Method, device and equipment for adjusting posture before vehicle collision and storage medium
By using an artificial intelligence decision-making model and a multi-level degradation strategy, the vehicle's collision posture is actively adjusted, which solves the problem of reduced occupant protection performance in real collisions and achieves improved safety performance and occupant protection under extreme conditions.
Patent Information
- Application Number
- CN202511460273.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-13
- Publication Date
- 2026-01-02
AI Technical Summary
Existing vehicle safety systems suffer from reduced occupant protection performance in real-world collisions due to the random and variable nature of the collision posture, and lack the ability to actively intervene in the final collision posture of the vehicle.
By employing an artificial intelligence decision-making model combined with a multi-level degradation strategy, collision situation data is collected and prioritized according to protection effectiveness. The vehicle attitude is proactively adjusted to optimize the collision attitude, ensuring the generation of feasible strategies under conditions of limited computing resources and time constraints.
It significantly improves the safety performance of vehicles in real-world collision scenarios, reduces the risk of occupant injury, reflects the "choose the lesser of two evils" decision-making logic in emergency situations, and ensures the robustness and practicality of the system's decision-making in extreme scenarios.
Smart Images
Figure CN121246784A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle control technology, specifically to a method, device, equipment, and storage medium for adjusting the attitude of a vehicle before a collision. Background Technology
[0002] Standard crash tests are conducted under idealized laboratory conditions, assuming that the vehicle is in a specific, pre-defined, standardized posture at the time of the collision (e.g., 100% frontal overlap rigid wall collision, 40% offset collision, etc.). However, in the complex conditions of real-world driving, potential collision scenarios have a high degree of uncertainty and randomness.
[0003] Existing vehicle safety systems lack the ability to actively intervene in the final collision posture of a vehicle when a collision is unavoidable. Due to the significant differences between actual collision conditions and standard laboratory test conditions, vehicles often collide in non-ideal, non-standard, and high-risk postures, causing their inherent passive safety designs to fail to perform as expected in real accidents, ultimately leading to an increased risk of personal injury. Summary of the Invention
[0004] In view of the above problems, embodiments of the present invention provide a method, apparatus, device and storage medium for adjusting the attitude of a vehicle before a collision, which is used to solve the problem of decreased occupant protection performance caused by the random and variable collision attitude of a vehicle in a real collision in the prior art.
[0005] According to one aspect of the present invention, a method for adjusting the attitude of a vehicle before a collision is provided, the method comprising:
[0006] In response to the determination that a collision between the vehicle and the target object is inevitable, the system collects current collision situation data.
[0007] Based on the collision situation data and the preset artificial intelligence decision model, a collision attitude adjustment strategy is determined. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which includes multiple collision attitude strategies sorted by protection effectiveness priority.
[0008] Based on the collision attitude adjustment strategy output by the preset artificial intelligence decision model, the vehicle is controlled to make corresponding attitude adjustments;
[0009] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0010] According to another aspect of the present invention, a vehicle pre-collision attitude adjustment device is provided, comprising:
[0011] The first control module is used to control the acquisition of current collision situation data in response to the determination that a collision between the vehicle and the target is inevitable.
[0012] The determination module is used to determine the collision attitude adjustment strategy based on the collision situation data and the preset artificial intelligence decision model. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which includes multiple collision attitude strategies sorted by protection effectiveness priority.
[0013] The second control module is used to control the vehicle to make corresponding attitude adjustments based on the collision attitude adjustment strategy output by the preset artificial intelligence decision model.
[0014] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0015] According to another aspect of the present invention, a vehicle pre-collision attitude adjustment device is provided, comprising:
[0016] The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus.
[0017] The memory is used to store at least one executable instruction that causes the processor to perform the vehicle attitude adjustment method as described above.
[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing at least one executable instruction that causes the vehicle pre-collision attitude adjustment device / apparatus to operate the vehicle pre-collision attitude adjustment method as described above.
[0019] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0020] The above description is merely an overview of the technical solutions of the embodiments of the present invention. In order to better understand the technical means of the embodiments of the present invention and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description
[0021] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:
[0022] Figure 1 A flowchart illustrating a first embodiment of the vehicle pre-collision attitude adjustment method provided by the present invention is shown.
[0023] Figure 2 A flowchart illustrating a second embodiment of the vehicle pre-collision attitude adjustment method provided by the present invention is shown.
[0024] Figure 3 The diagram illustrates a typical application scenario of the vehicle pre-collision attitude adjustment method provided by the present invention.
[0025] Figure 4 a shows a schematic diagram of the collision posture of the frontal offset collision strategy provided by the present invention;
[0026] Figure 4 b shows a collision posture diagram of the collision strategy provided by the present invention that enables control of the vehicle's trajectory after a collision.
[0027] Figure 4 c shows a schematic diagram of the collision posture of the collision strategy provided by the present invention, which involves impacting the part of the vehicle structure with the highest strength.
[0028] Figure 5 A flowchart of the vehicle attitude adjustment method before collision provided by the present invention is shown;
[0029] Figure 6 A schematic diagram of an embodiment of the vehicle pre-collision attitude adjustment device provided by the present invention is shown;
[0030] Figure 7 A schematic diagram of an embodiment of the vehicle pre-collision attitude adjustment device provided by the present invention is shown.
[0031] Figure 8 A structural schematic diagram of an embodiment of the vehicle provided by the present invention is shown. Detailed Implementation
[0032] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.
[0033] Figure 1 A flowchart illustrating a first embodiment of the vehicle pre-collision attitude adjustment method of the present invention is shown, which can be executed by a domain controller in the vehicle. Figure 1 As shown, the method includes the following steps:
[0034] Step 110: In response to the determination that a collision between the vehicle and the target object is inevitable, control the acquisition of current collision situation data.
[0035] Collision situation data refers to the set of all input information required for an artificial intelligence decision-making model to calculate and generate the optimal collision attitude adjustment strategy after a collision is deemed unavoidable and is triggered. This data set provides a complete input space for the AI decision-making model and is the data foundation for generating reliable and executable collision attitude adjustment strategies.
[0036] Step 120: Based on the collision situation data and the preset artificial intelligence decision model, determine the collision attitude adjustment strategy. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which contains multiple collision attitude strategies sorted by protection effectiveness priority.
[0037] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0038] Among them, the first collision posture strategy that is determined to be feasible can be identified as the target collision posture.
[0039] In this step, the collected real-time collision situation data is input into a pre-trained artificial intelligence decision-making model. This model then performs step-by-step reasoning and decision-making based on a built-in multi-level degradation strategy set. This multi-level degradation strategy set is a sequence of strategies pre-constructed based on collision biomechanics, vehicle dynamics, and numerous accident scenario simulations, with each collision attitude strategy ranked from highest to lowest protective effectiveness. The core of this step lies in using the artificial intelligence model to map complex real-time data to the optimal collision attitude strategy and automatically activating degradation schemes when the optimal collision attitude strategy is infeasible, thereby ensuring the system's decision-making robustness under extreme conditions.
[0040] Step 130: Control the vehicle to make corresponding attitude adjustments based on the collision attitude adjustment strategy output by the preset artificial intelligence decision-making model.
[0041] In this step, the determined strategy is translated into specific vehicle control commands, and the chassis, powertrain, and steering systems are coordinated to perform the corresponding actions.
[0042] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0043] Figure 2 A flowchart illustrating another embodiment of the vehicle pre-collision attitude adjustment method of the present invention is shown, which can be collaboratively executed by one or more domain controllers deployed in the vehicle. Figure 2 As shown, the method includes the following steps:
[0044] Step 210: In response to the determination that a collision between the vehicle and the target object is inevitable, control the acquisition of current collision situation data.
[0045] In this step, when the collision is determined to be unavoidable by conventional obstacle avoidance methods based on sensor fusion and prediction algorithms, a high-precision, multimodal data acquisition mechanism is triggered.
[0046] In one alternative approach, the collision situation data includes at least one of the following: target type data, vehicle trajectory, target trajectory, vehicle occupant status, vehicle motion status, and surrounding environment data.
[0047] Specifically, the target object type data is used to distinguish the physical attributes and categories of the collision object, such as whether it is a vehicle (further distinguishing between heavy trucks, medium-sized buses, light passenger cars, etc.), pedestrian, non-motorized vehicle driver, motorcycle, or special vehicle (such as ambulance, fire truck). The identification result directly affects the collision energy assessment and strategy selection.
[0048] The vehicle's trajectory and the target object's trajectory include information such as their respective positions, speeds, accelerations, and yaw rates during the current and predicted time periods. This information is used to construct the kinematic relationship between the two objects and predict the collision point and collision pattern.
[0049] The occupant status of this vehicle includes the occupancy status of each seat, occupant body type classification, seat belt usage status, and seat position. This information is used to assess the risk of personal injury under different collision strategies in decision-making, prioritizing occupant safety.
[0050] The vehicle's motion status includes current speed, heading angle, wheel speed, suspension height, and body attitude angle, which are used to assess the vehicle's controllability and the available margin of the actuators in real time.
[0051] The surrounding environment data includes the current road adhesion coefficient, road topology (such as curves and slopes), lane line information, roadside avoidable areas (such as soft soil slopes), weather conditions (such as rain, snow, and fog), and light intensity. This data is used to determine the feasibility and safety of a specific strategy in the current environment.
[0052] In one alternative approach, in response to the determination that a collision between the vehicle and the target object is inevitable, before controlling the acquisition of current collision situation data, the vehicle pre-collision attitude adjustment method of the present invention may further include the following steps:
[0053] Control and monitor the movement trajectory of objects around the vehicle;
[0054] Identify surrounding objects whose moving trajectories pose a collision risk with the vehicle's predicted trajectory, and send the identification results to the cloud server for risk level classification to obtain the classification results;
[0055] High-risk objects in the classification results are identified as targets that need to be continuously tracked.
[0056] In this embodiment, the movement trajectories of objects around the vehicle are continuously monitored, and objects that intersect or conflict with the vehicle's predicted trajectory are identified. Their features are then uploaded to a cloud server, where the more powerful computing power of the cloud is used for behavior prediction and risk level classification. Finally, objects classified as high-risk are identified as the tracking targets of this method, thereby achieving optimized allocation of sensing resources and early focusing on key threats.
[0057] Step 220: Based on the input of collision situation data, in the preset artificial intelligence decision model, perform the following feasibility assessments of collision attitude strategies in sequence: assess the feasibility of the frontal collision strategy; if not feasible, assess the feasibility of the frontal offset collision strategy with an offset angle less than a preset threshold; if not feasible, assess the feasibility of the collision strategy that can control the vehicle's trajectory after the collision; if not feasible, assess the feasibility of the collision strategy that hits the vehicle with the strongest part of its structure; and determine the first feasible collision attitude strategy as the target collision attitude.
[0058] Based on the input of collision situation data, the feasibility assessment of collision attitude strategies is sequentially performed in the preset artificial intelligence decision-making model, and the first feasible strategy is determined as the target collision attitude. This step is the decision-making link of this method. The preset artificial intelligence decision-making model has a multi-level set of degradation strategies ordered by protection effectiveness priority. It evaluates the feasibility of the following typical strategies in order of priority: first, the frontal collision strategy is evaluated; if it is determined to be infeasible due to insufficient execution conditions, the frontal offset collision strategy with an offset angle less than a preset threshold is downgraded for evaluation; if it is still infeasible, the collision strategy that can control the vehicle's trajectory after the collision is further evaluated (e.g., by controlling the lateral and longitudinal dynamics to avoid secondary collisions); if none of the above strategies are feasible, the strategy of colliding with the part of the vehicle with the highest structural strength is finally evaluated as the backup plan.
[0059] In one alternative approach, a collision strategy that controls the trajectory of a vehicle after a collision aims to prevent secondary collisions or entry into dangerous areas through proactive intervention.
[0060] This strategy can be further configured to guide the vehicle's trajectory after a collision to avoid secondary collisions with third-party obstacles or entering dangerous areas, ensuring that the vehicle eventually comes to a stable and controllable stop.
[0061] In one alternative approach, evaluating the feasibility of a collision attitude strategy may include the following steps:
[0062] Real-time simulation tests are conducted based on vehicle dynamics models and collision situation data to calculate the vehicle control quantities required to execute the corresponding collision attitude strategies.
[0063] Determine whether the vehicle control quantity exceeds the physical limits of the vehicle's actuators, and / or determine whether there is sufficient time remaining before the collision to complete the execution of the vehicle control quantity;
[0064] If the physical limits of the vehicle's actuators are not exceeded and the execution of vehicle control quantities is sufficient, the assessment is feasible; otherwise, the assessment is infeasible.
[0065] In this embodiment, simulations are performed based on the vehicle dynamics model and real-time collision situation data to calculate the specific control quantities (such as braking torque, steering angle, and drive torque) required to execute a particular strategy. Then, it is determined whether these control quantities exceed the physical operating limits of the vehicle's actuators (such as the braking system, steering motor, and power unit), and whether they can be stably executed within the remaining time window before the collision. The strategy is only evaluated as feasible when both physical feasibility and time adequacy constraints are simultaneously met.
[0066] Step 230: Control the vehicle to make corresponding attitude adjustments based on the collision attitude adjustment strategy output by the preset artificial intelligence decision-making model.
[0067] In this step, the output collision attitude adjustment strategy is parsed into specific chassis and powertrain control commands, and sent to each actuator via the vehicle bus. Through coordinated control of the braking, steering, drive, and suspension systems, precise and rapid adjustment of the vehicle's lateral and longitudinal attitudes is achieved, bringing it to the target collision attitude.
[0068] In one alternative embodiment, the vehicle attitude adjustment method before a collision of the present invention may further include the following steps:
[0069] After a collision occurs, the system collects process data and result data related to the collision.
[0070] Process data and result data are uploaded to the cloud server to iteratively train and optimize the global AI decision-making model in the cloud.
[0071] In this embodiment, data from the entire collision process (including decision inputs, control commands, vehicle responses, and the final collision result) is collected and uploaded to a cloud server. The cloud server uses massive amounts of real-vehicle data to iteratively train and optimize the parameters of the global artificial intelligence decision-making model, and then distributes the improved model to the vehicle via OTA (Over-The-Air).
[0072] Reference Figure 3 The diagram shown illustrates a typical application scenario of the vehicle pre-collision attitude adjustment method provided by this invention. As shown, vehicle A and vehicle B are about to collide, and the collision has been determined to be unavoidable. Vehicle A is equipped with the attitude adjustment system of this invention, and its execution control flow is as follows:
[0073] Vehicle A continuously monitors the relative motion state of vehicle B and identifies its vehicle model, load, speed and acceleration changes in real time.
[0074] When a collision is unavoidable, vehicle A, based on surrounding environmental data and the real-time motion status of both vehicles, uses a pre-set artificial intelligence decision-making model to sequentially assess the feasibility of multiple collision posture strategies to determine the optimal collision solution:
[0075] First, assess the feasibility of a frontal collision strategy.
[0076] If this is not feasible, then evaluate the feasibility of a frontal offset collision strategy with an offset angle of less than 15°. For example... Figure 4 Figure a shows a schematic diagram of the collision posture of the frontal offset collision strategy provided by the present invention, the goal of which is to control the collision angle between the two vehicles within 15°. Calculation results show that, under current conditions, it is impossible to adjust the vehicles to this posture before the collision.
[0077] Subsequently, the feasibility of collision strategies that could control the vehicle's trajectory after a collision is assessed. For example... Figure 4 Figure b shows a collision posture diagram of the collision strategy provided by the present invention, which enables control of the vehicle's trajectory after a collision, and predicts the trajectory after the collision. Calculation results show that if a collision occurs in the current state, vehicle A will rotate to the left and shift its position, and its trajectory will lead to a secondary uncontrollable collision with the oncoming truck.
[0078] Next, the feasibility of a collision strategy that targets the vehicle at its structurally strongest point is assessed. For example... Figure 4 Figure c shows a schematic diagram of the collision posture of the collision strategy provided by this invention, which involves impacting the vehicle at its structurally strongest point. Based on the structural characteristics of vehicle B, calculations show that the collision point should be concentrated in the B-pillar area on the right side of the vehicle. This position can fully utilize the design effectiveness of the vehicle's energy-absorbing structure and rigid support, while also guiding the vehicle to move to the left after the collision, allowing for a controlled secondary contact with the oncoming truck in the B-pillar area, thereby minimizing the overall risk.
[0079] Ultimately, based on Figure 4 The strategy determined by c generates control commands to coordinate the chassis and powertrain for deceleration control, precisely adjusting the vehicle's position so that the right B-pillar is positioned at the target impact point. Simultaneously, the left and right airbags are deployed in advance to provide pre-emptive protection for the occupants.
[0080] Reference Figure 5The diagram shows a flowchart of the vehicle pre-collision attitude adjustment method provided by this invention. First, it acquires environmental data surrounding the vehicle, monitors objects around the vehicle, determines the physical characteristics of the detected objects, analyzes and determines the collision mass characteristics of the objects, and tracks the motion trajectories of these objects to identify high-risk objects. It further evaluates the physical characteristics of high-risk objects, analyzes their collision mass characteristics, and calculates the potential impact of the collision on the vehicle structure. High-risk objects are tracked in real time, and the target collision object is determined when a collision is unavoidable. Collision situation data is collected, and a collision attitude adjustment strategy is determined based on this data. Before the collision occurs, the vehicle is controlled to perform attitude adjustments according to the collision attitude adjustment strategy to reduce the impact of the collision. Data from the entire collision process is collected and uploaded to a cloud server for iterative upgrades of the global artificial intelligence decision-making model, which can then be deployed to the vehicle.
[0081] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0082] Figure 6 A schematic diagram of an embodiment of the vehicle pre-collision attitude adjustment device of the present invention is shown. Figure 6 As shown, the device 600 includes: a first control module 610, a determination module 620, and a second control module 630.
[0083] The first control module is used to control the acquisition of current collision situation data in response to the determination that a collision between the vehicle and the target is inevitable.
[0084] The determination module is used to determine the collision attitude adjustment strategy based on collision situation data and a preset artificial intelligence decision model. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which contains multiple collision attitude strategies sorted by protection effectiveness priority.
[0085] The second control module is used to control the vehicle to make corresponding attitude adjustments based on the collision attitude adjustment strategy output by the preset artificial intelligence decision model.
[0086] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0087] In one alternative approach, the module is specifically used for:
[0088] Based on the input of collision situation data, the feasibility assessment of the following collision attitude strategies is performed sequentially in a pre-defined artificial intelligence decision-making model:
[0089] Assess the feasibility of a frontal collision strategy;
[0090] If not feasible, evaluate the feasibility of a frontal offset collision strategy with an offset angle less than a preset threshold.
[0091] If not feasible, assess the feasibility of collision strategies that can control the vehicle's trajectory after a collision.
[0092] If not feasible, assess the feasibility of a collision strategy that targets the part of the vehicle structure with the highest strength.
[0093] The first collision attitude strategy that is assessed as feasible is determined as the target collision attitude.
[0094] In one alternative approach, the module is specifically used for:
[0095] Real-time simulation tests are conducted based on vehicle dynamics models and collision situation data to calculate the vehicle control quantities required to execute the corresponding collision attitude strategies.
[0096] Determine whether the vehicle control quantity exceeds the physical limits of the vehicle's actuators, and / or determine whether there is sufficient time remaining before the collision to complete the execution of the vehicle control quantity;
[0097] If the physical limits of the vehicle's actuators are not exceeded and the execution of vehicle control quantities is sufficient, the assessment is feasible; otherwise, the assessment is infeasible.
[0098] In one alternative embodiment, the vehicle pre-collision attitude adjustment device is further used for:
[0099] Control and monitor the movement trajectory of objects around the vehicle;
[0100] Identify surrounding objects whose moving trajectories pose a collision risk with the vehicle's predicted trajectory, and send the identification results to the cloud server for risk level classification to obtain the classification results;
[0101] High-risk objects in the classification results are identified as targets that need to be continuously tracked.
[0102] In one alternative embodiment, the vehicle pre-collision attitude adjustment device is further used for:
[0103] After a collision occurs, the system collects process data and result data related to the collision.
[0104] Process data and result data are uploaded to the cloud server to iteratively train and optimize the global AI decision-making model in the cloud.
[0105] In one alternative approach, the collision situation data includes at least one of the following: target type data, vehicle trajectory, target trajectory, vehicle occupant status, vehicle motion status, and surrounding environment data.
[0106] In one alternative approach, a collision strategy that controls the trajectory of a vehicle after a collision aims to prevent secondary collisions or entry into dangerous areas through proactive intervention.
[0107] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0108] Figure 7 The diagram shows a structural schematic of an embodiment of the vehicle attitude adjustment device before a collision according to the present invention. The specific embodiments of the present invention do not limit the specific implementation of the vehicle attitude adjustment device before a collision.
[0109] like Figure 7 As shown, the vehicle's pre-collision attitude adjustment device may include: a processor 702, a communications interface 704, a memory 706, and a communications bus 708.
[0110] The processor 702, communication interface 704, and memory 706 communicate with each other via communication bus 708. Communication interface 704 is used to communicate with other network elements such as clients or other servers. Processor 702 executes program 710, specifically performing the relevant steps described in the embodiment of the vehicle collision pre-collision attitude adjustment method.
[0111] Specifically, program 710 may include program code, which includes computer-executable instructions.
[0112] The processor 702 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The vehicle pre-collision attitude adjustment device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.
[0113] Memory 706 is used to store program 710. Memory 706 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.
[0114] Specifically, program 710 can be called by processor 702 to cause the vehicle pre-collision attitude adjustment device to perform the following operations:
[0115] In response to the determination that a collision between the vehicle and the target object is inevitable, the system collects current collision situation data.
[0116] Based on collision situation data and a preset artificial intelligence decision-making model, a collision attitude adjustment strategy is determined. The preset artificial intelligence decision-making model has a built-in multi-level degradation strategy set, which contains multiple collision attitude strategies sorted by protection effectiveness priority.
[0117] Based on the collision attitude adjustment strategy output by the preset artificial intelligence decision-making model, the vehicle is controlled to make corresponding attitude adjustments.
[0118] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0119] In an alternative manner, program 710 is invoked by processor 702 to cause the vehicle pre-collision attitude adjustment device to perform the following operations:
[0120] Based on the input of collision situation data, the feasibility assessment of the following collision attitude strategies is performed sequentially in a pre-defined artificial intelligence decision-making model:
[0121] Assess the feasibility of a frontal collision strategy;
[0122] If not feasible, evaluate the feasibility of a frontal offset collision strategy with an offset angle less than a preset threshold.
[0123] If not feasible, assess the feasibility of collision strategies that can control the vehicle's trajectory after a collision.
[0124] If not feasible, assess the feasibility of a collision strategy that targets the part of the vehicle structure with the highest strength.
[0125] The first collision attitude strategy that is assessed as feasible is determined as the target collision attitude.
[0126] In an alternative manner, program 710 is invoked by processor 702 to cause the vehicle pre-collision attitude adjustment device to perform the following operations:
[0127] Real-time simulation tests are conducted based on vehicle dynamics models and collision situation data to calculate the vehicle control quantities required to execute the corresponding collision attitude strategies.
[0128] Determine whether the vehicle control quantity exceeds the physical limits of the vehicle's actuators, and / or determine whether there is sufficient time remaining before the collision to complete the execution of the vehicle control quantity;
[0129] If the physical limits of the vehicle's actuators are not exceeded and the execution of vehicle control quantities is sufficient, the assessment is feasible; otherwise, the assessment is infeasible.
[0130] In an alternative manner, program 710 is invoked by processor 702 to cause the vehicle pre-collision attitude adjustment device to perform the following operations:
[0131] Control and monitor the movement trajectory of objects around the vehicle;
[0132] Identify surrounding objects whose moving trajectories pose a collision risk with the vehicle's predicted trajectory, and send the identification results to the cloud server for risk level classification to obtain the classification results;
[0133] High-risk objects in the classification results are identified as targets that need to be continuously tracked.
[0134] In an alternative manner, program 710 is invoked by processor 702 to cause the vehicle pre-collision attitude adjustment device to perform the following operations:
[0135] After a collision occurs, the system collects process data and result data related to the collision.
[0136] Process data and result data are uploaded to the cloud server to iteratively train and optimize the global AI decision-making model in the cloud.
[0137] In one alternative approach, the collision situation data includes at least one of the following: target type data, vehicle trajectory, target trajectory, vehicle occupant status, vehicle motion status, and surrounding environment data.
[0138] In one alternative approach, a collision strategy that controls the trajectory of a vehicle after a collision aims to prevent secondary collisions or entry into dangerous areas through proactive intervention.
[0139] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0140] Figure 8 A structural schematic diagram of an embodiment of the vehicle of the present invention is shown. (As shown) Figure 8 As shown, the vehicle 800 includes: sensors, one or more processors, and a communication interface;
[0141] The sensor is used to collect collision situation data;
[0142] The processor is used to execute the steps in the above embodiments of the vehicle attitude adjustment method before collision.
[0143] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0144] This invention provides a computer-readable storage medium storing at least one executable instruction. When the executable instruction is executed on a vehicle pre-collision attitude adjustment device / apparatus, it causes the vehicle pre-collision attitude adjustment device / apparatus to perform the vehicle pre-collision attitude adjustment method in any of the above method embodiments.
[0145] Specifically, the executable instructions can be used to cause the vehicle pre-collision attitude adjustment device / mechanism to perform the following operations:
[0146] In response to the determination that a collision between the vehicle and the target object is inevitable, the system collects current collision situation data.
[0147] Based on collision situation data and a preset artificial intelligence decision-making model, a collision attitude adjustment strategy is determined. The preset artificial intelligence decision-making model has a built-in multi-level degradation strategy set, which contains multiple collision attitude strategies sorted by protection effectiveness priority.
[0148] Based on the collision attitude adjustment strategy output by the preset artificial intelligence decision-making model, the vehicle is controlled to make corresponding attitude adjustments.
[0149] The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
[0150] In one alternative approach, executable instructions cause the vehicle pre-collision attitude adjustment device / apparatus to perform the following operations:
[0151] Based on the input of collision situation data, the feasibility assessment of the following collision attitude strategies is performed sequentially in a pre-defined artificial intelligence decision-making model:
[0152] Assess the feasibility of a frontal collision strategy;
[0153] If not feasible, evaluate the feasibility of a frontal offset collision strategy with an offset angle less than a preset threshold.
[0154] If not feasible, assess the feasibility of collision strategies that can control the vehicle's trajectory after a collision.
[0155] If not feasible, assess the feasibility of a collision strategy that targets the part of the vehicle structure with the highest strength.
[0156] The first collision attitude strategy that is assessed as feasible is determined as the target collision attitude.
[0157] In one alternative approach, executable instructions cause the vehicle pre-collision attitude adjustment device / apparatus to perform the following operations:
[0158] Real-time simulation tests are conducted based on vehicle dynamics models and collision situation data to calculate the vehicle control quantities required to execute the corresponding collision attitude strategies.
[0159] Determine whether the vehicle control quantity exceeds the physical limits of the vehicle's actuators, and / or determine whether there is sufficient time remaining before the collision to complete the execution of the vehicle control quantity;
[0160] If the physical limits of the vehicle's actuators are not exceeded and the execution of vehicle control quantities is sufficient, the assessment is feasible; otherwise, the assessment is infeasible.
[0161] In one alternative approach, executable instructions cause the vehicle pre-collision attitude adjustment device / apparatus to perform the following operations:
[0162] Control and monitor the movement trajectory of objects around the vehicle;
[0163] Identify surrounding objects whose moving trajectories pose a collision risk with the vehicle's predicted trajectory, and send the identification results to the cloud server for risk level classification to obtain the classification results;
[0164] High-risk objects in the classification results are identified as targets that need to be continuously tracked.
[0165] In one alternative approach, executable instructions cause the vehicle pre-collision attitude adjustment device / apparatus to perform the following operations:
[0166] After a collision occurs, the system collects process data and result data related to the collision.
[0167] Process data and result data are uploaded to the cloud server to iteratively train and optimize the global AI decision-making model in the cloud.
[0168] In one alternative approach, the collision situation data includes at least one of the following: target type data, vehicle trajectory, target trajectory, vehicle occupant status, vehicle motion status, and surrounding environment data.
[0169] In one alternative approach, a collision strategy that controls the trajectory of a vehicle after a collision aims to prevent secondary collisions or entry into dangerous areas through proactive intervention.
[0170] This invention proactively adjusts the vehicle to a collision posture with superior protective effectiveness, enabling the vehicle's inherent passive safety design to better fulfill its intended protective function in real-world accidents, thereby significantly reducing the risk of occupant injury. By employing a pre-defined set of multi-level degradation strategies prioritized by protective effectiveness, and using an artificial intelligence model to sequentially evaluate their feasibility, this method reflects the human driver's decision-making logic of "choosing the lesser of two evils" in emergency situations. This design ensures that the system can quickly and reliably generate feasible strategies even under demanding conditions of limited computing resources and time constraints. Even if the optimal strategy cannot be implemented due to limitations of the actuator or insufficient time, the system can automatically degrade to a suboptimal strategy, thus guaranteeing the system's robustness and practicality in various extreme scenarios. By introducing artificial intelligence decision-making and a multi-level degradation strategy mechanism, this solution achieves proactive and intelligent control of the vehicle's posture in unavoidable collision scenarios, significantly improving the vehicle's safety performance in real-world collision scenarios.
[0171] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Furthermore, the embodiments of this invention are not directed to any particular programming language.
[0172] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. Similarly, for the sake of brevity and to aid in understanding one or more aspects of the invention, in the description of exemplary embodiments of the invention above, various features of the embodiments are sometimes grouped together in a single embodiment, figure, or description thereof. The claims, which follow the detailed description, are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of the invention.
[0173] Those skilled in the art will understand that the modules in the device of the embodiment can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiment can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components, except that at least some of such features and / or processes or units are mutually exclusive.
[0174] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.
Claims
1. A method for adjusting the attitude of a vehicle before a collision, characterized in that, The method includes: In response to the determination that a collision between the vehicle and the target object is inevitable, the system collects current collision situation data. Based on the collision situation data and the preset artificial intelligence decision model, a collision attitude adjustment strategy is determined. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which includes multiple collision attitude strategies sorted by protection effectiveness priority. Based on the collision attitude adjustment strategy output by the preset artificial intelligence decision-making model, the vehicle is controlled to make corresponding attitude adjustments; The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
2. The method according to claim 1, characterized in that, The step of determining the collision attitude adjustment strategy based on the collision situation data and a preset artificial intelligence decision-making model includes: Based on the input of the collision situation data, the following feasibility assessments of collision attitude strategies are performed sequentially in the preset artificial intelligence decision-making model: Assess the feasibility of a frontal collision strategy; If not feasible, evaluate the feasibility of a frontal offset collision strategy with an offset angle less than a preset threshold. If not feasible, assess the feasibility of collision strategies that can control the vehicle's trajectory after a collision. If not feasible, assess the feasibility of a collision strategy that targets the part of the vehicle structure with the highest strength. The first collision attitude strategy that is assessed as feasible is determined as the target collision attitude.
3. The method according to claim 2, characterized in that, The evaluation of the feasibility of a collision attitude strategy includes: Real-time simulation tests are performed based on the vehicle dynamics model and the collision situation data to calculate the vehicle control quantity required to execute the corresponding collision attitude strategy. Determine whether the vehicle control quantity exceeds the physical limits of the vehicle's actuator, and / or determine whether the remaining time before the collision is sufficient to complete the execution of the vehicle control quantity; If the physical limits of the vehicle's actuators are not exceeded and the execution of the vehicle control quantities is sufficient, the assessment is feasible; otherwise, the assessment is infeasible.
4. The method according to any one of claims 1-3, characterized in that, Before controlling the acquisition of current collision situation data in response to determining that a collision between the vehicle and the target object is unavoidable, the method further includes: Control and monitor the movement trajectory of objects around the vehicle; Identify surrounding objects whose moving trajectories pose a collision risk with the vehicle's predicted trajectory, and send the identification results to the cloud server for risk level classification to obtain the classification results; High-risk objects in the classification results are identified as target objects that need to be continuously tracked.
5. The method according to any one of claims 1-3, characterized in that, The method further includes: After a collision occurs, the system collects process data and result data related to the collision. The process data and result data are uploaded to the cloud server to iteratively train and optimize the cloud-based global artificial intelligence decision-making model.
6. The method according to any one of claims 1-3, characterized in that, The collision situation data includes at least one of the following: target object type data, vehicle trajectory, target object trajectory, vehicle occupant status, vehicle motion status, and surrounding environment data.
7. The method according to claim 2 or 3, characterized in that, The collision strategy described herein, which enables control over the vehicle's trajectory after a collision, aims to prevent secondary collisions or vehicles from entering dangerous areas through proactive intervention and control.
8. A vehicle attitude adjustment device before a collision, characterized in that, The device includes: The first control module is used to control the acquisition of current collision situation data in response to the determination that a collision between the vehicle and the target is inevitable. The determination module is used to determine the collision attitude adjustment strategy based on the collision situation data and the preset artificial intelligence decision model. The preset artificial intelligence decision model has a built-in multi-level degradation strategy set, which includes multiple collision attitude strategies sorted by protection effectiveness priority. The second control module is used to control the vehicle to make corresponding attitude adjustments based on the collision attitude adjustment strategy output by the preset artificial intelligence decision model. The preset artificial intelligence decision-making model is configured to: evaluate the feasibility of each collision posture strategy in order of priority, determine the target collision posture based on the feasibility evaluation results, and generate a collision posture adjustment strategy corresponding to the posture.
9. A vehicle attitude adjustment device before a collision, characterized in that, include: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation of the vehicle pre-collision attitude adjustment method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The storage medium stores at least one executable instruction, which, when executed on the vehicle pre-collision attitude adjustment device / apparatus, causes the vehicle pre-collision attitude adjustment device / apparatus to perform the operation of the vehicle pre-collision attitude adjustment method as described in any one of claims 1-7.