Vehicle adaptive cruise control method with steering collision avoidance function

Abstract

The present invention relates to a vehicle adaptive cruise control method with a steering collision avoidance function, comprising the following steps: constructing a critical longitudinal safety distance model for collision avoidance, for determining the critical longitudinal safety distance for collision avoidance corresponding to different leading vehicle driving states; obtaining the current actual vehicle-to-vehicle distance, comparing the actual vehicle-to-vehicle distance with the critical longitudinal safety distance for collision avoidance; if the actual vehicle-to-vehicle distance is less than the critical longitudinal safety distance for collision avoidance, implementing vehicle steering collision avoidance control by adopting a steering collision avoidance trajectory planning algorithm based on a quintic polynomial, combined with a steering collision avoidance trajectory tracking control method based on an MPC; otherwise, implementing vehicle following control by adopting a vehicle following control method based on adaptive adjustment of the performance indicator weight coefficients of the MPC. Compared with the prior art, the present invention can dynamically coordinate the safety, following performance, comfort, and economy of the vehicle in the following mode, and can promptly switch to the steering collision avoidance mode when the following braking distance is insufficient to avoid a collision accident.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control, and in particular to a vehicle adaptive cruise control method with a steering collision avoidance function. Background Art

[0002] ACC (Adaptive Cruise Control) is an assisted driving system that can automatically adjust the speed of the vehicle according to the speed of the vehicle in front to ensure driving safety. It obtains the driving status of surrounding vehicles through sensors and controls the acceleration and deceleration of the vehicle to achieve following driving. At present, the ACC system is based on longitudinal following control in most cases to ensure the safe driving of the vehicle, which requires a sufficient safety distance between the front and rear vehicles. However, in emergency situations such as if the vehicle in front suddenly stops or a vehicle suddenly drives out of the intersection in the blind spot, the longitudinal following braking safety distance is insufficient, and it is difficult to avoid a collision accident through following control, but it is possible through steering collision avoidance control. Therefore, in order to expand the use scenarios of the ACC system and increase active safety, the ACC system needs to be improved so that it can perform steering collision avoidance driving when the longitudinal following braking safety distance is insufficient to ensure the safe driving of the vehicle.

[0003] To ensure that the vehicle can make accurate decisions regarding following and steering avoidance modes, the rationality of the critical safety model is a crucial prerequisite. Currently, commonly used safety models are divided into those based on collision time and those based on relative vehicle distance. Compared to the collision time-based model, the relative vehicle distance-based model, which relies less on extensive experimental data and primarily analyzes the braking process, vehicle distance, and the driving conditions of both vehicles, is currently a research focus. Currently, the difficulty in designing control strategies for following mode lies in optimizing various performance indicators (safety, following, comfort, and economy) to adapt to complex and changing driving environments. The difficulty in steering avoidance mode lies in planning a smooth and continuous steering avoidance trajectory in real time, as well as designing a control method that balances tracking accuracy and real-time performance. Summary of the Invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a vehicle adaptive cruise control method with a steering collision avoidance function, which can dynamically coordinate multiple performance indicators of the vehicle in the following mode. When the following braking distance is insufficient, it can switch to the steering collision avoidance mode in time to avoid the occurrence of collision accidents.

[0005] The object of the present invention can be achieved by the following technical solution: A vehicle adaptive cruise control method with a steering collision avoidance function comprises the following steps:

[0006] Construct a critical longitudinal safety distance model for collision avoidance to determine the critical longitudinal safety distance for collision avoidance corresponding to different driving states of the preceding vehicle;

[0007] The current actual vehicle distance is obtained and compared with the critical longitudinal safety distance for collision avoidance. If the actual vehicle distance is less than the critical longitudinal safety distance, a steering collision avoidance trajectory planning algorithm based on a quintic polynomial is used, combined with a steering collision avoidance trajectory tracking control method based on MPC (Model Predictive Control) to achieve vehicle steering collision avoidance control.

[0008] Otherwise, the vehicle following control method based on the adaptive adjustment of the performance index weight coefficient of MPC is used to realize vehicle following control.

[0009] Furthermore, the process of constructing the collision avoidance critical longitudinal safety distance model includes:

[0010] Taking into account the motion state of the front and rear vehicles, the actual vehicle-to-vehicle distance, the road adhesion coefficient, and the brake system's pressure-building limit, a critical longitudinal safety distance model for collision avoidance is established according to different leading vehicle driving states. The leading vehicle driving states include stationary, constant speed, and emergency braking.

[0011] Furthermore, when the leading vehicle is stationary, the self-vehicle follows the vehicle at the maximum deceleration and decelerates to 0. The corresponding critical longitudinal safety distance model for collision avoidance is:

[0012]

[0013] Among them, τ′ b is the brake dead time, τ″ b is the deceleration response time, d min is the minimum safety distance, a bmax is the maximum braking deceleration, v e is the vehicle speed, when the actual vehicle distance d act Greater than S e_min When the vehicle does not need to make an emergency turn, it can make full use of d act Following the vehicle, the expected braking deceleration of the vehicle is:

[0014]

[0015] Among them, a des_s is the expected braking deceleration of the vehicle when the preceding vehicle is stationary.

[0016] Furthermore, when the leading vehicle is at a constant speed, the vehicle adjusts its speed to the same as that of the leading vehicle, and the distance traveled is less than the actual vehicle-to-vehicle distance. The corresponding critical longitudinal safety distance model for collision avoidance is:

[0017]

[0018] In this case, the ego vehicle uses the actual vehicle-to-vehicle distance to complete the following process. At this time, the expected braking deceleration of the ego vehicle is:

[0019]

[0020] Among them, v p is the speed of the preceding vehicle, a des_c is the expected braking deceleration of the vehicle when the preceding vehicle is at a constant speed.

[0021] Furthermore, when the preceding vehicle is in emergency braking state, the preceding vehicle brakes at the maximum deceleration a pmax During emergency braking, when the vehicle decelerates to 0, the relative distance between the vehicle and the preceding vehicle is greater than 0. At this time, the distance traveled by the preceding vehicle is:

[0022]

[0023] The corresponding critical longitudinal safety distance model for collision avoidance is:

[0024]

[0025] At this time, the expected braking deceleration corresponding to the actual vehicle-to-vehicle distance is:

[0026]

[0027] Among them, a des_e It is the expected braking deceleration of the vehicle when the preceding vehicle brakes suddenly.

[0028] Furthermore, the steering collision avoidance trajectory planning algorithm based on the quintic polynomial is used to solve the desired steering collision avoidance trajectory that takes into account lateral motion constraints, heading angle constraints, displacement constraints, and lateral acceleration constraints.

[0029] Furthermore, the desired trajectory for steering collision avoidance is specifically:

[0030]

[0031] Among them, y e is the lateral displacement corresponding to the collision avoidance termination point, t e is the time required to complete the collision avoidance, and t is the time independent variable.

[0032] Furthermore, the MPC-based steering collision avoidance trajectory tracking control method is used to solve the optimal control variable for tracking the desired trajectory of steering collision avoidance through rolling optimization. The corresponding objective function and constraints are:

[0033]

[0034] in, H is the control increment weight coefficient matrix, G1 is the control amount weight coefficient matrix, e k is the error between the prediction model and the reference value in the prediction time domain, y1(k+i) is the actual output of the system, Q1 and R1 are both weight matrices, ρ is the weight coefficient of the relaxation factor, ε1 is the relaxation factor, u1 and Δu1 are the actual control quantity and actual control quantity error of the system, respectively.

[0035] Furthermore, the vehicle-following control method with adaptive adjustment of the performance index weight coefficient based on MPC is used to optimize the desired deceleration of the vehicle-following in real time, and the corresponding objective function and constraint conditions are:

[0036]

[0037] in, U is the set of control variables, including the front wheel angle and the desired acceleration. ε is the vector relaxation factor. A H is the coefficient matrix of the constraint conditions, which are the upper and lower limits of the front wheel angle and the desired acceleration, b H is the constant term matrix of the constraints.

[0038] Furthermore, in the vehicle-following control method with adaptive adjustment of performance index weight coefficients based on MPC, performance index adaptive fuzzy logic rules are designed according to the following four scenarios:

[0039] (1) When the ego vehicle's speed is greater than that of the preceding vehicle and the actual inter-vehicle distance is less than the expected value, there is a risk of collision. The ego vehicle should be quickly decelerated. At this time, the weight coefficients of the relative speed and inter-vehicle distance error need to be increased, the restriction on the ego vehicle's acceleration needs to be removed, and the weight coefficient of the ego vehicle's acceleration needs to be reduced, thereby increasing the response speed of the following control.

[0040] (2) When the ego vehicle's speed is greater than that of the leading vehicle and the actual inter-vehicle distance is greater than the expected value, in order to ensure the following performance of the vehicle, the ego vehicle should first be controlled to decelerate and then accelerate to follow the leading vehicle. As the relative speed of the two vehicles approaches 0, the weight coefficient of the relative speed should be gradually reduced, while the weight coefficient of the ego vehicle's acceleration should be increased.

[0041] (3) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is lower than the expected value, the weight coefficient of the distance error should be increased first, and the weight coefficient of the ego vehicle acceleration should be increased at the same time. As the relative speed approaches zero, the weight coefficient of the relative speed should be gradually reduced.

[0042] (4) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is greater than the expected value, there is no collision risk. The weight coefficient of the ego vehicle acceleration should be appropriately increased. In this scenario, the weight coefficients of the relative speed and distance error should be gradually reduced.

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] The present invention comprehensively considers multiple factors such as the driving state of the leading vehicle, the road adhesion coefficient, the pressure building limit capacity of the braking system, etc., and establishes a critical longitudinal safety distance model for collision avoidance corresponding to different driving states of the leading vehicle. According to the model, the driving mode is divided into a following mode and a steering collision avoidance mode, so as to accurately decide the driving state that the vehicle should adopt; for the two driving states, a following control method with adaptive adjustment of performance index weight coefficients and a steering collision avoidance trajectory tracking control method are proposed under the MPC framework, which can dynamically coordinate the safety, following performance, comfort and economy of the vehicle in the following mode. When the following braking distance is insufficient, it can switch to the steering collision avoidance mode in time to avoid the occurrence of collision accidents, thereby improving the collision avoidance capability and safety in emergency situations, expanding the application scenarios of the ACC system, and helping to improve the overall safety of road traffic.

[0045] The present invention designs a vehicle to steer for collision avoidance when the actual vehicle-to-vehicle distance is less than the critical longitudinal safety distance for collision avoidance. In combination with the lateral collision avoidance constraint conditions that take into account vehicle dynamics and road adhesion characteristics, a steering collision avoidance trajectory planning algorithm based on quintic polynomials is proposed. On this basis, a steering collision avoidance trajectory tracking control method based on MPC is proposed. Its objective function comprehensively considers trajectory tracking control accuracy, actuator smooth transition characteristics, collision avoidance output boundary limits and actuator capabilities, and can ensure real-time and accurate steering collision avoidance control.

[0046] The present invention is designed to follow the vehicle when the actual vehicle-to-vehicle distance is greater than the critical longitudinal safety distance for collision avoidance. Taking into account the vehicle's followability, safety, comfort and economy, a following control method with adaptive adjustment of the performance index weight coefficient based on MPC is proposed to optimize the expected deceleration of the following vehicle in real time to meet the driver's following needs under different working conditions. In addition, taking into account the actual following driving process, performance index adaptive fuzzy logic rules are designed according to four scenarios, thereby effectively improving the reliability of following control. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Schematic diagram of the method flow of the present invention;

[0048] Figure 2 Schematic diagram of the application process of the embodiment;

[0049] Figure 3Schematic diagram of the critical longitudinal safety distance model for collision avoidance in the embodiment;

[0050] Figures 4a to 4d Schematic diagram for comparing the simulation results of the following vehicle control in the embodiment;

[0051] Figures 5a to 5f Schematic diagram for comparing simulation results of steering collision avoidance control in the embodiment. DETAILED DESCRIPTION

[0052] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0053] Example

[0054] like Figure 1 As shown, a vehicle adaptive cruise control method with a steering collision avoidance function includes the following steps:

[0055] Construct a critical longitudinal safety distance model for collision avoidance to determine the critical longitudinal safety distance for collision avoidance corresponding to different driving states of the preceding vehicle;

[0056] The current actual vehicle distance is obtained and compared with the critical longitudinal safety distance for collision avoidance. If the actual vehicle distance is less than the critical longitudinal safety distance, a steering collision avoidance trajectory planning algorithm based on a quintic polynomial is used, combined with a steering collision avoidance trajectory tracking control method based on MPC, to achieve vehicle steering collision avoidance control.

[0057] Otherwise, the vehicle following control method based on the adaptive adjustment of the performance index weight coefficient of MPC is used to realize vehicle following control.

[0058] This embodiment applies the above solution, such as Figure 2 As shown, the main contents are:

[0059] Step 1: Comprehensively considering the motion state of the front and rear vehicles, the actual inter-vehicle distance, the road adhesion coefficient, and the braking system's pressure-building limit, three different critical longitudinal safety distance models for collision avoidance are established based on the driving state of the leading vehicle.

[0060] Step 2: When the actual vehicle-to-vehicle distance is less than the critical longitudinal safety distance for collision avoidance, the vehicle steers to avoid collision. A steering collision avoidance trajectory planning algorithm based on a quintic polynomial is proposed, combining lateral collision avoidance constraints that consider vehicle dynamics and road adhesion characteristics. Based on this, a steering collision avoidance trajectory tracking control method based on MPC is designed. The objective function design comprehensively considers trajectory tracking control accuracy, actuator smooth transition characteristics, collision avoidance output boundary constraints, and actuator capabilities.

[0061] Step 3: When the actual vehicle-to-vehicle distance is greater than the critical longitudinal safety distance for collision avoidance, the vehicle follows the vehicle. Taking into account the vehicle's followability, safety, comfort, and economy, a following control method with adaptive adjustment of the performance index weight coefficient based on MPC is proposed to optimize the desired deceleration of the following vehicle in real time.

[0062] In step 1, the motion state of the front and rear vehicles, the actual vehicle distance, the road adhesion coefficient, and the braking system pressure limit are comprehensively considered. Assuming that the driving state of the front vehicle remains unchanged, three situations are considered: the front vehicle is stationary, the front vehicle is moving at a constant speed, and the front vehicle brakes suddenly. Under the three different driving states of the front vehicle, the corresponding collision avoidance critical longitudinal safety distance model (such as Figure 3 As shown) are as follows:

[0063] (1) The vehicle ahead is stationary

[0064] When the leading vehicle is stationary, the ego vehicle follows the vehicle at the maximum deceleration and decelerates to 0. At this time, the critical longitudinal safety distance model for collision avoidance is:

[0065]

[0066] Among them, τ′ b is the brake dead time, τ″ b is the deceleration response time, when the actual vehicle distance d act Greater than S e_min When the vehicle does not need to make an emergency turn, it can make full use of d act Follow the vehicle. The expected braking deceleration a of the vehicle at this time des_s :

[0067]

[0068] (2) The vehicle ahead has a constant speed

[0069] When the leading vehicle is traveling at a constant speed, the vehicle can avoid a collision by simply adjusting its speed to the same as the preceding vehicle and traveling a distance less than the actual distance between vehicles. At this point, the critical longitudinal safety distance model for collision avoidance is:

[0070]

[0071] In this case, the ego vehicle can fully utilize the actual vehicle-to-vehicle distance to complete the following vehicle. At this time, the expected braking deceleration a of the ego vehicle is des_c :

[0072]

[0073] (3) Emergency braking of the vehicle ahead

[0074] The current vehicle is braking at maximum deceleration a pmaxDuring emergency braking, the relative distance between the vehicle and the preceding vehicle is greater than 0 when the vehicle decelerates to 0, which is the critical collision avoidance condition in this case. At this time, the distance between the preceding vehicle and the preceding vehicle is:

[0075]

[0076] The corresponding critical longitudinal safety distance model for collision avoidance is:

[0077]

[0078] At this time, the vehicle fully utilizes the expected braking deceleration a corresponding to the actual vehicle distance. dese for:

[0079]

[0080] Based on the above model, decisions and judgments are made on the vehicle's current driving state, and the driving state is divided into two types: when the actual vehicle distance is less than the critical longitudinal safety distance for collision avoidance, the vehicle steers to avoid collision; when the actual vehicle distance is greater than the critical longitudinal safety distance for collision avoidance, the vehicle follows the vehicle.

[0081] In step 2, when the actual vehicle-to-vehicle distance is less than the critical longitudinal safety distance for collision avoidance, the vehicle steers to avoid collision. Combining the steering avoidance constraints that take into account the vehicle dynamics characteristics and road adhesion coefficient, a steering avoidance trajectory planning algorithm based on quintic polynomials is proposed.

[0082] Define the steering collision avoidance trajectory form based on the quintic polynomial:

[0083] y(x)=b0+b1x+b2x 2 +b3x 3 +b4x 4 +b5x 5

[0084] Among them, b i is the unknown constant, x is the longitudinal displacement, and y is the lateral displacement.

[0085] By taking the first-order and second-order derivatives of the above formula, we can get the lateral velocity and lateral acceleration of the vehicle:

[0086]

[0087] According to the road constraints, the fifth-order polynomial collision avoidance trajectory coefficients b0=b1=b2=0 can be obtained, and b3, b4, b5 satisfy the following relationship:

[0088]

[0089] Among them, x eis the longitudinal displacement corresponding to the collision avoidance end point, y e is the lateral displacement corresponding to the collision avoidance termination point. At the same time, we can get

[0090] The quintic polynomial collision avoidance trajectory can be organized as:

[0091]

[0092] Assume that the vehicle maintains a constant speed during the entire collision avoidance process, that is, x = v e t, where t is the collision avoidance time. When the vehicle avoids the obstacle, x e =v e t e , where t e is the time required to complete the collision avoidance. Therefore, the above trajectory formula is rewritten as the collision avoidance trajectory with time t as the independent variable:

[0093]

[0094] Since the vehicle's heading angle is small during collision avoidance, the vehicle's lateral acceleration is approximately equal to the transverse acceleration. Therefore, the lateral acceleration constraint is set as:

[0095] |a ey (t)|≤a ey_max

[0096] Among them, a ey (t) and a ey_max The lateral acceleration and maximum lateral acceleration are respectively the lateral acceleration of the vehicle. The lateral acceleration generally does not exceed 0.35g. Taking into account the grading of lateral acceleration, the maximum lateral acceleration of the vehicle is set to 0.5μg.

[0097] By solving the value of t, we can get the lateral acceleration a of the vehicle ey (t) is:

[0098]

[0099] The time to avoid a collision is calculated by the lateral acceleration that satisfies the stability constraint. The time must meet the following conditions:

[0100]

[0101] On this basis, an MPC-based steering collision avoidance trajectory tracking control method is designed, which comprehensively considers the trajectory tracking control accuracy, actuator smooth transition characteristics, collision avoidance output boundary limit and actuator capability. According to the emergency steering obstacle avoidance trajectory model and vehicle dynamics model, the state space equation of the control system is established:

[0102]

[0103] in, is the system state variable, u1=[δ f a des ] is the system control variable, e y is the lateral distance deviation, e θ is the heading angle deviation, θ des is the vehicle reference heading angle, e v is the speed error, e x is the longitudinal position error.

[0104] The nonlinear system model is linearized, and the continuous system needs to be further converted into a discrete system to facilitate sampling and calculation of trajectory deviation. The optimization objective function is established as follows:

[0105]

[0106] Where N p1 and N c1 are the prediction time domain and control time domain of the system respectively, y1(k+i) is the actual output of the system, y 1_ref (k+i) is the system's reference output, Q1 and R1 are weight matrices, ρ is the weight coefficient of the relaxation factor, and ε1 is the relaxation factor. Here, the first term represents the tracking accuracy of the target trajectory, the second term indicates whether the ego vehicle can stably steer to avoid collisions, and the third term indicates the degree of relaxation of the solution.

[0107] The problem of solving the optimal control increment in the control domain is transformed into a constrained quadratic programming problem:

[0108]

[0109] Where, e k is the error between the prediction model and the reference value in the prediction time domain.

[0110] In each control cycle, the optimization solution is completed through quadratic programming to obtain a series of control increments and relaxation factors in the prediction time domain.

[0111]

[0112] Taking the first item in the sequence as the actual control increment and acting on the system, the actual control quantity at the current moment can be expressed as:

[0113] u1(k)=u1(k-1)+Δu1 * (k)

[0114] Repeat the above steps in subsequent control cycles to repeatedly solve the optimal control quantity to achieve rolling optimization.

[0115] In step 3, when the actual vehicle-to-vehicle distance is greater than the critical longitudinal safety distance for collision avoidance, the vehicle follows the vehicle. Taking into account the vehicle's followability, safety, comfort, and economy, a following control method with adaptive adjustment of the performance index weight coefficient based on MPC is proposed to optimize the expected deceleration of the following vehicle in real time.

[0116] By analyzing the system's optimization objectives, we design the objective function to be optimized:

[0117]

[0118] Among them, y p (k+i|k) is the predicted output value of the system at time k for the future i-th step, y ref (k+i|k) is the expected value of the system at time k for the future i-th step, u(k+i) is the control amount of the system at time k for the i-th step, are the weight factors of vehicle distance error, relative vehicle speed, ego vehicle speed and ego vehicle acceleration, respectively.

[0119] According to the designed objective function, performance indicators and related constraints, the multi-objective optimization control of the system is transformed into a quadratic programming problem, namely:

[0120]

[0121] in, U is the set of control variables, ε is the vector relaxation factor, A H is the coefficient matrix of the constraint conditions, b H is the constant term matrix of the constraints.

[0122] During the rolling optimization process, if the inputs d, v, a, jerk, and u exceed the upper and lower bounds of the hard constraints, the vector relaxation factor is gradually increased to relax the upper and lower bounds of the input and output constraints, further expanding the feasible region of the solution system and ensuring that an optimal solution exists for u(k+i|k). The first component of the control system input is used as the actual input of the system at the next moment and a rolling optimization and prediction is performed.

[0123] Considering the actual vehicle-following process, performance index adaptive fuzzy logic rules are designed based on the following four scenarios:

[0124] (1) When the ego vehicle's speed is greater than that of the preceding vehicle and the actual inter-vehicle distance is less than the expected value, there is a risk of collision. The ego vehicle should be quickly decelerated. At this time, the weight coefficients of the relative speed and inter-vehicle distance error need to be increased, the restriction on the ego vehicle's acceleration needs to be removed, and the weight coefficient of the ego vehicle's acceleration needs to be reduced, thereby increasing the response speed of the following control.

[0125] (2) When the ego vehicle's speed is greater than that of the leading vehicle and the actual inter-vehicle distance is greater than the expected value, in order to ensure the following performance of the vehicle, the ego vehicle should first be controlled to decelerate and then accelerate to follow the leading vehicle. As the relative speed of the two vehicles approaches 0, the weight coefficient of the relative speed should be gradually reduced, while the weight coefficient of the ego vehicle's acceleration should be increased.

[0126] (3) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is lower than the expected value, the weight coefficient of the distance error should be increased first, and the weight coefficient of the ego vehicle acceleration should be increased at the same time. As the relative speed approaches zero, the weight coefficient of the relative speed should be gradually reduced.

[0127] (4) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is greater than the expected value, there is no collision risk. The weight coefficient of the ego vehicle acceleration should be appropriately increased. In this scenario, the weight coefficients of the relative speed and distance error should be gradually reduced.

[0128] In order to prove the effectiveness of this solution, this embodiment verifies the effectiveness of the following control and steering collision avoidance control methods under two typical working conditions. Figures 4a to 4d The simulation results of the car-following control are shown in Figure 2. Figures 5a to 5f The simulation results of steering collision avoidance control.

[0129] 1. Vehicle-following control verification conditions

[0130] A vehicle traveling at a constant speed of 60 km / h suddenly cuts in 30 meters ahead. The initial speed of the vehicle is 80 km / h. After calculation, the actual distance between the two vehicles is greater than the critical longitudinal safety distance for collision avoidance. At this time, the following control is adopted, and the control effect is as follows: Figures 4a to 4d As shown in the figure, during the entire vehicle-following process, as the vehicle-to-vehicle distance error decreases, its weight also gradually decreases; the speed and acceleration of the ego vehicle can change smoothly while satisfying the multi-objective constraints.

[0131] 2. Steering Collision Avoidance Control Verification Conditions

[0132] The vehicle is traveling at a constant speed of 120 km / h and suddenly sees a vehicle traveling at a constant speed of 30 km / h cut in 40 meters ahead. After calculation, the critical longitudinal safety distance for collision avoidance is 46.08 meters. The actual distance between the two vehicles is less than this safety distance. The vehicle turns to avoid collision. The control effect is as follows: Figures 5a to 5f The vehicle's lateral displacement tracking control during the steering collision avoidance process is effective, and within the road width, the center of mass sideslip angle and front wheel turning angle do not vary much, and the vehicle body posture is good.

[0133] In summary, this solution improves the ACC system and proposes a vehicle adaptive cruise control method with steering collision avoidance function, which can dynamically coordinate the vehicle's safety, following performance, comfort and economy in following mode. When the following braking distance is insufficient, it can switch to steering collision avoidance mode in time to avoid collision accidents.

Claims

1. A vehicle adaptive cruise control method with a steering collision avoidance function, characterized in that: The following steps are involved: Construct a critical longitudinal safety distance model for collision avoidance to determine the critical longitudinal safety distance for collision avoidance corresponding to different driving states of the preceding vehicle; The current actual vehicle distance is obtained and compared with the critical longitudinal safety distance for collision avoidance. If the actual vehicle distance is less than the critical longitudinal safety distance, a steering collision avoidance trajectory planning algorithm based on a quintic polynomial is used, combined with a steering collision avoidance trajectory tracking control method based on MPC, to achieve vehicle steering collision avoidance control. Otherwise, the vehicle following control method based on the adaptive adjustment of the performance index weight coefficient of MPC is used to realize the vehicle following control; The process of constructing the collision avoidance critical longitudinal safety distance model includes: Taking into account the motion state of the front and rear vehicles, the actual vehicle-to-vehicle distance, the road adhesion coefficient, and the brake system's pressure-building capacity, a critical longitudinal safety distance model for collision avoidance is established for each leading vehicle's driving state. The leading vehicle's driving states include stationary, constant speed, and emergency braking. When the leading vehicle is stationary, the self-vehicle follows the vehicle at the maximum deceleration and decelerates to 0. The corresponding critical longitudinal safety distance model for collision avoidance is: ， in, is the brake dead time, is the deceleration response time, is the minimum safe distance, is the maximum braking deceleration, is the vehicle speed, when the actual vehicle distance Greater than When the vehicle does not need to make an emergency turn, make full use of Following the vehicle, the expected braking deceleration of the vehicle is: ， in, is the expected braking deceleration of the vehicle when the preceding vehicle is stationary; When the leading vehicle is at a constant speed, the vehicle adjusts its speed to the same as that of the leading vehicle, and the distance traveled is less than the actual vehicle-to-vehicle distance. The corresponding critical longitudinal safety distance model for collision avoidance is: ， In this case, the ego vehicle uses the actual vehicle-to-vehicle distance to complete the following process. At this time, the expected braking deceleration of the ego vehicle is: ， in, is the speed of the preceding vehicle, is the expected braking deceleration of the vehicle under the constant speed of the preceding vehicle; When the preceding vehicle is in emergency braking state, the preceding vehicle is braking at the maximum deceleration During emergency braking, when the vehicle decelerates to 0, the relative distance between the vehicle and the preceding vehicle is greater than 0. At this time, the distance traveled by the preceding vehicle is: ， The corresponding critical longitudinal safety distance model for collision avoidance is: ， At this time, the expected braking deceleration corresponding to the actual vehicle-to-vehicle distance is: ， in, is the expected braking deceleration of the vehicle in the event of emergency braking of the preceding vehicle; The MPC-based steering collision avoidance trajectory tracking control method is used to solve the optimal control variable for tracking the desired trajectory of steering collision avoidance through rolling optimization. The corresponding objective function and constraints are: ， ， in, , , is the control increment weight coefficient matrix, is the control quantity weight coefficient matrix, is the error between the prediction model and the reference value in the prediction time domain, is the actual output of the system, and are all weight matrices, is the weight coefficient of the relaxation factor, is the relaxation factor, and are the actual control quantity and actual control quantity error of the system respectively; The vehicle-following control method with adaptive adjustment of performance index weight coefficients based on MPC is used to optimize the desired deceleration of the vehicle-following in real time. The corresponding objective function and constraints are: ， in, , is the set of control variables, including the front wheel angle and the desired acceleration. is the vector relaxation factor, is the coefficient matrix of the constraint conditions, which are the upper and lower limits of the front wheel angle and the desired acceleration. is the constant term matrix of the constraints.

2. The vehicle adaptive cruise control method with steering collision avoidance function according to claim 1, characterized in that: The steering collision avoidance trajectory planning algorithm based on quintic polynomial is used to solve the desired steering collision avoidance trajectory taking into account lateral motion constraints, heading angle constraints, displacement constraints, and lateral acceleration constraints.

3. The vehicle adaptive cruise control method with steering collision avoidance function according to claim 2, characterized in that: The desired trajectory of steering collision avoidance is specifically: ， in, is the lateral displacement corresponding to the collision avoidance end point, The time required to complete the collision avoidance process. is the time independent variable.

4. The vehicle adaptive cruise control method with steering collision avoidance function according to claim 1, characterized in that: In the vehicle-following control method with adaptive adjustment of performance index weight coefficients based on MPC, performance index adaptive fuzzy logic rules are designed according to the following four scenarios: (1) When the ego vehicle's speed is greater than that of the preceding vehicle and the actual inter-vehicle distance is less than the expected value, there is a risk of collision. The ego vehicle should be quickly decelerated. At this time, the weight coefficients of the relative speed and inter-vehicle distance error need to be increased, the restriction on the ego vehicle's acceleration needs to be removed, and the weight coefficient of the ego vehicle's acceleration needs to be reduced, thereby increasing the response speed of the following control. (2) When the ego vehicle's speed is greater than that of the leading vehicle and the actual inter-vehicle distance is greater than the expected value, in order to ensure the following performance of the vehicle, the ego vehicle should first be controlled to decelerate and then accelerate to follow the leading vehicle. As the relative speed of the two vehicles approaches 0, the weight coefficient of the relative speed should be gradually reduced, while the weight coefficient of the ego vehicle's acceleration should be increased. (3) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is lower than the expected value, the weight coefficient of the distance error should be increased first, and the weight coefficient of the ego vehicle acceleration should be increased at the same time. As the relative speed approaches zero, the weight coefficient of the relative speed should be gradually reduced. (4) When the speed of the ego vehicle is lower than that of the preceding vehicle and the actual inter-vehicle distance is greater than the expected value, there is no collision risk. The weight coefficient of the ego vehicle acceleration should be appropriately increased. In this scenario, the weight coefficients of the relative speed and distance error should be gradually reduced.

Images