Automatic emergency steering control method considering comfort of driver

Abstract

The invention discloses an automatic emergency steering control method considering driver comfort. The method comprises the steps that a vehicle-mounted sensor obtains vehicle information, and a camera, a millimeter wave radar and an angle radar sensor obtain dangerous target information of three lanes in front; judging the dangerous target with the minimum TTC, the minimum relative longitudinal distance or the minimum relative transverse distance as a main dangerous target according to the dangerous target information arbitration, and outputting the information of the main dangerous target; planning a path cluster according to the position of the vehicle by combining a quintic polynomial and a scattering point method, and screening out an optimal path through collision inspection; according to the vehicle information, longitudinal speed planning is carried out, and an optimal longitudinal speed curve is planned; for the optimal path, far and near preview points are adopted, steady-state yaw velocity gain processing is adopted for the far point, PID control is adopted for the near point, and a final steering wheel turning angle is obtained through superposition; and aiming at the optimal longitudinal speed planning curve, calculating by adopting a feedforward and feedback combined method to obtain a final expected acceleration.

Description

Technical Field

[0001] The invention belongs to the field of vehicle steering control, and in particular relates to an automatic emergency steering control method taking driver comfort into consideration. Background Art

[0002] The automatic emergency steering system belongs to the field of active safety assisted driving. It is generally based on the detection and recognition functions of millimeter-wave radar, camera and corner radar. When a dangerous target is detected and there is a risk of collision, the automatic emergency braking system's deceleration and braking cannot avoid the collision. The vehicle prompts the driver to turn to avoid the collision through sound, greatly reducing the risk of rear-end collision.

[0003] Through the study of a large number of traffic accidents, it is found that emergency braking alone cannot avoid collisions in some accident cases. When a pedestrian suddenly crosses the road in front of the vehicle while the vehicle is driving at high speed, the vehicle adopts automatic emergency braking. The sharp change in speed in a short period of time can reduce the risk of collision, but there is still a possibility of injuring pedestrians. Many current studies have shown that in these working conditions where emergency braking cannot avoid collisions, the driver can still avoid collisions through emergency steering. This is because the latest turning point of emergency steering is later than the latest braking point, which can further reduce the risk of collision. Emergency steering often brings a large lateral acceleration. On the premise of ensuring successful obstacle avoidance, how to reduce the driver's discomfort will become the focus of future research. Most control strategies only consider the robustness and rapid response of control, without considering reducing the driver's discomfort. Summary of the invention

[0004] The purpose of the present invention is to provide an automatic emergency steering control method taking into account the driver's comfort, reduce the lateral acceleration of emergency steering, ensure the vehicle's safe obstacle avoidance, and alleviate the driver's discomfort.

[0005] The technical solution of the present invention is:

[0006] An automatic emergency steering control method considering driver comfort comprises the following steps:

[0007] Step (1): Obtain vehicle information through on-board sensors, including vehicle speed, lateral and longitudinal acceleration of the vehicle, etc.

[0008] Step (2): Obtain information about dangerous vehicles in the left, middle and right lanes ahead through sensors such as cameras, millimeter-wave radars and corner radars;

[0009] Step (3): Based on the dangerous vehicle information of the left, middle and right lanes obtained in step (2), the dangerous target with the smallest TTC (time to collision), the smallest relative longitudinal distance or the smallest relative lateral distance is determined as the main dangerous target, and the speed, acceleration, relative lateral and longitudinal distances of the main dangerous target, as well as the predicted trajectory points of the main dangerous target within a certain period of time in the future are output;

[0010] Step (4): Based on the main dangerous target information obtained in step (3), according to the position of the vehicle, a planning method combining quintic polynomial and scattered points is used to plan an emergency turning path cluster, and a collision check method is used to determine whether the turning path collides with the main dangerous target, and the path with the smoothest curvature on the outermost side is selected as the optimal path;

[0011] Step (5): Further, according to the current information of the vehicle, longitudinal speed planning is added to design a speed cluster whose speed decreases over time. It is specifically divided into two parts:

[0012] Part 1: Process the speed, acceleration and other information at the beginning and end of speed planning. Considering that the maximum speed is set as the starting speed and the first speed curve has acceleration, the maximum speed is processed. The following formula is the maximum speed and minimum speed of the planning end point. The speed interval is calculated according to the number of sampling curves, and the speed of a series of planning end points is calculated cyclically:

[0013] v endmin =v start +TTC·a max ;

[0014] v endmax =v start +k·TTC·a max ;

[0015] In the formula, v endmin is the minimum speed at the planning end point, v endmax is the maximum speed of the planned endpoint, v start is the initial speed at the planning starting point, k is the adjustment coefficient, TTC is the collision time, a max It is the maximum deceleration that the vehicle can achieve without skidding.

[0016] Part 2: Calculate the quartic polynomial. Based on the initial position, initial velocity, initial acceleration, a series of terminal velocities, and terminal accelerations, the quartic polynomial is calculated. The specific solution process is shown in the following formula:

[0017] x=C0+C1·t+C2·t 2 +C3·t 3 +C4·t 4 ;

[0018] v=C1+2·C2·t+3·C3·t 2 +4·C4·t 3 ;

[0019] a=2·C2+6·C3·t+12·C4·t 2 ;

[0020] Substituting the above information, we can solve for the values ​​of C0, C1, C2, C3, and C4:

[0021] C0=x start ;

[0022] C1=v start ;

[0023]

[0024]

[0025]

[0026] In the formula, x start is the initial position, v start is the initial velocity, a start is the initial acceleration, v endi is a series of terminal velocities, a end is the terminal acceleration;

[0027] According to the collision checking method, the curve with the smallest speed drop is selected as the optimal longitudinal speed planning curve.

[0028] Step (6): Next, the optimal path of step (4) is lateral controlled;

[0029] For tracking the optimal path, two pre-aiming points, near and far, are selected in the optimal path. The lateral distance deviation of the far and near points can be obtained by substituting the quintic polynomial of the optimal path. The heading angle deviation can be obtained by taking the derivative of the quintic polynomial. The deviation at the far point is mainly processed by the steady-state yaw rate gain, and the deviation at the near point is controlled by the PID control strategy. The two are superimposed to obtain the final steering wheel angle. The specific method is as follows:

[0030] 1) Calculate the radius of curvature R at the far point:

[0031]

[0032] Where R is the radius of curvature at the far point, y far is the lateral distance deviation, x far is the relative vertical distance.

[0033] Because the steady-state yaw rate gain formula is:

[0034]

[0035] In the formula, w r is the steady-state yaw rate, v x is the longitudinal speed of the vehicle, L is the wheelbase, δ is the front wheel turning angle, m is the vehicle mass, a is the distance from the front axle to the center of mass, b is the distance from the rear axle to the center of mass, k2 is the rear wheel cornering stiffness, k1 is the front wheel cornering stiffness;

[0036] According to the above formula, the steering wheel angle θ corresponding to the far point can be solved far :

[0037]

[0038] Where k is the coefficient from the front wheel angle to the steering wheel angle, θ far is the steering wheel angle.

[0039] 2) Calculate the lateral velocity error v at the near point error :

[0040] v error = k·x near -v x sinθ loaclhead ;

[0041] In the formula, v error is the lateral velocity error, k is the lateral distance deviation to lateral velocity coefficient, θ loaclhead is the heading angle deviation;

[0042] Next, the PID control strategy is used to obtain the lateral acceleration a at the near point. lat :

[0043] a lat =k p ·v error +k i ·v error +k d ·v error ;

[0044] Then the steering wheel angle θ at the near point can be calculated near :

[0045]

[0046] The final steering wheel angle is obtained by superimposing the steering wheel angles of the far point and the near point. last :

[0047] θ last =w1·θ near +w2·θ far ;

[0048] Where w1 and w2 are the weight coefficients of the near point and the far point.

[0049] Taking into account that directly using the superimposed steering wheel angle will result in excessive lateral acceleration and the driver's driving experience will be very uncomfortable, the weight coefficients w1 and w2 are allocated and adjusted.

[0050] The specific allocation is based on the vehicle speed. The higher the vehicle speed, the higher the corresponding far point weight coefficient w2 should be, that is, the farther the preview is, the larger the steering wheel angle at the far point should be. The lower the vehicle speed, the higher the corresponding near point weight coefficient w1 should be, that is, the closer the preview is, the larger the steering wheel angle at the near point should be.

[0051] Step (7): Next, the optimal longitudinal speed planning curve of step (5) is longitudinally controlled.

[0052] Longitudinal control is specifically divided into two parts, one is the calculation of acceleration in the feedforward part, and the other is the calculation of expected acceleration in the feedback part.

[0053] (1) Feedforward acceleration calculation:

[0054] The calculation of the feedforward acceleration is based on the driving resistance equation. The specific formula is:

[0055] ∑F=F i +F w +F f +F j ;

[0056] In the formula, ∑F is the sum of resistance, F i is the slope resistance, F w is the air resistance, F f is the rolling resistance, F j is the acceleration resistance:

[0057] F i = mgi;

[0058]

[0059] Ff = mgf;

[0060]

[0061] Where i is the road slope, f is the rolling resistance coefficient, m is the total mass of the car (kg), and g is the acceleration due to gravity, 9.81 m / s 2 , C D is the air resistance coefficient; A is the windward area (m 2 ), v a is the vehicle speed (km / h), δ is the vehicle rotation mass conversion coefficient, is the driving acceleration.

[0062] The feed-forward acceleration is calculated based on the sum of the driving resistances calculated above.

[0063]

[0064] In the formula, a feedforward is the feedforward acceleration, ∑F is the sum of the four resistances calculated above, and m is the vehicle mass.

[0065] (2) Calculation of expected acceleration:

[0066] The expected acceleration is obtained by selecting PID control based on the error between the optimal longitudinal speed planning speed and the vehicle speed, as shown in the following formula:

[0067] a des =k p ·(v ref -v x )+k i ·(v ref -v x )+k d ·(v ref -v x );

[0068] The final acceleration is obtained by adding the feedforward acceleration and the desired acceleration, and is sent to the controller simultaneously with the steering wheel angle just obtained, thus achieving complete lateral and longitudinal control.

[0069] The advantages of the present invention are:

[0070] The present invention proposes an automatic emergency steering control method that takes into account the driver's comfort, formulates a weight coefficient scheme for the near and far point steering wheel angle distribution, designs an optimal longitudinal speed planning method and establishes a mathematical expression. According to the actual situation, selecting an accurate distribution weight coefficient and a suitable longitudinal speed curve can reduce the lateral acceleration of emergency steering, ensure the vehicle's safe obstacle avoidance, and reduce the driver's discomfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The present invention will be further described below in conjunction with the accompanying drawings and embodiments:

[0072] Figure 1 The figure is a flow chart of the automatic emergency steering control method of the present invention. DETAILED DESCRIPTION

[0073] like Figure 1 As shown, the automatic emergency steering control method considering driver comfort proposed by the present invention comprises the following steps:

[0074] S1: Obtain vehicle information through on-board sensors, including vehicle speed, lateral and longitudinal acceleration, etc.

[0075] S2: Obtain information about dangerous vehicles in the three lanes ahead, left, middle and right, through sensors such as cameras, millimeter-wave radars and corner radars;

[0076] S3: Based on the dangerous vehicle information of the left, middle and right lanes obtained in S2, the dangerous target with the smallest TTC (collision time), the smallest relative longitudinal distance or the smallest relative lateral distance is determined as the main dangerous target, and the speed, acceleration, relative lateral and longitudinal distance of the main dangerous target, and the predicted trajectory points of the main dangerous target within a certain period of time in the future are output;

[0077] S4: Based on the main dangerous target information obtained in S3 and the position of the vehicle, the emergency turning path cluster is planned by combining the fifth-order polynomial and the scattered point planning method. The collision check method is used to determine whether the turning path collides with the main dangerous target, and the path with the smoothest curvature on the outermost side is selected as the optimal path.

[0078] S 5: Based on the current information of the vehicle, add longitudinal speed planning and design a quartic polynomial speed cluster, which is divided into two parts:

[0079] Part 1: Process the speed, acceleration and other information at the beginning and end of speed planning. Calculate the speed interval according to the number of sampling curves, and calculate the speed of a series of planning end points in a loop:

[0080] v endmin =v start +TTC·a max ;

[0081] v endmax =v start +k·TTC·a max ;

[0082] In the formula, v endmin is the minimum speed at the planning end point, v endmax is the maximum speed of the planned endpoint, v start is the initial speed at the planning starting point, k is the adjustment coefficient, TTC is the collision time, a max It is the maximum deceleration that the vehicle can achieve without skidding.

[0083] Part 2: Calculate the quartic polynomial. Based on the initial position, initial velocity, initial acceleration, a series of terminal velocities, and terminal accelerations, the quartic polynomial is calculated. The specific solution process is shown in the following formula:

[0084] x=C0+C1·t+C2·t 2 +C3·t 3 +C4·t4 ;

[0085] v=C1+2·C2·t+3·C3·t 2 +4·C4·t 3 ;

[0086] a=2·C2+6·C3·t+12·C4·t 2 ;

[0087] Substituting the above information, we can solve for the values ​​of C0, C1, C2, C3, and C4.

[0088] C0=x start ;

[0089] C1=v start ;

[0090]

[0091]

[0092]

[0093] In the formula, x start is the initial position, v start is the initial velocity, a start is the initial acceleration, v endi is a series of terminal velocities, a end is the terminal acceleration;

[0094] The curve with the smallest speed drop is selected as the optimal longitudinal speed planning curve.

[0095] S 6: For tracking the optimal path, two pre-aiming points, near and far, are selected in the optimal path. The lateral distance deviation of the far and near points can be obtained by substituting the quintic polynomial of the optimal path. The heading angle deviation can be obtained by taking the derivative of the quintic polynomial. The deviation at the far point is mainly processed by the steady-state yaw rate gain, and the deviation at the near point is controlled by the PID control strategy. The two are superimposed to obtain the final steering wheel angle. The specific method is as follows:

[0096] 1) Calculate the radius of curvature R at the far point:

[0097]

[0098] Where R is the radius of curvature at the far point, y far is the lateral distance deviation, x far is the relative vertical distance.

[0099] The steady-state yaw rate gain formula is:

[0100]

[0101] In the formula, w r is the steady-state yaw rate, v x is the longitudinal velocity of the vehicle, L is the wheelbase, δ is the front wheel turning angle, m is the vehicle mass, a is the distance from the front axle to the center of mass, b is the distance from the rear axle to the center of mass, k2 is the rear wheel cornering stiffness, and k1 is the front wheel cornering stiffness.

[0102] According to the above formula, the steering wheel angle θ corresponding to the far point can be solved far :

[0103]

[0104] Where k is the coefficient from the front wheel angle to the steering wheel angle, θ far is the steering wheel angle.

[0105] 2) Calculate the lateral velocity error v at the near point error :

[0106] v error = k·x near -v x sinθ loaclhead ;

[0107] In the formula, v error is the lateral velocity error, k is the lateral distance deviation to lateral velocity coefficient, θ loaclhead is the heading angle deviation.

[0108] Next, the PID control strategy is used to obtain the lateral acceleration a at the near point. lat :

[0109] a lat =k p ·v error +k i ·v error +k d ·v error ;

[0110] Then the steering wheel angle θ at the near point can be calculated near :

[0111]

[0112] The final steering wheel angle is obtained by superimposing the steering wheel angles of the far point and the near point. last :

[0113] θ last =w1·θ near +w2·θ far ;

[0114] Where w1 and w2 are the weight coefficients of the near point and the far point.

[0115] Taking into account the situation that directly using the superimposed steering wheel angle will lead to excessive lateral acceleration, the weight coefficients w1 and w2 are allocated and adjusted.

[0116] The specific allocation is based on the vehicle speed. The higher the vehicle speed, the higher the corresponding far point weight coefficient w2 should be, that is, the farther the preview is, the larger the steering wheel angle at the far point should be. The lower the vehicle speed, the higher the corresponding near point weight coefficient w1 should be, that is, the closer the preview is, the larger the steering wheel angle at the near point should be.

[0117] S 7: Longitudinal control is specifically divided into two parts, one is the calculation of acceleration in the feedforward part, and the other is the calculation of expected acceleration in the feedback part.

[0118] (1) Feedforward acceleration calculation:

[0119] The calculation of the feedforward acceleration is based on the driving resistance equation. The specific formula is:

[0120] ∑F=F i +F w +F f +F j ;

[0121] In the formula, ∑F is the sum of resistance, F i is the slope resistance, F w is the air resistance, F f is the rolling resistance, F j is the acceleration resistance.

[0122] Fi = mgi;

[0123]

[0124] Ff = mgf;

[0125]

[0126] Where i is the road slope, f is the rolling resistance coefficient, m is the total mass of the car (kg), .g is the acceleration of gravity, 9.81m / s2, CD is the air resistance coefficient; A is the frontal area (m2), va is the vehicle speed (km / h), δ is the vehicle rotation mass conversion coefficient, is the driving acceleration.

[0127] The feed-forward acceleration is calculated based on the sum of the driving resistances calculated above.

[0128]

[0129] In the formula, afeedforward is the feedforward acceleration, ∑F is the sum of the four resistances calculated above, and m is the vehicle mass.

[0130] (2) Calculation of expected acceleration:

[0131] The expected acceleration is to select PID control based on the error between the optimal longitudinal speed planning speed and the vehicle speed to obtain the final expected acceleration, as shown in the following formula:

[0132] a des =k p ·(v ref -v x )+k i ·(v ref -v x )+k d ·(v ref -v x );

[0133] The final acceleration is obtained by adding the feedforward acceleration and the desired acceleration, and is sent to the controller simultaneously with the steering wheel angle of the S6, thus achieving complete lateral and longitudinal control.

[0134] According to the automatic emergency steering control method considering driver comfort described in the present invention, the entire algorithm includes three parts: collecting vehicle information, optimal path and optimal longitudinal speed curve planning, lateral path tracking control and longitudinal speed following control. First, the vehicle-mounted sensor collects the vehicle information and dangerous target information, and then plans the optimal path based on the fifth-order polynomial and the optimal longitudinal speed curve based on the fourth-order polynomial according to the vehicle information, and finally designs the lateral control strategy to track the optimal path and the longitudinal control strategy to follow the optimal longitudinal speed curve.

[0135] In summary, the present invention is based on an automatic emergency steering system. In order to improve the driver's comfort, an optimal longitudinal speed curve based on a quartic polynomial is proposed, and a near and far point steering wheel angle distribution strategy is designed. On the premise of ensuring obstacle avoidance stability, the driver's comfort is improved and the impact of large lateral acceleration on the driver is reduced.

[0136] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any modifications made according to the spirit of the main technical solution of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic emergency steering control method considering driver comfort, characterized in that: The steps include: Step (1): The vehicle-mounted sensor obtains the vehicle information, including the vehicle speed and the lateral and longitudinal acceleration information of the vehicle; Step (2): Obtain information about dangerous vehicles in the left, middle and right lanes ahead through cameras, millimeter-wave radars and corner radar sensors; Step (3): according to the dangerous vehicle information of the left, middle and right lanes obtained in step (2), the dangerous target with the shortest collision time, the shortest relative longitudinal distance or the shortest relative lateral distance is determined as the main dangerous target, and the information of the main dangerous target is outputted; Step (4): Based on the main dangerous target information obtained in step (3), according to the vehicle position, a planning method combining quintic polynomial and scattering is used to plan an emergency turning path cluster, and a collision check method is used to determine whether the turning path collides with the main dangerous target. The path with the smoothest outer curvature is selected as the optimal path. Step (5): Based on the current information of the vehicle, add longitudinal speed planning, design a speed cluster with a speed decrease over time, and select the curve with the smallest speed change as the optimal longitudinal speed curve; Step (6): performing lateral path following control according to the optimal path obtained in step (4); Step (7): Perform longitudinal speed following control according to the optimal longitudinal speed curve obtained in step (5).

2. The automatic emergency steering control method according to claim 1, characterized in that: The specific steps of planning the optimal longitudinal speed curve in step (5) are as follows: Process the speed and acceleration information at the beginning and end of speed planning, calculate the speed interval according to the number of sampling curves, and calculate the speed of a series of planning end points in a loop: v endmin =v start +TTC·a max ; v endmax =v start +k·TTC·a max ; In the formula, v endmin is the minimum speed at the planning end point, v endmax is the maximum speed of the planned endpoint, v start is the initial speed at the planning starting point, k is the adjustment coefficient, TTC is the collision time, a max The maximum deceleration that the vehicle can achieve without skidding; Compute the quartic polynomial for the velocity plan: x=C0+C1·t+C2·t 2 +C3·t 3 +C4·t 4 ; v=·C1+2·C2·t+3·C3·t 2 +4·C4·t 3 ; a=2·C2+6·C3·t+12·C4·t 2 ; Substitute the initial position, initial velocity, initial acceleration, a series of terminal velocities, and terminal accelerations to solve for the coefficients of the quartic polynomial: C0=x start ; C1=v start ; In the formula, x start is the initial position, v star is the initial velocity, a start is the initial acceleration, v endi is a series of terminal velocities, a end is the terminal acceleration; According to the collision checking method, the curve with the smallest speed drop is selected as the optimal longitudinal speed planning curve.

3. The automatic emergency steering control method according to claim 1, characterized in that: The specific steps of the lateral path following control in step (6) are: (1) Calculate the radius of curvature R at the far point: Where R is the radius of curvature at the far point, y far is the lateral distance deviation, x far is the relative longitudinal distance; Then solve the steering wheel angle θ at the far point according to the steady-state yaw rate gain far : Where k is the coefficient from the front wheel angle to the steering wheel angle, θ far is the steering wheel angle; (2) Calculate the lateral velocity error v at the near point error : v error =k x near -v x ·sinθ loaclhead ; In the formula, v error is the lateral velocity error, k is the lateral distance deviation to lateral velocity coefficient, θ loacl h ead is the heading angle deviation; Use PID control strategy to get the lateral acceleration a at the near point lat : a lat =k p ·v error +k i ·v error +k d ·v error ; Then the steering wheel angle θ at the near point can be calculated near : The final steering wheel angle is obtained by superimposing the steering wheel angles of the far point and the near point. last : i last =w1·θ near +w2·θ far ; Where w1 and w2 are the weight coefficients of the near point and the far point.

4. The automatic emergency steering control method according to claim 3, characterized in that: The weight coefficients w1 and w2 are adjusted based on the speed of the vehicle: when the speed of the vehicle is higher, the corresponding far point weight coefficient w2 should be higher, that is, the farther the preview is, the larger the steering wheel angle at the far point should be; when the speed of the vehicle is lower, the corresponding near point weight coefficient w1 should be higher, that is, the closer the preview is, the larger the steering wheel angle at the near point should be.

5. The automatic emergency steering control method according to claim 1, characterized in that: The specific steps of the longitudinal speed following control in step (7) are: (1) Calculate the feedforward acceleration: The calculation of the feedforward acceleration is based on the driving resistance equation. The specific formula is: ∑F=F i +F w +F f +F j ; In the formula, ∑F is the sum of resistance, F i is the slope resistance, F w is the air resistance, F f is the rolling resistance, F j is the acceleration resistance: F f =mgf; Where i is the road slope, f is the rolling resistance coefficient, m is the total mass of the car (kg), g is the acceleration of gravity, and C D is the air resistance coefficient; A is the windward area (m 2 ), v a is the vehicle speed (km / h), δ is the vehicle rotation mass conversion coefficient, is the driving acceleration; The feedforward acceleration is calculated based on the sum of the driving resistance calculated above: In the formula, a feedforward is the feedforward acceleration, ∑F is the sum of the four resistances calculated above, and m is the vehicle mass; (2) Calculate the expected acceleration: According to the error between the optimal longitudinal speed planning speed and the vehicle speed, PID control is selected to obtain the final expected acceleration, as shown in the following formula: a des =k p ·(v ref -v x )+k i ·(v ref -v x )+k d ·(v ref -v x ); The final acceleration is obtained by adding the feedforward acceleration and the desired acceleration.

Images