Vehicle and method and system for adjusting attitude thereof

Abstract

This invention proposes a vehicle and its attitude adjustment method and system. The vehicle is a vehicle with intelligent driving function. The method includes: when it is predicted that the vehicle will collide with a target vehicle, acquiring collision description information of the collision condition; generating a target attitude of the vehicle based on the collision description information; and adjusting the vehicle attitude based on the target attitude. The vehicle attitude adjustment method of this invention improves the safety of the occupants by adjusting the vehicle attitude and changing the initial boundary conditions of the collision before the vehicle collides and before the vehicle becomes unstable.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a vehicle and its attitude adjustment method and system. Background Technology

[0002] With the development of the automotive industry, vehicle safety performance has become one of the core indicators for measuring automobile quality. Currently, vehicle safety technologies are mainly divided into active safety technologies and passive safety technologies. Active safety technologies focus on risk avoidance before a collision, while passive safety technologies focus on occupant protection during a collision.

[0003] In related technologies, the protection target of safety technologies is the occupants, mainly for the post-accident protection of occupants. For example, the optimization scheme of the restraint system for the specific collision scenario of "car-truck rear-end collision" is based on the adaptive adjustment of restraint systems such as airbags after the collision object and working conditions are determined.

[0004] However, the above methods are not very effective in protecting occupants and cannot provide drivers and passengers with a low-injury and ideal driving environment. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0006] Therefore, the first objective of this invention is to provide a vehicle attitude adjustment method that improves the safety of occupants by adjusting the vehicle attitude and changing the initial boundary conditions of the collision before a collision occurs and before the vehicle becomes unstable.

[0007] Therefore, a second objective of the present invention is to provide a vehicle attitude adjustment system.

[0008] Therefore, a third objective of the present invention is to provide a vehicle.

[0009] To achieve the above objectives, an embodiment of the first aspect of the present invention provides a vehicle attitude adjustment method, the method comprising: when it is predicted that a collision will occur between the vehicle and a target vehicle, acquiring collision description information of the collision condition; generating a target attitude of the vehicle based on the collision description information; and adjusting the vehicle attitude based on the target attitude.

[0010] The vehicle attitude adjustment method according to embodiments of the present invention predicts an impending collision between the vehicle and a target vehicle and adjusts the vehicle's attitude accordingly. That is, before a collision occurs and before the vehicle becomes unstable, the vehicle's attitude is adjusted to change the initial boundary conditions of the collision, so that the vehicle is in a characteristic attitude that maximizes structural crashworthiness. This reduces the impact of the collision itself, is more forward-looking, and achieves a leap from passive instability prevention to active stability maintenance, resulting in a higher level of safety and providing a safer driving environment for passengers, thereby improving the safety of passengers inside the vehicle.

[0011] In some embodiments, the process of adjusting the vehicle attitude according to the target attitude further includes: adjusting the attitude of the vehicle's internal protective device according to the target attitude and the collision description information. This involves coordinated control of the active and passive safety systems to adjust both the vehicle's chassis attitude and internal protective attitude to the optimal collision avoidance state, actively shaping optimal initial collision conditions, thereby improving the safety of the vehicle's occupants.

[0012] In some embodiments, generating the target attitude of the vehicle based on the collision description information includes: when the collision condition is included in the preset collision condition-target attitude mapping database, determining the target attitude based on the collision description information and the preset collision condition-target attitude mapping database; when the collision condition is not included in the preset collision condition-target attitude mapping database, determining the target attitude by performing interpolation calculations on adjacent conditions based on a physical interpolation function and the preset collision condition-target attitude mapping database. By accessing and querying the aforementioned preset collision condition-target attitude mapping database and performing real-time interpolation calculations based on a physical model, the target attitude for the current precise collision condition is output.

[0013] In some embodiments, predicting a collision between a vehicle and a target vehicle includes: acquiring perception data within a preset range of the target vehicle; determining collision parameters based on target object information and geometric features in the perception data; classifying the collision parameters; and predicting a collision between the vehicle and the target vehicle when the classification result meets preset requirements. By processing the perception data accordingly, the system can predict an impending collision between the vehicle and the target vehicle, and only acquire collision description information when a collision is determined to occur, thus avoiding data waste.

[0014] In some embodiments, obtaining collision description information of a collision condition includes: determining motion geometry analysis results, collision parameters, and a target collision condition based on sensing data; and determining collision description information of the collision condition based on the motion geometry analysis results, the collision parameters, and the target collision condition. Parallel fusion processing of the sensing data improves the accuracy of obtaining the collision description information.

[0015] In some embodiments, adjusting the vehicle posture according to the target posture includes: determining a vehicle posture control command based on the target posture; and driving the actuators of the vehicle's chassis control system to perform coordinated actions in response to the vehicle posture control command. By acquiring the vehicle posture control command and controlling the coordinated actions of the actuators of the vehicle's chassis control system, the vehicle posture can be adjusted.

[0016] In some embodiments, adjusting the attitude of the vehicle interior protective device based on the target attitude and the collision description information includes: determining a control command for the vehicle interior protective device based on the equivalent velocity change in the target attitude and the collision description information; and controlling the linkage action of the vehicle interior protective device in response to the control command. By acquiring the vehicle attitude control command, the linkage action of various actuators in the vehicle's chassis control system is controlled to achieve a combined adjustment of active and passive safety.

[0017] In some embodiments, before determining the target attitude, the method further includes performing a safety verification on the target attitude. This ensures both the feasibility and safety of vehicle attitude adjustment.

[0018] To achieve the above objectives, a second aspect of the present invention provides a vehicle attitude adjustment system, the system comprising: The vehicle attitude adjustment system according to embodiments of the present invention predicts an impending collision between the vehicle and a target vehicle and adjusts the vehicle's attitude accordingly. That is, before a collision occurs and before the vehicle becomes unstable, the system adjusts the vehicle's attitude to change the initial boundary conditions of the collision, thereby placing the vehicle in a characteristic attitude that maximizes structural crashworthiness. This reduces the impact of the collision itself, is more proactive, and achieves a leap from passive instability prevention to active stability maintenance. It has a higher level of safety and provides a safer driving environment for passengers, thereby improving the safety of passengers inside the vehicle.

[0019] To achieve the above objectives, a third aspect of the present invention provides a vehicle that includes a posture adjustment device as described in the above embodiments. According to embodiments of the present invention, the vehicle, based on a vehicle attitude adjustment system, predicts an impending collision between the vehicle and a target vehicle and adjusts the vehicle's attitude accordingly. That is, before a collision occurs and before the vehicle becomes unstable, the vehicle's attitude is adjusted to change the initial boundary conditions of the collision, thereby placing the vehicle in a characteristic attitude that maximizes structural crashworthiness. This reduces the impact of the collision itself, is more proactive, and achieves a leap from passive instability prevention to active stability maintenance, resulting in a higher level of safety and providing a safer driving environment for passengers, thus improving the safety of occupants.

[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the structure of a vehicle attitude adjustment system according to an embodiment of the present invention; Figure 2 This is a flowchart of a vehicle attitude adjustment method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of data interaction controlled by an internal security protection device according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the construction of a database according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the hardware architecture of a collision-resistant attitude planning decision-maker according to an embodiment of the present invention; Figure 6 This is a flowchart illustrating the determination of a target posture based on a preset database according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the structure of a global sensing sensor array according to an embodiment of the present invention; Figure 8 This is a schematic diagram illustrating the acquisition of collision description information of a collision condition according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the linkage action of various actuators in a vehicle chassis control system according to an embodiment of the present invention. Figure 10 This is a schematic diagram of the coordinated timing of vehicle attitude adjustment according to an embodiment of the present invention; Figure 11 This is a flowchart of a method for adjusting the attitude of a vehicle according to an embodiment of the present invention; Figure 12 This is a block diagram of a vehicle according to an embodiment of the present invention. Detailed Implementation

[0022] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.

[0023] Related technologies, for example, integrate active and passive safety data, and use collision condition prediction models combined with real-time passenger characteristics to achieve personalized and precise control of airbag deployment time. Its main purpose is to optimize airbag deployment during a collision, providing protection when a collision occurs.

[0024] Understandably, current collision protection lacks a technology capable of proactively intervening and optimizing the vehicle's overall collision posture just before a collision. Vehicles typically enter a collision in a "random" posture, which may result in major energy-absorbing structures such as longitudinal beams failing to fully participate in energy absorption. The collision force may directly intrude into vulnerable areas such as the lower limit of the passenger compartment or the A-pillar, and restraint systems such as airbags and seat belts may become less effective under non-standard conditions.

[0025] Therefore, the vehicle attitude adjustment method based on the embodiments of the present invention optimizes the vehicle's collision attitude and position before a collision occurs, and changes the initial boundary conditions of the collision itself. That is, by utilizing the short time window before the collision, the vehicle is adjusted to the optimal crashworthiness attitude, thereby getting rid of dependence on the collision scenario and achieving universal protection against various collision scenarios, such as frontal collisions, offset collisions, and side collisions. At the same time, it reduces the severity of collision intrusion from the source, provides more favorable driving conditions for drivers and passengers, and achieves a deeper and more fundamental system-level safety gain.

[0026] It is understood that the vehicle attitude adjustment method of this invention, by creating optimal initial collision conditions, is safer than post-collision protection.

[0027] The vehicle attitude adjustment method of this invention is implemented based on a vehicle attitude adjustment system, such as... Figure 1 The diagram shown is a structural schematic of a vehicle attitude adjustment system according to an embodiment of the present invention. The vehicle attitude adjustment system 1 includes a global perception sensor array 10, a central fusion and prediction calculation module 11, a crashworthiness attitude planning decision unit 12, a chassis cooperative domain controller 13, an actuator 14, and a passive safety system linkage interface 15.

[0028] The system comprises a customized multi-source redundant sensor array 10, providing high-precision environmental information and vehicle status data, serving as the hardware carrier for signal input. The central fusion and predictive computing module 11, built upon a multi-core heterogeneous vehicle domain controller, is responsible for the synchronous fusion of data from the full-domain perception sensor array 10, collision risk quantification, and generation of structured condition descriptors such as collision description information; it is the core of data processing. The crashworthiness attitude planning decision-maker 12 is an embedded control unit integrating a dedicated storage unit and a hardware acceleration engine. Based on a fixed knowledge base, it outputs the vehicle's target attitude, realizing the hardware implementation of "data-driven decision-making." The chassis coordination domain controller 13, equipped with the MPC (Model Predictive Control) algorithm, is responsible for converting the target attitude into multiple coordinated control commands, serving as a bridge between decision-making and execution. The actuator 14 includes the active suspension, electronic stability control system, and active stabilizer bar, achieving vehicle attitude adjustment through high-precision mechanical movements; it is the physical execution terminal. The passive safety system linkage interface 15 is a dedicated communication gateway based on vehicle Ethernet, which realizes the coordinated control and parameter linkage of active safety systems and passive safety systems, so that the results of active attitude adjustment synchronously drive the pre-optimized configuration of passive safety devices.

[0029] The above modules are electrically interconnected through CAN FD bus (CAN with Flexible Data rate, such as 1Mbps), vehicle Ethernet, and analog / PWM (Pulse Width Modulation) signal lines. The overall response delay is controlled within a preset time, such as 230ms. The reserved time, such as less than 20ms, is used as redundancy for bus transmission, data frame parsing, and interface synchronization between modules. This is to adapt to complex operating conditions such as vehicle electromagnetic interference and bus load fluctuations, and to meet the timing requirements of collision warning and attitude adjustment.

[0030] The following is for reference. Figures 1-11 This invention describes a method for adjusting the posture of a vehicle, which has intelligent driving capabilities.

[0031] like Figure 2 The diagram shows a flowchart of a vehicle attitude adjustment method according to an embodiment of the present invention. The vehicle attitude adjustment method of this embodiment includes steps S1-S3.

[0032] Step S1: When it is predicted that a collision will occur between the vehicle and the target vehicle, obtain the collision description information of the collision condition.

[0033] Among them, the target vehicle is any vehicle within the information collection range of the full-domain sensor array that may collide with another vehicle.

[0034] Collision description information is used to describe the impending collision. This information includes descriptive data for fields such as collision direction, target type and key geometric features, collision type, and predicted equivalent collision velocity change. And the expected collision time.

[0035] Among them, the target type and key geometric features are as follows: for example, the target vehicle type is a truck, and the key geometric features are as follows: the cargo box height is H. The collision type is as follows: when a vehicle is about to collide with a target vehicle, the specific collision type is such as full-width frontal collision, small offset collision, undercut collision, or straddle collision.

[0036] In this embodiment, a full-range perception sensor array collects perception data within a preset range, such as 360°, of the vehicle. The central fusion and prediction calculation module receives the perception data and performs a full collision risk assessment, a classification judgment of the inevitability of collision, and generates collision description information for the collision conditions. Thus, when it is determined that a collision is about to occur between the vehicle and the target vehicle, collision description information is generated.

[0037] Understandably, by assessing the collision conditions between a vehicle and a target vehicle, and predicting an unavoidable collision risk, obtaining collision description information for the event facilitates proactive vehicle protection before a collision. Unavoidable collision risk refers to a situation confirmed through multiple levels of assessment that, given the vehicle's current dynamic capabilities, a collision cannot be avoided through braking or steering.

[0038] Step S2: Generate the target pose of the vehicle based on the collision description information.

[0039] The target attitude of the vehicle includes the vehicle's height, for example, denoted as... The vehicle's pitch angle, for example, is denoted as... The vehicle's roll angle, for example, is denoted as... The target attitude of the vehicle is the stable attitude shaped by an external collision, in which the vehicle has a high degree of crashworthiness.

[0040] In this embodiment, the crashworthiness attitude planning decision-maker receives the collision description information of the vehicle's collision conditions, realizes data-driven decision-making, and obtains the vehicle's target attitude by querying the internally fixed collision condition-target attitude mapping relationship database and inputting the collision description information of the collision conditions into the collision condition-target attitude mapping relationship database.

[0041] Understandably, the target attitude of the vehicle is generated based on the collision description information, providing adjustable data support for the vehicle's attitude adjustment before the collision.

[0042] Step S3: Adjust the vehicle attitude according to the target attitude.

[0043] In this embodiment, after generating the target vehicle posture, the vehicle is adjusted to that posture, enabling proactive posture adjustment just before a collision. By adjusting the vehicle posture, the vehicle chassis is positioned to maximize structural crashworthiness, thereby reducing the injury to occupants from a collision.

[0044] Understandably, adjusting vehicle posture based on the target posture, specifically when a collision is imminent, shifts the protection target from occupants to the vehicle structure itself. Compared to traditional passive safety technologies and their optimizations, such as optimizing airbag deployment, which focus on post-collision protection of occupants while assuming the vehicle's deformation pattern remains unchanged, the vehicle posture adjustment method of this invention prioritizes the crashworthiness of the vehicle structure itself. By optimizing the initial posture, it ensures that the vehicle's core energy-absorbing structures, such as longitudinal beams, participate in the collision most effectively, thereby altering and optimizing the physical process of the collision. This achieves "source control" of safety protection, fundamentally reducing collision energy at the physical level. This not only directly improves safety performance to a certain extent but also creates a lower-injury, more predictable ideal working environment for all occupant restraint systems, maximizing their protective effectiveness.

[0045] Furthermore, the vehicle attitude adjustment method of this invention breaks through the path dependence of safety technology based on "collision signal triggering" and establishes a new system of proactive optimization based on pre-collision conditions. Its technical approach shifts from traditional "passive response and remediation" to "proactive construction and prevention," thereby achieving effective intervention at the very front of the accident injury chain and leading vehicle safety technology into a new stage of pre-set safety.

[0046] According to the vehicle attitude adjustment method of the present invention, by predicting that a collision is about to occur between the vehicle and a target vehicle, the vehicle attitude is adjusted. That is, before the collision occurs and before the vehicle becomes unstable, the vehicle attitude is adjusted to change the initial boundary conditions of the collision, so that the vehicle is in a characteristic attitude that maximizes the structural crashworthiness, thereby reducing the impact of the collision itself. This method is more forward-looking and achieves a leap from passive instability prevention to active stability maintenance, resulting in a higher level of safety and providing a safer driving environment for passengers, thereby improving the safety of passengers inside the vehicle.

[0047] In related technologies, active safety systems and passive safety systems often operate independently. Active safety systems fail to adjust the vehicle's posture based on the optimal requirements of passive safety protection, and passive safety systems are unable to dynamically optimize protection strategies based on the posture after active intervention. This results in insufficient integration of active and passive safety technologies, making it difficult to achieve optimal safety protection effects under all operating conditions.

[0048] For example, integrating the perception, decision-making, and execution of active and passive safety systems into a single domain controller enables cross-stage collaborative control and dynamic threshold adjustment. However, this approach remains at the level of information sharing and command distribution, failing to delve into the level of actively optimizing the physical processes of a collision through vehicle dynamics control. Active safety refers to all technologies designed to prevent or avoid collisions. Its core function is to operate before an accident occurs. Examples include anti-lock braking systems, electronic stability programs, and automatic emergency braking. By monitoring, warning, and intervening in vehicle dynamics, it assists the driver in controlling the vehicle, thereby preventing accidents. Passive safety refers to technologies used to minimize occupant injury after a collision is unavoidable or has already occurred. It operates immediately after the accident. Examples include high-strength vehicle body structures, airbags, seat belts, and collapsible steering columns. They protect occupants by physically cushioning, restraining, and dispersing impact forces.

[0049] Therefore, the vehicle attitude adjustment method of this invention, in order to achieve the safety goal of optimal crashworthiness, performs coordinated control of active safety system and passive safety system under corresponding control commands, so that the vehicle chassis attitude and internal protection attitude are both adjusted to the optimal anti-collision state, actively shaping the optimal initial collision conditions, thereby improving the safety of the occupants.

[0050] The vehicle attitude adjustment method of this invention optimizes the attitude of the vehicle chassis before a collision occurs, reducing the collision intensity from the source. It not only absorbs more energy itself, but also creates a gentler and more ideal operating environment for the advanced airbag system. The two work together to achieve a system-level safety gain of "1+1>2".

[0051] In some embodiments, the process of adjusting the vehicle attitude according to the target attitude further includes: adjusting the attitude of the vehicle's internal protective devices according to the target attitude and collision description information.

[0052] In an embodiment, such as Figure 3 The diagram shown illustrates the data interaction of the internal safety protection device control according to an embodiment of the present invention. During the process of adjusting the vehicle's attitude based on the target attitude, corresponding control is performed on the vehicle's internal protection devices, such as the high-strength body structure, airbags, seat belts, and collapsible steering column. This enables the parameter preset of the passive safety devices, allowing them to enter an optimized working state before a collision occurs, thus shortening response delay.

[0053] For example, such as Figure 3 As shown, the passive security system linkage interface is the active hardware gateway, which is the core for realizing deep collaboration between active and passive security. It features forward-looking parameter presets, and the specific data interaction process can be found in [reference needed]. Figure 3During the process of adjusting the vehicle's attitude based on the target attitude, the linkage interface sends the target attitude and collision parameters, such as the equivalent speed collision change, to passive safety systems such as airbags and seat belts, thereby realizing the linkage control between passive and active safety systems.

[0054] The target attitude and collision parameters are broadcast via Ethernet to the ACU (Airbag Control Unit) and seatbelt pretensioner system. The ACU then calculates the impact parameters based on the attitude and equivalent velocity changes during the collision. The system dynamically adjusts the airbag ignition timing and inflation level (adjustment range ±30%), and the seatbelt pretensioner adjusts accordingly. Set two-stage preload thresholds (first-stage preload varies with...) Linear adjustment and secondary pretensioning timing are triggered based on actual collision signals, enabling the parameter presetting of passive safety devices to enter an optimized working state before a collision occurs, thus shortening response delay. By adjusting the attitude of the vehicle's internal protective devices according to the target attitude and collision parameters, the active and passive safety systems are linked for coordinated control, reducing collision intensity at the source. This not only allows the system to absorb more energy but also creates a gentler and more ideal operating environment for advanced passive safety systems such as airbags. The two work together to achieve safety gains for both active and passive safety systems.

[0055] In related technologies, rule-based or simple table lookup-based control strategies cannot cope with an infinite variety of real-world collision scenarios, nor can they guarantee that the decision result is globally optimal.

[0056] Therefore, the vehicle attitude adjustment method of this invention adopts a decision-making paradigm of offline data-driven global optimization + online hardware-accelerated retrieval. Specifically, the system constructs a globally optimal attitude knowledge base covering all working conditions through offline massive simulations and stores it in the vehicle's onboard memory. During online collision warning, dedicated hardware performs millisecond-level retrieval and physical interpolation on this stored knowledge base, directly outputting the optimal attitude command. This paradigm ensures that the complex calculations for global optimality are completed offline without resource constraints, simplifying the real-time onboard task into a single reliable hardware query, thereby fundamentally solving the problem of not being able to simultaneously guarantee decision optimality, real-time performance, and determinism in the extremely short time before a collision.

[0057] In some embodiments, generating the vehicle's target attitude based on collision description information includes: when the preset collision condition-target attitude mapping database includes collision conditions, determining the target attitude in the preset collision condition-target attitude mapping database based on the collision description information; when the preset collision condition-target attitude mapping database does not include collision conditions, determining the target attitude by interpolating adjacent conditions based on a physical interpolation function and the preset collision condition-target attitude mapping database. The crashworthiness attitude planning decision-maker internally integrates and solidifies the preset collision condition-target attitude mapping database for determining the target attitude. The collision description information consists of structured description parameters of the vehicle under the collision condition.

[0058] In this embodiment, the preset collision condition-target attitude mapping database is not a collection of manually derived empirical rules, but rather generated offline through massive parametric collision simulations and machine learning. Specifically, it is a solidified data set that uncovers the complex mapping relationship between "attitude parameters" and "structural crashworthiness indicators" such as longitudinal beam energy absorption and passenger compartment intrusion through systematic simulation sample construction and model training.

[0059] The process of determining the target attitude based on collision description information is divided into two stages. The first stage is offline modeling and data generation, the core task of which is to build an optimal crashworthiness attitude database for decision-making through system simulation analysis and machine learning. The second stage is online real-time prediction and control, the core task of which is to query the database based on real-time perception information during vehicle operation and drive the chassis co-controller to accurately execute the target attitude.

[0060] like Figure 4 The diagram shown illustrates the construction of a database according to an embodiment of the present invention. The construction and solidification of this database is an independent and offline engineering process. Specifically, it includes parametric simulation and scene generation, optimal solution mining through machine learning, and knowledge base solidification and in-vehicle solidification.

[0061] Parametric simulation and scene generation, on a computing cluster, establishes a target attitude model for the vehicle, including adjustable attitude parameters such as vehicle height. Pitch angle yaw angle A parameterized finite element model of the entire vehicle is constructed. A series of parameterized collision models are established, including sedans, SUVs, and trucks of different sizes, such as those with variable cargo box heights, guardrails, and pillars, and are given typical material and stiffness properties. By changing collision condition parameters such as collision angle, overlap ratio, relative velocity, and initial vehicle attitude parameters, tens of thousands to hundreds of thousands of virtual collision scenarios covering the entire collision space are generated. Nonlinear finite element collision simulations of all scenarios are executed in parallel on a high-performance computing cluster to form an original simulation data pool.

[0062] Machine learning-based optimal solution mining, or data-driven optimal solution mining and knowledge extraction, extracts key crashworthiness evaluation indicators from each simulation result, such as intrusion amount at key measurement points in the passenger compartment, energy absorption of the entire vehicle structure, and bending patterns of the chassis longitudinal beams. Based on this data, a high-dimensional response surface model is constructed, taking collision condition parameters and initial attitude as inputs and the crashworthiness indicator, i.e., the target attitude, as the output. For each fixed collision condition, within the allowable attitude parameter space, optimization algorithms such as gradient descent or genetic algorithms are used to search the response surface model to find a specific combination of attitude parameters that optimizes the crashworthiness indicator, such as minimizing intrusion. This set of parameters is defined as the "optimal crashworthiness attitude" under that condition, i.e., the target attitude.

[0063] Knowledge base formatting and vehicle-mounted persistence: The correspondence between all "collision condition features and optimal attitudes" is formatted and encoded, organized into a multi-dimensional lookup table or a highly simplified lightweight neural network model. This lookup table or model file is then persisted to the non-volatile Flash memory of the crashworthiness attitude planning decision-maker using a burning tool, becoming its internally queryable static database. This persistence process ensures the stability and reliability of the database throughout the vehicle's lifecycle.

[0064] After the database is built, vehicle online control is performed. For example, when adjusting the vehicle attitude, the crashworthiness attitude planning decision-maker receives the real-time operating condition description from the perception module. By accessing and querying the above-mentioned preset collision condition-target attitude mapping relationship database, and performing real-time interpolation calculation based on the physical model, the target attitude for the current precise collision condition is output.

[0065] Understandably, the online real-time prediction and control phase focuses on efficient querying and application. This, to some extent, avoids the stringent real-time requirements of decision-making in the immediate moments before a collision, and also differs from control strategies based on simple preset rules. Through this technology, while ensuring the scientific rigor and accuracy of decision-making, the system's response speed can be effectively improved, meeting the stringent real-time requirements of completing perception-decision-execution within a very short time window before a collision, thus combining advanced technology with feasibility.

[0066] For example, such as Figure 5The diagram shown illustrates the hardware architecture of a crashworthiness attitude planning decision-maker according to an embodiment of the present invention. The crashworthiness attitude planning decision-maker is a dedicated embedded control unit. Its core feature lies in accelerating data-driven decision-making through hardware storage and FPGA (Field Programmable Gate Array), avoiding the uncertainty of traditional software decision-making. By setting up a crashworthiness attitude planning decision-maker, the present invention accurately and uniquely limits the target of vehicle attitude control to improving the overall vehicle structural crashworthiness. This allows the vehicle to actively optimize its structural collision interface before a collision through the coordinated action of the chassis system, thus achieving a technological leap from information fusion control to proactive optimization of physical safety performance, resulting in significantly better safety protection than general attitude adjustment. It is understood that the crashworthiness attitude planning decision-maker adjusts the vehicle chassis to a specific attitude that maximizes structural crashworthiness. This is a highly targeted proactive intervention based on collision mechanics principles, rather than general attitude control.

[0067] Specifically, the crashworthiness attitude planning decision-maker includes a dedicated storage unit and a safety verification module. The dedicated storage unit, such as a non-volatile Flash memory, stores a "collision condition-target attitude mapping database" built through an offline optimization process. The safety verification module performs dual verification on the planned target attitude. It verifies whether the target attitude exceeds the physical motion limits of the chassis actuators, such as the active suspension, including travel and speed limits; simultaneously, it predicts whether there is a risk of vehicle instability during attitude adjustment. If the verification fails, it automatically calls a conservative default attitude strategy to ensure system safety.

[0068] In the storage unit, by matching collision description information with the database, when the preset collision condition-target attitude mapping relationship database includes collision conditions, the target attitude, such as vehicle height, is output by taking the collision description information features corresponding to the collision condition as input, such as direction, type, Δv, etc. Target pitch angle And target roll angle .

[0069] When the preset collision condition-target attitude mapping database does not include collision conditions, the crashworthiness attitude planning decision-maker receives the collision description information corresponding to the collision condition. The integrated FPGA hardware acceleration engine immediately starts the distance-weighted K-nearest neighbor search algorithm to achieve a millisecond-level database matching query of less than or equal to 2ms. When there is no exact matching collision condition, it calls the physical interpolation function embedded in the hardware, such as one built based on the nonlinear characteristics of the vehicle suspension, to interpolate the baseline attitude parameters of the adjacent collision conditions in the database and generate a target attitude that adapts to the parameters of the current collision condition, ensuring the continuity and physical rationality of the attitude adjustment.

[0070] The following is for reference. Figure 6 This invention describes how an embodiment of the present invention determines the target attitude based on a preset database.

[0071] like Figure 6 As shown, determining the target pose based on a preset database in this embodiment of the invention includes the following steps.

[0072] Step S10: Receive collision description information.

[0073] Step S11: Extract features and generate query vectors from the collision description information.

[0074] Step S12: Execute the K-nearest neighbor search algorithm.

[0075] Step S13: Determine whether the collision condition-target attitude mapping relationship database includes collision description information corresponding to the collision condition. If yes, proceed to step S14; otherwise, proceed to step S15.

[0076] Step S14: Read the target pose.

[0077] Step S15, nonlinear interpolation adaptation.

[0078] Step S16: Check whether the vehicle safety check is passed. If yes, proceed to step S18; otherwise, proceed to step S17.

[0079] Step S17: Output the default safe posture.

[0080] Step S18: Output the target pose.

[0081] In some embodiments, predicting a collision between a vehicle and a target vehicle includes: acquiring perception data within a preset range of the target vehicle; determining collision parameters based on target object information and geometric features in the perception data; classifying the collision parameters; and predicting a collision between the vehicle and the target vehicle when the classification results meet preset requirements. By processing the perception data accordingly, the system can predict an impending collision between the vehicle and the target vehicle, and only acquire collision description information when a collision is determined to occur, thus avoiding data waste.

[0082] In this embodiment, the central fusion and prediction calculation module is used to acquire perception data, perform omnidirectional collision risk assessment and collision inevitability progressive judgment based on the perception data, and predict that a collision will occur between the vehicle and the target vehicle.

[0083] For example, such as Figure 7The diagram shown is a schematic representation of a global perception sensor array according to an embodiment of the present invention. The global perception sensor array serves as a signal input device for vehicle attitude adjustment, and its key features include customized physical deployment of multiple source sensors, rigid fixation, and hardware-level synchronization design to ensure the integrity and timeliness of data acquisition.

[0084] The forward-facing main sensing unit includes at least one long-range millimeter-wave radar and one high-resolution forward-facing camera. The camera is mounted in the front roof area above the windshield, within the rearview mirror module assembly. The millimeter-wave radar is mounted in the center of the front bumper; its beam can penetrate a specific area of ​​the windshield, and the camera directly obtains its field of view through the glass. The spatial relative positions are fixed, and the unit is connected to the central platform via a video interface to ensure synchronized data flow. The millimeter-wave radar provides precise relative distance information for targets ahead. With relative velocity measurement The camera has a built-in convolutional neural network inference chip, which is equipped with target classification and geometric feature extraction algorithms. It can identify various obstacles such as cars, SUVs, trucks, and pedestrians in real time, and extract key parameters accordingly. For example, for trucks / vans, it can estimate the ground clearance of the cargo box (error ≤ 5cm) and the shape of the rear bumper beam. For passenger cars, it can identify the body outline and potential intrusion surfaces. The data stream transmission latency is ≤ 5ms.

[0085] To achieve all-around collision warning, target classification, and accurate perception of small offset collisions, the side and rear perception units employ a dual-redundant configuration of "high-precision millimeter-wave angular radar + surround-view camera," balancing accuracy, cost, and adaptability to adverse weather conditions. The specific hardware layout and collaborative logic are as follows: 1. Front Side Core Perception (Left Front, Right Front Corner): A high-precision mid-range millimeter-wave radar (hardware parameters: horizontal field of view ±75°, detection range ≥50m, response time ≤20ms, point cloud density ≥80,000 points / second) is deployed on the inner side of each of the left and right front fenders to enhance point cloud acquisition capabilities in the 30°-60° high-risk collision area (high-incidence area for small offset collisions and angled collisions); 2. Rear Side Basic Perception: The same model of mid-range millimeter-wave radar is deployed at the left and right rear corners to meet the needs of routine rear-end collision and side / rear approach warnings; 3. Visual Supplement: Three high-definition surround-view cameras (≥2 megapixels) are deployed below the left and right rearview mirrors and in the center of the trunk lid. Frame rate ≥25fps, night vision enhancement function), aligned with the radar position in the corresponding area; 4. Hardware collaborative logic: Four millimeter-wave radars form a 360° no-dead-angle distance, speed and point cloud detection field (adjacent radar overlap ≥10%), which solves the basic perception stability under bad weather, rain, fog and strong light. At the same time, the accuracy of target contour restoration is improved by multi-radar point cloud fusion; The surround view camera has a built-in lightweight classification algorithm, which is solidified in the local chip to complete the target type label (car, guardrail, pedestrian, etc.) and extract geometric features. After the data of the two are fused, it not only solves the problem that pure radar cannot classify, but also achieves accurate perception of high-risk side front scenes with high cost performance, avoiding the shortcomings of performance degradation in rain and fog.

[0086] Vehicle self-state perception unit: One high-precision six-axis inertial measurement unit (IMU) is fixed on a floor bracket directly below the vehicle's center of gravity (e.g., Figure 7 (Top view ellipse marked position), sampling frequency ≥100Hz, can measure ±50m / s² longitudinal / lateral acceleration and ±500° / s yaw rate, with measurement accuracy of ±0.01m / s² and ±0.1° / s respectively. It outputs data to the central fusion platform through the SPI interface to ensure accurate capture of the vehicle's dynamic status.

[0087] Hardware synchronization mechanism: All sensors receive a unified hardware trigger signal from the central fusion platform to achieve microsecond-level (≤1μs) time synchronization sampling, which completely solves the problem of insufficient fusion accuracy caused by the spatiotemporal asynchrony of traditional multi-sensor data.

[0088] The core innovation of the multi-core heterogeneous vehicle domain controller of the central fusion and predictive computing module lies in embedding the collision risk quantification algorithm into the hardware and outputting standardized, fixed-format structured collision condition descriptors to achieve determinism and universality in data processing.

[0089] The central fusion and predictive computing module needs to acquire perception data, conduct omnidirectional collision risk assessment, and progressively determine the inevitability of collisions. Perception data acquisition is based on hardware interfaces and data input, receiving perception data from all sensors through its 4 CAN FD interfaces, 2 Ethernet interfaces, and 1 GMSL2 interface. The internal kernel runs a target-level fusion algorithm to generate a dynamic environment model that includes the target's position, velocity, classification label, and predicted trajectory.

[0090] The enhanced omnidirectional collision risk quantification module is used for omnidirectional collision risk assessment. As a functional unit of the central fusion and prediction calculation module, it receives a dynamic environment model from the fusion and tracking unit within the module, containing the real-time status (position, speed, acceleration, and category) of the vehicle and surrounding targets. Based on the dynamic environment model, parallel risk assessments are performed for potential threats from different directions: for forward potential threats, the imminent time of collision with the vehicle is calculated; for lateral and rearward potential threats, the TCPA (time of closest point of approach) and DCPA (Distance at Closest Point of Approach) at that point are calculated respectively. Simultaneously, the maximum steering angle of the EPS (Electric Power Steering) and adjacent lane occupancy data are collected to provide a basis for verifying lateral / rearward avoidance capabilities.

[0091] Based on the calculation results of the enhanced risk parameters mentioned above, this collision inevitability progressive judgment module adopts a three-level progressive judgment strategy to accurately identify scenarios where "conventional obstacle avoidance methods are ineffective and collisions are unavoidable," thus avoiding false or delayed triggering.

[0092] Specifically, Level 1 Judgment: Threshold Triggering and Initial Screening of Obstacle Avoidance Capability. This level of judgment is based on risk parameter thresholds, combined with the effectiveness of conventional obstacle avoidance methods, to complete an initial prediction: Forward Scenario: When the enhanced TTC (Time To Collision) value falls within the 1.5s-3.0s warning window (which can be calibrated according to the vehicle's braking performance) and meets the following conditions: ① The vehicle has triggered the maximum intensity AEB automatic emergency braking, and the remaining braking distance is greater than the current relative distance; ② The target's trajectory is stable (multi-model adaptive filtering prediction trajectory deviation ≤ 0.5m), the forward collision risk is determined to have reached the trigger threshold, and it is initially determined that avoidance by braking is not possible. Lateral / Rearward Scenario: When TCPA ≤ 2.0s and DCPA ≤ 1.2 times the width of the vehicle (safe distance threshold), and meets the following conditions: ① There is an obstacle / vehicle in the adjacent lane of the vehicle, and the maximum steering angle of EPS cannot achieve effective avoidance; ② The target's radar point cloud trajectory deviation for 20 consecutive frames is ≤ 0.3m, the lateral / rearward collision risk is determined to have reached the trigger threshold, and it is initially determined that avoidance by steering is not possible. Exclusion criteria: If TTC > 3.0s (sufficient obstacle avoidance time) or TTC < 1.5s (insufficient attitude adjustment time), or if there is active avoidance action from lateral / rearward targets, it is directly judged as an "avoidable scenario" and the subsequent process is not initiated.

[0093] Level 2 Judgment: Multi-parameter Fusion Verification (Accurate Elimination of False Risk Scenarios) For scenarios that pass Level 1 judgment, multi-source sensor data is further integrated to complete three core verifications, improving the accuracy of the judgment: 1. Vehicle Extreme Obstacle Avoidance Capability Verification: Based on the yaw rate and longitudinal / lateral acceleration collected by the chassis IMU (inertial measurement units), combined with the vehicle dynamics model, the extreme obstacle avoidance trajectory of the vehicle under combined braking and steering intervention is calculated, and it is verified whether the trajectory overlaps with the target trajectory; 2. Target Type and Scene Compatibility Verification: Through a lightweight classification algorithm of the surround-view camera, the target type (fixed obstacle / moving vehicle / pedestrian) and scene (highway / urban intersection) are identified. If it is a "fixed obstacle + highway" scene, it is directly judged as a high-priority unavoidable scene; 3. Sensor Data Credibility Verification: Through redundant comparison of multi-sensor data (e.g., the target position deviation between radar point cloud and camera image is ≤0.2m), false risk signals caused by sensor noise are eliminated.

[0094] Level 3 Judgment: System-level Triggered Decision (Final Judgment and Engine Startup) This module determines "collision is unavoidable" only when the Level 1 judgment passes and all three verifications of the Level 2 judgment pass. It immediately sends a trigger command to the central fusion and collision prediction engine: start the structured collision condition descriptor generation process, and simultaneously send a pre-notification to the crashworthiness attitude decision-maker and the active and passive safety linkage interface. The entire judgment process takes ≤50ms. The time taken for the subsequent decision-control-execution-linkage process must be added to the judgment time to meet the system's total timing budget requirement of ≤230ms, ensuring that all attitude adjustment actions and passive safety preset actions are fully effective before the collision occurs.

[0095] In some embodiments, obtaining collision description information of a collision condition includes: determining motion geometry analysis results, collision parameters, and a target collision condition based on sensing data; and determining collision description information of the collision condition based on the motion geometry analysis results, collision parameters, and the target collision condition. Parallel fusion processing of the sensing data improves the accuracy of obtaining the collision description information.

[0096] In an embodiment, such as Figure 8 The diagram shown illustrates the acquisition of collision description information for a collision scenario according to an embodiment of the present invention. When a collision is predicted between a vehicle and a target vehicle, i.e., when the collision risk in any direction exceeds a preset threshold, the central fusion and collision prediction engine is immediately activated.

[0097] When a collision is predicted between a vehicle and a target vehicle, the system acquires perception information and simultaneously performs calculations on the perception information, including motion geometry calculations, collision dynamics calculations, and target feature-specific analysis.

[0098] Specifically, motion geometry calculations are used to output collision direction, predict overlap rate, and collision angle, such as outputting basic classifications like "oblique collision" and "small overlap offset collision." Collision dynamics calculations, based on parameters such as target relative velocity and estimated mass, calculate the predicted equivalent collision velocity change. The severity and urgency of a collision are quantified based on the expected collision time. Target feature-specific analysis is used to finely identify target types and assess special risks. Through hierarchical judgment logic, uncommon high-risk situations such as "drilling" collisions for high-chassis vehicles and "straddling / tripping" collisions for long, narrow objects like guardrails are identified. The results of the parallel processing above are then fused to generate collision condition description information, such as a structured collision condition descriptor with a fixed preset byte length. Complex perception data results are refined and standardized into a fixed-format binary message containing five key dimensions (direction, type, Δv, etc.). This collision condition descriptor becomes the unique and authoritative decision input for all downstream modules.

[0099] In some embodiments, adjusting the vehicle posture according to a target posture includes: determining a vehicle posture control command based on the target posture; and driving the various actuators of the vehicle's chassis control system to perform coordinated actions in response to the vehicle posture control command. By acquiring the vehicle posture control command and controlling the coordinated actions of the various actuators of the vehicle's chassis control system, the vehicle posture can be adjusted.

[0100] In this embodiment, the adjustment of the vehicle's attitude based on the target attitude is implemented by actuators of the vehicle's chassis control system. For example... Figure 9 The diagram illustrates the coordinated operation of various actuators in a vehicle chassis control system according to an embodiment of the present invention. Unlike related technologies that merely integrate information or allow individual chassis systems to perform stability control independently, the chassis cooperative domain controller requires the active suspension, active braking, and active steering systems to perform complex coordinated actions under the unified command of the chassis domain controller, rather than simple functional superposition or independent correction and stabilization. This method achieves rapid and precise control of multiple degrees of freedom (height, pitch, roll, yaw, etc.) through system-level coordination of actuators, overcoming the limitations of single actuator capabilities and ensuring the target posture is reliably achieved in a very short time. Its purpose is a deeply coordinated execution system designed to achieve the specific safety goal of "optimal crashworthiness posture."

[0101] For example, after the crashworthiness attitude planning decision-maker generates the vehicle's target attitude, the chassis cooperative domain controller and actuators achieve high accuracy in the target attitude through a dedicated "chassis cooperative domain controller-actuator" adaptation design. The hierarchical MPC (Model Predictive Control) cooperative control architecture incorporates a hierarchical model predictive control algorithm for collision attitude optimization within the chassis cooperative domain controller. The upper-level chassis cooperative domain controller solves a global optimization problem considering vehicle dynamics and the physical constraints of each actuator; the lower-level chassis cooperative domain controller performs precise tracking and disturbance compensation. Its optimization objective function J aims to balance attitude tracking accuracy, control energy consumption, and actuation smoothness.

[0102] Wherein, the state vector Reference state Control input This includes the power supply of each suspension component, the pressure of the brake wheel cylinders, etc. , and The weight matrix is ​​positive definite, and its value determines the degree of emphasis the chassis cooperative domain controller places on different performance indicators.

[0103] The chassis cooperative domain controller is equipped with dedicated interface circuitry to ensure the accuracy and real-time performance of command transmission, distributing commands to the actuator cluster through three types of interfaces. Specifically, four high-precision PWM (Pulse Width Modulation) signals are used to drive the electromagnetic or hydraulic actuators of the active suspension. Four analog voltage signals, with a voltage range of 0-5V, are used to control the brake pressure valves of each wheel in the ESC (Electronic Stability Control) system. One high-speed CAN FD bus (1Mbps) is used to send torque commands to the active stabilizer bar controller.

[0104] The actuators include an active suspension, an ESC system, and an active stabilizer bar. The active suspension, for example, has an actuation speed ≥100mm / s, a travel ±50mm, employs position-force dual-loop control, and a displacement tracking error ≤2mm. The ESC system, for example, has a brake pressure build-up rate ≥30MPa / s, a pressure control accuracy ≤0.5MPa, supports independent control of individual wheels, and a maximum single-wheel braking force ≥1500N. The active stabilizer bar, for example, has a maximum output torque ≥200Nm, a response time ≤50ms, and can quickly adjust the vehicle's roll attitude. By outputting multiple control signals, such as a high-precision PWM signal driving the active suspension, an analog voltage signal controlling the ESC braking pressure, and torque commands sent to the active stabilizer bar via the CAN FD bus, the coordinated actions of each actuator are achieved.

[0105] In some embodiments, adjusting the attitude of the vehicle interior protective device according to the target attitude and collision parameters includes: determining a control command for the vehicle interior protective device based on the equivalent velocity change in the target attitude and collision parameters; and controlling the linkage action of the vehicle interior protective device in response to the control command.

[0106] In this embodiment, when controlling the vehicle's passive safety system, the vehicle's internal protective devices are linked to each other by pre-setting forward parameters during the process of actively controlling the vehicle's attitude.

[0107] The collision-resistant attitude planning decision-maker outputs the target attitude. The equivalent velocity change is encapsulated as a 64-byte standard application APDU (Application Protocol Data Unit). The message format is fixed and includes parameter identifier, value, and checksum fields to ensure that passive security systems can be compatible with its parsing.

[0108] APDU messages are broadcast via Ethernet to the ACU (Airbag Control Unit) and seatbelt pretensioner system. The ACU dynamically adjusts the airbag ignition timing and inflation level (adjustment range ±30%) based on attitude and equivalent velocity changes. The seatbelt pretensioner sets a two-stage pretensioning force threshold based on the equivalent velocity change (the first-stage pretensioning force is linearly adjusted with Δv, and the second-stage pretensioning timing is triggered based on the actual collision signal), thereby enabling the passive safety device to preset its parameters and enter an optimized working state before a collision occurs, thus shortening the response delay.

[0109] Among them, the passive safety system linkage interface is a dedicated hardware gateway, which is the core of realizing deep collaboration between active and passive safety. Unlike the "post-event response" of traditional passive safety systems, its feature is forward-looking parameter pre-setting.

[0110] It adopts a standard-compliant automotive Ethernet switch with built-in isolation circuit, transmission delay ≤1ms, and anti-interference capability that meets the automotive electromagnetic compatibility standard. One end is connected to the crashworthiness attitude planning decision unit via Ethernet, and the other end is connected to the passive safety system network.

[0111] The vehicle attitude adjustment method of this invention is a deterministic closed loop based on strict time budget and collaborative work of various hardware modules. It covers four major stages: perception and prediction, decision planning, control execution, and linkage preset. The total timing is controlled within a preset time, such as 230ms. The execution time of each module is illustrated below. The time consumption of the core function stages is as follows: perception and prediction <100ms, decision planning <30ms, control execution <80ms, totaling <210ms. The remaining ≤20ms is the redundancy between modules, used to cope with signal delay fluctuations under complex vehicle conditions and ensure timing stability.

[0112] like Figure 10 The diagram shown is a schematic diagram of the coordinated timing of vehicle attitude adjustment according to an embodiment of the present invention.

[0113] Phase 1: Global Perception and Structured Description (T0-T0+100ms, total time <100ms): At time T0: The global perception sensor array synchronously acquires data under hardware trigger signals. From T0 to T0+80ms: The central fusion and prediction calculation module receives data and performs target fusion, enhanced TTC calculation, Δv prediction, and feature analysis in parallel. From T0+80ms to T0+100ms: The platform generates a pre-defined number of structured collision condition descriptors and sends them to the crashworthiness attitude planning decision-maker via the CAN FD bus.

[0114] Phase Two: Data-Driven Attitude Decision and Safety Verification (T0+100ms-T0+130ms, total time <30ms): T0+100ms-T0+102ms: The crashworthiness attitude planning decision-maker receives the descriptor of the collision description information, and its FPGA acceleration engine starts querying and interpolating the knowledge base in the embedded Flash memory. T0+102ms-T0+112ms: The crashworthiness attitude planning decision-maker calls the built-in dynamics model to perform dynamic safety boundary verification. T0+112ms-T0+130ms: The crashworthiness attitude planning decision-maker encapsulates the verified target attitude into instructions and sends them to the chassis cooperative domain controller via the CAN FD bus.

[0115] Phase 3: Multi-actuator Cooperative Control and Vehicle Body Adjustment (T0+130ms-T0+230ms, total time <80ms): T0+130ms-T0+150ms: The chassis cooperative domain controller, based on a hierarchical MPC algorithm, calculates the cooperative control sequence of each actuator within the next 80ms. T0+150ms-T0+230ms: The controller sends real-time commands to the chassis mechanical actuator cluster via PWM, analog voltage, and CAN FD interfaces. The actuators coordinately operate in sequence, for example: active suspension extension and retraction, ESC differential braking, and applying torque to the stabilizer bar, precisely adjusting the vehicle body attitude to the target value.

[0116] Phase Four: Proactive Linkage of Passive Safety Systems (starting from T0+112ms, running in parallel with Phase Three): From T0+112ms: Upon receiving the final target posture, the passive safety system's linkage interface immediately broadcasts a standardized APDU message via Ethernet. From T0+112ms until the collision: Based on the received posture and Δv information, the airbag controller, seatbelt pretensioners, etc., proactively pre-set their trigger parameters (such as ignition time, inflation level, and pretensioning force threshold), entering an optimized standby state. End of process: By T0+230ms, the vehicle has completed the adjustment to the optimal crashworthiness posture, and the passive safety system has completed parameter preloading. The system is ready to meet the inevitable collision with the optimal physical configuration.

[0117] The following specific embodiments illustrate the vehicle attitude adjustment method of the present invention.

[0118] Example 1: Rear-end collision with a heavy truck (undercut collision) Perception and prediction stage: When the TTC enters the 3.0s to 1.5s warning window, the vehicle's attitude adjustment system initiates a deep analysis of the target ahead. The forward-facing main sensing unit identifies the target as a "heavy truck" and extracts its key geometric feature, "cargo box height H=1.5m". The central fusion and prediction module calculates and continuously predicts, confirming a "dive collision" risk and estimating Δv=55km / h.

[0119] The system outputs a structured collision description within this warning window, with key fields including: collision direction: forward, target type: heavy truck (H=1.5m), collision type: penetration collision, and predicted Δv: 55km / h. The fixed format and specific content of this data message can be captured, parsed, and verified using an onboard CAN FD bus monitoring tool.

[0120] Decision-making and planning stage: The system's actions include receiving collision description information, querying the fixed database, and determining the optimal response strategy as "significantly raising the front of the vehicle to align the front longitudinal beam with the truck's cargo box."

[0121] The system outputs a target attitude command from the decision-maker: +75mm (front lift), +2.5° (nose tilt). This means the front wheel arches are vertically raised by 75mm, with a 2.5° angle between the front and rear to guide the front longitudinal beams into effective contact with the lower edge of the cargo box in a high-chassis truck. This target attitude command is transmitted via another CAN FD, and its parameters and decision logic constitute the core control output of the system.

[0122] Collaborative control and execution phase: System actions: The chassis co-controller calculates commands and drives: ① The front suspension actuators extend to full capacity, and the rear suspension retracts; ② The ESC applies moderate braking force to the rear wheels to assist stability. During the linkage preset phase, the system actions: The linkage interface transmits the final target attitude parameters—including the target vehicle height—to the target vehicle body. Target pitch angle The predicted equivalent collision velocity change Δv is encapsulated as a standardized Application Protocol Data Unit (APDU) and broadcast over Ethernet. Based on this, the airbag controller extends the front airbag ignition time by approximately 5ms (due to the occupant's backward lean).

[0123] System output: Standardized linkage APDU messages containing the above parameters can be captured on the vehicle Ethernet, and the parameters of the passive safety system are modified before the collision, demonstrating the system's forward-looking collaborative capabilities.

[0124] Example 2: Left side 25% small overlap offset collision Perception and prediction stage: The system action and output were determined by the central platform as a "small overlap offset collision on the left", with an estimated Δv = 40km / h, and were marked accordingly in the descriptor.

[0125] Decision-making and planning stage: System actions and outputs: The decision-maker outputs the target attitude command: = -4° (tilt to the left).

[0126] Collaborative control and execution phase: The system's actions and macroscopic response, along with the chassis controller's coordinated drive of the suspension, stabilizer bar, and ESC, enable the vehicle to actively tilt to the left and engage in coordinated yaw control, resulting in an optimized "side intrusion prevention posture."

[0127] Linkage Preset Stage: The system action triggers the linkage system, and presets the parameters of the passive safety device on the left.

[0128] Example 3: Rear-end collision on the right side Perception and prediction stage: The system actions and outputs are as follows: the central platform determines it as a "right rear-end collision", estimates Δv, and updates the descriptor.

[0129] Decision-making and planning stage: System actions and outputs: The decision-maker outputs the target attitude command: = -20mm (overall altitude reduction).

[0130] Collaborative control and execution phase: The system's actions and macroscopic response involve a rapid and synchronized descent of the vehicle's height to optimize the dynamic response under rear-end collision conditions.

[0131] Linkage Preset Stage: The system activates, the seatbelt pretensioner sets the pretension force based on the predicted Δv, and the seat makes adaptive adjustments.

[0132] In summary, this invention deeply integrates "offline massive optimization" with "online millisecond-level response" to construct a concrete, detectable, and enforceable active vehicle crashworthiness attitude control system. Its core innovations are as follows: A dedicated physical architecture of "six major hardware modules + three types of standardized buses" is proposed, clearly defining the physical location, electrical interface, and signal flow of each module. This materializes the abstract attitude optimization strategy into an integrable and verifiable hardware assembly. Unlike existing distributed safety systems, the interfaces and data formats between modules are standardized, resulting in extremely high engineering feasibility. It is not a simple aggregation of software functions, but a dedicated hardware system designed for the new paradigm of "active attitude adjustment before collision."

[0133] The perception-decision chain is hardware-based: core logic such as the enhanced TTC algorithm, FPGA-accelerated KNN search, and suspension nonlinear interpolation function are solidified into the descriptor encapsulation function of the central platform and the hardware query function of the decision-maker, respectively. This makes the algorithm an inherent attribute of the hardware, solves the uncertainty problem of traditional software decision-making, and improves the accuracy of data acquisition and processing through dedicated sensor deployment and synchronization mechanism.

[0134] Precise and controllable control-execution collaborative path: Construct a hierarchical MPC collaborative control architecture, clarify the complete conversion chain of "CAN FD command → dedicated interface signal → high dynamic actuator action", combine the actuator's dedicated performance parameter design to achieve high-precision attitude adjustment in a short time, and after superimposing the redundant design between modules, the timing determinism is stronger, reproducible and detectable.

[0135] Active and Passive Safety Proactive Linkage Paradigm: A dedicated linkage interface based on automotive Ethernet is designed to achieve cross-domain synchronization of attitude and Δv information through standardized APDU messages, upgrading the passive safety system from "post-event triggering" to "pre-event pre-setting," thus improving collision protection effectiveness. The proactive linkage design utilizes a dedicated linkage interface based on automotive Ethernet to broadcast the final attitude and predicted intensity to the passive system in real time before a collision occurs, allowing it to pre-adjust parameters. Deterministic Timing: A total time budget of ≤230ms and sub-budgets for each stage are defined for the entire process of "perception → decision → control → linkage," and this is guaranteed through hardware selection and system design.

[0136] A closed-loop knowledge system of simulation-solidification-invocation: This system proposes a technical chain of "offline simulation optimization of millions of scenarios → solidification of hardware knowledge base → online millisecond-level invocation", which transforms the complex collision and crashworthiness optimization problem into a fast query problem of vehicle-mounted embedded systems. It balances optimization depth and response timeliness, and provides core support for the implementation of active collision-resistant technology.

[0137] The vehicle attitude adjustment method of this invention elevates vehicle collision safety from a passive domain dependent on component performance and random collision attitude to a new level of active technology that can be systematically designed, actively controlled, and quantitatively optimized. It can proactively "shape" the vehicle into the physical form best suited to withstand the specific impact just before a collision, and work in conjunction with passive safety systems to construct a three-dimensional, precise, and forward-looking life protection defense. This is not only a significant supplement to existing safety technologies but also a redefinition of the future intelligent vehicle safety architecture, possessing significant technological advancements, a clear engineering practice path, and immense industrial application value.

[0138] It is understood that the vehicle attitude adjustment method of this invention, through an integrated architecture of "perception-planning-control-linkage," achieves proactive adjustment of the vehicle's attitude before a collision to optimally mitigate collision risks, and the technical solution possesses clear physical characteristics and protectability. This system is an integrated device with deeply bound hardware and software. Its core includes a global perception hardware array, a central fusion control unit, a chassis collaborative actuator cluster, and active and passive safety linkage interfaces. Each component forms a closed-loop control system through a dedicated bus architecture. Its overall architecture ensures millisecond-level response and precise execution from collision threat identification to attitude adjustment.

[0139] The following is for reference. Figure 11 A flowchart describing a method for adjusting the attitude of a vehicle according to an embodiment of the present invention.

[0140] like Figure 11 As shown, the method for adjusting the attitude of a vehicle according to an embodiment of the present invention includes the following steps.

[0141] Step S20: The global perception sensor acquires perception data within a preset range of the target vehicle.

[0142] Step S21: The perception data is transmitted to the central fusion and prediction calculation module to determine whether the vehicle and the target vehicle meet the collision conditions. When it is predicted that the vehicle and the target vehicle will collide, the collision description information under the collision condition, i.e., the collision condition descriptor, is obtained.

[0143] Step S22: The collision description information, i.e. the collision condition descriptor, is sent to the collision-resistant attitude planning decision-maker to plan the target attitude under the collision condition within milliseconds.

[0144] In step S23, the chassis cooperative domain controller calculates the target attitude into the same command sequence of the actuators, and determines the operation of the active suspension, electronic stability control system and active stabilizer bar.

[0145] In step S24, the linkage interface sends the target attitude and equivalent velocity change to passive systems such as airbags and seat belts.

[0146] In summary, the present invention achieves a paradigm shift in active and passive safety technology, transforming from traditional "post-collision remediation" and "information integration" to "pre-collision optimization." Its fundamental idea is that, when a collision is unavoidable, the vehicle no longer passively accepts predetermined physical conditions, but actively and intelligently adjusts itself to a pre-set "optimal crashworthiness posture" that maximizes the energy absorption potential of the vehicle's structure, thereby reducing the severity of the collision at its source.

[0147] The vehicle according to an embodiment of the present invention is described below.

[0148] like Figure 12As shown, the vehicle 2 in this embodiment of the invention includes a vehicle attitude adjustment system 1.

[0149] According to embodiments of the present invention, the vehicle, through an integrated "perception-planning-control-linkage" architecture, actively adjusts its attitude before a collision to optimally mitigate the risk of collision, and the technical solution possesses clear physical characteristics and protectability. This system is an integrated device with deeply integrated hardware and software. Its core includes a global perception hardware array, a central fusion control unit, a chassis collaborative actuator cluster, and active and passive safety linkage interfaces. Each component forms a closed-loop control system through a dedicated bus architecture. The overall architecture ensures millisecond-level response and precise execution from collision threat identification to attitude adjustment.

[0150] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0151] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method of adjusting the attitude of a vehicle, characterized by, The vehicle has an intelligent driving function, and the method comprises: When it is predicted that the vehicle will collide with a target vehicle, collision description information of a collision condition is obtained; A target attitude of the vehicle is generated according to the collision description information; The attitude of the vehicle is adjusted according to the target attitude.

2. The attitude adjustment method of a vehicle according to claim 1, characterized by, In the process of adjusting the attitude of the vehicle according to the target attitude, the process further comprises: The attitude of an internal protection device of the vehicle is adjusted according to the target attitude and the collision description information.

3. The attitude adjustment method of a vehicle according to claim 1, characterized by, The target attitude of the vehicle is generated according to the collision description information, which comprises: When the collision condition is included in a preset collision condition-target attitude mapping relationship database, the target attitude is determined according to the collision description information and the preset collision condition-target attitude mapping relationship database; When the collision condition is not included in the preset collision condition-target attitude mapping relationship database, the target attitude is determined by performing interpolation calculation on adjacent conditions of the collision condition based on a physical interpolation function and the preset collision condition-target attitude mapping relationship database.

4. The attitude adjustment method of a vehicle according to claim 1, characterized by, The process of predicting that the vehicle will collide with a target vehicle comprises: Perception data within a preset range of the target vehicle is obtained; Collision parameters are determined according to target object information and geometric features in the perception data; The collision parameters are judged in stages, and when the result of the stage judgment meets a preset requirement, it is predicted that the vehicle will collide with the target vehicle.

5. The attitude adjustment method of a vehicle according to claim 1 or 4, characterized by, The process of obtaining collision description information of a collision condition comprises: Motion geometry analysis results, collision parameters and a target collision condition are determined according to perception data; The collision description information of the collision condition is determined according to the motion geometry analysis results, the collision parameters and the target collision condition.

6. The attitude adjustment method of a vehicle according to claim 1, characterized by, The process of adjusting the attitude of the vehicle according to the target attitude comprises: Vehicle attitude control instructions are determined according to the target attitude; In response to the vehicle attitude control instructions, each actuator of a chassis control system of the vehicle is driven to move in linkage.

7. The attitude adjustment method of a vehicle according to claim 2, characterized by, The process of adjusting the attitude of the internal protection device of the vehicle according to the target attitude and the collision description information comprises: Control instructions of the internal protection device of the vehicle are determined according to equivalent speed changes in the target attitude and the collision description information; In response to the control instructions, the internal protection device of the vehicle is controlled to move in linkage.

8. The attitude adjustment method of a vehicle according to claim 3, characterized by, Before the target attitude is determined, the process further comprises: The target attitude is subjected to safety verification.

9. A posture adjustment system of a vehicle, characterized by, The system comprises: A central fusion and prediction calculation module, which is used to obtain collision description information of a collision condition when it is predicted that a vehicle will collide with a target vehicle; A collision-resistant attitude planning decision maker, which is connected with the central fusion and prediction calculation module and is used to generate a target attitude of the vehicle according to the collision description information; A chassis coordination domain controller, which is connected with the collision-resistant attitude planning decision maker and is used to adjust the attitude of the vehicle according to the target attitude.

10. A vehicle characterized by comprising: The system comprises: The attitude adjustment system of the vehicle according to claim 9.

Images