Pedestrian protection methods based on vehicle safety systems and vehicle safety systems

Abstract

This application proposes a pedestrian protection method based on a vehicle safety system. The vehicle safety system includes a perception module, a decision module, and an execution module. The decision module is connected to both the perception module and the execution module. The execution module includes a pneumatic lifting device and an active hinge. The method includes: acquiring surrounding environmental data of the vehicle through the perception module; generating collision prediction data of the vehicle based on the surrounding environmental data through the decision module, and generating a protection execution command in response to the collision prediction data meeting preset collision conditions; controlling the lifting device of the pneumatic lifting device according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the vehicle's engine hood to a preset height. Therefore, by actively perceiving and predicting pedestrian collision risks, the recognition accuracy is significantly improved and the false trigger rate is reduced. Furthermore, the use of a pneumatically resettable lifting device eliminates the risk of obstructed vision and reduces maintenance costs.

Description

Technical Field

[0001] This application relates to the field of vehicle safety technology, and in particular to a pedestrian protection method and vehicle safety system based on a vehicle safety system. Background Technology

[0002] Currently, to mitigate injuries caused by collisions between vehicles and pedestrians or cyclists, active pedestrian protection systems have emerged and are becoming increasingly widespread, such as active hood systems. Existing systems of this type typically rely on collision sensors installed in the front bumper of the vehicle, which determine whether to trigger the protective device by detecting the energy intensity generated at the moment of impact. This passive collision signal-based identification method dictates that the system must make a response decision in an extremely short time. To meet this instantaneous response requirement, existing solutions often employ a pyrotechnic lifter driven by gunpowder as the actuator to lift the hood at the moment of impact, creating a buffer space.

[0003] However, the aforementioned existing technical solutions have significant shortcomings in practical applications. First, because they rely on energy recognition after a collision, the system's response inherently has a lag and it is difficult to accurately distinguish between collisions with humans or other objects (such as animals), easily leading to a high probability of false triggering. Second, once the pyrotechnic hood lifter is triggered, it is irreversible. After the system is activated, not only will the lifted engine hood obstruct the driver's view, creating a driving safety hazard, but it also usually requires the replacement of the entire active engine hood assembly, resulting in high maintenance costs and seriously affecting the user experience. Summary of the Invention

[0004] This application aims to at least partially solve one of the technical problems in the aforementioned technologies.

[0005] Therefore, one objective of this application is to propose a pedestrian protection method based on a vehicle safety system, which significantly improves recognition accuracy and reduces false triggering rate by actively sensing and predicting pedestrian collision risks. At the same time, it adopts a pneumatically resettable lifting device to eliminate the hidden danger of obstructed vision and reduce maintenance costs.

[0006] Another objective of this application is to propose a vehicle safety system.

[0007] To achieve the above objectives, a first aspect of this application proposes a pedestrian protection method based on a vehicle safety system. The vehicle safety system includes a sensing module, a decision module, and an execution module. The decision module is connected to both the sensing module and the execution module. The execution module includes a pneumatic lifting device and an active hinge. The method includes: acquiring surrounding environmental data of the vehicle through the sensing module; generating collision prediction data of the vehicle based on the surrounding environmental data through the decision module, and generating a protection execution command in response to the collision prediction data meeting preset collision conditions; and controlling the lifting device of the pneumatic lifting device according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the engine hood of the vehicle to a preset height.

[0008] In addition, the pedestrian protection method based on a vehicle safety system proposed in the above embodiments of this application may also have the following additional technical features:

[0009] In one embodiment of this application, the sensing module includes a lidar, a millimeter-wave radar, and a camera. The lidar and the millimeter-wave radar are respectively mounted on the front bumper of the vehicle, and the camera is mounted on the upper part of the vehicle's windshield. The surrounding environment data includes surrounding image data, first radar surrounding data, and second radar surrounding data. The step of acquiring the vehicle's surrounding environment data through the sensing module includes: acquiring the vehicle's driving speed; and in response to the driving speed being within a preset speed range, acquiring the surrounding image data through the camera, acquiring the first radar surrounding data through the lidar, and acquiring the second radar surrounding data through the millimeter-wave radar.

[0010] In one embodiment of this application, the step of generating collision prediction data for the vehicle based on the surrounding environment data by the decision module includes: determining the target object of the vehicle based on the surrounding image data, the first radar surrounding data, and the second radar surrounding data; extracting features from the surrounding image data based on the target object to obtain a first behavioral feature of the target object; extracting features from the first radar surrounding data based on the target object to obtain a second behavioral feature of the target object; and extracting features from the second radar surrounding data based on the target object to obtain a third behavioral feature of the target object. The collision prediction data for the vehicle is then generated based on the first behavioral feature, the second behavioral feature, and the third behavioral feature.

[0011] In one embodiment of this application, determining the target object of the vehicle based on the surrounding image data, the first radar surrounding data, and the second radar surrounding data includes: acquiring the vehicle's driving data; extracting the object's state data from the surrounding image data, wherein the object includes one or more of pedestrians, two-wheeled vehicle riders, and three-wheeled vehicle riders, and the state data includes the object's posture, the object's facial orientation, the object's forward speed, and direction; extracting the object's behavior data from the first radar surrounding data and / or the second radar surrounding data, wherein the behavior data includes position data, movement direction data, and movement speed data; and determining the vehicle's target object based on the state data, the behavior data, and the driving data.

[0012] In one embodiment of this application, generating collision prediction data for a vehicle based on the first behavioral feature, the second behavioral feature, and the third behavioral feature includes: acquiring the predicted collision time and collision process time of the target object, and acquiring the system decision time and system deployment time of the vehicle; adding the predicted collision time and the collision process time to obtain a first time, and adding the system decision time and the system deployment time to obtain a second time; in response to the first time being greater than the second time, performing feature fusion on the first behavioral feature, the second behavioral feature, and the third behavioral feature to obtain a target behavioral feature; and inputting the target behavioral feature into a collision prediction model to obtain the collision prediction data, wherein the collision prediction data includes a collision probability value.

[0013] In one embodiment of this application, obtaining the predicted collision time of the target object includes: obtaining the distance between the vehicle and the target object, and calculating the predicted collision time based on the distance and the driving speed.

[0014] In one embodiment of this application, the preset collision condition includes the collision probability value being greater than a preset collision probability threshold.

[0015] In one embodiment of this application, the pneumatic lifting device includes: a high-pressure gas cylinder, an air guide pipe, an electronic control valve, and four lifting devices. The high-pressure gas cylinder is connected to the vehicle frame, and both ends of the air guide pipe are connected to the high-pressure gas cylinder and the four lifting devices, respectively. The four lifting devices are respectively disposed between the vehicle frame and the vehicle's engine hood, with two lifting devices located at the front of the vehicle's engine hood and the other two located at the rear of the vehicle's engine hood. The electronic control valve is disposed on the high-pressure gas cylinder and is connected to the vehicle controller. Each lifting device is equipped with a vent valve.

[0016] In one embodiment of this application, the active hinge includes: a first link, a second link, a third link, a fourth link, and a fifth link, wherein the first link, the second link, the third link, the fourth link, and the fifth link are respectively connected by connecting shafts, and the first link and the fifth link are respectively connected to the engine hood and the vehicle frame of the vehicle through connecting members; the first link is provided with a lifting part; the first link is provided with a limiting groove, the second link is provided with a limiting stop bar, and the limiting stop bar is slidably disposed in the limiting groove; the third link and the fourth link are respectively provided with limiting blocks.

[0017] To achieve the above objectives, a second aspect of this application proposes a vehicle safety system, including a sensing module, a decision-making module, and an execution module. The decision-making module is connected to both the sensing module and the execution module. The execution module includes a pneumatic lifting device and an active hinge. The sensing module acquires data about the vehicle's surrounding environment. The decision-making module generates collision prediction data for the vehicle based on the surrounding environment data and generates a protection execution command in response to the collision prediction data meeting preset collision conditions. The execution module controls the lifting mechanism of the pneumatic lifting device according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the vehicle's engine hood to a preset height.

[0018] Compared with the prior art, the technical solution provided in this application has the following beneficial effects: The pedestrian protection method and vehicle safety system based on the vehicle safety system in the embodiments of this application actively acquire surrounding environmental data through the perception module and generate collision prediction data by the decision module, realizing the predictive identification of vulnerable road users such as pedestrians. Compared with the traditional passive solution that relies on collision signal triggering, it significantly improves the accuracy of accident identification and effectively reduces the probability of false triggering, while also gaining more response time for execution. On this basis, the execution module adopts a cooperative structure of pneumatic lifting device and active hinge, and realizes the lifting, pressure holding and natural falling of the engine hood through programmed pneumatic control. This not only avoids the driving safety hazards caused by the engine hood obstructing the driver's vision for a long time, but also abandons the traditional irreversible pyrotechnic drive scheme. After the system is triggered, there is no need to replace the entire active engine hood assembly, thereby greatly reducing maintenance costs and improving user experience. It effectively solves the comprehensive technical problems of high false triggering rate, large safety hazards and high maintenance costs in the prior art.

[0019] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a pedestrian protection method based on a vehicle safety system according to an embodiment of this application; Figure 2 This is a block diagram of the vehicle safety system structure in a pedestrian protection method based on a vehicle safety system according to an embodiment of this application; Figure 3 This is a block diagram of a vehicle safety system and vehicle structure according to an embodiment of this application; Figure 4 This is a schematic diagram showing the installation location of a sensing module in a vehicle according to an embodiment of this application; Figure 5 This is a schematic diagram of a pneumatic lifting device according to an embodiment of this application; Figure 6 This is a schematic diagram of a high-pressure gas cylinder structure according to an embodiment of this application; Figure 7 This is a schematic diagram of the engine hood lifting position structure according to an embodiment of this application; Figure 8 This is a schematic diagram of an active hinge structure according to an embodiment of this application; Figure 9 This is a schematic diagram of an active hinge connection structure to a vehicle according to an embodiment of this application; Figure 10 This is a schematic diagram illustrating the functions and timing of a vehicle safety system according to an embodiment of this application; Figure 11 This is a schematic diagram of HIT simulation analysis according to an embodiment of this application; Figure 12 A schematic diagram of simulation analysis for determining the HIT time according to an embodiment of this application; Figure 13 This is a schematic diagram of the engine hood lifting height according to one embodiment of this application; Figure 14 This is a schematic cross-sectional view of a lifter according to an embodiment of this application; Figure 15 This is a schematic diagram illustrating the operating conditions for verifying the effectiveness of a vehicle safety system according to an embodiment of this application. Figure 16 This is a schematic diagram of the non-inactive damage color zone of a vehicle safety system according to an embodiment of this application; Figure 17 This is a schematic diagram of the damage color zone of a vehicle safety system according to an embodiment of this application; Figure 18This is a schematic diagram of the TTC distribution curves at different driving speeds according to an embodiment of this application; Figure 19 This is a schematic diagram of a typical characteristic curve of the air pressure of a lifter according to an embodiment of this application.

[0021] Reference numerals: 100, Vehicle safety system; 200, Perception module; 300, Decision module; 400, Execution module; 401, Pneumatic lifting device; 4011, High-pressure gas cylinder; 4012, Air duct; 4013, Lifter; 4014, On-board controller; 4015, Air release valve; 402, Active hinge; 4021, First link; 4022, Second link; 4023, Third link; 4024, Fourth link; 4025, Fifth link; 4026, Lifting part; 4027, Limiting slide; 4028, Limiting stop bar; 4029, Limiting block; 201, LiDAR; 202, Millimeter-wave radar; 203, Camera; 10, Child dummy; 20, Female dummy; 30, Medium-sized male dummy; 40, Large male dummy. Detailed Implementation

[0022] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0023] The pedestrian protection method and vehicle safety system based on a vehicle safety system according to embodiments of this application are described below with reference to the accompanying drawings. Regarding the related technologies mentioned in the background section, due to reliance on energy recognition after a collision, the system response inherently has a lag and is difficult to accurately distinguish between collisions with humans or other objects (such as animals), easily leading to a high probability of false triggering. Secondly, the pyrotechnic lifter is irreversible once triggered. After the system operates, it not only obstructs the driver's view due to the raised engine hood, posing a driving safety hazard, but also usually requires replacement of the entire active engine hood assembly, resulting in high maintenance costs and seriously affecting the user experience. This application provides a pedestrian protection method based on a vehicle safety system, which significantly improves recognition accuracy and reduces the false triggering rate by actively sensing and predicting pedestrian collision risks. Simultaneously, it employs a pneumatically resettable lift device, eliminating the risk of obstructed view and reducing maintenance costs. Thus, the above-mentioned technical problems are solved.

[0024] Figure 1 This is a flowchart of a pedestrian protection method based on a vehicle safety system according to an embodiment of this application; Figure 2This is a block diagram of the vehicle safety system structure in a pedestrian protection method based on a vehicle safety system according to an embodiment of this application; Figure 3 This is a block diagram of a vehicle safety system and vehicle structure according to an embodiment of this application.

[0025] like Figures 1-3 As shown in the embodiment of this application, the pedestrian protection method based on a vehicle safety system includes a vehicle safety system 100 comprising a sensing module 200, a decision module 300, and an execution module 400. The decision module 300 is connected to the sensing module 200 and the execution module 400, respectively. The execution module 400 includes a pneumatic lifting device 401 and an active hinge 402.

[0026] It should be noted that the pneumatic lifting device 401 in the perception module 200, decision module 300 and execution module 400 described in this embodiment is connected to the vehicle controller 4014, which may include a central processing unit (CPU), a graphics processing unit (GPU) and a vehicle networking module.

[0027] It should be noted that the central processing unit (CPU) and graphics processing unit (GPU) work together to rapidly process massive amounts of data from sensors, using complex algorithm models to predict pedestrian trajectories, and combining this with the vehicle's own driving status (speed, acceleration, steering angle, etc.) to calculate potential collision risks. For example, based on the pedestrian's current speed and walking direction, their route over the next few seconds is estimated and compared with the vehicle's expected trajectory.

[0028] Vehicle-to-everything (V2X) module: Through 5G or more advanced communication technologies, it interacts with surrounding vehicles and intelligent transportation infrastructure (such as smart streetlights and roadside sensors) in real time to obtain more comprehensive road condition and pedestrian information. If the vehicle in front detects a pedestrian in advance and brakes, this vehicle can receive an early warning and react in time to avoid the risk of chain collisions; it can also receive alerts from roadside sensors about pedestrians in blind spots.

[0029] The methods include: Step S101: Obtain the surrounding environment data of the vehicle through the perception module.

[0030] In one embodiment of this application, such as Figure 4As shown, the perception module 200 may include a lidar 201, a millimeter-wave radar 202, and a camera 203. The lidar 201 and millimeter-wave radar 202 are respectively mounted on the front bumper of the vehicle, and the camera 203 is mounted on the upper part of the windshield of the vehicle. The surrounding environment data includes surrounding image data, first radar surrounding data, and second radar surrounding data. Acquiring the vehicle's surrounding environment data through the perception module 200 includes: acquiring the vehicle's driving speed; in response to the driving speed being within a preset speed range, acquiring surrounding image data through the camera 203, acquiring first radar surrounding data through the lidar 201, and acquiring second radar surrounding data through the millimeter-wave radar 202. The preset speed range can be calibrated according to time conditions, for example: 20-80 km / h.

[0031] It should be noted that the lidar 201 described in this embodiment can scan the vehicle's surrounding environment accurately in real time and construct a high-precision three-dimensional point cloud map. The first radar's surrounding data can include the pedestrian's position, speed, and direction of movement, and can predict in advance whether the pedestrian intends to suddenly enter the vehicle's driving path. For example, when a pedestrian leans forward or accelerates their pace on the roadside, the lidar 201 can detect it in time.

[0032] Furthermore, the millimeter-wave radar 202 described in this embodiment focuses on detecting pedestrians under adverse weather conditions (such as rain, fog, and snow), compensating for the shortcomings of lidar 201 and camera 203 which are affected by weather. It integrates and complements the data from the former two to ensure stable operation of the system in any environment. The second radar's surrounding data may include the dynamics of pedestrians around the vehicle.

[0033] Furthermore, the camera 203 described in this embodiment may include multiple high-definition cameras to form a high-definition camera group, including front-view and side-view wide-angle cameras, etc., and uses deep learning algorithms to analyze the acquired images in real time to assist the lidar 201 in more accurately judging pedestrian behavior. Surrounding image data may include details such as pedestrian posture, facial orientation, and whether they are carrying items, for example, identifying features such as elderly people walking unsteadily and potentially moving slowly, and children running and playing with unpredictable routes.

[0034] As one possible scenario, by using a combination of LiDAR 201, millimeter-wave radar 202 and camera 203, the probability of an accident can be identified 3-5 seconds before the accident occurs, and the algorithm can be used to determine whether a collision with a pedestrian / two-wheeled vehicle has occurred.

[0035] The principle that needs to be ensured in the development of the vehicle safety system 100 is that when the head of a pedestrian / bicycle rider impacts the vehicle (generally landing on the outer surface of the engine hood), the execution module 400 should have already raised the engine hood to a preset position. If the engine hood is raised later than the time it is in position, a "head-on collision" will occur as the pedestrian's head falls and the engine hood rises. Since the direction of the pedestrian's head falling and the direction of the engine hood rising are opposite, this process not only fails to achieve the protective effect, but also causes additional injury to the head of the pedestrian / bicycle rider.

[0036] Step S102: The decision module generates collision prediction data for the vehicle based on surrounding environmental data, and generates a protection execution command in response to the collision prediction data meeting preset collision conditions. The preset collision conditions can be calibrated according to actual conditions.

[0037] In one embodiment of this application, the decision module 300 generates vehicle collision prediction data based on surrounding environmental data, including: determining the target object of the vehicle based on surrounding image data, first radar surrounding data, and second radar surrounding data; extracting features from the surrounding image data based on the target object to obtain a first behavioral feature of the target object; extracting features from the first radar surrounding data based on the target object to obtain a second behavioral feature of the target object; and extracting features from the second radar surrounding data based on the target object to obtain a third behavioral feature of the target object. The vehicle collision prediction data is then generated based on the first, second, and third behavioral features.

[0038] Further, based on surrounding image data, first radar surrounding data, and second radar surrounding data, the target object of the vehicle is determined, including: acquiring the vehicle's driving data; extracting the object's state data from the surrounding image data, wherein the object includes one or more of pedestrians, two-wheeled vehicle riders, and three-wheeled vehicle riders, and the state data includes the object's posture, the object's facial orientation, the object's forward speed, and direction; extracting the object's behavior data from the first radar surrounding data and / or the second radar surrounding data, wherein the behavior data includes position data, movement direction data, and movement speed data; and determining the vehicle's target object based on the state data, behavior data, and driving data.

[0039] To clearly illustrate the previous embodiment, in the embodiments of this application, vehicle collision prediction data is generated based on a first behavioral feature, a second behavioral feature, and a third behavioral feature, including: acquiring the predicted collision time and collision process time of the target object, and acquiring the system decision time and system deployment time of the vehicle; adding the predicted collision time and the collision process time to obtain a first time, and adding the system decision time and the system deployment time to obtain a second time; in response to the first time being greater than the second time, performing feature fusion on the first behavioral feature, the second behavioral feature, and the third behavioral feature to obtain a target behavioral feature; and inputting the target behavioral feature into a collision prediction model to obtain collision prediction data, wherein the collision prediction data includes a collision probability value.

[0040] In one embodiment of this application, obtaining the predicted collision time of the target object includes: obtaining the distance between the vehicle and the target object, and calculating the predicted collision time based on the distance and the driving speed.

[0041] In one embodiment of this application, the preset collision condition includes a collision probability value greater than a preset collision probability threshold. It should be noted that the preset collision probability threshold can be set according to actual conditions, for example, 80%.

[0042] Specifically, the decision module 300 outputs the collision probability value between the current target and the vehicle. When the collision probability value is greater than the preset collision probability threshold, that is, when the collision probability value is greater than 80%, the preset collision condition is met.

[0043] It should be noted that the predicted collision time (TTC) described in this embodiment refers to the time required for the vehicle to collide with an obstacle ahead while maintaining its current speed and direction. The vehicle needs to monitor the surrounding environment using various sensors (such as LiDAR 201, millimeter-wave radar 202, and camera 203) to calculate the TTC of the vehicle with other road users (including pedestrians, bicycles, and other vehicles) in real time. For example, when driving on urban roads, if a pedestrian suddenly appears crossing the road ahead, the autonomous driving system uses the aforementioned sensors to obtain the distance and relative speed to the pedestrian and calculates the TTC. If the TTC is less than a set threshold, the system will immediately trigger emergency braking and the pedestrian protection system to avoid or mitigate the impact of the collision.

[0044] It should be noted that the threshold described in this embodiment can be determined according to the actual situation, and no specific limitation is made here.

[0045] Understandably, fusing data from multiple sensors, such as LiDAR 201, millimeter-wave radar 202, and camera 203, can more accurately determine TTC. For example, LiDAR 201 and millimeter-wave radar 202 provide precise distance and velocity information, while camera 203 provides rich target object feature information. By using a fusion algorithm, the advantages of both are combined, improving the accuracy and reliability of TTC calculation. In complex road conditions, for instance, if radar detects multiple target objects but struggles to distinguish them accurately, combining the image recognition results from camera 203 can identify the truly relevant target object, thus enabling more precise TTC calculation.

[0046] As one possible approach, based on TTC calculation, this application provides a kinematic model simulation: A kinematic model of the vehicle and the target object is established, describing the changes in their position, velocity, acceleration, and other states over time. Based on the set initial conditions and motion parameters, the trajectories of the vehicle and the target object are predicted through mathematical calculations, thereby deriving their relative distance and relative velocity, and calculating the TTC. For example, assuming the vehicle and the target object undergo uniformly accelerated linear motion, the future position and velocity are calculated using kinematic formulas, thus obtaining the TTC.

[0047] As another possible approach, based on TTC calculation, this application provides a Monte Carlo simulation suitable for handling problems with uncertainty and randomness. In TTC simulation, factors such as sensor measurement errors and the uncertainty of target object motion can be considered. Through extensive random sampling, different scenarios and parameter combinations are simulated, and the TTC is calculated for each sampling. Finally, statistical analysis is performed to obtain the distribution of TTC, providing a basis for reliability assessment and risk analysis of the sensing system.

[0048] As another possible approach, based on TTC calculations, this application also provides a hardware-in-the-loop simulation: This method connects some real hardware devices, such as LiDAR 201, millimeter-wave radar 202, and cameras 203, to the simulation system. The sensors collect real data and input it into the simulation model, while the simulation model outputs control signals to the hardware devices, forming a closed-loop system. Through this loop simulation, the actual operation of the pedestrian protection system can be more realistically simulated, making TTC calculations closer to reality, effectively verifying the system's performance and reliability, and promptly identifying and resolving potential problems.

[0049] As another possible approach, this application can also be used in combination of the three methods, applicable to different stages of vehicle safety system 100 development. In the initial development phase, the first two models can be used for simulation. After the sensor information is determined, the third model is used to determine the final TTC (Total Traffic Conversion) index. Typical TTC curves at different vehicle speeds are shown below. Figure 18As shown, for pedestrian protection, the effective range of TTC is the minimum threshold speed to 80 km / h: below the minimum threshold speed set by the system, it is believed that a collision with the pedestrian's head will not cause injury; for impact energy above 80 km / h, even this system cannot prevent fatal injury.

[0050] It should be noted that the collision process time described in this embodiment, HIT (Head Impact Time), refers to the time required from the moment a human body contacts the vehicle after a collision until the pedestrian's head impacts the vehicle's surface. In the determination of the vehicle safety system 100, this is primarily achieved through virtual simulation. The method involves establishing finite element models of the target vehicle and the virtual simulated human body, and operating at a low threshold speed set by the vehicle safety system 100 (generally 20-25 kph; if the speed is below this, it indicates that the impact energy between the vehicle and the pedestrian / bicycle is insufficient to cause injury; if the speed is above this and a collision occurs, the vehicle safety system 100 needs to be triggered). This timeframe is generally within 20-80 milliseconds.

[0051] It should be noted that the finite element model described in this embodiment includes a child simulation dummy 10, a female simulation dummy 20, a medium-sized male simulation dummy 30, and a large-sized male simulation dummy 40, which collide with the human body through a set system low threshold speed (e.g., Figure 11 As shown), the contact time between the human head and the vehicle surface was recorded, and the minimum value between the contact time and the time from head to head landing on the vehicle was the HIT (e.g., ...). Figure 12 (As shown).

[0052] Based on the system design requirement of TTC+HIT>SDT+DT, specify the target settings for SDT and DT (pedestrian protection system deployment time).

[0053] It should be noted that the system decision time (SDT) described in this embodiment refers to the time when a collision scenario is received. The system automatically identifies the collision risk using an algorithm, and once the upper limit of the system's preset collision probability threshold is reached, a trigger signal is issued to activate the execution module 400. Generally, the SDT is very short, within 1-2 milliseconds.

[0054] It should be noted that the system deployment time (DT) described in this embodiment refers to the time it takes for the system to fully deploy to the designed state after receiving the trigger signal, which is generally within 10-100 milliseconds.

[0055] Therefore, the design principle of HIT+TTC>SDT+DT must be met in the development of vehicle safety system 100.

[0056] Understandably, this application employs a multimodal data fusion model to efficiently integrate radar distance and velocity information with visual image data from camera 203. A deep learning-based fusion algorithm is used, trained on a large amount of labeled data, enabling the model to automatically learn the complementary relationships between different modalities. For example, a convolutional neural network (CNN) is used to process images from camera 203 to extract pedestrian appearance features; simultaneously, a specially designed radar data processing network is used to analyze radar point cloud data to obtain the pedestrian's position and motion state.

[0057] Then, these two sets of features are fused at higher levels of the network to more comprehensively and accurately identify pedestrians. For complex and ever-changing road scenarios, an adaptive perception strategy is introduced. Through real-time monitoring of environmental factors (such as weather, lighting, and road conditions), the parameter settings of the radar and camera 203 are automatically adjusted. For example, in rainy or foggy weather, the radar performance is relatively stable, and the weight of radar data in the fusion decision can be appropriately increased; while on sunny days with sufficient light, the high-resolution advantage of the camera 203 is fully utilized, focusing on using image information for pedestrian identification.

[0058] Simultaneously, machine learning algorithms are used to analyze historical environmental data and perception results to predict potential changes in pedestrian characteristics under different environments, thus optimizing the perception model in advance. When camera 203 detects a pedestrian, it transmits the pedestrian's initial position information to radar. Radar, utilizing its high-precision distance measurement capabilities, tracks the pedestrian's trajectory in real time and feeds back the updated position information to camera 203. This closed-loop collaborative mechanism improves the accuracy and stability of pedestrian trajectory tracking. Furthermore, combined with a deep learning prediction model, based on the pedestrian's current movement state and historical trajectory, the system predicts the pedestrian's next action in advance, providing the vehicle safety system 100 with more decision-making time.

[0059] Step S103: The pneumatic lifting device controls the lifting components according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the vehicle's engine hood to a preset height. The preset height can be calibrated according to the actual situation.

[0060] In one embodiment of this application, such as Figure 5 and Figure 6As shown, the pneumatic lifting device 401 includes: a high-pressure gas cylinder 4011, an air guide pipe 4012, an electronic control valve, and four lifting devices 4013. The high-pressure gas cylinder 4011 is connected to the vehicle frame, and both ends of the air guide pipe 4012 are connected to the high-pressure gas cylinder 4011 and the four lifting devices 4013, respectively. The four lifting devices 4013 are respectively located between the vehicle frame and the vehicle's engine hood, with two lifting devices 4013 located at the front of the vehicle's engine hood and the other two at the rear of the vehicle's engine hood. The electronic control valve is located on the high-pressure gas cylinder 4011 and is connected to the vehicle controller 4014. Each lifting device 4013 is equipped with a vent valve 4015.

[0061] It should be noted that the high-pressure gas cylinder 4011 described in this embodiment stores high-pressure inert gas for driving the lifting device 4013. The high-pressure gas cylinder 4011 can be filled with gas using a dedicated filling device. After the system has been triggered a specified number of times, the gas can be replenished via the filling device. Generally, the pressure of a vehicle-mounted gas cylinder is 30-100 MPa, and it is reusable.

[0062] As a possible scenario, a check valve is arranged in the middle of the gas guide pipe 4012. When gas is input to the lifter 4013, it can be connected, but not vice versa, so as to ensure the unidirectional flow of inert gas.

[0063] It should be noted that the vent valve 4015 installed at the lower end of the jack 4013 opens to release gas when the gas flow exceeds the set threshold, causing the jack rod of the jack 4013 to fall back to the initial position, thus achieving automatic reset of the jacking height.

[0064] Understandably, the difference between this pneumatic lifting device 401 and traditional lifting solutions lies in the fact that traditional active lifting systems only lift the rear of the engine hood. Therefore, the improved space results in a slight reduction in the front space of the engine hood (with the front lock of the engine hood as the rotation point) and an increase in thickness. This pneumatic lifting device 401 has four lifting positions, aiming to lift the entire engine hood area. The two front lifting points are on the inner panel of the engine hood, and the two rear lifting points are on the active hinge 402. The corresponding positions of the engine hood area are as follows: Figure 7 As shown.

[0065] In one embodiment of this application, such as Figure 8 and Figure 9As shown, the active hinge 402 includes: a first link 4021, a second link 4022, a third link 4023, a fourth link 4024, and a fifth link 4025. The first link 4021, second link 4022, third link 4023, fourth link 4024, and fifth link 4025 are connected by connecting shafts. The first link 4021 and fifth link 4025 are connected to the vehicle's engine hood and frame by connecting parts. The first link 4021 has a lifting portion 4026. The first link 4021 has a limiting groove 4027, and the second link 4022 has a limiting stop bar 4028, which is slidably disposed within the limiting groove 4027. The third link 4023 and fourth link 4024 each have a limiting stop block 4029.

[0066] It should be noted that the limiting stop bar 4028 described in this embodiment is slidably disposed in the limiting groove 4027, the purpose of which is to limit the lifting height of the first link 4021. The third link 4023 and the fourth link 4024 are respectively provided with limiting blocks 4029. The two limiting blocks 4029 can contact each other to prevent the active hinge 402 structure from continuously deforming (when impacted, the two limiting blocks 4029 form a resisting structure to prevent the head from falling further and touching the rigid structure inside the cabin).

[0067] As a possible scenario, displacement sensors are placed at the jacking position. The measurement system measures the moment when the jacking unit 4013 pushes the jacking position to the designed height, with DT recording the maximum value of the four jacking position arrival times. Once the vehicle styling is determined, TTC is determined through the perception system development, and HIT is determined through simulation modeling. Therefore, by confirming SDT through the system, the DT target can be set, and the corresponding jacking height and jacking air pressure can be designed based on the DT time.

[0068] As another possible scenario, a simulation method is used to determine DT: a simulation model of the pneumatic lifting device 401 is established, and the set internal pressure curve is input to drive the lifting rod to move upward, lifting the engine hood assembly to a specified height and maintaining pressure. After a period of lifting (programmed depressurization after a pedestrian head impact), the engine hood assembly returns to its pre-lift position (to avoid prolonged obstruction of the driver's view). DT is defined as the time from receiving the ignition signal to the first reaching of the designed height.

[0069] The internal pressure curve of the jack 4013 is as follows: Figure 19 As shown, after receiving the ignition command from the vehicle computer, the program controls the high-pressure gas cylinder 4011 to quickly release the gas pressure and lift the push rod; after reaching the set height, it performs pressure holding to keep the height position at the effective design height; when the sensing module 200 determines the collision event, the vehicle controller 4014 issues a deflation command, and the push rod returns to the initial position.

[0070] Furthermore, the determination of the lifting height is primarily based on the injury requirements of the affected area: since the source of head impact injury in vehicles is insufficient head cushioning space, the system's improvement method is to lift the contact area (mainly referring to the engine hood) to a certain height so that the cushioning space is sufficient to absorb the impact when the head strikes. Therefore, the minimum lifting space must be determined based on the original injury distribution of the head impact area.

[0071] This vehicle safety system 100 provides four lift spaces, by Figure 7 It is known that two of the lifting devices 4013 are located at the leading edge of the engine hood; the other two lifting devices 4013 are located at the position of the active hinge 402. Generally, the two front lifting devices are required to have the same lifting height, and the two rear lifting devices are required to have the same lifting height. Generally, lifting injuries can be determined using the following empirical formula (here, the impact point is located in the vertical space from the outer surface of the engine hood to the lower rigid structure of the engine compartment): If the space is greater than 100mm, it is predicted as a green dot and can be awarded 1 point (full marks). If 100mm or more is greater than 80mm, the prediction is a yellow dot, which can be scored as 0.75 points. If the space is greater than or equal to 80mm, it is predicted as an orange dot and can be scored 0.5 points. If the space is 60mm or more and the area is greater than 40mm, it is predicted to be a brown dot and can be scored 0.25 points. If the space is ≤40mm, the prediction is a red dot, no points can be scored, and it is recorded as 0 points.

[0072] The total score for the entire impact area is calculated by dividing the total score by the number of impact points. Four lifting heights can be designed specifically to meet the protection requirements of the entire area, based on the overall score requirements.

[0073] The lifting height and time are controlled by air pressure. The lifting device 4013 used in this application has a cylindrical structure, as shown in the figure below. Figure 14 As shown. Due to the need to withstand high pressure, the internal pressure of the outer cylinder of the jacking device 4013 needs to be checked and determined. The intensity of the internal pressure is shown in the following formula:

[0074] in, t is the strength borne by the outer diameter of the cylinder; P is the internal pressure of the jack 4013; r is the radius of the cylinder; t is the wall thickness of the cylinder.

[0075] After verification, the safety factor k of the outer cylinder is as follows: ; Where K is the safety factor for the outer diameter of the cylinder; This represents the maximum tensile strength of the outer cylinder material. This refers to the strength that the outer diameter of the cylinder can withstand.

[0076] Generally, k≥2 is required to select the outer cylinder material.

[0077] As a possible scenario, after the vehicle safety system 100 is developed, an effectiveness verification procedure is required. The method for verifying the effectiveness of this system is as follows: the vehicle on which the system is installed needs to undergo simulation and real-vehicle road testing with the following matrix: The test primarily examines whether the system can effectively detonate. The following 12 operating conditions must be tested (e.g., ...). Figure 15 As shown), the system effectiveness coefficient θ is calculated: ; Where θ is the system effectiveness coefficient; n e n represents the number of valid operating conditions. t This represents the total number of test conditions.

[0078] After the vehicle safety system 100 is installed in a vehicle, its effectiveness in protecting the head of pedestrians / bicycle riders needs to be evaluated. The head injury score is calculated as follows: Generally, the overall head injury assessment involves individually scoring all test points within the impact area and then converting this score into a head injury score for the vehicle model. This score is typically recorded using a score area map. Figure 16 This is a score diagram for heads that do not have a protection system. Figure 17 The image shows the score after the system is expanded. The blue box indicates the range of system impact assessment points, and point X is the verification point selected for the experiment.

[0079] To ensure system robustness, the head shape scoring formula is designed as follows:

[0080] ~(0,1) Among them, S p S is the final score after the pedestrian protection system is installed. u To exclude the area outside the protection range of the protection system, A represents the score for the affected area without system protection; B represents the score for the affected area with system protection; C represents the score for the affected area outside the protection range of the protection system. i The score is the score of a single point after the system protection is implemented; N is the number of verification points selected.

[0081] It should be noted that the design is confirmed through simulation, but after the actual vehicle is produced, it is necessary to test and verify whether the score is consistent with the simulation. The final score is based on the test and verification score. However, since there are too many points in the impact area, it is impossible to test every point. Therefore, a portion of the points are selected for testing. In this example, N impact points are selected, and the remaining unselected points are executed according to the simulation prediction score.

[0082] Finally, the system's final score is derived by combining its effectiveness and protection effect: S e =S p ×θ Combination Figure 1 , Figure 2 and Figure 3 The present application will describe in detail the pedestrian protection method based on a vehicle safety system according to an embodiment.

[0083] Specifically, during vehicle operation, the perception module 200 continuously monitors the surrounding environment. A lidar 201 and a millimeter-wave radar 202 are mounted on the front bumper, while a high-definition camera 203 is positioned at the top of the windshield. When the vehicle's speed is within a preset range of 20-80 km / h, the system activates a multi-sensor collaborative acquisition mode: the lidar 201 emits a laser beam to scan the environment in front of the vehicle, constructing a high-precision 3D point cloud map to acquire real-time data on the pedestrian's precise location, speed, and direction of movement; the millimeter-wave radar 202 continues to operate under adverse weather conditions, acquiring data on the second radar's surroundings to compensate for the weather-dependent limitations of the lidar 201 and camera 203; and the high-definition camera 203 acquires surrounding image data, using deep learning algorithms to obtain detailed information such as the pedestrian's posture, facial orientation, and whether they are carrying items. Through this combined sensor perception, the system can identify the probability of an accident 3-5 seconds before it occurs.

[0084] After the multimodal data collected by the perception module 200 is transmitted to the decision module 300, the central processing unit and the graphics processor in the vehicle controller 4014 work together to process the massive amount of data quickly. The system first extracts the state data of objects from the surrounding image data, including the posture, facial orientation, forward speed, and direction of pedestrians, two-wheeled cyclists, etc.; at the same time, it extracts the position, direction of movement, and speed of objects from the radar data, and combines the vehicle's own driving data to comprehensively determine the entity with the greatest collision risk in the current scene as the target object. Subsequently, for the determined target object, the system extracts behavioral features from sensor data of different modalities: the first behavioral feature is extracted from the surrounding image data, the second behavioral feature is extracted from the LiDAR 201 data, and the third behavioral feature is extracted from the millimeter-wave radar 202 data.

[0085] After feature extraction, the system enters the collision risk quantification assessment stage. The decision module 300 acquires the predicted collision time (TTC) of the target object, which is the time required for the vehicle to collide with the target object while maintaining the current speed and direction. TTC calculation employs a multi-sensor data fusion method, using LiDAR 201 and millimeter-wave radar 202 to obtain precise distance and relative speed information, combined with target feature information provided by camera 203, and processed by a fusion algorithm to obtain a more accurate TTC value. Simultaneously, the system acquires the collision process time (HIT), which is the time from when a person makes contact with the vehicle to when the pedestrian's head impacts the vehicle surface. HIT is determined through virtual simulation, establishing finite element models covering children, women, medium-sized men, and large-sized individuals. Collision simulations are performed at a low threshold speed of 20-25 kph, recording the contact time between the head and vehicle surface for each model and taking the minimum value. In addition, the system also acquires the system decision time (SDT) (typically 1-2 milliseconds) and the system deployment time (DT) (typically 10-100 milliseconds).

[0086] The system follows the design principle of HIT+TTC>SDT+DT for time verification. Only when the first time interval is greater than the second time interval is the system confirmed to have a sufficient response time window to continue subsequent collision probability calculations. After the time condition is met, the decision module 300 uses a multimodal data fusion model to deeply fuse the first, second, and third behavioral features. It uses a convolutional neural network to process images from camera 203 to extract pedestrian appearance features, and simultaneously uses a radar data processing network to analyze radar point cloud data to obtain the pedestrian's position and motion state. Then, the two sets of features are fused at a higher level of the network to obtain the target behavioral features. The system also introduces an adaptive perception strategy, automatically adjusting the fusion weights of different sensor data through real-time monitoring of environmental factors—appropriately increasing the weight of radar data in rainy or foggy weather, and focusing on image information in well-lit sunny weather. The fused target behavioral features are input into the collision prediction model, which outputs the current collision probability value between the target and the vehicle, as well as the predicted trajectory of the target in the next few seconds. Simultaneously, the system interacts with surrounding vehicles and intelligent transportation infrastructure in real time through the vehicle-to-everything (V2X) module to obtain more comprehensive road condition and blind spot pedestrian information as supplementary data for collision prediction.

[0087] The decision module 300 compares the collision probability value output by the collision prediction model with a preset collision probability threshold (e.g., 80%). This threshold is dynamically and adaptively adjusted based on the vehicle's current speed, the type of target object, and the risk level of the external environment. When the collision probability value is greater than the adaptive threshold for the current scenario, and this state is maintained for multiple consecutive decision cycles, the decision module 300 determines that the preset collision condition is met, generates a protection execution command, and sends it to the execution module 400.

[0088] After receiving the protection execution command, the execution module 400 initiates the operation of the pneumatic lifting device 401. A high-pressure gas cylinder 4011, connected to the vehicle frame, stores 30-100 MPa high-pressure inert gas. The two ends of the gas guide pipe 4012 are connected to the high-pressure gas cylinder 4011 and the four lifting devices 4013, respectively, with a check valve in the middle to ensure unidirectional gas flow. An electronic control valve is located on the high-pressure gas cylinder 4011 and connected to the vehicle controller 4014. After the vehicle controller 4014 issues an ignition command, the electronic control valve programmatically controls the high-pressure gas cylinder 4011 to rapidly release high-pressure gas. The gas is then transported through the gas guide pipe 4012 to the four lifting devices 4013. Two of the four lifting devices 4013 are located at the front of the engine hood (lifting from the inner panel of the engine hood), and two are located at the rear of the engine hood (lifting from the active hinge 402), with the aim of lifting the entire engine hood area to maximize the buffer space.

[0089] High-pressure gas enters the lifter 4013 and rapidly lifts the push rod, which then impacts the active hinge 402. Upon impact from the lifter 4013, the first connecting rod 4021 separates from the second connecting rod 4022, causing the active hinge 402 to open in the reverse direction. The system lifts the engine hood to a preset height within a preset system deployment time DT. DT is determined by establishing a simulation model of the pneumatic lifting device 401 and inputting a set internal pressure curve to drive the push rod upward.

[0090] Once the hood has risen to the designed height, the electronic control valve maintains appropriate air pressure to keep the lifting height at the effective design height for a certain period of time, ensuring that sufficient buffer space has been formed in the event of a pedestrian head impact. When a head impact occurs, the limiting blocks 4029 on the third link 4023 and the fourth link 4024 of the active hinge 402 contact each other, forming a resisting structure to prevent the head from falling further and hitting the rigid structure inside the engine compartment, thus avoiding a "head-on collision". After the sensing module 200 has completed continuous monitoring and determination of the collision event, the on-board controller 4014 issues a deflation command, and the electronic control valve controls the deflation valve 4015 at the lower end of the lifter 4013 to open and release pressure, causing the lift rod to fall back to its initial position under air pressure, and the hood returns to its pre-trigger state, preventing prolonged obstruction of the driver's view.

[0091] It is understandable that the technical solution provided in this application significantly improves recognition accuracy and reduces false triggering rate by actively sensing and predicting pedestrian collision risks. At the same time, the use of a pneumatically resettable lifting device eliminates the risk of obstructed vision and reduces maintenance costs.

[0092] This application embodiment also provides a vehicle safety system 100, including a sensing module 200, a decision module 300, and an execution module 400. The decision module 300 is connected to the sensing module 200 and the execution module 400, respectively. The execution module 400 includes a pneumatic lifting device 401 and an active hinge 402. The sensing module 200 is used to acquire data of the vehicle's surrounding environment. The decision module 300 is used to generate collision prediction data of the vehicle based on the surrounding environment data, and generate a protection execution command in response to the collision prediction data meeting preset collision conditions. The execution module 400 is used to control the lifter 4013 of the pneumatic lifting device 401 according to the protection execution command, so that the pneumatic lifting device 401 cooperates with the active hinge 402 to lift the vehicle's engine hood to a preset height.

[0093] In summary, the pedestrian protection method and vehicle safety system based on the vehicle safety system in this application actively acquire surrounding environmental data through the perception module and generate collision prediction data through the decision module, realizing predictive identification of vulnerable road users such as pedestrians. Compared with the traditional passive solution that relies on collision signal triggering, it significantly improves the accuracy of accident identification and effectively reduces the probability of false triggering, while also gaining more response time for execution. On this basis, the execution module adopts a combination structure of pneumatic lifting device and active hinge, and realizes the lifting, pressure holding and natural falling of the engine hood through programmed pneumatic control. This not only avoids the driving safety hazards caused by the engine hood obstructing the driver's vision for a long time, but also abandons the traditional irreversible pyrotechnic drive scheme. After the system is triggered, there is no need to replace the entire active engine hood assembly, thereby greatly reducing maintenance costs and improving user experience. It effectively solves the comprehensive technical problems of high false triggering rate, large safety hazards and high maintenance costs in the prior art.

[0094] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc., indicating the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0095] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0096] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a communication connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0097] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0098] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A pedestrian protection method based on a vehicle safety system, characterized in that, The vehicle safety system includes a perception module, a decision module, and an execution module. The decision module is connected to both the perception module and the execution module. The execution module includes a pneumatic lifting device and an active hinge. The method includes: The sensing module acquires data about the vehicle's surrounding environment. The decision module generates collision prediction data for the vehicle based on the surrounding environment data, and generates a protection execution command in response to the collision prediction data meeting the preset collision conditions. The pneumatic lifting device controls the lifting components of the pneumatic lifting device according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the engine cover of the vehicle to a preset height.

2. The pedestrian protection method based on a vehicle safety system according to claim 1, characterized in that, The sensing module includes a lidar, a millimeter-wave radar, and a camera. The lidar and the millimeter-wave radar are respectively mounted on the front bumper of the vehicle, and the camera is mounted on the upper part of the windshield of the vehicle. The surrounding environment data includes surrounding image data, first radar surrounding data, and second radar surrounding data. The acquisition of the vehicle's surrounding environment data through the sensing module includes: Obtain the vehicle's speed; In response to the driving speed being within a preset speed range, the system acquires surrounding image data via the camera, acquires surrounding data of the first radar via the lidar, and acquires surrounding data of the second radar via the millimeter-wave radar.

3. The pedestrian protection method based on a vehicle safety system according to claim 2, characterized in that, The step of generating collision prediction data for the vehicle based on the surrounding environment data through the decision module includes: Based on the surrounding image data, the first radar surrounding data, and the second radar surrounding data, the target object of the vehicle is determined; Based on the target object, feature extraction is performed on the surrounding image data to obtain the first behavioral feature of the target object; Based on the target object, feature extraction is performed on the surrounding data of the first radar to obtain the second behavioral features of the target object; Based on the target object, feature extraction is performed on the surrounding data of the second radar to obtain the third behavioral feature of the target object; Based on the first behavioral feature, the second behavioral feature, and the third behavioral feature, collision prediction data for the vehicle is generated.

4. The pedestrian protection method based on a vehicle safety system according to claim 3, characterized in that, The step of determining the target object of the vehicle based on the surrounding image data, the first radar surrounding data, and the second radar surrounding data includes: Obtain the vehicle's driving data; Extract the state data of the object from the surrounding image data, wherein the object includes one or more of pedestrians, two-wheeled vehicle riders and three-wheeled vehicle riders, and the state data includes the object's posture, the object's facial orientation, the object's forward speed and direction; The behavior data of the object is extracted from the peripheral data of the first radar and / or the peripheral data of the second radar, wherein the behavior data includes position data, direction of movement data and speed of movement data; Based on the status data, the behavior data, and the driving data, the target object of the vehicle is determined.

5. The pedestrian protection method based on a vehicle safety system according to claim 3, characterized in that, The step of generating vehicle collision prediction data based on the first behavioral feature, the second behavioral feature, and the third behavioral feature includes: The predicted collision time and collision process time of the target object are obtained, as well as the system decision time and system deployment time of the vehicle are obtained. The first time is obtained by adding the predicted collision time to the collision process time, and the second time is obtained by adding the system decision time to the system deployment time. In response to the first time being greater than the second time, feature fusion is performed on the first behavioral feature, the second behavioral feature, and the third behavioral feature to obtain the target behavioral feature; The target behavior features are input into the collision prediction model to obtain the collision prediction data, wherein the collision prediction data includes collision probability values.

6. The pedestrian protection method based on a vehicle safety system according to claim 5, characterized in that, The step of obtaining the predicted collision time of the target object includes: The distance between the vehicle and the target object is obtained, and the predicted collision time is calculated based on the distance and the driving speed.

7. The pedestrian protection method based on a vehicle safety system according to claim 5, characterized in that, The preset collision conditions include the collision probability value being greater than a preset collision probability threshold.

8. The pedestrian protection method based on a vehicle safety system according to claim 1, characterized in that, The pneumatic lifting device includes: a high-pressure gas cylinder, a gas delivery pipe, an electronic control valve, and four lifting devices, wherein... The high-pressure gas cylinder is connected to the vehicle frame, and the two ends of the air guide tube are respectively connected to the high-pressure gas cylinder and the four lifting devices; The four lifting devices are respectively disposed between the vehicle frame and the engine hood of the vehicle, wherein two of the lifting devices are located at the front of the engine hood of the vehicle, and the other two of the lifting devices are located at the rear of the engine hood of the vehicle. The electronic control valve is installed on the high-pressure gas cylinder and is connected to the vehicle controller; The lifting device is equipped with a vent valve.

9. The pedestrian protection method based on a vehicle safety system according to claim 8, characterized in that, The active hinge includes: a first link, a second link, a third link, a fourth link, and a fifth link, wherein, The first link, the second link, the third link, the fourth link, and the fifth link are connected by connecting shafts, and the first link and the fifth link are connected to the engine cover and the vehicle frame by connecting parts, respectively. The first connecting rod is provided with a lifting part; The first connecting rod has a limiting groove, and the second connecting rod has a limiting stop rod, which is slidably disposed in the limiting groove. Limiting blocks are provided on the third link and the fourth link respectively.

10. A vehicle safety system, characterized in that, It includes a sensing module, a decision-making module, and an execution module. The decision-making module is connected to both the sensing module and the execution module. The execution module includes a pneumatic lifting device and an active hinge. The sensing module is used to acquire data about the vehicle's surrounding environment. The decision module is used to generate collision prediction data for the vehicle based on the surrounding environment data, and to generate a protection execution command in response to the collision prediction data meeting preset collision conditions. The execution module is used to control the lifting device of the pneumatic lifting device according to the protection execution command, so that the pneumatic lifting device cooperates with the active hinge to lift the engine cover of the vehicle to a preset height.

Images