A vehicle hierarchical high-voltage protection method, device and system based on collision detection

Abstract

The application belongs to the technical field of vehicle safety, and specifically provides a vehicle hierarchical high-voltage protection method, device and system based on collision detection, which comprises the following steps: constructing a collision detection model based on vehicle information to confirm the type of a collision scene; constructing a classification decision tree according to the collision scene and vehicle acceleration information to grade the collision degree; and outputting control information for the protection of high-voltage components according to the grade of the collision degree and the message interaction between the vehicle control unit and the vehicle high-voltage battery management system and the distribution box controller. The collision degree classification decision tree is constructed, the collision degree is graded by comparing the size relationship between the three-axis acceleration change rate under different collision scenes and the empirical threshold, and finally the VCU triggers the high-voltage component protection mechanism according to the collision degree and the CAN line communication timeout condition, thereby effectively increasing the safety of the hybrid vehicle in the event of a collision.

Description

Technical Field

[0001] This invention relates to the field of vehicle safety technology, specifically to a method, device, and system for graded high-voltage protection of vehicles based on collision detection. Background Technology

[0002] With the increasing demands for vehicle range, efforts are being made to improve battery storage capacity and enhance vehicle performance by increasing the voltage of the high-voltage system. Consequently, in the event of a collision, higher requirements are needed for the high-voltage safety of the entire vehicle to prevent more severe explosions and combustion accidents, which could result in serious vehicle damage and personal injury. In this context, vehicle collision detection and collision-based protection of the battery and other high-voltage components are crucial for vehicles with large-capacity batteries. Accurate collision detection can both trigger protection promptly in the event of a collision and prevent false alarms that could cause the vehicle to stall, resulting in unnecessary trouble.

[0003] Existing vehicle collision detection methods are generally divided into traditional methods and neural network methods. Traditional collision detection typically detects vehicle acceleration, velocity, and angular acceleration, determining whether a collision has occurred based on the reasonableness of instantaneous changes in acceleration, but this method has low accuracy. There are also examples of installing deformation detection devices at collision-prone locations on the vehicle to determine the severity of the collision based on deformation; these methods require additional sensors, resulting in higher costs and are closely related to the sensor installation location. Existing neural network-based vehicle collision detection methods generally extract collision-related feature parameters, including vehicle operating status information, road condition information, and surrounding vehicle motion information, and then train the model using a large amount of data. To improve model accuracy, existing methods often select features that include road condition information, frequently requiring additional image sensing equipment and vehicle-to-everything (V2X) devices, placing higher demands on hardware and increasing costs.

[0004] In related technologies, the state information of the target vehicle is obtained, and the event type of the target vehicle is determined based on the state information and the trained convolutional neural network. Although the collision event type is divided into collision event, near collision event and baseline event, it is not possible to further distinguish scenarios such as rear-end collision, being rear-ended, and side collision; nor are there any public high-voltage protection measures when a collision occurs. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, the present invention provides a vehicle graded high-voltage protection method, device and system based on collision detection.

[0006] In a first aspect, the present invention provides a vehicle graded high-voltage protection method based on collision detection, comprising the following steps: A collision detection model is built based on vehicle information to confirm the type of collision scenario; A classification decision tree is constructed based on the collision scenario and vehicle acceleration information to classify the degree of collision. Based on the collision severity level and the message exchange between the vehicle controller, the vehicle's high-voltage battery management system, and the distribution box controller, control information is output to protect high-voltage components.

[0007] As a further limitation of the technical solution of the present invention, the step of confirming the type of collision scenario by constructing a collision detection model based on vehicle information includes: The vehicle information is obtained, specifically including: vehicle speed signal from the instrument panel; braking status, steering angle and angular acceleration from the EBS; gear position, current slope and vehicle weight from the TCU; and acceleration information of the vehicle on the x, y and z axes through the acceleration sensor.

[0008] As a further limitation of the technical solution of the present invention, the step of confirming the type of collision scenario by constructing a collision detection model based on vehicle information includes: Collision scenarios are categorized into types including head-on collision with other vehicles, side collision with other vehicles, being hit on the side by other vehicles, rear-ending other vehicles, being rear-ended by other vehicles, scraping with other vehicles, and no collision. Construct a neural network model and use the neural network output layer. The function outputs the predicted probability of various collision scene categories, and confirms the type with the highest probability as the corresponding collision type; Construct a loss function based on the probability of the confirmed collision type; The collision detection model is obtained by iteratively training the neural network model using the loss function as a guide. Obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scene type signal.

[0009] As a further limitation of the technical solution of the present invention, the step of obtaining the vehicle feature vector at time k and inputting it into the trained collision detection model to obtain the collision scene type signal includes: Acquire vehicle information at time K and process the acquired vehicle information to generate the golden emergency time t for high-voltage protection after a collision. 3D feature vector ; The obtained feature vector Input the trained collision detection model to obtain the collision scene type signal.

[0010] As a further limitation of the technical solution of the present invention, the step of constructing a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision includes: A classification decision tree is constructed based on the collision scenario and the vehicle's three-axis acceleration change rate information; The acquired information on the rate of change of vehicle three-axis acceleration and the obtained collision scenario type signal are input into a classification decision tree to obtain a collision severity level signal.

[0011] As a further limitation of the technical solution of the present invention, in the step of constructing a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information, the classification logic includes: In scenarios where no collision occurs or in scenarios involving minor collisions with other vehicles, the output is at the first level. In a frontal collision scenario, the threshold for the rate of change of vehicle x-axis acceleration is set as the first threshold. If the rate of change of vehicle x-axis acceleration is greater than the first threshold, the output is the third level; otherwise, the output is the second level. In scenarios where a vehicle is rear-ended by another vehicle and in scenarios where a vehicle rear-ends another vehicle, the threshold for the rate of change of the vehicle's x-axis acceleration is set to the third threshold and the fourth threshold, respectively. If the rate of change of the vehicle's x-axis acceleration is greater than the corresponding threshold, the output is at the third level; otherwise, the output is at the second level. In the scenario of colliding with the side of other vehicles, the threshold for the rate of change of vehicle x-axis acceleration is set to the fifth threshold. If the rate of change of vehicle x-axis acceleration is greater than the fifth threshold, the output is the third level; otherwise, the output is the second level. In the scenario of being hit from the side by another vehicle, the threshold for the rate of change of the vehicle's y-axis acceleration is set to the sixth threshold. If the rate of change of the vehicle's y-axis acceleration is greater than the sixth threshold, the output is at the third level; otherwise, the output is at the second level. Among the various thresholds, the sixth threshold is the smallest.

[0012] As a further limitation of the technical solution of the present invention, the step of protecting high-voltage components by outputting control information based on the degree of collision and the message interaction between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller includes: Real-time monitoring of message exchanges between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller; After a vehicle collision, monitor whether a message timeout occurs; If so, the emergency high-voltage protection process after a collision is initiated; specifically, the connection between each high-voltage component and the high-voltage battery is forcibly cut off by disconnecting the relay, and the key wake-up signal of the vehicle's high-voltage battery management system and the distribution box controller is cut off. If not, the normal high-voltage protection process under different collision levels will be initiated based on the collision severity level signal.

[0013] As a further limitation of the technical solution of the present invention, the steps of activating the normal high-voltage protection process under different collision levels according to the collision severity level signal include: After a vehicle collision, if the collision severity level signal obtained from the classification decision tree is level two, and the vehicle controller communicates normally with the vehicle high-voltage battery management system and the power distribution box controller, the vehicle's power output is limited by limiting the vehicle's output power to a set power. After a vehicle collision, if the collision severity level signal obtained from the classification decision tree is level three, and the vehicle controller communicates normally with the vehicle high-voltage battery management system and the distribution box controller, an instruction is sent to cut off the enable signals of each high-voltage component, causing the high-voltage component to stop working. Issue relay commands to disconnect each high-voltage component, severing the connection between each high-voltage component and the high-voltage battery; The control system directly switches to the discharge circuit after the high-voltage component disconnects the relay, actively releasing the voltage of the high-voltage component to a safe voltage range for the human body. Controls disconnecting the low-voltage power from the vehicle's high-voltage battery management system and the distribution box controller.

[0014] Secondly, the technical solution of the present invention provides a vehicle graded high-voltage protection device based on collision detection, including a collision scenario confirmation module, a collision level classification module, and a high-voltage component protection control module. The collision scenario confirmation module is used to confirm the type of collision scenario by building a collision detection model based on vehicle information; The collision level classification module is used to construct a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision. The high-voltage component protection control module is used to output control information to protect high-voltage components based on the collision severity level and the message exchange between the vehicle controller, the vehicle's high-voltage battery management system, and the distribution box controller.

[0015] As a further limitation of the technical solution of the present invention, the device includes a data acquisition module for acquiring vehicle information, specifically including: acquiring vehicle speed signal from the instrument panel; acquiring braking status, steering angle and angular acceleration from the EBS; acquiring gear position, current slope and vehicle weight from the TCU; and acquiring vehicle acceleration information on the x, y and z axes through an acceleration sensor.

[0016] As a further limitation of the technical solution of the present invention, the collision scene confirmation module includes a scene division unit, a model construction unit, a model training unit, and a scene type confirmation unit; Scene segmentation unit is used to classify collision scenarios; the types include frontal collision with other vehicles, side collision with other vehicles, being hit on the side by other vehicles, rear-ending other vehicles, being rear-ended by other vehicles, scraping with other vehicles, and no collision. Model building units are used to build neural network models, employing the output layer of the neural network. The function outputs the predicted probability of various collision scene categories, and confirms the type with the highest probability as the corresponding collision type; The model training unit is used to construct a loss function based on the probability of the confirmed collision type, and to iteratively train the neural network model using the loss function as a guide to obtain the collision detection model. The scene type confirmation unit is used to obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scene type signal.

[0017] As a further limitation of the technical solution of the present invention, the scene type confirmation unit is specifically used to acquire vehicle information at time K and process the acquired vehicle information to generate the golden emergency time t for high-voltage protection after a collision. 3D feature vector The obtained feature vector Input the trained collision detection model to obtain the collision scene type signal.

[0018] As a further limitation of the technical solution of the present invention, the collision level classification module includes a decision tree construction unit and a level acquisition unit; The decision tree construction unit is used to construct a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information. The level acquisition unit is used to input the acquired vehicle three-axis acceleration change rate information and the obtained collision scene type signal into the classification decision tree to obtain the collision severity level signal.

[0019] As a further limitation of the technical solution of the present invention, the classification logic for constructing the decision tree includes: In scenarios where no collision occurs or in scenarios involving minor collisions with other vehicles, the output is at the first level. In a frontal collision scenario, the threshold for the rate of change of vehicle x-axis acceleration is set as the first threshold. If the rate of change of vehicle x-axis acceleration is greater than the first threshold, the output is the third level; otherwise, the output is the second level. In scenarios where a vehicle is rear-ended by another vehicle and in scenarios where a vehicle rear-ends another vehicle, the threshold for the rate of change of the vehicle's x-axis acceleration is set to the third threshold and the fourth threshold, respectively. If the rate of change of the vehicle's x-axis acceleration is greater than the corresponding threshold, the output is at the third level; otherwise, the output is at the second level. In the scenario of colliding with the side of other vehicles, the threshold for the rate of change of vehicle x-axis acceleration is set to the fifth threshold. If the rate of change of vehicle x-axis acceleration is greater than the fifth threshold, the output is the third level; otherwise, the output is the second level. In the scenario of being hit from the side by another vehicle, the threshold for the rate of change of the vehicle's y-axis acceleration is set to the sixth threshold. If the rate of change of the vehicle's y-axis acceleration is greater than the sixth threshold, the output is at the third level; otherwise, the output is at the second level. Among the various thresholds, the sixth threshold is the smallest.

[0020] As a further limitation of the technical solution of the present invention, the high-voltage component protection control module includes a monitoring unit and a protection control unit; The monitoring unit is used to monitor the message exchange between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller in real time; after a vehicle collision, it monitors whether a message timeout occurs. The protection control unit is used to initiate the emergency high-voltage protection process after a collision; specifically, it forcibly disconnects the connection between each high-voltage component and the high-voltage battery by disconnecting the relay, and cuts off the key wake-up signal of the vehicle's high-voltage battery management system and the distribution box controller; it is also used to initiate the normal high-voltage protection process under different collision levels according to the collision severity level signal.

[0021] As a further limitation of the technical solution of the present invention, the protection control unit is specifically used to limit the vehicle's power output by limiting the vehicle's output power to a set power when the collision severity level signal obtained according to the classification decision tree is at the second level and the vehicle controller, the vehicle high-voltage battery management system, and the distribution box controller are communicating normally after a vehicle collision; when the collision severity level signal obtained according to the classification decision tree is at the third level and the vehicle controller, the vehicle high-voltage battery management system, and the distribution box controller are communicating normally after a vehicle collision, the control unit sends an instruction to cut off the enable signal of each high-voltage component, causing the high-voltage component to stop working; issues a relay instruction to disconnect each high-voltage component, cutting off the connection between each high-voltage component and the high-voltage battery; controls the direct switching into the discharge circuit after the high-voltage component disconnects the relay, actively releasing the voltage of the high-voltage component to within the range of human safety voltage; and controls the disconnection of the low-voltage power of the vehicle high-voltage battery management system and the distribution box controller.

[0022] Thirdly, the technical solution of the present invention also provides a vehicle graded high-voltage protection system based on collision detection, including a vehicle controller and a high-voltage battery management system, a power distribution box controller, an instrument, an electronic braking system, a transmission controller and a motor controller connected to the vehicle controller via a CAN line; The vehicle controller performs the method as described in the first aspect.

[0023] As can be seen from the above technical solutions, the present invention has the following advantages: it classifies vehicle collision scenarios, establishes a vehicle collision scenario classification system, then constructs a neural network model, and trains it using data from each collision scenario. It detects collisions using vehicle operating status data within t seconds, then constructs a collision severity classification decision tree, and classifies the collision severity by comparing the relationship between the rate of change of triaxial acceleration and the empirical threshold under different collision scenarios. Finally, the VCU triggers the high-voltage component protection mechanism in stages according to the collision severity and CAN line communication timeout, effectively increasing the safety of hybrid vehicles when they collide.

[0024] Furthermore, the design principle of this invention is reliable, the structure is simple, and it has a very wide range of application prospects.

[0025] Therefore, it is evident that the present invention has outstanding substantive features and significant progress compared with the prior art, and the beneficial effects of its implementation are also obvious. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a schematic flowchart of a method according to an embodiment of the present invention.

[0028] Figure 2 A schematic diagram of a vehicle collision scenario classification system.

[0029] Figure 3 This is a schematic diagram of a decision tree for classifying collision severity levels.

[0030] Figure 4 This is a schematic diagram of a high-voltage component protection method based on vehicle collision detection.

[0031] Figure 5 This is a schematic block diagram of an apparatus according to an embodiment of the present invention.

[0032] Figure 6 This is a schematic diagram of a graded high-voltage protection system for hybrid vehicles based on collision detection. Detailed Implementation

[0033] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0034] like Figure 1 As shown, this embodiment of the invention provides a vehicle graded high-voltage protection method based on collision detection, including the following steps: Step 1: Confirm the type of collision scenario by building a collision detection model based on vehicle information; Step 2: Construct a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision. Step 3: Based on the collision severity level and the message exchange between the vehicle controller, the vehicle high-voltage battery management system, and the distribution box controller, output control information to protect the high-voltage components.

[0035] In this embodiment of the invention, it should be noted that the method described in this invention is a system-based method, the system including a vehicle controller (VCU), a CAN bus, a high-voltage battery management system (BMS), a power distribution box controller (PDU), an instrument cluster, an electronic braking system (EBS), a transmission controller (TCU), a motor controller (MCU), and hardwired components.

[0036] The VCU is connected to the instrument cluster, electronic braking system (EBS), transmission control unit (TCU), battery management system (BMS), power distribution unit (PDU), and motor control unit (MCU) via a CAN bus. The VCU obtains the vehicle speed signal from the instrument cluster via the CAN bus. Obtain braking status from EBS Steering angle and angular acceleration ; Get gear from TCU Current slope and vehicle weight The VCU interacts with the BMS, PDU, and MCU via the CAN bus to power on / off and enable high-voltage components. The VCU is also connected to the accelerometer via a hardwired connection to acquire vehicle acceleration information along the x, y, and z axes. , , The VCU controls the key-activated wake-up signal of the BMS and PDU by driving relays to turn them on and off.

[0037] In some embodiments, the step of constructing a collision detection model based on vehicle information to confirm the type of collision scenario includes: Step 11: Classify the collision scenarios; such as Figure 2 As shown, the types include head-on collision with other vehicles, side collision with other vehicles, being hit on the side by other vehicles, rear-ending other vehicles, being rear-ended by other vehicles, scraping with other vehicles, and no collision. Step 12: Construct a neural network model, using the neural network output layer. The function outputs the predicted probability of various collision scene categories, and confirms the type with the highest probability as the corresponding collision type; Assuming the golden emergency response time for high-voltage protection after a vehicle collision is t, then for each time k, construct... 3D eigenvectors: ], Among them, braking state Current gear Vehicle weight It is discrete data, other variables If one-hot encoding is used to quantize the above 7 types of collision scenarios, the output category can be vectorized as follows: Among them, the first In the vector corresponding to the class scenario, All other elements are 0; the input layer of the neural network model has a total of There are 7 nodes in the output layer; Step 13: Construct a loss function based on the probability of the confirmed collision type; guided by the loss function, iteratively train the neural network model to obtain the collision detection model; Loss function selection: cross loss, entropy (cross entropy) : in, For the number of categories, For the sign function, the sample The true category is If the value is 1, then 1 is used; otherwise, 0 is used. Time observation samples Category The predicted probability, which is generated by the output layer of the neural network. The result of the function calculation is obtained.

[0038] Collect real vehicle data and process it into feature vectors In this manner, after data collection and processing, the data is divided into training, validation, and test sets, and then input into the model for iterative training. Guided by the cross-loss entropy loss function, a collision detection model with a certain prediction accuracy is finally obtained.

[0039] Step 14: Obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scenario type signal. Specifically, obtain the vehicle information at time k and process the obtained vehicle information to generate the golden emergency time t for high-voltage protection after a collision. 3D feature vector The obtained feature vector Input the trained collision detection model to obtain the collision scene type signal.

[0040] In some embodiments, the step of constructing a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision includes: Step 21: Construct a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information; Step 22: Input the acquired vehicle three-axis acceleration change rate information and the obtained collision scene type signal into the classification decision tree to obtain the collision severity level signal.

[0041] It should be noted that in the step of constructing a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information, such as... Figure 3 As shown, the classification logic includes: A classification decision tree is constructed based on the collision scenario and the vehicle's three-axis acceleration change rate information. The classification logic is as follows: In the "no collision" scenario, the collision level is directly output as "no collision or minor collision"; in the "scratching with other vehicles" scenario, the output is directly "no collision or minor collision"; in the "frontal collision" scenario, the threshold for the vehicle's x-axis acceleration change rate is set to... If greater than If the condition is met, the output is "Severe Collision"; otherwise, it is "Normal Collision". In the scenarios of "Being rear-ended by another vehicle" and "Rear-ending another vehicle", the threshold for the vehicle's x-axis acceleration rate of change is set to [value missing]. and If the value exceeds the threshold, the output is "Severe Collision"; otherwise, the output is "Normal Collision". In the "Collision with the Side of Other Vehicles" scenario, a threshold for the x-axis acceleration rate of change is set. If the value exceeds the threshold, the output is "Severe Collision"; otherwise, it is "Normal Collision". In the "side impact by another vehicle" scenario, a threshold for the y-axis acceleration change rate is set. If the acceleration rate exceeds the threshold, the output is "Severe Collision"; otherwise, it is "Normal Collision". The acceleration rate threshold for each scenario should meet the following requirements: This means that the system should be more sensitive when hit from the side by other vehicles, because the fuel tank is usually located on the left and right sides of the vehicle, making it more dangerous in a collision.

[0042] The vehicle's three-axis acceleration change rate information and the collision scenario type signal are input into the decision tree mentioned above to obtain the collision severity level signal. .

[0043] The VCU monitors the message interactions with the BMS and PDU in real time and determines the level of collision based on the signal. The graded high-voltage protection procedure is executed. If, after a vehicle collision, message exchange between the VCU, BMS, and PDU continues without message timeout or high-voltage component failure, then the collision severity level signal is applied. The normal high-voltage protection process is initiated under different collision conditions, namely, the normal high-voltage protection process after no collision or minor collision, the normal high-voltage protection process after a general collision, and the normal high-voltage protection process after a severe collision. If a message exchange timeout occurs, it is determined that the collision caused a timeout fault in the CAN line, and the emergency high-voltage protection process after the collision is initiated directly.

[0044] In some embodiments, such as Figure 4 As shown, the steps for protecting high-voltage components by outputting control information based on the degree of collision and the message exchange between the vehicle controller, the vehicle high-voltage battery management system, and the distribution box controller include: Step 31: Monitor the message exchange between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller in real time; Step 32: After a vehicle collision, monitor whether a message timeout occurs; If yes, proceed to step 33; if no, proceed to step 34. Step 33: Initiate the emergency high-voltage protection process after a collision; Emergency high-voltage protection process after a collision: If a timeout fault occurs in the CAN line after a collision, the vehicle controller VCU will directly force a power-off by disconnecting the relay, cutting off the key wake-up signal of the BMS and PDU. The BMS will cut off the high-voltage output to external high-voltage components before cutting off the low-voltage power, avoiding the safety hazard of high-voltage leakage caused by the collision.

[0045] Step 34: Initiate the normal high-voltage protection process for different collision levels based on the collision severity level signal. This specifically includes: Normal high-voltage protection process after no-collision or minor collision: If a collision occurs, the VCU will use the collision severity level signal obtained from the decision tree. If there is no collision or a minor collision, the CAN bus does not experience a timeout fault, and the message exchange between the VCU, BMS, and PDU is normal, then the VCU will not perform any power-down process to ensure the vehicle's normal power output afterwards.

[0046] Normal high-voltage protection process after a collision: If a vehicle is involved in a collision, the VCU will determine the collision severity level signal based on the decision tree. In a typical collision, if the CAN bus does not experience a timeout fault and the message exchange between the VCU, BMS, and PDU is normal, the VCU will limit the vehicle's output power to a first output power threshold. Limiting the vehicle's power output to a range of approximately 60% of the vehicle's maximum output power can alleviate the working pressure on high-voltage components, prevent high-voltage components from working under heavy load due to high power output after a collision, and further protect the high-voltage system.

[0047] Normal high-voltage protection process after a severe collision: If a vehicle is involved in a collision, the VCU will use the collision severity level signal obtained from the decision tree. If a severe collision occurs, but the CAN line does not experience a timeout fault and the message exchange between the VCU, BMS, and PDU is normal, then the following high-voltage protection procedure will be performed: First, the VCU sends a command to cut off the enable signals of each high-voltage component, gradually disconnecting the enable signals of the high-voltage components and causing them to stop working. Secondly, the VCU issues relay commands to disconnect each high-voltage component, severing the connection between each high-voltage component and the high-voltage battery; Then, because the high-voltage components have large capacitance, high voltage still exists in the large capacitance of the high-voltage components after the high-voltage relay is disconnected, requiring active discharge. The VCU interacts with the controller of the high-voltage components (such as the motor controller MCU) via the CAN line, and sends the active discharge command to each controller. The corresponding high-voltage component controller responds and completes the relevant discharge action, that is, after the high-voltage component relay is disconnected, it directly switches to a dedicated discharge circuit to actively release the voltage of the high-voltage components to a safe voltage range for the human body; thus avoiding the high voltage safety hazard that still exists in the high-voltage components after a vehicle collision. Finally, the low-voltage power to the BMS and PDU is disconnected, completing the multiple high-voltage power-off protection process after receiving the collision signal.

[0048] like Figure 5 As shown, this embodiment of the invention provides a vehicle graded high-voltage protection device based on collision detection, including a collision scenario confirmation module, a collision level classification module, and a high-voltage component protection control module. The collision scenario confirmation module is used to confirm the type of collision scenario by building a collision detection model based on vehicle information; The collision level classification module is used to construct a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision. The high-voltage component protection control module is used to output control information to protect high-voltage components based on the collision severity level and the message exchange between the vehicle controller, the vehicle's high-voltage battery management system, and the distribution box controller.

[0049] It should be noted that the device includes a data acquisition module for acquiring vehicle information, specifically including: acquiring vehicle speed signals from the instrument panel; acquiring braking status, steering angle, and angular acceleration from the EBS; acquiring gear position, current gradient, and vehicle weight from the TCU; and acquiring vehicle acceleration information along the x, y, and z axes through an acceleration sensor.

[0050] In some embodiments, the collision scene confirmation module includes a scene segmentation unit, a model building unit, a model training unit, and a scene type confirmation unit; Scene segmentation unit is used to classify collision scenarios; the types include frontal collision with other vehicles, side collision with other vehicles, being hit on the side by other vehicles, rear-ending other vehicles, being rear-ended by other vehicles, scraping with other vehicles, and no collision. Model building units are used to build neural network models, employing the output layer of the neural network. The function outputs the predicted probability of various collision scene categories, and confirms the type with the highest probability as the corresponding collision type; The model training unit is used to construct a loss function based on the probability of the confirmed collision type, and to iteratively train the neural network model using the loss function as a guide to obtain the collision detection model. The scene type confirmation unit is used to obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scene type signal.

[0051] In some embodiments, the scenario type confirmation unit is specifically used to acquire vehicle information at time K and process the acquired vehicle information to generate the golden emergency time t for high-voltage protection after a collision relative to time k. 3D feature vector The obtained feature vector Input the trained collision detection model to obtain the collision scene type signal.

[0052] In some embodiments, the collision level classification module includes a decision tree construction unit and a level acquisition unit; The decision tree construction unit is used to construct a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information. The level acquisition unit is used to input the acquired vehicle three-axis acceleration change rate information and the obtained collision scene type signal into the classification decision tree to obtain the collision severity level signal.

[0053] In some embodiments, the classification logic for constructing a decision tree includes: the classification logic includes: A classification decision tree is constructed based on the collision scenario and the vehicle's three-axis acceleration change rate information. The classification logic is as follows: In the "no collision" scenario, the collision level is directly output as "no collision or minor collision"; in the "scratching with other vehicles" scenario, the output is directly "no collision or minor collision"; in the "frontal collision" scenario, the threshold for the vehicle's x-axis acceleration change rate is set to... If greater than If the condition is met, the output is "Severe Collision"; otherwise, it is "Normal Collision". In the scenarios of "Being rear-ended by another vehicle" and "Rear-ending another vehicle", the threshold for the vehicle's x-axis acceleration rate of change is set to [value missing]. and If the value exceeds the threshold, the output is "Severe Collision"; otherwise, the output is "Normal Collision". In the "Collision with the Side of Other Vehicles" scenario, a threshold for the x-axis acceleration rate of change is set. If the value exceeds the threshold, the output is "Severe Collision"; otherwise, it is "Normal Collision". In the "side impact by another vehicle" scenario, a threshold for the y-axis acceleration change rate is set. If the acceleration rate exceeds the threshold, the output is "Severe Collision"; otherwise, it is "Normal Collision". The acceleration rate threshold for each scenario should meet the following requirements: This means that the system should be more sensitive when hit from the side by other vehicles, because the fuel tank is usually located on the left and right sides of the vehicle, making it more dangerous in a collision.

[0054] In some embodiments, the high-voltage component protection control module includes a monitoring unit and a protection control unit; The monitoring unit is used to monitor the message exchange between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller in real time; after a vehicle collision, it monitors whether a message timeout occurs. The protection control unit is used to initiate the emergency high-voltage protection process after a collision; specifically, it forcibly disconnects the connection between each high-voltage component and the high-voltage battery by disconnecting the relay, and cuts off the key wake-up signal of the vehicle's high-voltage battery management system and the distribution box controller; it is also used to initiate the normal high-voltage protection process under different collision levels according to the collision severity level signal.

[0055] In some embodiments, the protection control unit is specifically used to limit the vehicle's power output by limiting the vehicle's output power to a set power level when the collision severity level signal obtained according to the classification decision tree is at the second level and the vehicle controller, vehicle high-voltage battery management system, and distribution box controller are communicating normally after a vehicle collision; when the collision severity level signal obtained according to the classification decision tree is at the third level and the vehicle controller, vehicle high-voltage battery management system, and distribution box controller are communicating normally after a vehicle collision, it sends an instruction to cut off the enable signal of each high-voltage component, causing the high-voltage component to stop working; it issues a relay instruction to disconnect each high-voltage component, cutting off the connection between each high-voltage component and the high-voltage battery; it controls the direct switching into the discharge circuit after the high-voltage component disconnects the relay, actively releasing the voltage of the high-voltage component to within the range of human safety voltage; and it controls the disconnection of the low-voltage power of the vehicle high-voltage battery management system and distribution box controller.

[0056] like Figure 6 As shown, this embodiment of the invention also provides a vehicle graded high-voltage protection system based on collision detection, including a vehicle controller and a high-voltage battery management system, a power distribution box controller, an instrument panel, an electronic braking system, a transmission controller, and a motor controller connected to the vehicle controller via a CAN line; The vehicle controller performs the method described in the above embodiments.

[0057] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the invention should also be covered within the protection scope of the invention. Therefore, the protection scope of the invention should be determined by the scope of the claims.

Claims

1. A vehicle graded high-voltage protection method based on collision detection, characterized in that, Includes the following steps: A collision detection model is built based on vehicle information to confirm the type of collision scenario; A classification decision tree is constructed based on the collision scenario and vehicle acceleration information to classify the collision severity level. The steps include: constructing a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information; inputting the obtained vehicle's three-axis acceleration change rate information and the obtained collision scenario type signal into the classification decision tree to obtain the collision severity level signal. The classification logic in the step of constructing a classification decision tree based on the collision scenario and the vehicle's three-axis acceleration change rate information includes: In scenarios where no collision occurs or in scenarios involving minor collisions with other vehicles, the output is at the first level. In a frontal collision scenario, the threshold for the rate of change of vehicle x-axis acceleration is set as the first threshold. If the rate of change of vehicle x-axis acceleration is greater than the first threshold, the output is the third level; otherwise, the output is the second level. In scenarios where a vehicle is rear-ended by another vehicle and in scenarios where a vehicle rear-ends another vehicle, the threshold for the rate of change of the vehicle's x-axis acceleration is set to the third threshold and the fourth threshold, respectively. If the rate of change of the vehicle's x-axis acceleration is greater than the corresponding threshold, the output is at the third level; otherwise, the output is at the second level. In the scenario of colliding with the side of other vehicles, the threshold for the rate of change of vehicle x-axis acceleration is set to the fifth threshold. If the rate of change of vehicle x-axis acceleration is greater than the fifth threshold, the output is the third level; otherwise, the output is the second level. In the scenario of being hit from the side by another vehicle, the threshold for the rate of change of the vehicle's y-axis acceleration is set to the sixth threshold. If the rate of change of the vehicle's y-axis acceleration is greater than the sixth threshold, the output is the third level; otherwise, the output is the second level. Among the various thresholds, the sixth threshold is the smallest. Based on the collision severity level and the message exchange between the vehicle controller and the vehicle's high-voltage battery management system and distribution box controller, control information is output to protect high-voltage components. The steps include: Real-time monitoring of message exchanges between the vehicle controller and the vehicle high-voltage battery management system and the distribution box controller; After a vehicle collision, monitor whether a message timeout occurs; If so, the emergency high-voltage protection process after a collision is initiated; specifically, the connection between each high-voltage component and the high-voltage battery is forcibly cut off by disconnecting the relay, and the key wake-up signal of the vehicle's high-voltage battery management system and the distribution box controller is cut off. If not, the normal high-voltage protection process under different collision levels will be initiated based on the collision severity level signal.

2. The vehicle graded high-voltage protection method based on collision detection according to claim 1, characterized in that, Before confirming the type of collision scenario by building a collision detection model based on vehicle information, the following steps are included: The vehicle information is obtained, specifically including: vehicle speed signal from the instrument panel; braking status, steering angle and angular acceleration from the EBS; gear position, current slope and vehicle weight from the TCU; and acceleration information of the vehicle on the x, y and z axes through the acceleration sensor.

3. The vehicle graded high-voltage protection method based on collision detection according to claim 2, characterized in that, The steps for identifying the type of collision scenario by building a collision detection model based on vehicle information include: Collision scenarios are categorized into types including head-on collision with other vehicles, side collision with other vehicles, being hit on the side by other vehicles, rear-ending other vehicles, being rear-ended by other vehicles, scraping with other vehicles, and no collision. Construct a neural network model and use the neural network output layer. The function outputs the predicted probability of various collision scene categories, and confirms the type with the highest probability as the corresponding collision type; Construct a loss function based on the probability of the confirmed collision type; The collision detection model is obtained by iteratively training the neural network model using the loss function as a guide. Obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scene type signal.

4. The vehicle graded high-voltage protection method based on collision detection according to claim 3, characterized in that, The steps to obtain the vehicle feature vector at time k and input it into the trained collision detection model to obtain the collision scene type signal include: Acquire vehicle information at time K and process the acquired vehicle information to generate the golden emergency time t for high-voltage protection after a collision. 3D feature vector ; The obtained feature vector Input the trained collision detection model to obtain the collision scene type signal.

5. The vehicle graded high-voltage protection method based on collision detection according to claim 1, characterized in that, The steps for initiating the normal high-voltage protection process under different collision levels based on the collision severity level signal include: After a vehicle collision, if the collision severity level signal obtained from the classification decision tree is level two, and the vehicle controller communicates normally with the vehicle high-voltage battery management system and the power distribution box controller, the vehicle's power output is limited by limiting the vehicle's output power to a set power. After a vehicle collision, if the collision severity level signal obtained from the classification decision tree is level three, and the vehicle controller communicates normally with the vehicle high-voltage battery management system and the distribution box controller, an instruction is sent to cut off the enable signals of each high-voltage component, causing the high-voltage component to stop working. Issue relay commands to disconnect each high-voltage component, severing the connection between each high-voltage component and the high-voltage battery; The control system directly switches to the discharge circuit after the high-voltage component disconnects the relay, actively releasing the voltage of the high-voltage component to a safe voltage range for the human body. Controls disconnecting the low-voltage power from the vehicle's high-voltage battery management system and the distribution box controller.

6. A collision detection-based vehicle graded high-voltage protection device, applicable to the collision detection-based vehicle graded high-voltage protection method according to any one of claims 1-5, characterized in that, This includes a collision scenario confirmation module, a collision level classification module, and a high-voltage component protection and control module; The collision scenario confirmation module is used to confirm the type of collision scenario by building a collision detection model based on vehicle information. The collision level classification module is used to construct a classification decision tree based on the collision scenario and vehicle acceleration information to classify the degree of collision. The high-voltage component protection control module is used to output control information to protect high-voltage components based on the collision severity level and the message exchange between the vehicle controller, the vehicle's high-voltage battery management system, and the distribution box controller.

7. A vehicle graded high-voltage protection system based on collision detection, characterized in that, This includes the vehicle controller and the high-voltage battery management system, distribution box controller, instrument panel, electronic braking system, transmission controller, and motor controller connected to the vehicle controller via a CAN bus. The vehicle controller performs the method as described in any one of claims 1-5.

Images