A longitudinal and lateral AEB collision avoidance method combining steering with visual impairment scenes

Abstract

The present invention proposes a longitudinal and lateral AEB collision avoidance method combined with steering in a visually impaired scene, including: step 1: detecting the speed information of other vehicles in adjacent lanes through sensors, and determining the potential collision risk within the blind spot of the main vehicle based on the speed information; step 2: when the acceleration of the adjacent lane vehicle is less than zero, taking measures to eliminate the braking gap to reduce the braking reaction time; step 3: obtaining the relevant motion information of the main vehicle and the vulnerable road user VRU through sensors, calculating the time TTC of the main vehicle and the VRU arriving at the intersection of the trajectories, and using the difference in TTC between the two to determine whether there is a collision risk; step 4: if the longitudinal and lateral TTC difference is greater than or equal to and less than or equal to, executing collision avoidance planning; step 5: according to the cost function calculation, using the optimization algorithm to find the optimal braking deceleration and steering angle to avoid collision. The present invention proposes a more effective collision avoidance strategy to reduce the risk of collision between the vehicle and the VRU driving out of the blind spot in a visually impaired scene.

Description

Technical Field

[0001] The present invention belongs to the technical field of automobile active safety, and in particular relates to a longitudinal and lateral AEB collision avoidance method combining visually impaired scenes with steering. Background Art

[0002] With the continuous development of automobile technology, the systems used for autonomous automobile navigation are becoming increasingly mature. These systems use sensors to determine the location and driving direction of the vehicle, use mathematical analysis methods to establish the driving path, and compare it with the roads on the electronic map to determine the direction and remaining distance of the vehicle to the destination, thereby playing the role of navigation and guidance. In this process, when other vehicles or pedestrians appear on the driving route, the car can detect the relative distance, relative speed, and time to collide with other vehicles or pedestrians on the road through laser radar, cameras, sensors, etc., and use the detected driving conditions to trigger constraints, such as limiting the maximum driving speed of the main vehicle, the maximum distance to other vehicles or pedestrians, etc., and avoid obstacles through automatic emergency braking, steering and lane changing, etc., to ensure that the vehicle reaches the desired destination safely and accurately.

[0003] Although most cars are currently equipped with automatic emergency braking (AEB) systems, based on a safety distance model, which compares the detected distance to obstacles with the warning distance and safety distance, and issues an alarm when the distance is less than the warning distance, the AEB system initiates automatic emergency braking when the distance is less than the safety distance, thus providing protection for safe driving of the car. However, in some complex road conditions, especially when the main vehicle has visual impairments, this technology may not fully consider the dynamic interaction between the vehicle and vulnerable road users (Vulnerable Road Users, VRU), which in turn makes it impossible to accurately predict the potential risk of collision between the car and the VRU.

[0004] In addition, past technologies have failed to effectively combine the vehicle's steering conditions. When the main vehicle has visual impairments and faces the sudden appearance of a VRU, it is difficult to avoid a collision only through the triggering strategy of the traditional AEB system. Therefore, there is an urgent need to propose a more effective collision avoidance strategy to reduce the risk of a vehicle colliding with a VRU that drives out of the blind spot in a visual impairment scenario. Summary of the invention

[0005] The main purpose of the present invention is to provide a longitudinal and lateral AEB collision avoidance method combined with steering in visually impaired scenarios, aiming to reduce the risk of a car colliding with a vulnerable road user (VRU) in a visually impaired situation, thereby enhancing the reliability of the car's AEB system in avoiding collisions.

[0006] According to one aspect of the present invention, a longitudinal and lateral AEB collision avoidance method in a visually impaired scene combined with steering is provided, the method comprising the following steps:

[0007] Step 1: Detect the speed information of other vehicles in the adjacent lanes through sensors, and determine the potential collision risk within the blind spot of the main vehicle according to the acceleration changes of the vehicles in the adjacent lanes;

[0008] Step 2: When the acceleration of the adjacent lane vehicle is less than zero, taking measures to eliminate the braking gap to reduce the braking reaction time;

[0009] Step 3: Obtain the relevant motion information of the main vehicle and VRU through sensors, calculate the time TTC when the main vehicle and VRU reach the intersection of the trajectories, and use the difference between the two TTCs to determine whether there is a collision risk. Among them, the main vehicle travels along the X-axis direction, and the VRU travels along the Y-axis direction. (x0, y0) is the intersection of the trajectories of the main vehicle and the VRU. The moment when the VRU first appears in the main vehicle's field of vision after the blind spot is recorded as the initial moment. At this time, the main vehicle speed is detected to be v x , VRU speed is v y , the distance between them is s, and the azimuth is θ. The calculation formula of the longitudinal and transverse TTC difference is:

[0010] ΔTTC=TTC x -TTC y

[0011]

[0012]

[0013] Where ΔTTC is the difference between the longitudinal and transverse TTC, which indicates the time difference between the main vehicle and the VRU reaching the intersection of the trajectories along the longitudinal direction and the transverse direction. x TTC is the time it takes for the vehicle to reach the intersection of the track on the longitudinal X axis. y It is the time when the VRU reaches the track intersection point on the horizontal Y axis.

[0014] Step 4: If the longitudinal and transverse TTC difference is greater than or equal to ΔTTC min and less than or equal to ΔTTC max , there is a risk of collision, and collision avoidance planning is performed;

[0015] Step 5: Based on the cost function calculation, an optimization algorithm is used to find the optimal braking deceleration and steering angle to avoid collision.

[0016] According to another aspect of the present invention, step 5 includes braking at a deceleration of a, Turn and avoid at an angle. Establish a cost function to reflect the mapping relationship between the collision avoidance planning cost and the relative distance between the two targets, the deceleration provided when the AEB function intervenes, and the steering angle. The calculation formula of the cost function is as follows:

[0017]

[0018] In the formula, cost function is the collision avoidance planning cost, f(s) is the safety cost, f(a) is the acceleration cost, is the steering cost, s is the relative distance between the vehicle and the VRU, a is the braking deceleration provided by the AEB system, is the vehicle steering angle, and α i is the weight coefficient of each cost.

[0019] According to another aspect of the present invention, the definition of the safety cost is as follows:

[0020]

[0021] where s is the relative distance between the vehicle and the VRU, [s0, s n is the distance threshold for whether to perform collision avoidance planning. Within the threshold range, the safety cost is determined by the attenuation function g(s).

[0022] According to another aspect of the present invention, the attenuation function can be an exponential attenuation function, a linear function, a reciprocal attenuation function, etc.

[0023] According to another aspect of the present invention, step 5 further includes detecting the relative distance s between the host vehicle and the VRU through the perception system. When s0 ≤ s ≤ s n , perform collision avoidance planning, substitute the distance s into the cost function to solve, and use the optimization algorithm to find the optimal braking deceleration a and steering angle Solve for the minimum value of cost function, which is the optimal route in the collision avoidance planning.

[0024] According to another aspect of the present invention, the optimization algorithm can be the gradient descent method or the genetic algorithm.

[0025] According to another aspect of the present invention, step 5 further includes that when s < s0 or the minimum value cannot be obtained, it means that the collision is inevitable. The maximum friction coefficient of the urban road surface is 0.8. At this time, brake with a braking deceleration of a max = 0.8g to reduce the damage caused by the collision between the vehicle and the VRU.

[0026] According to another aspect of the present invention, step 5 further includes that when s > s n , this cost is 0, indicating that the collision will not occur.

[0027] According to another aspect of the present invention, the collision avoidance method is applied to the fields of intelligent driving and automatic driving.

[0028] The present invention aims to reduce the risk of a car colliding with a vulnerable road user (VRU) in a visually impaired scene, and enhances the reliability of the car AEB system in avoiding collisions through a longitudinal and lateral AEB (automatic emergency braking) collision avoidance method combined with steering. The innovation of the present invention lies in the integration of visually impaired scenes and steering considerations, and the use of longitudinal and lateral TTC differences to determine the risk of collision. In addition, when there is a risk of collision, the optimal braking deceleration and steering angle are calculated by a cost function, thereby providing a method for avoiding collisions in visually impaired scenes. This method can improve the protection performance of the car for vulnerable road users in visually impaired scenes, and enhance the reliability and practicality of the AEB system. It is easy to understand that in addition to being applied to the fields of intelligent driving and automatic driving, in specific scenarios, the collision avoidance method can detect the motion information of surrounding vehicles through sensors, determine the potential collision risk between the car and the VRU, and implement collision avoidance measures. It is also applicable to various scenarios such as intelligent safety systems and environmental perception systems, including but not limited to traffic safety, smart devices, robots, and automation industries, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings, which are incorporated into the specification and constitute a part of the specification, illustrate embodiments of the present invention and are used to explain the principles of the present invention together with the relevant text description. In these drawings, similar reference numerals are used to represent similar elements. The drawings described below are some embodiments of the present invention, but not all embodiments. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A schematic diagram of a visual impairment working condition in the prior art is shown;

[0031] Figure 2 A flow chart of a longitudinal and lateral AEB collision avoidance method combined with steering in a visually impaired scene in one embodiment of the present invention is shown;

[0032] Figure 3 A simplified model of a car and a VRU pre-collision in one embodiment of the present invention is shown;

[0033] Figure 4 ΔTTC in one embodiment of the present invention is shown max Schematic diagram of relative motion condition;

[0034] Figure 5 ΔTTC in one embodiment of the present invention is shownmin Schematic diagram of relative motion condition. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. It should be noted that the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily without conflict.

[0036] In the existing methods, although the automatic emergency braking (AEB) system installed in most cars currently compares the detected distance to obstacles with the warning distance and the safety distance based on the safety distance model, an alarm will be given when the distance is less than the warning distance, and the AEB system will start automatic emergency braking when the distance is less than the safety distance, thus providing protection for the safe driving of the car. However, in some complex road conditions, this technology may not fully consider the dynamic interaction between the vehicle and vulnerable road users (Vulnerable Road Users, VRU), which may lead to the inability to accurately predict the potential collision risk between the car and the VRU. For example, in a scene with visual impairment, such as Figure 1 As shown in the figure, when the vehicle is driving, other vehicles or obstacles on the side road will block the view of the main vehicle. The driving trajectory of the VRU is uncertain and it is easy to appear in the visual blind spot. At this time, the vehicle cannot identify the VRU because the obstacle blocks its view. In this dangerous scenario, the relative distance and collision time (TTC) between the vehicle and the VRU are both small, and there is a greater risk of collision.

[0037] In addition, past technologies have failed to effectively incorporate the vehicle's steering conditions. When a VRU suddenly appears, it is difficult to avoid a collision using only the traditional AEB system's triggering strategy.

[0038] In order to overcome these problems, it is urgent to propose a more effective collision avoidance strategy to reduce the risk of collision between vehicles and VRUs that exit from the blind spot in visually impaired scenarios.

[0039] An embodiment of the present invention provides a longitudinal and lateral AEB collision avoidance method combined with steering in a visually impaired scene, such as Figure 2 As shown, in practical application, the steps of the above method may include:

[0040] S1: Detect the speed information of other vehicles in the adjacent lane through sensors, and determine the potential collision risk within the blind spot of the main vehicle based on the acceleration changes of the vehicles in the adjacent lane.

[0041] In the first step (S1), for visually impaired scenes or complex traffic scenes, detecting the speed information of other vehicles in adjacent lanes through sensors is a key starting point. The purpose of this step is to obtain dynamic information about the surrounding environment of the main vehicle, especially the motion state of vehicles in the adjacent lanes. For example, by analyzing the changes in their speed and acceleration, it is possible to estimate whether there is a potential collision risk in the blind spot of the main vehicle. Among them, the sensor can use, for example, radar, camera or lidar to obtain the speed information of surrounding vehicles. By detecting the acceleration information of vehicles in the adjacent lanes, more insights about the current behavior of the vehicle and the actions to be taken can be provided, so that timely measures can be taken to reduce the probability of collision between the main vehicle and the VRU in the blind spot.

[0042] S2: When the acceleration of the adjacent lane vehicle is less than zero, measures are taken to eliminate the braking gap to reduce the braking reaction time;

[0043] In the second step (S2), when the acceleration of the adjacent vehicle is detected to be less than zero, the system takes measures to eliminate the brake gap in order to reduce the braking reaction time and improve the response efficiency of the AEB system. Specifically, during driving, when the sensor detects that the speed change of the adjacent vehicle is negative, that is, the adjacent vehicle may be slowing down or stopping, it indicates that there is a potential collision risk within the blind spot of the main vehicle, and the system responds immediately.

[0044] In order to reduce the potential risk of collision between this car and VRU, the system takes measures to eliminate the brake clearance. The brake clearance refers to the distance between the brake shoe and the wheel hub. When the brake pedal is pressed or the system brakes, the car will not brake immediately. There is a distance between the brake shoe and the wheel hub, and the system has a reaction time. Eliminating the brake clearance can reduce this reaction time, and the system can brake in time when encountering an emergency later. This not only improves the system's timely responsiveness to potential collision situations, but also enables the system to successfully avoid collisions with VRUs, thereby further improving the performance of the AEB system in visually impaired scenarios, especially when the status of adjacent lane vehicles changes dynamically, the system can make decisions more quickly and accurately. If the adjacent lane vehicle does not warn the main vehicle of danger, the main vehicle continues to drive normally.

[0045] S3: Obtain the relevant motion information of the main vehicle and VRU through sensors, calculate the time TTC when the main vehicle and VRU reach the intersection of the trajectories, and use the difference between the two TTCs to determine whether there is a collision risk. Figure 3The simplified model of the car and VRU pre-collision is shown in the figure, where the main car moves along the X-axis direction and the VRU moves along the Y-axis direction. (x0, y0) is the intersection of the trajectories of the main car and the VRU. The moment when the VRU first appears in the sight range of the main car after the blind spot is recorded as the initial moment. At this time, the main car speed is detected to be v x , VRU speed is v y , the distance between them is s, and the azimuth is θ. The calculation formula of the longitudinal and transverse TTC difference is:

[0046] ΔTTC=TTC x -TTC y

[0047]

[0048]

[0049] Where ΔTTC is the difference between the longitudinal and transverse TTC, which indicates the time difference between the main vehicle and the VRU reaching the intersection of the trajectories along the longitudinal direction and the transverse direction. x TTC is the time it takes for the vehicle to reach the intersection of the track on the longitudinal X axis. y It is the time when the VRU reaches the track intersection point on the horizontal Y axis.

[0050] In the third step (S3), when the vulnerable road user (VRU) drives out of the blind spot, the relevant motion information of the main vehicle and the VRU is obtained through sensors, and the time taken for the two to reach the trajectory intersection (Trajectory Crossing Point) is calculated to assess the risk of collision. Specifically, the system obtains the relevant motion information of the main vehicle and the VRU in real time, including speed, direction, etc., through the sensors it carries, which may be radars, cameras, etc. By analyzing the motion trajectories of the main vehicle and the VRU, the system determines the point where they may intersect in the future, that is, the trajectory intersection point. This is obtained by solving the intersection of the trajectories of the main vehicle and the VRU, represented by (x0, y0). The system calculates the time taken by the main vehicle and the VRU to reach the trajectory intersection in the longitudinal direction (X-axis) and the lateral direction (Y-axis), which are TTC and y0, respectively. x and TTC y This is obtained by considering the current position, speed and other motion information, using kinematics and trajectory prediction methods. The longitudinal and lateral TTC difference calculation formula is used to calculate the longitudinal and lateral TTC difference between the main vehicle and the VRU. This difference reflects the time relationship between the main vehicle and the VRU near the intersection and is the basis for judging whether there is a collision risk.

[0051] By accurately calculating the motion trajectory and time information of the main vehicle and VRU, the system can timely predict whether there is a collision risk. Using the longitudinal and lateral TTC differences as decision-making indicators provides a sensitive judgment of possible collisions in the driving environment, thereby providing accurate input for subsequent collision avoidance planning, helping the system to more intelligently and proactively avoid collisions between cars and VRUs.

[0052] S4: If the longitudinal and transverse TTC difference is greater than or equal to ΔTTC min and less than or equal to ΔTTC max , there is a risk of collision, and collision avoidance planning is performed;

[0053] In the fourth step (S4), the system determines the collision risk based on the range of the longitudinal and lateral TTC differences. If the difference is greater than or equal to a certain threshold and less than or equal to another threshold, that is, there is a collision risk, then the collision avoidance plan is executed. Specifically, the system pre-sets two thresholds, a larger threshold and a smaller threshold. The selection of these two thresholds may depend on the system design, performance and safety requirements, and can be adjusted according to the actual scenario.

[0054] The system determines whether there is a collision risk by comparing the longitudinal and lateral TTC differences with the set thresholds. Figure 4 , Figure 5 As shown in the figure, when the car just reaches the track intersection, the VRU has already passed the track intersection, and the front end of the car collides with the rear end of the VRU. At this time, there is ΔTTC max When the car has passed the track intersection and the VRU has just arrived at the track intersection, the front end of the VRU collides with the rear end of the car. At this time, there is ΔTTC min , if the difference between the longitudinal and transverse TTC is greater than or equal to ΔTTC min At the same time, it is less than or equal to ΔTTC max , the system determines that there is a collision risk, otherwise the main vehicle drives normally. When the system confirms that there is a collision risk, it triggers collision avoidance planning, which may include formulating new vehicle movement strategies, such as adjusting braking deceleration and steering angle, in order to effectively avoid collision with the VRU.

[0055] The system dynamically determines whether there is a collision risk based on the real-time longitudinal and lateral TTC differences, so that it can take proactive avoidance measures. This risk judgment based on the TTC difference allows the system to respond more flexibly to different traffic scenarios and vehicle motion states, improving the timeliness and accuracy of collision avoidance decisions. By executing collision avoidance planning, the system can effectively reduce the risk of collision with VRUs and enhance the reliability of the AEB system in longitudinal and lateral collision avoidance.

[0056] S5: Based on the cost function calculation, an optimization algorithm is used to find the optimal braking deceleration and steering angle to avoid collision.

[0057] In the fifth step (S5), the collision avoidance planning process is further refined. Specifically, after the system determines that there is a risk of collision between the host vehicle and the VRU in this step, the collision avoidance planning process is immediately initiated. This ensures the real-time response of the system to minimize the possibility of collision.

[0058] In order to find the optimal braking deceleration and steering angle, the system uses a cost function for calculation. The cost function is an evaluation function that takes into account multiple factors, including collision risk, vehicle stability, passenger comfort, etc. The goal of the cost function is to find the optimal collision avoidance planning strategy by adjusting decision variables (such as braking deceleration and steering angle) to minimize or maximize the combination of cost terms. By weighing these factors, the system can quantify the pros and cons of different decision plans and find the optimal solution. Based on the calculation results of the cost function, the system finds the optimal braking deceleration and steering angle. This decision is intended to minimize the risk of collision while considering the overall safety of the vehicle and passengers.

[0059] Due to the ever-changing traffic scene, the system may need to adjust decisions and provide real-time feedback, which ensures that the system can flexibly respond to different traffic conditions and make adaptive adjustments at any time. Ultimately, the system implements a collision avoidance strategy by executing the calculated optimal braking deceleration and steering angle, which may include adjusting vehicle speed, steering, or taking other maneuvering measures to ensure safe avoidance of collisions.

[0060] Through the above methods, the system can find a balance between real-time, accuracy and adaptability, so that the AEB system can effectively deal with the risk of collision between the car and the VRU, thereby improving overall safety. This intelligent collision avoidance planning allows the host vehicle to avoid collisions more reliably in complex traffic environments.

[0061] In another embodiment of the present invention, S5 includes braking at a deceleration of a. The cost function is established to reflect the mapping relationship between the collision avoidance planning cost and the relative distance between the two targets, the deceleration and the steering angle provided when the AEB function intervenes. The cost function calculation formula is:

[0062]

[0063] Where, is the collision avoidance planning cost, f(s) is the safety cost, f(a) is the acceleration cost, is the steering cost, s is the relative distance between the car and the VRU, a is the braking deceleration provided by the AEB system, is the steering angle of the car, α1, α2, α3 are the weight coefficients of each cost.

[0064] The following is a detailed description of the collision avoidance planning process. When the main vehicle faces the risk of collision, it adopts an active collision avoidance strategy. The following are the steps of the collision avoidance planning:

[0065] The main vehicle needs to brake at a deceleration of a. In order to achieve the optimal collision avoidance strategy, in this embodiment, a cost function is established to reflect the collision avoidance planning cost and the relative distance s between the two targets, the braking deceleration a provided by the AEB system, and the steering angle of the car. The mapping relationship between them. The calculation formula of the cost function is as follows:

[0066]

[0067] Among them, f(s) is the safety cost, which reflects the relative distance between the car and the VRU, and f(a) is the acceleration cost, taking into account the braking deceleration. is the steering cost, taking into account the steering angle, and α1, α2, and α3 are the weight coefficients of each cost.

[0068] In the embodiments of the present invention, the setting of weight coefficients is crucial because they directly affect the calculation results of the cost function. The setting of weight coefficients needs to take into account the relative importance of different factors. In practical applications, in order to make the collision avoidance planning closer to the actual driving scene, it is necessary to reasonably set the weight coefficients of each item in the cost function. The selection of weight coefficients can be based on the following considerations:

[0069] Weight α1 of safety cost f(s): Safety is the most important consideration, so the impact on relative distance s should generally have a higher weight. If maintaining a safe distance when avoiding collisions is critical in a specific scenario, the value of α1 can be adjusted to a larger positive number.

[0070] The weight α2 of the acceleration cost f(a): The impact of braking deceleration is also a key factor. In actual driving, excessive braking may cause discomfort or affect the driving of the following vehicles. Therefore, the value of α2 can be adjusted according to the needs of comfort and safety of the following vehicles.

[0071] Turning cost Weight α3: The influence of steering angle also needs to be considered. If in some cases, such as on low-adhesion roads or high-speed driving, the steering angle of the vehicle needs to be minimized, the value of α3 can be appropriately increased.

[0072] By setting these weight coefficients reasonably, the collision avoidance planning can be more in line with the actual driving scenarios, especially the needs of visual impairment scenarios, so that the cost function better reflects the optimal collision avoidance decision in complex traffic environments. By considering the relative importance of different factors, the setting logic of the weight coefficients can make the collision avoidance planning more intelligent, better adapt to the needs of different driving scenarios, especially visual impairment scenarios, and improve the practicality and reliability of collision avoidance planning.

[0073] Using the cost function, the system uses an optimization algorithm. In another embodiment of the present invention, the optimization algorithm can use a gradient descent method or a genetic algorithm to calculate the braking deceleration a and the steering angle that minimize the cost function. To achieve the best collision avoidance plan under the current circumstances.

[0074] Through this embodiment, when there is a risk of collision between the main vehicle and the VRU, the main vehicle can actively take measures such as braking and steering to avoid the collision. By establishing and optimizing the cost function, intelligent collision avoidance planning in visually obstructed scenes is realized. This active collision avoidance strategy effectively reduces the risk of collision between the vehicle and the VRU and improves the reliability of the vehicle's AEB system in complex traffic environments.

[0075] In another embodiment of the present invention, the security cost is defined as follows:

[0076]

[0077] Where s is the relative distance between the car and the VRU, [s0,s n ] is the distance threshold for whether to perform collision avoidance planning. Within the threshold range, the safety cost is determined by the attenuation function g(s).

[0078] The setting of relative distance s plays a vital role in the entire AEB system because it directly reflects the distance between the vehicle and the surrounding environment and is an important basis for determining whether collision avoidance planning is needed.

[0079] In particular, s in this embodiment is also related to the distance threshold [s0,s n ] is related. This distance threshold can be regarded as a decision-making demarcation point, which is used to determine whether the distance between the vehicle and the VRU needs to trigger collision avoidance planning. When s is within the threshold range, the safety cost will be determined by a decay function. In another embodiment of the present invention, the decay function can adopt an exponential decay function, a linear function or a reciprocal decay function.

[0080] This design takes into account the impact of distance on safety and makes the system respond more intelligently to situations at different distances by setting thresholds. This helps the system to more flexibly determine when to trigger collision avoidance planning in different driving scenarios, thereby improving the practical applicability and reliability of the entire AEB system.

[0081] In another embodiment of the present invention, S5 further includes detecting the relative distance s between the host vehicle and the VRU through a perception system. When s0 ≤ s ≤ s n , collision avoidance planning is carried out, substituting the distance s into the cost function to solve, and finding the optimal braking deceleration a and steering angle The minimum value obtained by solving is the optimal route in the collision avoidance planning.

[0082] Specifically, the perception system is usually composed of sensors. Data of the external environment is sensed or acquired through the sensors, which can be lidar, cameras, ultrasonic sensors, radars, etc. These sensors can detect information such as objects around the vehicle, road conditions, and lighting. The perception system integrates and processes the data collected by the sensors and analyzes them through algorithms to form an understanding of the environment. This understanding can be used for driving decisions, such as avoiding collisions, maintaining lanes, and pedestrian recognition. Therefore, the perception system relies on the information provided by the sensors to enable the automated system to make corresponding responses.

[0083] Detecting the relative distance s between the host vehicle and the VRU through the perception system. When the detected relative distance s is between s0 and s n , this distance information is substituted into the cost function. The cost function comprehensively considers multiple factors such as the relative distance, the braking deceleration provided by the AEB system, and the steering angle of the vehicle. By solving the cost function, the system can find the braking deceleration a and steering angle that minimize the cost function This distance-related collision avoidance planning strategy enables the system to more flexibly respond to collision risks in different situations.

[0084] In another embodiment of the present invention, S5 further includes that when s < s0 or the minimum value cannot be obtained, it indicates that the collision is inevitable. At this time, braking is carried out with a braking deceleration of a max = 0.8g to reduce the damage caused by the collision between the vehicle and the VRU.

[0085] Specifically, when the perception system detects that the relative distance s between the host vehicle and the VRU is less than the set threshold s0, or the minimum value cannot be obtained by solving the cost function due to various reasons, the system determines that the collision is inevitable. In this case, in order to minimize the damage caused by the collision to the greatest extent, the system adopts a special emergency braking deceleration a max .

[0086] The maximum friction coefficient of urban road surfaces is usually 0.8. The friction coefficient indicates the resistance provided by the road surface to the vehicle's travel. The maximum friction coefficient indicates that the road surface provides greater resistance. max It is set to 0.8 times the acceleration due to gravity (g) to fully utilize the maximum braking capacity provided by the road surface.

[0087] The implementation of this special braking strategy aims to quickly slow down the vehicle through emergency braking to reduce the impact of the collision. The braking force in this case is relatively large, which is a proactive response to the emergency situation. This intelligent braking measure is designed to minimize the damage caused by the collision when the collision cannot be avoided, thereby improving the life safety protection level of the entire AEB system in emergency situations.

[0088] In another embodiment of the present invention, S5 further includes, when s>s n , the cost is 0, indicating that no collision will occur.

[0089] When the perception system detects that the relative distance between the host vehicle and the VRU is s>s n This means that the current relative distance is outside a safe range and a collision will not occur immediately. In this case, the collision avoidance plan can set this cost to 0, indicating that there is no current risk of collision and no additional collision avoidance measures are required.

[0090] This design helps the system handle a variety of traffic scenarios more effectively. By setting this item in the cost function to 0, the system will pay more attention to other cost factors when evaluating collision avoidance strategies, such as the optimal choice of safe distance, braking deceleration, and steering angle. This comprehensive consideration makes collision avoidance planning more intelligent and flexible, and can make more reasonable and safe decisions under different traffic conditions.

[0091] By dynamically adjusting the collision avoidance planning strategy, the system can flexibly respond to different traffic situations based on real-time relative distance information, thereby improving the practicality and reliability of the system. This avoids unnecessary collision avoidance measures outside the safe distance range, reduces the frequency of system intervention, and ensures that timely and accurate responses can be made when there is a real risk of collision, thus minimizing the probability of traffic accidents.

[0092] In another embodiment of the present invention, the collision avoidance method can be applied to the field of intelligent driving and automatic driving. Intelligent driving systems and automatic driving technologies are intended to improve the autonomy and intelligence of vehicles in different traffic scenarios, thereby enhancing driving safety and efficiency. By integrating the collision avoidance method into intelligent driving and automatic driving systems, a higher level of traffic safety and driving comfort can be achieved in scenarios such as unmanned driving and semi-automatic driving.

[0093] In intelligent driving and autonomous driving applications, this collision avoidance method can obtain information about the vehicle and the surrounding environment through the perception system, calculate the relative distance between the main vehicle and the VRU in real time, and perform collision avoidance planning based on this information. For example, in autonomous driving mode, when the system detects a collision risk between the car and the VRU, the vehicle's braking deceleration and steering angle can be automatically adjusted to minimize the risk of collision and ensure safe interaction between the vehicle and other road users.

[0094] By introducing collision avoidance methods into the field of intelligent driving and autonomous driving, the overall safety of the transportation system can be improved, accidents can be reduced, and traffic flow can be optimized. Through the application of collision avoidance methods, intelligent driving and autonomous driving systems can respond to complex driving situations more intelligently, thus laying the foundation for the development of future transportation systems. The expansion of this application field will play a positive role in promoting the development of intelligent transportation technology and improving the driving experience.

[0095] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A longitudinal and lateral AEB collision avoidance method in a visually impaired scene combined with steering, characterized in that: The following steps are involved: Step 1: Detect the speed information of other vehicles in the adjacent lanes through sensors, and determine the potential collision risk within the blind spot of the main vehicle based on the acceleration changes of the vehicles in the adjacent lanes; Step 2: When the acceleration of the adjacent lane vehicle is less than zero, measures are taken to eliminate the braking gap to reduce the braking reaction time; Step 3: Obtain the relevant motion information of the host vehicle and the vulnerable road user VRU through the sensor, calculate the time TTC when the host vehicle and the VRU reach the intersection of the trajectories, and use the difference between the two TTCs to determine whether there is a collision risk. The host vehicle travels along the X-axis direction, and the VRU travels along the Y-axis direction. (x0, y0) is the intersection of the trajectories of the host vehicle and the VRU. The moment when the VRU behind the blind spot first appears in the field of vision of the host vehicle is recorded as the initial moment. At this time, the speed of the host vehicle is detected to be , VRU speed is , the distance between them is s, the azimuth is θ, and the calculation formula of the longitudinal and transverse TTC difference is: In the formula is the longitudinal and transverse TTC difference, which indicates the time difference between the main vehicle along the longitudinal direction and the VRU along the transverse direction reaching the track intersection point. is the time when the car reaches the intersection of the track on the longitudinal X axis, is the time when the VRU reaches the track intersection point on the horizontal Y axis; Step 4: If the longitudinal and transverse TTC differences are greater than or equal to and less than or equal to , there is a risk of collision, and collision avoidance planning is performed; Step 5: According to the cost function calculation, an optimization algorithm is used to find the optimal braking deceleration and steering angle to avoid collision, where The deceleration rate is used to brake. The cost function is established to reflect the mapping relationship between the collision avoidance planning cost and the relative distance between the two targets, the deceleration and the steering angle provided when the AEB function intervenes; the cost function calculation formula is: In the formula, To avoid collisions, For the cost of safety, is the acceleration cost, For the cost of turning, is the relative distance between the car and the VRU, The braking deceleration provided by the AEB system, is the steering angle of the car, is the weight coefficient of each cost.

2. The collision avoidance method according to claim 1, characterized in that: The security cost is defined as follows: in, is the relative distance between the car and the VRU, The distance threshold for whether to perform collision avoidance planning. Within the threshold range, the safety cost is determined by the attenuation function Sure.

3. The collision avoidance method according to claim 2, characterized in that: The decay function is an exponential decay function, a linear function or a reciprocal decay function.

4. The collision avoidance method according to claim 2, characterized in that: The step 5 also includes detecting the relative distance between the main vehicle and the VRU through the sensing system. ,when When Substitute the cost function to solve and use the optimization algorithm to find the optimal braking deceleration and steering angle , solve for The minimum value of is the optimal route in collision avoidance planning.

5. The collision avoidance method according to claim 2, characterized in that: The optimization algorithm is a gradient descent method or a genetic algorithm.

6. The collision avoidance method according to claim 2, characterized in that: The step 5 also includes, when Or if the minimum value cannot be found, it means that collision is inevitable. The vehicle is braked at a braking deceleration to reduce the damage caused by the collision between the vehicle and the VRU.

7. The collision avoidance method according to claim 2, characterized in that: The step 5 also includes, when , the cost is 0, indicating that no collision will occur.

8. A collision avoidance method according to any one of claims 1 to 7, characterized in that: The collision avoidance method is applied to the fields of intelligent driving and automatic driving.

Images