An emergency obstacle avoidance system and control method based on characteristic human safety and comfort

Abstract

The present invention discloses an emergency obstacle avoidance system and method based on characteristic human body safety and comfort. The emergency obstacle avoidance system comprises a characteristic human body damage prediction module for predicting multi-working condition damage for characteristic human body, a multi-objective optimization module for comprehensively processing multi-source information under vehicle driving state and establishing an emergency obstacle avoidance system constraint model, and an intelligent obstacle avoidance decision module for performing path planning when the vehicle encounters a situation requiring emergency obstacle avoidance. The present invention can fully guarantee the safety and comfort of human body in emergency obstacle avoidance by performing safety and comfort constraints on human body with different physical signs, and the obstacle avoidance method proposed by the present invention can perform adaptive transformation according to the physical sign parameter characteristics of human body with different physical signs, and has stronger robustness.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle active safety, and in particular to an emergency obstacle avoidance system and a control method based on characteristic human safety and comfort. Background Art

[0002] At present, the assisted driving systems of intelligent vehicles for path planning problems often set the initial point and the target point, and then implement the processes such as following the vehicle and changing lanes based on a predetermined algorithm. This method makes path planning decisions based on the status of the vehicle itself and the surrounding vehicles, ignoring the safety and comfort of the occupants themselves. Moreover, this method often cannot make accurate lane keeping and lane changing decisions when the vehicle encounters emergency obstacle avoidance needs. In addition, the protection of occupants is often tested by placing dummies and digital human models in the cockpit. This method can only test specific human bodies and cannot reflect the safety and comfort of human bodies with multiple physical signs. Summary of the invention

[0003] The purpose of the present invention is to provide an emergency obstacle avoidance system and control method based on characteristic human safety and comfort, so as to solve the problem that the existing method can only test a specific human body and cannot reflect the safety and comfort of multiple signs of the human body.

[0004] The above technical objectives of the present invention are achieved through the following technical solutions:

[0005] An emergency obstacle avoidance system based on characteristic human body safety and comfort, the emergency obstacle avoidance system comprising a characteristic human body injury prediction module, a multi-objective optimization module and an intelligent obstacle avoidance decision module;

[0006] The characteristic human body injury prediction module is used to perform multi-condition injury prediction for the characteristic human body;

[0007] The multi-objective optimization module is used to comprehensively process multi-source information under the vehicle driving state and establish a constraint model for the emergency obstacle avoidance system;

[0008] The intelligent obstacle avoidance decision module is used to plan a path when the vehicle encounters an emergency obstacle avoidance situation.

[0009] Further preferably, the characteristic human body injury prediction module is established by the following steps:

[0010] Step 1.1, input different human body sign parameters into the human body statistics database to obtain human body surface point clouds under different physical sign parameters, wherein the human body sign parameters include gender, age, height, weight, etc.;

[0011] Step 1.2, based on the acquired human body surface point cloud, generate a human body finite element model with different physical characteristics through RBF-TPS mesh deformation technology;

[0012] Step 1.3, the human body finite element model is coupled with the vehicle model, and DOE simulation is performed on front collision, rear collision, side collision and other working conditions under different accelerations to obtain human body damage parameters of different working conditions under different human body characteristics. The damage parameters are the maximum compression ratios of the human head, upper limbs, chest cavity and lower limbs. The damage parameters are shown in the following formula:

[0013]

[0014] In the formula, C N is the maximum compression ratio, D I is the initial position of the measuring point, D D is the deformation position of the measuring point;

[0015] Step 1.4, combine the injury parameters at different positions according to the following formula to obtain the human body injury level;

[0016] C=k 1 c 1 +k 2 c 2 +k 3 c 3 +k 4 c 4

[0017] Where C is the level of human injury, c 1 、c 2 、c 3 、c 4 are the maximum compression ratios of the human head, upper limbs, chest cavity, and lower limbs, respectively, 1 , k 2 , k 3 , k 4 are weight coefficients respectively;

[0018] Step 1.5: Taking human body characteristic parameters, vehicle speed, and collision form as input and human body injury level as output, a neural network model is established and trained to obtain a characteristic human body injury prediction model.

[0019] Further preferably, the multi-objective optimization module is established by the following steps:

[0020] Step 2.1: Use the vehicle-mounted millimeter-wave radar to obtain the real-time location information of the vehicle itself and the coordinate information of the nearest vehicles in the vicinity, and establish a safe location constraint, as shown in the following formula:

[0021]

[0022] In the formula, x e With y e are the lateral and longitudinal information of the vehicle, respectively,i With y i are the lateral and longitudinal information of the surrounding i-th vehicle respectively;

[0023] Step 2.2, arrange pressure measuring pads on the seats in the cockpit to measure the body pressure distribution of the human body, select the human buttocks pressure, thigh pressure, back pressure, and backrest pressure distribution uniformity as comfort indicators, and establish a comfort model;

[0024] The comfort model is shown as follows:

[0025] S=α 1 p 1 +α 2 p 2 +α 3 p 3 +α 4 p 4

[0026] Where S is the overall comfort level of the human body, p 1 , p 2 , p 3 , p 4 They are the pressure on the buttocks, thighs, back and backrest pressure distribution uniformity, α 1 , α 2 , α 3 , α 4 are weight coefficients respectively;

[0027] The comfort constraint model is established as shown below:

[0028] Max S

[0029] Step 2.3: Use the on-board millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, and calculate the time t when the two vehicles collide. max , and then establish the time window constraint, as shown in the following formula:

[0030] 0≤t≤t max

[0031] In the formula, t is the time after the emergency obstacle avoidance system is activated;

[0032] Step 2.4: Based on the human body injury level obtained in step 1.4, establish an injury constraint model as shown in the following formula.

[0033] Min C

[0034] Further preferably, the intelligent obstacle avoidance decision module is established by the following steps:

[0035] Step 3.1, parameter initialization, first convert the coordinate nodes in the traditional ant colony algorithm into braking torque and steering wheel angle, then define the maximum number of iterations as 50 and the number of ants as 20;

[0036] Step 3.2, establish an obstacle avoidance decision diagram, with the ordinate being the braking torque and the abscissa being the steering wheel angle. Define the target point as the position where the braking torque is 0 and the steering wheel angle is 0, and the ant starting position is the intersection of the current vehicle braking torque and the steering wheel angle;

[0037] Step 3.3: Create a node selection table Tab N , used to record the unvisited nodes in the obstacle avoidance decision graph, and to establish a node record table Tab Y ;

[0038] Step 3.4, traverse Tab N The nodes in the ant colony algorithm are calculated and selected according to the ant colony algorithm model formula. i ;

[0039] Step 3.5: Determine node N i Check whether the safe position constraint, comfort constraint, time window constraint and damage constraint established in step 2 are satisfied. If so, execute step 3.6. If not, repeat step 3.4.

[0040] Step 3.6: N i Record to Tab Y , while updating Tab N , and continue to select nodes in the obstacle avoidance decision graph until the target point is reached;

[0041] Step 3.7, update the pheromone between each node according to the pheromone update formula of the ant colony algorithm;

[0042] Step 3.8: Determine whether the maximum number of iterations has been reached. If so, output the final decision path for the steering wheel angle and braking torque. If not, repeat step 3.2.

[0043] A control method for an emergency obstacle avoidance system based on characteristic human safety and comfort, specifically comprising the following steps:

[0044] Step 4.1, take the RGB image of the passenger through the vehicle camera, and then use Vggnet to identify the characteristics of the passenger in the image;

[0045] Step 4.2, obtain the real-time vehicle speed through the vehicle-mounted sensor, and input the occupant characteristics, vehicle speed, and collision form into the characteristic human injury prediction model, output the human injury level, and input it into the multi-objective optimization module to establish the damage constraint;

[0046] Step 4.3, obtain the body pressure distribution of the human body through the pressure cushion, and input it into the multi-objective optimization module to establish the comfort constraint;

[0047] Step 4.4, obtain the position information of surrounding vehicles through the vehicle-mounted millimeter-wave radar, input it into the multi-objective optimization module, and establish a safe position constraint;

[0048] Step 4.5: Use the on-board millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, input them into the multi-objective optimization module, and establish a time window constraint;

[0049] Step 4.6: Input the damage constraint, comfort constraint, safe position constraint, and time window constraint into the intelligent obstacle avoidance decision module to obtain the braking torque and steering wheel angle decision route.

[0050] In summary, the present invention has the following beneficial effects:

[0051] First, the present invention can fully guarantee the safety and comfort of the human body during emergency obstacle avoidance by imposing safety and comfort constraints on the human body with different physical signs. In addition, the obstacle avoidance method proposed in the present invention can be adaptively transformed according to the physical parameter characteristics of the human body with different physical signs, and is more robust.

[0052] Secondly, the obstacle avoidance decision algorithm proposed in the present invention constructs an obstacle avoidance decision diagram based on the steering wheel angle and braking torque, thereby achieving accurate control of the vehicle in emergency situations. The architecture is highly lightweight and is constrained based on human safety and comfort, so the control effect is more precise.

[0053] Third, the characteristic human injury prediction module proposed in the present invention can be adaptively updated according to the real-time vehicle speed, combined with the real-time time window and the surrounding vehicle information update, so as to realize the continuous updating of the steering wheel angle and braking torque decision route in an emergency situation, thus enhancing safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of the architecture of the emergency obstacle avoidance system based on characteristic human safety and comfort in the present invention;

[0055] Figure 2 It is a schematic diagram of the construction of the intelligent obstacle avoidance decision module in the present invention;

[0056] Figure 3 It is the obstacle avoidance decision diagram of the steering wheel angle and braking torque in the present invention. DETAILED DESCRIPTION

[0057] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0058] Embodiment, an emergency obstacle avoidance system based on characteristic human body safety and comfort, the emergency obstacle avoidance system comprises a characteristic human body injury prediction module, a multi-objective optimization module and an intelligent obstacle avoidance decision module, such as Figure 1 shown.

[0059] The characteristic human body damage prediction module is used to predict multi-condition damage for the characteristic human body. The module is established through the following steps:

[0060] Step 1.1, input different human body sign parameters into the human body statistics database to obtain human body surface point clouds under different physical sign parameters, wherein the human body sign parameters include gender, age, height, weight, etc.;

[0061] Step 1.2, based on the acquired human body surface point cloud, generate a human body finite element model with different physical characteristics through RBF-TPS mesh deformation technology;

[0062] Step 1.3, the human body finite element model is coupled with the vehicle model, and DOE simulation is performed on front collision, rear collision, side collision and other working conditions under different accelerations to obtain human body damage parameters of different working conditions under different human body characteristics. The damage parameters are the maximum compression ratios of the human head, upper limbs, chest cavity and lower limbs. The damage parameters are shown in the following formula:

[0063]

[0064] In the formula, C N is the maximum compression ratio, D I is the initial position of the measuring point, D D is the deformation position of the measuring point;

[0065] Step 1.4, combine the injury parameters at different positions according to the following formula to obtain the human body injury level;

[0066] C=k 1 c 1 +k 2 c 2 +k 3 c 3 +k 4 c 4

[0067] Where C is the human body injury level, c 1 、c 2 、c 3 、c 4 are the maximum compression ratios of the human head, upper limbs, chest cavity, and lower limbs, respectively, 1 , k 2 , k 3 , k 4 are weight coefficients respectively;

[0068] Step 1.5: Taking human body characteristic parameters, vehicle speed, and collision form as input and human body injury level as output, a neural network model is established and trained to obtain a characteristic human body injury prediction model.

[0069] The multi-objective optimization module is used to comprehensively process the multi-source information under the vehicle driving state and establish the constraint model of the emergency obstacle avoidance system. The module is established through the following steps:

[0070] Step 2.1: Use the vehicle-mounted millimeter-wave radar to obtain the real-time location information of the vehicle itself and the coordinate information of the nearest vehicles in the vicinity, and establish a safe location constraint, as shown in the following formula:

[0071]

[0072] In the formula, x e With y e are the lateral and longitudinal information of the vehicle, respectively, i With y i are the lateral and longitudinal information of the surrounding i-th vehicle respectively;

[0073] Step 2.2, arrange pressure measuring pads on the seats in the cockpit to measure the body pressure distribution of the human body, select the human buttocks pressure, thigh pressure, back pressure, and backrest pressure distribution uniformity as comfort indicators, and establish a comfort model;

[0074] The comfort model is shown as follows:

[0075] S=α 1 p 1 +α 2 p 2 +α 3 p 3 +α 4 p 4

[0076] Where S is the overall comfort level of the human body, p 1 , p 2 , p 3 , p 4 They are the pressure on the buttocks, thighs, back and backrest pressure distribution uniformity, α 1 , α 2 , α 3 , α 4 are weight coefficients respectively;

[0077] The comfort constraint model is established as shown below:

[0078] MaxS

[0079] Step 2.3: Use the on-board millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, and calculate the time t when the two vehicles collide. max Then, the time window constraint is established as shown in the following formula:

[0080] 0≤t≤t max

[0081] In the formula, t is the time after the emergency obstacle avoidance system is activated;

[0082] Step 2.4: Based on the human body injury level obtained in step 1.4, establish an injury constraint model as shown in the following formula.

[0083] Min C

[0084] The intelligent obstacle avoidance decision module is used to plan the path when the vehicle encounters an emergency obstacle avoidance situation. This module is based on the variation of the ant colony algorithm. The process is as follows Figure 2 shown.

[0085] Step 3.1, parameter initialization, first convert the coordinate nodes in the traditional ant colony algorithm into braking torque and steering wheel angle, then define the maximum number of iterations as 50 and the number of ants as 20;

[0086] Step 3.2: Create an obstacle avoidance decision diagram, with the ordinate being the braking torque and the abscissa being the steering wheel angle. Define the target point as the position where the braking torque is 0 and the steering wheel angle is 0, such as Figure 3 As shown, the starting position of the ant is the intersection of the current vehicle braking torque and the steering wheel angle;

[0087] Step 3.3: Create a node selection table tab N , used to record the unvisited nodes in the obstacle avoidance decision graph, and to establish a node record table Tab Y ;

[0088] Step 3.4, traverse Tab N The nodes in the ant colony algorithm are calculated and selected according to the ant colony algorithm model formula. i ;

[0089] Step 3.5: Determine node N i Check whether the safe position constraint, comfort constraint, time window constraint and damage constraint established in step 2 are satisfied. If so, execute step 3.6. If not, repeat step 3.4.

[0090] Step 3.6: N i Record to Tab Y , while updating Tab N , and continue to select nodes in the obstacle avoidance decision graph until the target point is reached;

[0091] Step 3.7, update the pheromone between each node according to the pheromone update formula of the ant colony algorithm;

[0092] Step 3.8: Determine whether the maximum number of iterations has been reached. If so, output the final decision path for the steering wheel angle and braking torque. If not, repeat step 3.2.

[0093] Furthermore, the present invention also provides a control method for the emergency obstacle avoidance system, such as Figure 1 As shown, the following steps are included:

[0094] Step 4.1, take the RGB image of the passenger through the vehicle camera, and then use Vgg net to identify the characteristics of the passenger in the image;

[0095] Step 4.2, obtain the real-time vehicle speed through the vehicle-mounted sensor, and input the occupant characteristics, vehicle speed, and collision form into the characteristic human injury prediction model, output the human injury level, and input it into the multi-objective optimization module to establish the damage constraint;

[0096] Step 4.3, obtain the body pressure distribution of the human body through the pressure cushion, and input it into the multi-objective optimization module to establish the comfort constraint;

[0097] Step 4.4, obtain the position information of surrounding vehicles through the vehicle-mounted millimeter-wave radar, input it into the multi-objective optimization module, and establish a safe position constraint;

[0098] Step 4.5: Use the on-board millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, input them into the multi-objective optimization module, and establish a time window constraint;

[0099] Step 4.6: Input the damage constraint, comfort constraint, safe position constraint, and time window constraint into the intelligent obstacle avoidance decision module to obtain the braking torque and steering wheel angle decision route.

[0100] This specific embodiment is merely an explanation of the present invention and is not a limitation of the present invention. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed. However, as long as they are within the scope of the claims of the present invention, they are protected by the patent law.

Claims

1. An emergency obstacle avoidance system based on characteristic human safety and comfort, characterized by: The emergency obstacle avoidance system includes a characteristic human injury prediction module, a multi-objective optimization module and an intelligent obstacle avoidance decision module; The characteristic human body injury prediction module is used to perform multi-condition injury prediction for the characteristic human body; The multi-objective optimization module is used to comprehensively process multi-source information under the vehicle driving state and establish a constraint model for the emergency obstacle avoidance system; The intelligent obstacle avoidance decision module is used to plan a path when the vehicle encounters an emergency obstacle avoidance situation; The characteristic human body injury prediction module is established by the following steps: Step 1.1, input different human body sign parameters into the human body statistics database to obtain human body surface point clouds under different physical sign parameters, wherein the human body sign parameters include gender, age, height, and weight; Step 1.2, based on the acquired human body surface point cloud, generate a human body finite element model with different physical characteristics through RBF-TPS mesh deformation technology; Step 1.3, the human body finite element model is coupled with the vehicle model, and DOE simulation is performed on the front collision, rear collision and side collision working conditions under different accelerations to obtain the human body damage parameters of different working conditions under different human body characteristics. The damage parameters are the maximum compression ratios of the human head, upper limbs, chest cavity and lower limbs. The damage parameters are shown in the following formula: In the formula, C N is the maximum compression ratio, D I is the initial position of the measuring point, D D is the deformation position of the measuring point; Step 1.4, combine the injury parameters at different positions according to the following formula to obtain the human body injury level; C=k1c1+k2c2+k3c3+k4c4 Where C is the human body injury level, c1, c2, c3, c4 are the maximum compression ratios of the human head, upper limbs, chest, and lower limbs, respectively, and k1, k2, k3, k4 are weight coefficients, respectively; Step 1.5, using human body characteristic parameters, vehicle speed, and collision form as inputs and human body injury level as output, a neural network model is established, and the model is trained to obtain a characteristic human body injury prediction model; The multi-objective optimization module is established by the following steps: Step 2.1: Use the vehicle-mounted millimeter-wave radar to obtain the real-time location information of the vehicle and the coordinate information of the nearest vehicles in the vicinity, and establish a safe location constraint, as shown in the following formula: In the formula, x e With y e are the lateral and longitudinal information of the vehicle, respectively, i With y i are the lateral and longitudinal information of the surrounding i-th vehicle respectively; Step 2.2, arrange pressure measuring pads on the seats in the cockpit to measure the body pressure distribution of the human body, select the human buttocks pressure, thigh pressure, back pressure, and backrest pressure distribution uniformity as comfort indicators, and establish a comfort model; The comfort model is shown as follows: S=α1p1+α2p2+α3p3+α4p4 Where S is the overall comfort level of the human body, p1, p2, p3, p4 are the uniformity of the pressure distribution of the buttocks, thighs, back and backrest respectively, and α1, α2, α3, α4 are the weight coefficients respectively; The comfort constraint model is established as shown below: Max S Step 2.3: Use the on-board millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, and calculate the time t when the two vehicles collide. max , and then establish the time window constraint, as shown in the following formula: 0≤t≤t max Where, t is the time after the emergency obstacle avoidance system is activated; Step 2.4: Based on the human body injury level obtained in step 1.4, establish an injury constraint model as shown in the following formula: Min C The intelligent obstacle avoidance decision module is established by the following steps: Step 3.1, parameter initialization, first convert the coordinate nodes in the traditional ant colony algorithm into braking torque and steering wheel angle, then define the maximum number of iterations as 50 and the number of ants as 20; Step 3.2, establish an obstacle avoidance decision diagram, with the ordinate being the braking torque and the abscissa being the steering wheel angle. Define the target point as the position where the braking torque is 0 and the steering wheel angle is 0, and the ant starting position is the intersection of the current vehicle braking torque and the steering wheel angle; Step 3.3: Create a node selection table Tab N , used to record the unvisited nodes in the obstacle avoidance decision graph, and to establish a node record table Tab Y ; Step 3.4, traverse Tab N The nodes in the ant colony algorithm are calculated and selected according to the ant colony algorithm model formula. i ; Step 3.5: Determine node N i Check whether the safe position constraint, comfort constraint, time window constraint and damage constraint established in step 2 are satisfied. If so, execute step 3.

6. If not, repeat step 3.

4. Step 3.6: N i Record to Tab Y , while updating Tab N , and continue to select nodes in the obstacle avoidance decision graph until the target point is reached; Step 3.7, update the pheromone between each node according to the pheromone update formula of the ant colony algorithm; Step 3.8: Determine whether the maximum number of iterations has been reached. If so, output the final decision path for the steering wheel angle and braking torque. If not, repeat step 3.

2.

2. The control method of the emergency obstacle avoidance system based on characteristic human safety and comfort according to claim 1, characterized in that: The specific steps include: Step 4.1, take the RGB image of the passenger through the vehicle camera, and then identify the passenger features in the image through Vgg net; Step 4.2, obtain the real-time vehicle speed through the vehicle-mounted sensor, and input the occupant characteristics, vehicle speed, and collision form into the characteristic human injury prediction model, output the human injury level, and input it into the multi-objective optimization module to establish the damage constraint; Step 4.3, obtain the body pressure distribution of the human body through the pressure cushion, and input it into the multi-objective optimization module to establish the comfort constraint; Step 4.4, obtain the position information of surrounding vehicles through the vehicle-mounted millimeter-wave radar, input it into the multi-objective optimization module, and establish a safe position constraint; Step 4.5: Use the vehicle-mounted millimeter-wave radar to test the real-time relative speed and relative distance between the vehicle and the nearest vehicle in front, input them into the multi-objective optimization module, and establish a time window constraint; Step 4.6: Input the damage constraint, comfort constraint, safe position constraint, and time window constraint into the intelligent obstacle avoidance decision module to obtain the braking torque and steering wheel angle decision route.

Images