Development method of active and passive fusion pedestrian protection system based on ADAS perception
Through 3D modeling and simulation analysis, the problem of 2D modeling identifying pedestrian delay in simulated ADAS signal test is solved, and the accurate function trigger time judgment of the active hood system is realized, which improves pedestrian monitoring accuracy and system safety.
Patent Information
- Application Number
- CN202411136335.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-19
AI Technical Summary
The existing 2D modeling cannot accurately reflect the barrier vehicle hood area is lower than the total height of the vehicle in simulated ADAS signal test, resulting in delays in the recognition of pedestrians by the camera, and the calculation time accuracy is not high, affecting the function of the active hood system to trigger time judgment.
Using 3D modeling method, we establish test vehicles, obstacle vehicles and pedestrian models, use LS-DYNA and other software for simulation analysis, define the time and collision time required for cameras to identify pedestrian characteristics, output the TTC value of each working condition, and support the controller to Accurately judge the trigger time of the active hair cover function under cross-collision conditions.
Through the 3D modeling method, the related problems of 2D modeling model analysis are solved, the accuracy and time accuracy of the on-board camera for pedestrian monitoring are improved, and the accurate function triggering of the active hair cover system is supported, reducing the risk of false triggering.
Smart Images

Figure CN118656996B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of active hood systems, and in particular relates to a method for developing an active-passive fusion pedestrian protection system based on ADAS perception. Background Art
[0002] At present, the widely used signal sensing solution for active pedestrian protection systems in the industry is the pressure tube sensor and acceleration sensor solution, also known as the contact sensing solution. In the actual application scenario of the contact sensing solution, the ACU only recognizes whether it is a pedestrian through the collision signal, which will cause a certain proportion of false explosion problems.
[0003] In view of the fact that the configuration rate of active safety configuration AEB in automobiles has greatly increased, the number of hardware equipment such as cameras, millimeter-wave radars and other radars used has increased, which has provided a certain hardware foundation for the active and passive fusion of pedestrian protection active hood systems, making major domestic and foreign automakers study active and passive fusion solutions to reduce the false trigger rate. The active and passive fusion solution mainly uses camera recognition to assist in determining whether the object being collided is a pedestrian. If it is a pedestrian, an ignition command is issued according to the normal ignition threshold to lift the hood, increasing the energy absorption space after the pedestrian's head hits the hood; if it is not a pedestrian, the ignition threshold is increased so that the active hood system does not ignite to avoid false triggering problems.
[0004] Starting from 2023, the 2023 version of the China Insurance Automobile Safety Index Regulations officially proposed a test method for active hood systems using ADAS signals as a perception solution. This test method will propose new development ideas for the research and development of active hood systems in the domestic automobile industry in the future.
[0005] When simulating the ADAS signal test plan, it is necessary to calculate the distance between the pedestrian dummy / cycling dummy and the estimated collision point M between the on-board camera in the first frame during the collision between the test vehicle and the pedestrian. In the case of a cross-collision between a vehicle and a pedestrian, this distance is divided by the vehicle's driving speed to output the time TH1 between the pedestrian dummy / cycling dummy and the collision between the pedestrian dummy / cycling dummy and the vehicle in the first frame. In the case of a cross-collision between a vehicle and a pedestrian, TH1=ST+TTC. However, the above process model is modeled in 2D, and the obstacle vehicle is simplified into a rectangle, which cannot reflect that the obstacle vehicle's hood area is lower than the total height of the vehicle so that the camera can recognize the characteristics of an adult or cyclist in advance. Moreover, the above 2D modeling simplification leads to a delay in the time when the test vehicle camera sees the pedestrian, and the calculation time is not accurate. In the subsequent algorithm development, it cannot accurately support the judgment of the trigger time / TTC minimum threshold of the active hood function. Summary of the invention
[0006] The purpose of the present invention is to provide a method for developing an active and passive fusion pedestrian protection system based on ADAS perception to solve the related problems existing in 2D modeling model analysis.
[0007] The objective of the present invention is achieved through the following technical solutions:
[0008] A method for developing an active and passive fusion pedestrian protection system based on ADAS perception includes the following steps:
[0009] A. Set the ADAS signal solution test conditions used by the active hood system;
[0010] B. Establish the test vehicle model. The non-turning test vehicle settings are as follows:
[0011] B1. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data;
[0012] B2. The camera of the test vehicle, including the camera data, including the field of view verification data, needs to be divided into grids at the same time;
[0013] B3. Define all meshes, materials and properties;
[0014] C. Establish the test vehicle model. The turning condition test vehicle settings are as follows:
[0015] C1. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data;
[0016] C2, the camera of the test vehicle, where the camera data, including the field of view verification data, needs to be divided into grids at the same time;
[0017] C3. Define all meshes, materials and properties;
[0018] C4, modeling includes creating suspension hard point locations, components, elastic elements and bushings, hinges;
[0019] C5. Establish 12 connection points between the suspension and the chassis, B1-B12;
[0020] C6, B1-B12 connection points are connected to the vehicle body;
[0021] D. Establish an obstacle vehicle model using a simplified modeling method. The specific method is consistent with the test vehicle model in the second step, and the vehicle size is consistent with the size of the obstacle vehicle in the actual working condition;
[0022] E. Establish pedestrian dummy models, including adult pedestrian dummy models and child pedestrian dummy models. The sizes of pedestrian dummies should be consistent with those of pedestrian dummies in actual working conditions.
[0023] F. Establish rider models, including bicycle rider models and electric scooter rider models. The rider model sizes should be consistent with the rider sizes in the actual working conditions.
[0024] G. Simulation test condition analysis, define pedestrian feature recognition standards to determine the t0 time, establish simulation test conditions 1-10 according to test scenario conditions 1-10, and calculate the TTC values of 10 conditions;
[0025] H. Output results, providing the controller with basic TTC values for each operating condition.
[0026] Further, in step A, the ADAS signal scheme test conditions include 5 test conditions and 5 extended conditions, conditions 1-3 are pedestrians and vehicles crossing each other, and there is an obstructed vehicle; condition 4 is pedestrians and vehicles driving in opposite directions; condition 5 is pedestrians and vehicles crossing each other, and there are obstructed vehicles in both directions; condition 6 is pedestrians and vehicles crossing each other, without obstacles; condition 7 is cyclists and vehicles driving in opposite directions; conditions 8-9 are cyclists and vehicles crossing each other, without obstacles; condition 10 is scooter riders driving straight and vehicles turning left, without obstacles.
[0027] Furthermore, in step B3, all meshes need to be defined as 1 part, the material definition is rigid material MAT20, and the attribute definition is thickness 3mm.
[0028] Further, in step C4, the hard point part is established according to the hard point information of the suspension; each component of the suspension is simplified to a rigid beam unit, and each component of the rigid beam unit is simulated using the MAT 20 material card of LS-DYNA, and the *Part_Inertia keyword is used to give the component real mass, center of mass position, moment of inertia and inertia product; the suspension spring, shock absorber and limit buffer block are modeled using the spring unit in DYNA, and the spring stiffness and shock absorber damping can be set; the spring pre-compression amount is set by the DYNA keyword *Element_Discrete, and the spring material uses the DYNA MAT S04 material model to simulate the nonlinear characteristics of the spring; the shock absorber uses the DYNA MAT S05 material model to simulate the damping characteristics of the shock absorber; the upper and lower buffer blocks of the shock absorber are also modeled by the spring unit of DYNA, and the working clearance of the buffer block is set by the keyword *Section_Discrete; the bushing is modeled using a zero-length beam unit, and the MAT Material model No. 67 can simulate the nonlinear elastic or viscoelastic properties of rubber materials; the bushing unit can set the stiffness and damping of the bushing in six directions through the local coordinate system; the hinge connection, the *Constrained_Joint keyword in DYNA can be used to set various kinematic pairs; beam units are established between the inner points of the steering rod to replace the simplified steering system, and a translational pair is established between the simplified steering system and the subframe to simulate the movement of the steering gear rack when the vehicle turns, and the rack stroke can be calculated through the transmission ratio of the steering wheel angle and the steering line angle.
[0029] Furthermore, in step C5, each connection point is characterized by a rigid point.
[0030] Furthermore, in step C6, the B1-B12 connection points are connected to the vehicle body using the *extra_nodes method.
[0031] Furthermore, in step G, the time t0 when the camera first detects 50% of the body features of the pedestrian dummy model / cyclist model is set.
[0032] Furthermore, the camera fully recognizes the physical features of the pedestrian dummy model / cyclist model as t1, t1-t0=TS.
[0033] Furthermore, TS is defined as the length of time required for the camera to recognize pedestrian features, the moment when the controller issues an active hood lifter trigger command is t2, t2-t0=ST, and ST is defined as the perception time required for the ADAS signal perception solution, which is a fixed value.
[0034] Furthermore, the contact time between the pedestrian dummy model / cyclist model and the test vehicle is t3, and t3-t2=TTC is defined, and the TTC value for each working condition is different; the detonation lower limit speed is set to 25km / h.
[0035] Compared with the prior art, the present invention has the following beneficial effects:
[0036] 1. The present invention adopts a 3D modeling method. The sizes of the test vehicle model, obstacle vehicle model, pedestrian model and cyclist model used in the modeling are consistent with the real obstacle vehicle, test vehicle, pedestrian entity collider and cyclist entity collider used in the test scene, which can completely solve the related problems existing in the 2D modeling model analysis;
[0037] 2. This 3D simulation analysis method can output the TH1 and TH2 values of various test schemes and newly added extended test scenarios. In the present invention, in the case of a cross-car collision, the first frame of the on-board camera can fully monitor the time TH1 between the pedestrian dummy / cyclist dummy and the pedestrian dummy / cyclist dummy and the car collision. TH1 is used to determine whether the system algorithm minimum response requirement (ST+DT) is met. <TTC+HIT);
[0038] 3. Test conditions: When the first main vehicle is 150m away from the pedestrian dummy's driving path (CPLA-25), and the main vehicle is 100m away from the bicycle (CBLA-50) / the longitudinal distance between the main vehicle and the scooter rider is 200m (CSFtap-50), when data recording begins, when a person and a vehicle collide in opposite directions, the time TH2 between the pedestrian dummy / cyclist dummy and the detection of the collision between the pedestrian dummy / cyclist dummy and the vehicle can be monitored earlier. TH2 is used to limit the TTC threshold for triggering the active hood function, and supports the formulation of the minimum threshold for the positive triggering of the controller's real active hood pop-up demand and the development of algorithms. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0040] Figure 1 A flowchart of the steps of the active and passive fusion pedestrian protection system development method based on ADAS perception of the present invention;
[0041] Figure 2 Test vehicle turning system models;
[0042] Figure 3 Schematic diagram of working condition 1 and working condition 2;
[0043] Figure 4 Schematic diagram of working condition 3;
[0044] Figure 5 Schematic diagram of working condition 4;
[0045] Figure 6 Schematic diagram of working condition 5;
[0046] Figure 7 Schematic diagram of working condition 6;
[0047] Figure 8 Schematic diagram of working condition 7;
[0048] Fig. 9 Schematic diagram of working condition 8;
[0049] Fig.10 Schematic diagram of working condition 9;
[0050] Fig.11 Schematic diagram of working condition 10. DETAILED DESCRIPTION
[0051] The present invention will be further described below in conjunction with embodiments:
[0052] The present invention takes the active hood system using ADAS signals as a perception scheme as a premise, and proposes a development method for an active and passive fusion pedestrian protection system based on ADAS perception. The method adopts a 3D modeling method, and the sizes of the test vehicle type, obstacle vehicle type, pedestrian model and cyclist model used in the modeling are consistent with the real obstacle vehicle, test vehicle, pedestrian entity collider and cyclist entity collider used in the test scene, which can completely solve the related problems existing in 2D modeling model analysis.
[0053] like Figure 1 As shown, the active and passive fusion pedestrian protection system development method based on ADAS perception of the present invention includes the following steps:
[0054] The first step is to set the ADAS signal test scheme used by the active hood system to cover the test scenarios. The ADAS signal scheme test condition setting includes 5 test conditions and 5 extended conditions.
[0055] The second step is to build a test vehicle model. The non-turning test vehicle settings are as follows:
[0056] 1. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data; 2. The test vehicle is equipped with a camera, and the camera data (including field of view verification data) needs to be meshed at the same time; 3. All the above meshes need to be defined as 1 part, the material definition is rigid material MAT20, and the attribute definition is thickness 3mm.
[0057] The third step is to build a test vehicle model. The turning test vehicle settings are as follows:
[0058] Step 1 to step 3 are the same as step 1 to step 3 in the above second step;
[0059] 4. Modeling includes creating suspension hard point locations, components, elastic elements and bushings, and hinges.
[0060] 5. Establish 12 connection points between the suspension and the chassis, B1~B12, and each connection point is characterized by a rigid point;
[0061] 6. The B1-B12 connection points are connected to the vehicle body using the *extra_nodes method;
[0062] The fourth step is to establish an obstacle vehicle model using a simplified modeling method. The specific method is consistent with the test vehicle model in the second step, and the vehicle size is consistent with the size of the obstacle vehicle in the actual working condition;
[0063] Step 5: Establish a pedestrian dummy model, including an adult pedestrian dummy model and a child pedestrian dummy model. The size of the pedestrian dummy is consistent with the size of the pedestrian dummy in the actual working condition.
[0064] Step 6: Establish rider models, including bicycle rider models and electric scooter rider models. The rider model sizes are consistent with the rider sizes in the actual working condition.
[0065] Step 7: Simulate the test condition analysis. Set the time when the camera first detects 50% of the physical characteristics of the pedestrian dummy model / cyclist model as t0. The time when the camera fully recognizes the physical characteristics of the pedestrian dummy model / cyclist model is t1. t1-t0=TS. TS is defined as the length of time required for the camera to recognize the pedestrian characteristics. The time when the controller sends the active hood lifter trigger command is t2. t2-t0=ST. ST is defined as the perception time required for the ADAS signal perception solution, which is a fixed value. The time when the pedestrian dummy model / cyclist model contacts the test vehicle is t3. t3-t2=TTC is defined. The TTC value for each condition is different. Set the lower detonation speed limit to 25km / h.
[0066] According to the test scenario conditions 1 to 10, simulation test conditions 1 to 10 are established, and a summary table of 10 simulation test conditions TTC1 to TTC10 is provided;
[0067] Step 8: Output the results to provide the controller with basic TTC values for each operating condition to support algorithm development.
[0068] In the present invention, TTC is Time To Collision, which is the remaining time before the test vehicle hits the test target. TRT is Total Response Time, which is the response time of the active hood system. ST is Sensing Time, which is the sensing time of the active hood system. DT is Deployment Time, which is the deployment time of the active hood system. HIT is Head Impact Time, which is the head collision time. CPLA-25 is Car-to-Pedestrian Longitudinal Adult 25%, which is a 25% longitudinal rear-end collision of an adult. CBLA-50 is Car-to-Bicyclist Longitudinal Adult 50%, which is a 50% longitudinal rear-end collision of a cyclist. CSFtap-50 is Car-to-Scooter Front Turn Across Path, which is a 50% condition of the main vehicle turning left and the scooter rider going straight in the opposite direction. CPNSOC is Car-to-Pedestrian Nearside Single-side Obstructed Child, which is a near-side single-side obstructed child condition. CPFOA is Car-to-Pedestrian Farside Obstructed Adult. CPNDOC is Car-to-Pedestrian Nearside Double-side ObstructedChild. CPNA is Car-to-Pedestrian Nearside Adult. CBNA is Car-to-Bicyclist Nearside Adult. CSFA is Car-to-Scooter Farside Adult.
[0069] Embodiment 1: A method for developing an active and passive fusion pedestrian protection system based on ADAS perception, comprising the following steps:
[0070] 1. Set the ADAS signal test solution used by the active hood system to cover the test scenarios, including 5 test conditions and 5 extended conditions 6 to 10;
[0071] Table 1 Insurance regulations ADAS signal solution test conditions
[0072] Condition 1: People and vehicles are crossing each other, and there is an obstacle vehicle
[0073] Condition 2: People and vehicles are crossing each other, and there is an obstacle vehicle
[0074] Condition 3: People and vehicles are crossing each other, and there is an obstacle vehicle
[0075] Condition 4: People and vehicles driving in opposite directions
[0076] Condition 5: People and vehicles are crossing each other, and there are vehicles with obstacles in both directions
[0077] Table 2 Expansion conditions
[0078] Condition 6: People and vehicles crossing each other, no obstacles
[0079] Condition 7: Bicyclist and vehicle traveling in opposite directions
[0080] Condition 8: Bicyclists and cars crossing each other, no obstacles
[0081] Condition 9: Scooter riders and cars cross each other, no obstacles
[0082] Condition 10: Scooter rider goes straight, car turns left, no obstacles
[0083] Table 3 Left turn path setting for test vehicles
[0084]
[0085] 2. Establish the test vehicle model. The non-turning test vehicle settings are as follows:
[0086] 21. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data; 22. The test vehicle is equipped with a camera, and the camera data, including the field of view verification data, needs to be meshed at the same time; 23. All the above meshes need to be defined as 1 part, the material definition is rigid material MAT20, and the attribute definition is 3mm thick;
[0087] 3. Build a test vehicle model. The turning test vehicle settings are as follows:
[0088] 31. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data; 32. The test vehicle is equipped with a camera, and the camera data, including the field of view verification data, needs to be meshed at the same time; 33. All the above meshes need to be defined as 1 part, the material definition is rigid material MAT20, and the attribute definition is 3mm thick;
[0089] 34. Modeling includes creating suspension hard point locations, components, elastic elements and bushings, and hinges.
[0090] Among them, 1. The hard point part is established according to the hard point information of the suspension; 2. The various parts of the suspension are simplified into rigid beam units, and the various components of the rigid beam unit are simulated using the MAT 20 material card of LS-DYNA, and the *Part_Inertia keyword is used to give the components real mass, center of mass position, moment of inertia and product of inertia; 3. The suspension spring, shock absorber and limit buffer block are modeled using the spring unit in DYNA, and the spring stiffness and shock absorber damping can be set; the spring pre-compression is set by the DYNA keyword *Element_Discrete, and the spring material uses the DYNA MAT S04 material model to simulate the nonlinear characteristics of the spring; 4. The shock absorber uses the DYNA MAT S05 material model, simulating the damping characteristics of the shock absorber; 5. The upper and lower buffer blocks of the shock absorber are also modeled by the spring unit of DYNA, and the working clearance of the buffer blocks is set by the keyword *Section_Discrete; 6. The bushing is modeled using a zero-length (two-point coincidence) beam unit, using the MAT67 material model in DYNA, which can simulate the nonlinear elastic or viscoelastic characteristics of rubber materials. The bushing unit can set the stiffness and damping of the bushing in 6 directions through the local coordinate system; 7. Hinge connection. There is a *Constrained_Joint keyword in DYNA that can be used to set various kinematic pairs; 8. A beam unit is established between the inner points of the steering rod to replace the simplified steering system, and a translational moving pair is established between the simplified steering system and the subframe to simulate the movement of the steering gear rack when the vehicle turns. The rack travel can be calculated by the transmission ratio of the steering wheel angle and the steering line angle.
[0091] 35. Establish 12 connection points between the suspension and the chassis, B1~B12, and each connection point is characterized by a rigid point;
[0092] 36. The B1-B12 connection point is connected to the vehicle body using the *extra_nodes method;
[0093] 4. Establish an obstacle vehicle model using a simplified modeling method. The specific method is consistent with the test vehicle model in the second step, and the vehicle size is consistent with the size of the obstacle vehicle in the actual working condition;
[0094] 5. Build pedestrian dummy models, including adult pedestrian dummy models and child pedestrian dummy models. The sizes of pedestrian dummies should be consistent with those of pedestrian dummies in actual working conditions.
[0095] 6. Establish rider models, including bicycle rider models and electric scooter rider models. The rider model size should be consistent with the rider size in the actual working condition;
[0096] 7. Analysis of simulation test conditions: set the moment when the camera first detects 50% of the physical characteristics of the pedestrian dummy model / cyclist model as t0, the moment when the camera fully recognizes the physical characteristics of the pedestrian dummy model / cyclist model as t1, t1-t0=TS, define TS as the length of time required for the camera to recognize the pedestrian characteristics, and the moment when the controller sends the active hood lifter trigger command is t2, t2-t0=ST, define ST as the perception time required for the ADAS signal perception solution, which is a fixed value. The moment when the pedestrian dummy model / cyclist model contacts the test vehicle is t3, define t3-t2=TTC, and the TTC value for each condition is different. Set the detonation lower limit speed to 25km / h. Establish simulation test conditions 1 to 10 according to the test scenario conditions 1 to 10, and the condition settings are as follows:
[0097] Working condition 1: CPNSOC working condition, the test vehicle SV1 is assigned a speed VC1 of the detonation lower limit speed of 7.5m / s (25+2km / h), the child pedestrian is assigned a speed of 1.389m / s (5km / h), and it is assumed that the test vehicle SV1 contacts the pedestrian at point M. The pedestrian dummy model moves at a constant speed from point R to point M at a distance of D11=4-1=3m. The walking time of the pedestrian dummy model is t11=3 / 1.389=2.16s. The initial position of the test vehicle SV1 is D12=VC1*t11=7.5*2.16=16.199m away from point M. The initial position of the child pedestrian dummy model is at point R. The placement of the obstacle vehicle is shown in the figure of working condition 1. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV1 and point M is D13, D13 / VC1=ST+TTC1, ST is a fixed value, TTC1=D23 / VC1-ST.
[0098] Working condition 2: CPNSOC working condition, the test vehicle SV1 is assigned a speed VC2 of 17.222m / s (60+2km / h), and the child pedestrian is assigned a speed of 1.389m / s (5km / h). It is assumed that the test vehicle SV1 contacts the pedestrian at point M. The pedestrian dummy model moves at a constant speed from point R to point M at a distance of D21=3m. The walking time of the pedestrian dummy model is t21=3 / 1.389=2.16s. The initial position of the test vehicle SV1 is D22=VC2*t21=17.222*2.16=37.2m away from point M. The initial position of the child pedestrian dummy model is at point R. The placement of the obstacle vehicle is shown in the figure of working condition 2. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV1 and point M is D23, D23 / VC2=ST+TTC2, ST is a fixed value, and TTC2=D23 / VC2-ST.
[0099] Working condition 3: CPFOA working condition, the test vehicle SV2 is assigned a speed VC3 of the detonation lower limit speed of 7.5m / s (25+2km / h), the adult pedestrian is assigned a speed of 1.389m / s (5km / h), and it is assumed that the test vehicle SV2 contacts the pedestrian at point M. The adult pedestrian dummy model moves at a constant speed from point R to point M at a distance of D31=5m. The walking time of the pedestrian dummy model is t31=5 / 1.389=3.6s. The initial position of the test vehicle SV2 is D32=VC3*t31=7.5*3.6=27m away from point M. The initial position of the adult pedestrian dummy model is at point R. The placement of the obstacle vehicle is shown in the figure of working condition 3. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV2 and point M is D33, D33 / VC3=ST+TTC3, ST is a fixed value, and TTC3=D33 / VC3-ST.
[0100] Condition 4: CPLA condition, the test vehicle SV3 is assigned a speed VC4 of 15.833m / s (55+2km / h), and the adult pedestrian dummy is assigned a speed of 1.389m / s (5km / h). This condition is a condition in which the test vehicle SV3 and the adult pedestrian dummy are driving in opposite directions. The test vehicle SV3 starts recording data 150m away from the adult dummy's driving path. It is assumed that the test vehicle SV3 and the pedestrian are in contact at point M. The contact time between the adult pedestrian dummy model and the test vehicle SV3 is t41=150 / (15.833+1.389)=8.71s, and the initial position of the adult pedestrian dummy is The distance from point M is D41=8.71*1.389=12.098m, the initial position of the test vehicle SV3 is D42=VC4*t41=15.833*8.71=137.902m away from point M, the working condition arrangement is shown in the working condition 4 diagram, the camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0, the distance between the test vehicle SV3 and point M is D43, D43 / VC4=ST+TTC4, ST is a fixed value, TTC4=D43 / VC4-ST, TTC4 is a larger value, which is mainly used to set the minimum threshold for the controller to actively trigger the hood pop-up demand.
[0101] Working condition 5: CPNDOC working condition, the test vehicle SV4 is assigned a speed VC5 of 11.667m / s (40+2km / h), and the child pedestrian is assigned a speed of 1.389m / s (5km / h). It is assumed that the test vehicle SV4 contacts the pedestrian at point M. The child pedestrian dummy model moves at a constant speed from point R to point M at a distance of D51=3m. The walking time of the pedestrian dummy model is t51=3 / 1.389=2.16s. The initial position of the test vehicle SV4 is D52=VC5*t51=11.667*2.16=25.2m away from point M. The initial position of the child pedestrian dummy model is at point R. The obstacle vehicle is placed as shown in the figure of working condition 5. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The test vehicle SV4 is D53 away from point M, D53 / VC5=ST+TTC5, ST is a fixed value, and TTC5=D53 / VC5-ST.
[0102] Condition 6: CPNA condition, the test vehicle SV5 is assigned a speed VC6 of 17.222m / s (60+2km / h), the adult pedestrian is assigned a speed of 1.389m / s (5km / h), the test vehicle SV5 is in contact with the pedestrian at point M, assuming that the test vehicle SV5 is 2m wide, H=25% vehicle width=0.5m, the adult pedestrian dummy model moves at a constant speed from point R to point M at a distance of D61=2.5m, and the pedestrian dummy model walks for a time t61=2 .5 / 1.389=1.8s, the initial position of the test vehicle SV5 is D62=VC6*t51=17.222*1.8=31m away from point M, the initial position of the adult pedestrian dummy model is at point R, the working condition setting is shown in the working condition 6 diagram, the camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0, the distance between the test vehicle SV5 and point M is D63, D63 / VC6=ST+TTC6, ST is a fixed value, TTC6=D63 / VC6-ST.
[0103] Condition 7: CBLA condition, the test vehicle SV6 is assigned a speed VC7 of 18.611 m / s (65 + 2 km / h), the cyclist is assigned a speed of 4.167 m / s (15 km / h), the test vehicle SV6 and the cyclist are in contact at point M, the cyclist travels a distance of D71 = 28 m to point M, the cyclist travels for t71 = 28 / 4.167 = 6.72 s, and the test vehicle SV6 is at a distance of D71 from point M. 72=VC7*t71=18.611*6.72=125.01m. The working condition setting is shown in the working condition 7 figure. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV6 and point M is D73. D73 / VC7=ST+TTC7, ST is a fixed value, TTC7=D73 / VC7-ST, and the calculated result of TTC7 is a larger value, which is mainly used to set the minimum threshold for the controller to actively trigger the hood pop-up demand.
[0104] Working condition 8: CBNA working condition, the test vehicle SV7 is assigned a speed VC8 of 17.222m / s (60+2km / h), and the cyclist is assigned a speed of 4.167m / s (15km / h). It is assumed that the test vehicle SV7 contacts the cyclist at point M. The distance from the cyclist to point M is D81=17m, and the cyclist's driving time is t81=17 / 4.167=4.08s. The initial position of the test vehicle SV7 is D82=VC8*t81=17.222*4.08=70.26m away from point M. The working condition setting is shown in the figure of working condition 8. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV7 and point M is D83, D83 / VC8=ST+TTC8, ST is a fixed value, and TTC8=D83 / VC8-ST.
[0105] Working condition 9: CSFA working condition, the test vehicle SV8 is assigned a speed VC9 of 17.222m / s (60+2km / h), and the scooter rider is assigned a speed of 5.556m / s (20km / h). It is assumed that the test vehicle SV8 and the scooter rider are in contact at point M. The distance from the scooter rider to point M is D91=22m, and the scooter rider's driving time is t91=22 / 5.556=3.96s. The initial position of the test vehicle SV8 is D92=VC9*t91=17.222*3.96=68.2m away from point M. The working condition setting is shown in the figure of working condition 9. The camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0. The distance between the test vehicle SV8 and point M is D93, D93 / VC9=ST+TTC9, ST is a fixed value, and TTC9=D93 / VC9-ST.
[0106] Working condition 10: CSFtap working condition, the test vehicle SV9 is assigned a speed VC10 of 4.722m / s (15+2km / h) and drives forward along the test path BB shown in working condition 10, and turns left along the track. At this time, VC10 is divided into a variable fraction VC10 on the A3A3 axis. The scooter rider is assigned a speed of 5.556m / s (20km / h). It is assumed that the test vehicle SV9 and the scooter rider are in contact at point M. The distance from the scooter rider to point M is D101=35m, and the scooter rider's driving time is t91=35 / 5.556=6.3s. The distance from the initial position of the test vehicle SV9 to point M D102 needs to be deduced according to working condition 10 and Table 3. For specific working condition settings, see working condition Fig.10 As shown, the camera first monitors 50% of the body characteristics of the pedestrian dummy model at time t0, the distance between the test vehicle SV9 and point M is D103, and the time it takes for the test vehicle to travel to point M is recorded as ST+TTC10. ST is a fixed value, and the value of TTC10 can be calculated.
[0107] 8. Summary table of 10 simulation test conditions TTC1~TTC10;
[0108] 9. The result output provides the controller with the basic TTC value of each working condition to support algorithm development.
[0109] like Figure 3-Figure 11 The figure shows the schematic diagram of working condition 1 to working condition 10. The letter numbers in the figure are as follows: AA-pedestrian motion trajectory; A1A1-test vehicle center axis; A2A2-bicyclist motion trajectory, test vehicle center axis; A3A3-scooter rider motion trajectory; BB-test vehicle center axis; B1B1-pedestrian center axis; CC-barrier vehicle A / B center axis; E-distance from pedestrian starting point to test vehicle center axis; E1-bicyclist stable state distance; E2-scooter stable state distance; I-distance from barrier vehicle A / C to pedestrian / The distance between obstacle vehicle A and obstacle vehicle B; I1-the distance between the center axis of the test vehicle and the center axis of the obstacle vehicle; J-the distance between the test vehicle and the larger obstacle vehicle A / B; J1-the distance between the obstacle vehicle and the pedestrian (changes with the test vehicle speed); point M-collision position; F-the distance between the center axis of the test vehicle and the center axis of the pedestrian; K-the distance between obstacle vehicles B and C; H-25% overlap rate collision point offset; L-the distance between the test vehicle and the scooter rider; N-the zero point of the turning trajectory coordinate of the test vehicle; dot-dash line-camera field of view angle; R-the initial position of the pedestrian.
[0110] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims
1. A method for developing an active and passive fusion pedestrian protection system based on ADAS perception, characterized in that: The following steps are involved: A. Set the ADAS signal solution test conditions used by the active hood system; B. Establish the test vehicle model. The non-turning test vehicle settings are as follows: B1. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data; B2. The camera of the test vehicle, including the camera data, including the field of view verification data, needs to be divided into grids at the same time; B3. Define all meshes, materials and properties; C. Establish the test vehicle model. The turning condition test vehicle settings are as follows: C1. The test vehicle adopts a simplified modeling method, retaining only one layer of test vehicle outer surface data or vehicle CAS data; C2, the camera of the test vehicle, where the camera data, including the field of view verification data, needs to be divided into grids at the same time; C3. Define all meshes, materials and properties; C4, modeling includes creating suspension hard point locations, components, elastic elements and bushings, hinges; C5. Establish 12 connection points between the suspension and the chassis, B1-B12; C6, B1-B12 connection points are connected to the vehicle body; D. Establish an obstacle vehicle model using a simplified modeling method. The specific method is consistent with the test vehicle model in step B, and the vehicle size is consistent with the size of the obstacle vehicle in the actual working condition; E. Establish pedestrian dummy models, including adult pedestrian dummy models and child pedestrian dummy models. The sizes of pedestrian dummies should be consistent with those of pedestrian dummies in actual working conditions. F. Establish rider models, including bicycle rider models and electric scooter rider models. The rider model sizes should be consistent with the rider sizes in the actual working conditions. G. Simulation test condition analysis, define pedestrian feature recognition standards to determine t0, establish simulation test conditions 1-10 according to test scenario conditions 1-10, and calculate the TTC values of the 10 conditions; TTC is the remaining time before the test vehicle hits the test target; set the time when the camera first detects 50% of the body features of the pedestrian dummy model / cyclist model as t0; H. Output results, providing the controller with basic TTC values for each operating condition.
2. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: Step A, the ADAS signal solution test conditions include 5 test conditions and 5 extended conditions, conditions 1-3 are cross-driving of people and vehicles, and there are obstacle vehicles; Condition 4 is when people and vehicles are driving in opposite directions; Condition 5 is when people and vehicles are driving in an intersecting manner, and there are vehicles with obstacles in both directions; Working condition 6 is when people and vehicles are driving at the same time, without any obstacles; working condition 7 is when cyclists and vehicles are driving in opposite directions; working condition 8 is when cyclists and vehicles are driving at the same time, without any obstacles; working condition 9 is when scooter riders and vehicles are driving at the same time, without any obstacles; working condition 10 is when scooter riders are driving straight and vehicles are turning left, without any obstacles.
3. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: In step B3, all meshes need to be defined as 1 part, the material definition is rigid material MAT20, and the attribute definition is thickness 3mm.
4. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: Step C4, the hard point part is established according to the hard point information of the suspension; the various parts of the suspension are simplified into rigid beam units, and the various components of the rigid beam unit are simulated using the MAT20 material card of LS-DYNA, and the *Part_Inertia keyword is used to give the components real mass, center of mass position, moment of inertia and inertia product; the suspension spring, shock absorber and limit buffer block are modeled using the spring unit in DYNA, and the spring stiffness and shock absorber damping can be set; the spring pre-compression amount is set by the DYNA keyword *Element_Discrete, and the spring material uses the DYNA MAT S04 material model to simulate the nonlinear characteristics of the spring; the shock absorber uses the DYNA MAT S05 material model to simulate the damping characteristics of the shock absorber; the upper and lower buffer blocks of the shock absorber are also modeled by the DYNA spring unit, and the working clearance of the buffer block is set by the keyword *Section_Discrete; the bushing is modeled using the zero-length beam unit, and the MAT Material model No. 67 can simulate the nonlinear elastic or viscoelastic properties of rubber materials; the bushing unit can set the stiffness and damping of the bushing in six directions through the local coordinate system; the hinge connection, the *Constrained_Joint keyword in DYNA can be used to set various kinematic pairs; beam units are established between the inner points of the steering rod to replace the simplified steering system, and a translational pair is established between the simplified steering system and the subframe to simulate the movement of the steering gear rack when the vehicle turns, and the rack stroke can be calculated through the transmission ratio of the steering wheel angle and the steering line angle.
5. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: In step C5, each connection point is characterized by a rigid point.
6. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: Step C6, B1-B12 connection points are connected to the vehicle body using the *extra_nodes method.
7. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 1, characterized in that: The camera fully recognizes the physical features of the pedestrian dummy model / cyclist model as t1, t1-t0=TS.
8. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 7, characterized in that: TS is defined as the length of time required for the camera to recognize pedestrian features. The moment when the controller sends out the active hood lifter trigger command is t2, t2-t0=ST. ST is defined as the perception time required for the ADAS signal perception solution, which is a fixed value and is the sensing time of the active hood system.
9. The method for developing an active and passive fusion pedestrian protection system based on ADAS perception according to claim 8, characterized in that: The contact moment between the pedestrian dummy model / cyclist model and the test vehicle is t3, and t3-t2=TTC is defined. The TTC value for each working condition is different. The lower limit speed for detonation is set to 25km / h.
