target vehicles do not overlap, then let .

[0203] Based on this, the equivalent collision times of all target vehicles within the intersection are combined, and the minimum value is taken as the overall equivalent collision time of this vehicle facing inter-vehicle conflict at the intersection under the current control cycle, that is:

[0204]

[0205] Meanwhile, the target vehicle that meets the minimum equivalent collision time is designated as the critical conflict vehicle, and its index is represented as:

[0206]

[0207] To further describe the approach characteristics of the key conflict vehicle relative to this vehicle, in the current control cycle Below, the position of the vehicle body reference point in the road fixed coordinate system is: The position of the reference point of the key conflict vehicle body in the road fixed coordinate system is as follows: The current yaw angle of this vehicle is The position increment of the critical conflict vehicle relative to the main vehicle is defined as:

[0208]

[0209]

[0210] Transforming it to the vehicle's own coordinate system, we get:

[0211]

[0212]

[0213] Therefore, the angle of approach of the critical conflict vehicle relative to itself is defined as:

[0214]

[0215] in, ,when When the value is close to 0, it indicates that the key conflict vehicle is mainly located in front of this vehicle from the direction of approach; when When, it indicates that the key conflict vehicle is mainly located on the left side of this vehicle from the direction of approach; when At that time, it indicated that the key conflict vehicle was mainly located on the right side of this vehicle from the direction of approach.

[0216] Therefore, step S23 finally outputs the equivalent collision time for each target vehicle. Overall equivalent collision time Key Conflict Vehicle Index and the angle of approach of the key conflict vehicles This provides a basis for subsequent collision risk assessment and basic AEB intervention requirements output.

[0217] Step S24: Obtain the overall equivalent collision time in step S23. Then, based on the preset warning threshold and two-stage braking threshold , The risk of collisions between turning vehicles at the current intersection is classified, and the corresponding basic AEB intervention requirements are output. Among these, , , .

[0218] Define the current basic risk status for:

[0219]

[0220] in, Indicates a safe state; Indicates a warning status; Indicates the first-level braking state; This indicates a level two braking state.

[0221] Based on the basic risk status Output the basic target deceleration And the corresponding intervention requirements. Among them, the target deceleration for primary braking is... The target deceleration for secondary braking is .

[0222] Then we have:

[0223]

[0224] when When the system does not trigger AEB, it only maintains monitoring; when When, the system outputs an audible and visual warning signal; when At this point, the system enters the first-level braking phase, outputting the basic target deceleration. Prioritize collision mitigation with a gentler braking intensity; when At this point, the system enters the second-level braking phase, outputting the basic target deceleration. To address the risk of imminent collisions with higher braking intensity.

[0225] Furthermore, in this embodiment, step S3 specifically includes the following steps:

[0226] Step S31. Obtain the acceleration response of each representative seat in step S12. , , and the rate of change of acceleration , Then, at a length of Within the sliding evaluation window, the first... The indicators for acceleration impact, acceleration vibration, and pitch angle vibration are representative of the seats. .

[0227] No. The acceleration impact index for each representative seat is defined as follows:

[0228]

[0229] in, This is the weighting coefficient for the rate of change of lateral acceleration.

[0230] No. The acceleration vibration index for each representative seat is defined as follows:

[0231]

[0232] in, and These are the weighting coefficients for longitudinal acceleration and lateral acceleration, respectively. Used to characterize the discomfort caused by braking and deceleration. Used to characterize discomfort caused by lateral swaying during cornering.

[0233] Considering the differences in comfort perceived by occupants at different positions due to vehicle pitch vibration, the vehicle pitch angle acceleration is used as the metric. Representing the longitudinal position of the seat Construct the first The pitch angle vibration index represents the seat position:

[0234]

[0235] Normalizing the above three sub-indicators, we get:

[0236]

[0237] in, , and These are the normalized acceleration impact index, acceleration vibration index, and pitch angle vibration index, respectively. , and These are the reference values ​​for acceleration impact, acceleration vibration, and pitch angle vibration, respectively.

[0238] Furthermore, construct the first The overall comfort index of each representative seat:

[0239]

[0240] in, , and These are the weighting coefficients for acceleration impact, acceleration vibration, and pitch angle vibration, respectively. Considering the impact of sudden longitudinal deceleration, in-plane acceleration fluctuations, and vehicle pitch response on occupant comfort in AEB scenarios at intersections, the preferred method is... , , .

[0241] Step S32. In step S31, obtain the comprehensive comfort index of the six representative seats. Then, the spatial distribution characteristics of the most unfavorable occupant position in the vehicle at the current moment and the comfort response of each representative seat are compared and statistically analyzed to evaluate the comprehensive comfort index of the six representative seats.

[0242] The most unfavorable representative seat index at the current moment is defined as:

[0243]

[0244] The corresponding most unfavorable representative seat's overall comfort index is:

[0245]

[0246] in, , The larger the value, the higher the level of discomfort for the occupant in the most unfavorable seat at that moment.

[0247] Furthermore, to describe the overall level of the six representative seat comfort indices, the average in-vehicle comfort index is defined as:

[0248]

[0249] To describe the consistency of comfort response across representative seats, the in-vehicle comfort distribution dispersion index is defined as:

[0250]

[0251] in, The larger the value, the more significant the difference in comfort between the representative seats, meaning the greater the unevenness in the perception of the current braking conditions among occupants in different positions within the vehicle. The smaller the value, the more uniform the comfort response of each representative seat.

[0252] Furthermore, to reflect the coupled influence of the approach characteristics of key conflict vehicles, the current turning direction of the vehicle, and the representative seat position inside the vehicle on occupant comfort, a direction-sensitive comfort index is constructed in step S32. Let the index obtained in step S31 be... The overall comfort index for each representative seat is: In step S12 The longitudinal and lateral coordinates of each representative seat relative to the vehicle reference point are respectively... and Define the seat longitudinal position normalization factor and the lateral position normalization factor as follows:

[0253]

[0254] when When, it indicates that the seat is located in front of the vehicle's reference point; when When, it indicates that the seat is located behind the vehicle's reference point; when When, it indicates that the seat is located on the left side of the vehicle; when When the position is indicated, it means that the seat is located on the right side of the vehicle. Using the above normalization method, position variables in different directions can be uniformly mapped to a dimensionless interval without changing the relative spatial distribution of the seats.

[0255] The vehicle steering direction coefficient is defined based on the sign of the vehicle's current yaw rate:

[0256]

[0257] in, This indicates that the current trend is to turn left. This indicates a current trend towards a right turn. Combined with the critical conflict vehicle collision angle obtained in step S23... Construct the first The original weights for the orientation sensitivity of each representative seat are:

[0258]

[0259] Among them, the second term ( ) is used to describe the moderating effect of the forward and backward orientation of key conflict vehicles on the weighting of representative seats, the third term ( ) is used to describe the moderating effect of the left and right approach directions of key conflict vehicles on the weighting of representative seats, the fourth item ( This is used to describe the moderating effect of the vehicle's turning outer edge effect on the weights of representative seats, thereby giving higher directional sensitivity weights to representative seats that are more relevant to the current critical conflict direction and vehicle steering state. To ensure that the weights of each representative seat are positive and to satisfy the normalization condition, the first... The orientation sensitivity weight for each representative seat is:

[0260]

[0261] in, To prevent the numerator or denominator from having extremely small positive numbers that are not positive.

[0262] Therefore, the current control cycle is defined. The overall comfort index for direction sensitivity is:

[0263]

[0264] Therefore, the overall comfort index of direction sensitivity It can comprehensively characterize the coupled effects of the direction of approach of key conflict vehicles, the direction of vehicle turning, and the distribution of representative seating space inside the vehicle on occupant comfort.

[0265] Step S33. Obtain the overall equivalent collision time in step S24. Basic risk status Warning threshold Level 1 braking threshold and secondary braking threshold In step S32, the most unfavorable representative seat's overall comfort index is obtained. In-vehicle comfort distribution dispersion index And after the direction-sensitive comprehensive comfort index To balance collision risks and passenger comfort in intersection turning scenarios, this invention constructs a risk partition agent optimization model for AEB trigger threshold correction, adaptively corrects the AEB trigger threshold, and determines the current control mode.

[0266] Let the current state vector be:

[0267]

[0268] Set the AEB trigger threshold (early warning threshold) Two-stage braking threshold , The correction vector is:

[0269]

[0270] in, , and These represent the comfort correction amounts for the warning threshold, the first-level braking threshold, and the second-level braking threshold, respectively.

[0271] Considering that the trade-off between collision risk and comfort differs under different basic risk states, the state space is divided according to the basic risk states obtained in step S24. Partitioning is performed, and risk sub-regions are defined as follows:

[0272]

[0273] in, Indicates a warning status. Indicates the first-level braking state. Indicates a level 2 braking state; when At this time, the system is in a safe state and no threshold correction optimization is performed.

[0274] For each risk sub-region A Latin hypercube experimental design was employed to sample state variables and threshold corrections. Combined with the vehicle dynamics model established in steps S11 to S32, the representative seat response mapping relationship, comfort evaluation indices, and collision risk assessment results, the comprehensive comfort index of the most unfavorable representative seat, the dispersion of in-vehicle comfort distribution, the orientation-sensitive comfort index, and the minimum equivalent collision time in the predicted time domain were calculated under different correction values. Based on this, surrogate models were established for each risk sub-region.

[0275]

[0276]

[0277]

[0278]

[0279] in, This represents the predicted value of the most unfavorable representative seat's overall comfort index after adjusting for the threshold effect. This represents the predicted dispersion value of the in-vehicle comfort distribution after applying the corrected threshold. This represents the predicted value of the direction-sensitive comfort index after adjusting for the threshold effect. This represents the predicted minimum equivalent collision time in the time domain after the threshold adjustment.

[0280] Furthermore, in each risk sub-region Within this framework, we construct the following multi-objective optimization problem:

[0281]

[0282]

[0283]

[0284]

[0285] Wherein, objective function The objective function is used to reduce discomfort at the most unfavorable representative seat. The objective function is used to reduce the unevenness in comfort experienced by occupants in different positions within the vehicle. The objective function is used to reduce the discomfort experienced by occupants in sensitive areas under the influence of critical conflicts. This is used to ensure that the collision avoidance safety margin is still met after the correction.

[0286] The constraints are:

[0287]

[0288]

[0289]

[0290]

[0291] in, , and These are the lower bounds of the trigger thresholds at each level to meet basic safety requirements.

[0292] Based on the agent model for each risk sub-region, the NSGA-II algorithm is used to solve for the Pareto optimal solution set, denoted as . .

[0293] To avoid the inadequacy of traditional fixed-weight solution selection methods in scenarios with rapidly changing risks at intersections, this invention further constructs risk adaptive coefficients and comfort distribution adaptive coefficients.

[0294] Define the risk adaptive coefficient as:

[0295]

[0296] in, It is a saturation function, and its output range is .when near hour, A smaller value indicates a relatively low current collision risk, allowing for greater flexibility in comfort adjustments; when near hour, A larger value indicates an increased risk of collision, and safety should be prioritized.

[0297] Define the adaptive coefficient of comfort distribution as:

[0298]

[0299] in, To prevent extremely small positive numbers with a denominator of zero. The larger the value, the more significant the difference in comfort between different seats inside the vehicle; in this case, the comfort level should be increased. Optimize the weights; The smaller the value, the lower the level of imbalance in comfort within the vehicle. In this case, priority should be given to reducing the overall discomfort of the most unfavorable seat. .

[0300] Define the direction-sensitive adaptive coefficient as:

[0301]

[0302] in, The larger the value, the closer the discomfort level of occupants in the sensitive area under the influence of the critical conflict is to the most unfavorable representative seat discomfort level. At this point, the directional sensitivity comfort index should be increased. Optimize the weights; The smaller the value, the weaker the current orientation sensitivity effect, and the more likely it is to reduce the overall discomfort level of the most unfavorable representative seat.

[0303] The objective function values ​​of each candidate solution in the Pareto optimal solution set are normalized, denoted as . , , and The optimal threshold correction at the current moment is determined by the following online adaptive solution selection criterion:

[0304]

[0305] in, .

[0306] Therefore, the corrected trigger threshold is:

[0307]

[0308]

[0309]

[0310] When the risk is low, When the risk is relatively high, the threshold correction effect is more significant, and the timing of AEB intervention can be appropriately delayed to improve comfort; when the risk increases, The correction is gradually reduced, and the magnitude of the correction automatically shrinks, causing the trigger threshold to revert to the basic safety threshold.

[0311] Furthermore, the comfort correction risk status is determined based on the revised trigger threshold. :

[0312]

[0313] Based on the basic risk status With comfort correction risk status Determine the current control cycle Control mode under for:

[0314]

[0315] in, This indicates the comfort-priority AEB control mode. This indicates the AEB control mode with safety as the priority.

[0316] when Furthermore, when the risk level decreases after threshold correction, it indicates that, under the premise of meeting basic safety constraints, the timing of warning or braking intervention can be appropriately delayed, and the comfort-priority AEB control mode should be adopted at this time; when If the risk level does not decrease after threshold correction, it indicates that the current collision risk is high or there is insufficient space for comfort adjustment. In this case, the safety-first AEB control mode is adopted.

[0317] Step S33 finally outputs the corrected warning threshold. Level 1 braking threshold Secondary braking threshold Comfort-related risk status and control mode This is used for target braking demand optimization and closed-loop correction control in subsequent step S4.

[0318] Furthermore, in this embodiment, step S4 specifically includes the following steps:

[0319] Step S41. Obtain the basic risk status in step S24. and basic target deceleration And obtain the current control mode in step S33. Then, the generation path of the target braking demand is determined based on the current control mode.

[0320] when When this indicates that there is still some room for comfort adjustment under the premise of meeting basic safety constraints, the current collision risk is entered into the comfort-first target deceleration generation path, and in step S42, the comprehensive comfort index of the most unfavorable representative seat is combined. In-vehicle comfort distribution dispersion index In addition to collision risk constraints, a comfort-oriented target deceleration optimization problem is established and solved to obtain the current control cycle. Comfort-corrected deceleration .

[0321] when If the current collision risk is high, or there is insufficient room for comfort adjustment under the current operating conditions, then comfort optimization will no longer be performed. Instead, the basic target deceleration output in step S24 will be directly adopted. As the current control cycle The safety-first objective is to reduce speed to ensure that the system prioritizes collision mitigation requirements.

[0322] Step S42. When the current control mode is determined in step S41. This indicates that the current collision risk still has some room for comfort adjustment while meeting basic safety constraints. At this point, deceleration is adjusted for comfort. As the decision variable, the basic target deceleration output in step S24 Within the scope of this study, the commercial bus turning and braking dynamics model and representative seat response mapping relationship established in steps S11 and S12, the future trajectory prediction obtained in steps S22 and S23, the overall equivalent collision time and the calculation results of the collision angle of the key conflict vehicles, and the comprehensive comfort index of the most unfavorable representative seat obtained in steps S31, S32 and S33 are combined. In-vehicle comfort distribution dispersion index Direction-sensitive comfort index Corrected trigger threshold , and In addition to comfort-oriented risk management, a comfort-oriented target deceleration optimization problem is established.

[0323] Preferably, in the interval The set of candidate decelerations is constructed internally. For any candidate deceleration... Substituting these values ​​into the vehicle dynamics model and seat point response mapping relationship corresponding to steps S11 and S12, the response results of six representative seats under the candidate deceleration are obtained. Then, following the calculation method established in steps S31 and S32, the comprehensive comfort index and in-vehicle comfort distribution dispersion index of the most unfavorable representative seat corresponding to the candidate deceleration are directly calculated. Simultaneously, the candidate deceleration is substituted into the trajectory prediction and dynamic safety envelope judgment process corresponding to steps S22 and S23 to recalculate its corresponding overall equivalent collision time. Combined with the corrected warning threshold, first-level braking threshold, and second-level braking threshold output in step S33, the comfort correction risk state under the candidate deceleration is determined.

[0324] Candidate decelerations that meet the following conditions The following conditions are retained as feasible solutions: First, the overall equivalent collision time under the candidate deceleration satisfies the risk threshold constraint corrected in step S33; second, the comfort-corrected risk state under the candidate deceleration does not transition to the safety-first AEB control mode. Candidate decelerations that do not meet the above conditions are eliminated. Therefore, the current control cycle can be obtained. The set of feasible candidate decelerations that satisfy the safety constraints.

[0325] When the set of feasible candidate decelerations is empty, it indicates that there is no deceleration solution that simultaneously satisfies comfort optimization and safety constraints within the current control cycle. In this case, simply set... This allows the target deceleration to automatically return to the basic safety braking requirements. Conversely, within the set of feasible candidate decelerations, the current control cycle is constructed with the objectives of minimizing the most unfavorable representative's overall seat comfort index, minimizing the in-vehicle comfort distribution dispersion index, minimizing the direction-sensitive comfort index, and maximizing the safety margin. Multi-objective optimization problem:

[0326]

[0327]

[0328]

[0329]

[0330] Wherein, objective function The objective function is used to reduce discomfort at the most unfavorable representative seat. The objective function is used to reduce the unevenness in comfort experienced by occupants in different positions within the vehicle. The objective function is used to reduce the discomfort experienced by occupants in sensitive areas under the influence of critical conflicts. This is used to ensure that the current candidate deceleration still has a sufficient safety margin.

[0331] Furthermore, the current control period is constructed for all feasible candidate decelerations. The Pareto non-dominated solution set is used, and the risk adaptive coefficients established in step S33 are adopted. Adaptive coefficient of comfort distribution and direction-sensitive adaptive coefficient The solution is selected using an online adaptive selection criterion, thereby obtaining the current control cycle. Optimal comfort-corrected deceleration:

[0332] ;

[0333] in, For the current control cycle The Pareto non-dominated solution set corresponding to the next feasible candidate deceleration. , , and These are the normalized results for each objective function value.

[0334] Step S43. Determine the current control mode based on the basic target deceleration output in steps S41 and S42. and comfort correction deceleration Determine the target deceleration for the current control cycle. ,Right now:

[0335]

[0336] in, This indicates that the current mode is comfort-priority AEB control mode. This indicates that the system is currently in the safety-priority AEB control mode.

[0337] To decelerate the target Converted into a control quantity that can be directly invoked by the execution layer, preferably, the target braking pressure. As a control signal for the braking system, it determines the required braking force for the current control cycle based on the vehicle's longitudinal braking demand. It can be represented as:

[0338]

[0339] in, This indicates the equivalent braking force that the braking system needs to provide during the current control cycle, excluding the rolling resistance and air resistance naturally present during vehicle movement. For vehicle quality, For vehicle weight, The rolling resistance coefficient, The air drag coefficient, The vehicle's frontal area. air density, The current vehicle speed. Based on the mapping relationship between braking force and braking pressure: The feedforward braking pressure, obtained by inversely calculating the target deceleration, is:

[0340]

[0341] in, This is the proportionality coefficient between braking force and braking pressure. The feedforward braking pressure is obtained from the inverse model for the current control cycle.

[0342] Considering that the feedforward braking pressure obtained solely from the inverse model cannot completely eliminate the impact of modeling errors, parameter disturbances, and changes in external operating conditions on deceleration tracking accuracy, a feedback correction controller based on the deceleration error is further constructed. The deceleration tracking error for the current control cycle is defined as:

[0343]

[0344] in, This represents the vehicle's current actual longitudinal deceleration. Preferably, a PID feedback control law is used to correct the braking pressure, and the feedback correction amount is... It can be represented as:

[0345]

[0346] in, , and These are the proportional coefficient, integral coefficient, and differential coefficient, respectively. To control the cycle.

[0347] The feedforward braking pressure and the feedback correction pressure are superimposed to obtain the original braking control signal for the current control cycle:

[0348]

[0349] Furthermore, to avoid excessively large sudden changes in the braking control signal, [the following measures are taken]... By applying amplitude constraints, the final output braking system control signal is obtained:

[0350]

[0351] in, This is the maximum braking pressure that the braking system is allowed to output.

[0352] Preferably, in this embodiment, the PID controller parameters , and Offline tuning can be performed using particle swarm optimization algorithms to improve target deceleration tracking accuracy and braking smoothness under different operating conditions.

[0353] Therefore, step S43 determines the target deceleration based on the current control mode. The feedforward braking pressure was obtained through the inverse model of the braking system. Combined with PID feedback correction based on deceleration tracking error Ultimately, this generates the braking system control signal. It is then output to the braking actuator for execution, thereby achieving accurate tracking of the target deceleration and reducing sudden changes in commands and passenger discomfort during the braking process.

[0354] The present invention also provides an electronic device, comprising: one or more processors and a memory; wherein the memory is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the above-described method for AEB control of commercial buses turning at intersections, taking into account passenger comfort.

[0355] The present invention also provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for AEB control of commercial buses turning at intersections, taking into account passenger comfort.

[0356] Those skilled in the art will understand that all or part of the functions of the various methods / modules in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved.

[0357] In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the programs can also be stored in storage media such as servers, other computers, disks, optical discs, flash drives, or portable hard drives. They can be downloaded or copied to the memory of the local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be implemented.

[0358] The above describes specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for AEB (Automatic Emergency Braking) control of commercial buses turning at intersections, considering passenger comfort, characterized in that: The method includes the following steps: Step S1. Establish a vehicle dynamics model of a commercial bus under the turning condition at an intersection; at the same time, based on the vehicle body reference point, construct a position coordinate model representing the seats and the seat point response mapping relationship; Step S2. Obtain information about the vehicle and the target vehicle or obstacle, assess the potential collision risk in the intersection turning scenario, identify the key conflict vehicle, and output the basic risk status and basic AEB decision. Step S3. Based on the vehicle dynamics model and the position coordinate model of the representative seats established in Step S1, calculate the response parameters of the representative seats during the cornering and braking process, construct the comprehensive comfort index of each representative seat, and determine the comfort state of the most unfavorable representative seat; at the same time, combined with the key conflict vehicle and its conflict direction information relative to the vehicle obtained in Step S2, construct a direction-sensitive comfort index that considers the coupling relationship between conflict direction, vehicle steering direction, and representative seat spatial position; then, based on the collision risk results, the comfort state of the most unfavorable representative seat, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, correct the AEB trigger threshold; finally, combine the basic risk state determined based on the trigger threshold before correction and the comfort-corrected risk state determined based on the trigger threshold after correction to determine the current control mode; Step S4. Based on the control mode determined in step S3, the target braking deceleration is optimized online by comprehensively considering the collision risk level, the most unfavorable representative seat comfort state, the in-vehicle comfort distribution characteristics, and the direction-sensitive comfort index, and braking control commands are generated.

2. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 1, is characterized in that... The vehicle dynamics model described in step S1 adopts a nonlinear dynamics model, establishing the vehicle's longitudinal motion equation, lateral motion equation, vertical motion equation, and so on. The equations of motion for the unsprung mass of each wheel position are as follows: vertical motion equation for vehicle body pitch, roll, and yaw. The models of tire force, suspension force, and normal load for each wheel position are also established. The tire slip angle model and wheel rotation dynamics equation for each wheel position are also established.

3. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 2, is characterized in that... In step S1, the vehicle body reference point is used. Establish a position coordinate model of six representative seats inside the vehicle based on the reference; wherein, the six representative seats are the front left seat Front right seat Middle left seat Middle right seat Rear left seat and the rear right seat ; The seat point response mapping relationship is a mapping relationship from vehicle body movement to representative seat response, used to obtain the first... Acceleration response of each representative seat in the longitudinal, lateral, and vertical directions. and the The rate of change of longitudinal acceleration for each representative seat and lateral acceleration rate of change .

4. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 3, is characterized in that... Step S2 includes the following steps: Step S21. Based on the vehicle dynamics model established in step S1 and the on-board sensors, obtain the vehicle's operating status information in real time, and directly measure or estimate the target vehicle's operating status information through the environmental perception module. Step S22. Based on the two-dimensional minimum jerk turning trajectory prediction model, predict the future smooth turning trajectories of the vehicle and each target vehicle respectively; Step S23. Based on the position and attitude changes of each vehicle in the prediction time domain, construct the dynamic safety envelopes of the vehicle and each target vehicle respectively, and determine whether they overlap in the prediction time domain. Then calculate the equivalent collision time between the vehicle and each target vehicle, and take the minimum equivalent collision time as the overall equivalent collision time of the vehicle facing the intersection vehicle-to-vehicle conflict in the current control cycle. The target vehicle that meets the minimum equivalent collision time is designated as the critical conflict vehicle; at the same time, the angle of approach of the critical conflict vehicle relative to the vehicle itself is determined. ; Step S24. Based on the overall equivalent collision time Preset warning threshold Level 1 braking threshold and secondary braking threshold The basic risk status of current intersection turning vehicle conflicts. R The system is categorized and the corresponding basic AEB intervention requirements are output; when When, AEB is not triggered; when When, output an audible and visual warning signal; when When entering the first-level braking phase, the first-level braking target deceleration is output and used as the base target deceleration; when At that time, it enters the second-level braking stage, outputs the second-level braking target deceleration, and uses it as the basic target deceleration.

5. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 4, is characterized in that... The two-dimensional minimum jerk turning trajectory prediction model described in step S2 minimizes the prediction time domain. The vehicle is constructed using the square integral of the third derivative of the inner trajectory as the objective. The minimum jerk performance index is: ； in, and Representing vehicles Under the fixed coordinate system of the road direction and The jerk of direction; minimized under given initial position boundary, initial acceleration boundary, final position boundary, and final acceleration boundary condition constraints. Obtain the vehicle in the prediction time domain The smooth turning trajectory within the range, i.e., the two-dimensional minimum jerk prediction trajectory. .

6. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 5, is characterized in that... Step S3 includes the following steps: Step S31. Based on the acceleration response of each representative seat in the longitudinal, lateral, and vertical directions. and the rate of change of longitudinal acceleration and lateral acceleration rate of change , construct the first The system first generates three sub-indices: acceleration impact index, acceleration vibration index, and pitch angle vibration index. Then, these three sub-indices are normalized to obtain normalized acceleration impact index, acceleration vibration index, and pitch angle vibration index. Finally, a weighted fusion method is used to construct the first... A comprehensive comfort index representing each seat; Step S32. Based on the comprehensive comfort index of the six representative seats obtained in step S31, determine the most unfavorable representative seat at the current moment and its corresponding comprehensive comfort index, and statistically analyze the average comprehensive comfort index and the dispersion index of comfort distribution in the vehicle; at the same time, combined with the key conflict vehicle and its conflict direction information relative to the vehicle obtained in step S2, construct a direction-sensitive comfort index that considers the coupling relationship between conflict direction, vehicle turning direction and representative seat spatial position. Step S33. Construct a risk zoning proxy model for AEB trigger threshold correction, adaptively correct the AEB trigger threshold, and determine the current control mode. The trigger threshold includes a warning threshold, a first-level braking threshold, and a second-level braking threshold. The risk zoning proxy model first distinguishes basic risk states based on the state space. For each risk sub-region, a proxy model corresponding to each risk sub-region is established. The goal is to reduce the discomfort at the most unfavorable representative seat, reduce the imbalance in comfort experience among occupants in different positions within the vehicle, reduce the discomfort of occupants in sensitive areas under the influence of critical collisions, and ensure that the collision avoidance safety margin is still met after correction. Under constraints, the NSGA-II algorithm is used to solve the Pareto optimal solution set to initially determine the trigger threshold correction amount. Then, the objective function values ​​of each candidate solution in the Pareto optimal solution set are normalized, and the risk adaptive coefficient, comfort distribution adaptive coefficient, and direction sensitivity adaptive coefficient are comprehensively considered to calculate the optimal threshold correction amount at the current moment, thereby obtaining the corrected trigger threshold. The comfort correction risk status is then determined based on the revised trigger threshold. And based on the basic risk status With comfort correction risk status Determine the control mode under the current control cycle. for: ； in, This indicates the comfort-priority AEB control mode. This indicates the AEB control mode with safety as the priority.

7. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 6, is characterized in that... In step S3 Acceleration impact index for each representative seat for: ； in, The weighting coefficient for the rate of change of lateral acceleration; No. An acceleration vibration index representing a seat. for: ； in, and These are the weighting coefficients for longitudinal acceleration and lateral acceleration, respectively. No. Pitch angle vibration index of a representative seat for: ； in, The acceleration is the pitch angle of the vehicle body. This represents the longitudinal position of the seat; Direction-sensitive comprehensive comfort index for: ； ； ； in, For the first Each representative seat has a direction-sensitive weight. For the first Each representative seat represents an overall comfort index. To prevent positive numbers from appearing in the numerator or denominator that are not positive, For the first The original weights of the seats are sensitive to orientation. The normalized values ​​of the longitudinal and lateral coordinates of each representative seat relative to the vehicle reference point are respectively... , The vehicle's steering direction coefficient is the vehicle's current yaw rate. , Conversely, it is 0.

8. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 7, is characterized in that... The optimal threshold correction amount at the current time in step S3 Determined through the following online adaptive solution selection criteria: ； Revised warning threshold Level 1 braking threshold Secondary braking threshold for: ； ； ； in, , These are the optimal correction values ​​for the warning threshold, the first-level braking threshold, and the second-level braking threshold, respectively. For risk adaptive coefficient, For direction-sensitive adaptive coefficients, For the comfort distribution adaptive coefficient, , , and For the Pareto optimal solution set The result is the normalized result of the four objective function values ​​corresponding to each candidate solution.

9. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 8, is characterized in that... Step S4 determines the target braking demand based on the current control mode. At that time, the basic target deceleration output in step S2 is directly used. As the current control cycle The safety priority objective is deceleration; when At that time, the deceleration is corrected for comfort. For decision variables, in the interval The internal structure includes a candidate deceleration set. Candidate decelerations must satisfy the modified risk threshold constraint that the overall equivalent collision time under the action of the candidate decelerations must be within the required range. The comfort-corrected risk state must not transition to the safety-first AEB control mode. When the candidate deceleration set is empty, let... ; Conversely, within the candidate deceleration set, with the objectives of reducing discomfort at the most unfavorable representative seat, minimizing the imbalance in comfort experience among occupants in different positions within the vehicle, reducing discomfort in sensitive areas under the influence of critical conflicts, and ensuring sufficient safety margin for the current candidate deceleration, the Pareto nondominated solution set is solved. Taking into account risk adaptive coefficients, comfort distribution adaptive coefficients, and direction sensitivity adaptive coefficients, the optimal comfort-corrected deceleration for the current control cycle is calculated. This optimal comfort-corrected deceleration is then used as the target deceleration and converted into a control variable directly invoked by the execution layer. Finally, it is corrected based on a PID feedback control law to obtain the final braking control signal.

10. The AEB control method for commercial buses turning at intersections, considering passenger comfort, as described in claim 9, is characterized in that... Current control cycle in step S4 Optimal comfort-corrected deceleration: ； in, For the current control cycle Next candidate deceleration The corresponding Pareto non-dominated solution set, , , and These are the normalized results of the objective function values ​​for each candidate solution in the Pareto non-dominated solution set. , , The current control cycle The adaptive coefficients for risk, direction sensitivity, and comfort distribution are as follows.

Images