A speed planning method for autonomous driving vehicle during braking
By simulating the segmented deceleration method of human drivers, the problem of unsmooth speed planning during braking of autonomous vehicles is solved, and smoother deceleration changes and higher driving comfort are achieved.
Patent Information
- Application Number
- CN202210170409.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-02-24
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-02-24
AI Technical Summary
The existing speed planning method during braking of autonomous vehicles relies on the precise information output by the perception module, resulting in unsmoothing speed and deceleration, and there are mutations, affecting driving comfort.
Using a segmented deceleration method that simulates the deceleration habits of human drivers, the time and speed information of each deceleration stage are calculated by setting multiple braking stages and corresponding deceleration and time proportions, and the results are smoothed to ensure deceleration. Continuity of curves.
It effectively avoids frequent changes in deceleration during braking, improves the smoothness of speed and deceleration, reduces the tracking difficulty of the control module, and improves the driving comfort of the vehicle.
Smart Images

Figure CN114537433B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of autonomous driving, and in particular to a speed planning method for the braking process of an autonomous vehicle. Background Art
[0002] During the driving process of an autonomous vehicle, if a target point that requires the autonomous vehicle to brake appears, such as a moving (static) obstacle in front, a traffic light intersection, or when reaching the destination, the speed planning module will plan the expected speed, expected deceleration, and other information based on the current vehicle speed, target point distance, target point speed, and other information. The control module will then calculate the deceleration to be executed and send it to the vehicle's braking system for braking. Currently, some speed planning methods for autonomous vehicles during braking rely more on the perception module outputting accurate target speed, distance, and other information, and then combining it with the vehicle's expected driving path to perform speed planning through a dynamic planning algorithm. Figure 1 However, during actual driving, due to factors such as sensor measurement errors and vehicle braking system response errors, the planned expected speed and deceleration may not be smooth and may have sudden changes, thus affecting driving comfort.
[0003] Chinese invention patent application "Speed planning guide line for gentle deceleration" (application number: 202010475869.7, publication date: 2021.07.02) discloses a method, apparatus and system for planning deceleration trajectory based on the natural deceleration scheme of an automated vehicle (ADV). Based on the current speed of the ADV, a speed guide is generated based on a predetermined natural deceleration scheme associated with the ADV. Speed planning is achieved by optimizing the total cost function through the speed guide to determine the speed of multiple trajectory points along the planned trajectory. Chinese invention patent application "A method, electronic device and storage medium for deceleration and parking of an autonomous driving system" (application number: 202010292448.0, publication date: 2020.08.18) provides a method, electronic device and storage medium for deceleration and parking of an autonomous driving system. The method includes: establishing discrete sets of vehicle speed, distance and deceleration respectively; constructing an input-output sample set of a neural network model based on the discrete sets of vehicle speed, distance and deceleration, and training the neural network model offline using the input-output sample set; obtaining the real-time speed of the current vehicle and the real-time distance to the target position, inputting the real-time speed and real-time distance into the trained neural network model, and sending the real-time calculated deceleration to the driving control module for braking and deceleration. Chinese invention patent application "A method and apparatus for speed planning of autonomous driving" (application number: 202110351230.2, publication date: 2021.06.15) discloses a method and apparatus for speed planning of autonomous driving, relating to the field of autonomous driving system technology. The method includes updating the speed curve from the current moment to the target moment in real time based on the current vehicle pose information, the behavior instructions output by the autonomous driving system, and the trajectory information output by the trajectory planning layer; configuring the speed and acceleration at the current moment and the speed and acceleration at the target moment, and generating speed curves under different configurations based on a polynomial speed curve algorithm, and using the generated multiple speed curves as candidate speed curves; solving for the parameters of each candidate speed curve by differentiation, the parameters including real-time speed, real-time acceleration, and real-time jerk; and filtering the candidate speed curves based on vehicle kinematic constraints and the parameters of the candidate speed curves to obtain the optimal speed curve. Chinese invention patent application "Speed Planning Method, Apparatus, Electronic Device, and Storage Medium" (Application No.: 202011526407.X, Publication Date: 2021.03.19) discloses a speed planning method, apparatus, electronic device, and storage medium, relating to the fields of autonomous driving and artificial intelligence. Based on candidate speed trajectories of a target vehicle, the method obtains the position information of the target vehicle at least at one moment within a planning period; based on the position information of at least one target obstacle and the position information of the target vehicle at at least one moment, it determines the passage space of the target vehicle at at least one moment; if the passage space at at least one moment meets preset conditions, the candidate speed trajectory is determined as the speed trajectory of the target vehicle.
[0004] Regarding the above-mentioned public technologies: Existing speed planning methods for autonomous driving vehicles during braking mostly use cost functions or neural network model training methods to obtain the planned speed curve. The configuration parameters involved are numerous and difficult to adjust to the optimal value, and are highly dependent on the perception results output by the perception module.
[0005] The present invention simulates the deceleration habits of human drivers and adopts segmented deceleration based on a kinematic model. The speed and acceleration information of each deceleration stage are calculated according to the current vehicle speed, and the obtained speed and acceleration information are smoothed to achieve safe and comfortable parking of the vehicle. Summary of the Invention
[0006] The purpose of the present invention is: for the braking conditions of an autonomous vehicle, the present invention provides a speed planning method during the braking process of an autonomous vehicle, thereby avoiding frequent changes in deceleration during braking, improving the smoothness of the expected speed and deceleration, reducing the tracking difficulty of the control module, and thus improving vehicle driving comfort.
[0007] The technical solution of the present invention is:
[0008] A speed planning method for the braking process of an autonomous vehicle includes:
[0009] Set n braking stages, n≥2, and set the deceleration value (A1, A2...A1) for each braking stage. n ), and the duration of each braking phase (P1, P2...P n );
[0010] When approaching the destination or sensing an obstacle ahead, the autonomous vehicle plans its speed during braking, based on the set safe distance S between the final stopping position and the destination or obstacle. min Given the current speed V of the autonomous vehicle, calculate the duration of each deceleration phase and the distance S between the autonomous vehicle and the destination or obstacle when braking begins.
[0011] Preferably, the method for calculating the duration of each deceleration phase includes:
[0012] S01. Assume that the duration of each deceleration phase is T1, T2...T n ;
[0013] P1+P2+...+P n =1 (Formula 1.1)
[0014] From (Equation 1.1) we can get:
[0015] T2=T1*(P2 / P1) (Formula 1.2) ...
[0017] T n =T1*(P n / P1) (Formula 1.n)
[0018] S02, assuming that the initial speed of each deceleration stage is V1, V2...V n ;
[0019] V1=V (Equation 2.1)
[0020] V2=V1+A1*T1 (Formula 2.2) ...
[0022] V n =V n-1 +A n-1 *T n-1 (Formula 2.n)
[0023] V n =-A n *T n (Formula 2.n+1)
[0024] Combining (Equation 2.1) to (Equation 2.n+1) yields:
[0025] V+A1*T1+A2*T2+...+A n-1 *T n-1 =-A n *T n (Formula 2.n+2)
[0026] Combining (Equation 1.2) to (Equation 1.n) and (Equation 2.n+2) yields:
[0027] V+A1*T1+A2*T1*(P2 / P1)+...+A n *T1*(P n / P1)=0 (Formula 2.n+3)
[0028] S03, calculate the duration of each deceleration stage T1, T2...T n ;
[0029] The value of T1 is calculated by (Formula 2.n+3):
[0030] T1=-V / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) (Equation 3.1)
[0031] The duration of each deceleration stage can be calculated from (Equation 3.1) and (Equation 1.2) to (Equation 1.n):
[0032] T2=-V*(P2 / P1) / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) (Equation 3.2) ...
[0034] T n =-V*(P n / P1) / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) (Formula 3.n).
[0035] Preferably, the method for calculating the distance S between the autonomous driving vehicle and the end point or obstacle when braking begins includes:
[0036] S04, calculate the initial speed V1, V2...V of each deceleration stage n ;
[0037] From (Equation 2.2) we get
[0038] V1 = V (Equation 4.1)
[0039] Combining (Equation 2.2), (Equation 3.1), and (Equation 4.1) yields:
[0040] V2=V1-A1*V / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) (Equation 4.2) ...
[0042] Combining equations (2.n+1), (3.n), and (4.1), we obtain:
[0043] V n =A n *V*(P n / P1) / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) (Equation 4.n)
[0044] S05, calculate the braking distance S1, S2...S in each deceleration stage n ;
[0045] S1 = V1 * T1 + 0.5 * A1 * T1 2 (Equation 5.1)
[0046] S2=V2*T2+0.5*A2*T2 2 (Equation 5.2) ...
[0048] S n =V n *T n +0.5*A n *T n 2 (Formula 5.n)
[0049] S06. Calculate the distance S between the vehicle and the destination or obstacle when braking begins;
[0050] S1+S2+...+S n =SS min (Equation 6.1)
[0051] By combining equations (3.1) to (3.n), (4.1) to (4.n), (5.1) to (5.n), and (6.1), the value of S can be calculated:
[0052] S=S min +V1*T1+0.5*A1*T1 2 +V2*T2+0.5*A2*T2 2 +...+V n *T n +0.5*A n *T n 2 (Equation 6.2)
[0053] When the distance between the vehicle and the destination or obstacle is less than or equal to the S value calculated by (Equation 6.2) during the vehicle's journey, braking begins.
[0054] Preferably, after calculating the duration of each deceleration stage and the initial deceleration distance, the deceleration curve is also smoothed.
[0055] Discretize the deceleration process using Δt as the unit of time, and calculate the velocity values at all discrete velocity points:
[0056] T m =m*Δt (Equation 7)
[0057] V m =V m -1+A x *Δt (Equation 8)
[0058] Combine (Equation 7) and (Equation 8), and take m = 1, 2, 3, ...;
[0059] Among them, when T m <=T1, A x =A1; when T1 <T m <=T2, Ax =A2,..., when T n-1 <T m <=T n At that time, A x =A n . The velocity values of all velocity discrete points are calculated accordingly;
[0060] Curve fitting is performed on the discrete points of velocity so that the first and second order deceleration curves can be derived, that is, the acceleration and its rate of change are continuous.
[0061] Preferably, when the distance between the vehicle and the destination or obstacle is less than the safety distance S min When the vehicle speed is greater than the set safe speed V safe Then, with the set maximum deceleration A max Apply the brakes.
[0062] Preferably, when the vehicle speed is reduced to a safe vehicle speed V safe When the distance between the vehicle and the destination or obstacle is greater than S min Buffer distance S from the set safety distance buff The sum of the two values will reduce the deceleration.
[0063] Preferably, the curve fitting of the velocity discrete points includes one of polynomial fitting, B-spline curve fitting and Bezier curve fitting.
[0064] The advantages of the present invention are:
[0065] Compared with the existing technology, the speed planning method proposed in the present invention can effectively avoid frequent changes in deceleration during braking, improve the smoothness of the expected speed and deceleration, and reduce the tracking difficulty of the control module; by designing several deceleration stages, setting a reasonable deceleration and deceleration time ratio and debugging, it can effectively improve the driving comfort of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS
[0066] The present invention will be further described below with reference to the accompanying drawings and embodiments:
[0067] Figure 1 This is a principle block diagram of a trajectory planning module in the prior art;
[0068] Figure 2 It is the flow chart of speed planning calculation of the present invention;
[0069] Figure 3 This is a schematic diagram of speed planning configuration parameters in the embodiment;
[0070] Figure 4 This is a schematic diagram of the speed planning calculation results in the embodiment:
[0071] Figure 5Schematic diagram of a bus entering a station in the embodiment;
[0072] Figure 6 is the acceleration curve of each deceleration stage in the embodiment;
[0073] Figure 7 The speed discrete points obtained by discretizing each deceleration stage in the embodiment are:
[0074] Figure 8 The curve obtained by fitting the discrete velocity points using a fifth-order polynomial in the embodiment;
[0075] Figure 9 is the complete planning trajectory obtained in the embodiment. DETAILED DESCRIPTION
[0076] like Figure 2 As shown, the speed planning method for the braking process of an autonomous vehicle in this invention adopts a segmented braking scheme based on a kinematic model. The current vehicle speed V and configured deceleration planning parameters are received in real time. When approaching the destination or sensing an obstacle ahead, the autonomous vehicle performs speed planning during the braking process. The speed is determined based on the set safe distance S between the final stopping position and the destination or obstacle. min Given the current speed V of the autonomous vehicle, the system sequentially calculates the duration of each deceleration phase, the initial speed of each deceleration phase, the braking distance of each deceleration phase, and the distance S between the autonomous vehicle and the destination or obstacle at the start of braking. Braking is triggered when the distance S is reached, and finally, a safe stopping judgment is made.
[0077] For the braking and stopping scenario when the bus arrives at the station, such as Figure 3 As shown, this embodiment designs three deceleration stages (ie, n=3), and the deceleration rate of each stage is -0.4m / s 2 、-0.6m / s 2 、-0.8m / s 2 That is, A1 = -0.5, A2 = -0.8, A3 = -1.0; the duration of each deceleration stage accounts for 50%, 30%, and 20% respectively, i.e., P1 = 0.5, P2 = 0.3, P3 = 0.2; the bus speed is 25 km / h when approaching the stop, i.e., V = 6.944 m / s, such as Figure 4 As shown; the safe distance for buses to arrive at the station is set to 1m, that is, S min =1, such as Figure 5 shown.
[0078] 1. Calculate the duration of each deceleration stage T1, T2, and T3:
[0079] From T1=-V / (A1+A2*(P2 / P1)+...+A n*(P n / P1)) (Equation 3.1) yields:
[0080] T1=-6.944 / (-0.5-0.8*(3 / 5)-1.0*(2 / 5)) (Formula s1)
[0081] According to (Formula s1), T1=5.03 can be calculated.
[0082] From (Equation 1.2) to (Equation 1.n), we can get:
[0083] T2=T1*3 / 5 (Formula s2)
[0084] T3=T1*2 / 5 (Formula s3)
[0085] According to (Formula s2), (Formula s3) and T1=5.03, T2=3.02 and T3=2.01 can be calculated.
[0086] 2. Calculate the initial speeds V1, V2, and V3 of each deceleration stage:
[0087] From (Equation 4.1) to (Equation 4.n), we can obtain:
[0088] V1=V=6.9444 (Formula s4)
[0089] V2=6.9444+0.5*6.9444 / (-0.5-0.8*(3 / 5)-1.0*(2 / 5)) (Formula s5)
[0090] V3=-1.0*6.9444*(2 / 5) / (-0.5-0.8*(3 / 5)-1.0*(2 / 5)) (Formula s6)
[0091] According to (Formula s4), (Formula s5), (Formula s6) and T1, T2, and T3, V2 = 4.43 and V3 = 2.01 can be calculated.
[0092] 3. Calculate the distance S between the vehicle and the end point or obstacle when braking begins:
[0093] From (Equation 6.2) and the above calculated T1, T2, T3, V1, V2, V3, we can get:
[0094] S1=6.944*5.03-0.5*0.5*5.03^2=28.615 (Formula s7)
[0095] S2=4.43*3.01-0.5*0.8*3.02^2=9.724 (Formula s8)
[0096] S3=2.01*2.01-0.5*1.0*2.01^2=2.026 (Formula s9)
[0097] S=1+6.944*5.03-0.5*0.5*5.03^2+4.43*3.01-0.5*0.8*3.02^2+2.01*2.01-0.5*1.0*2.01^2 (Formula s10)
[0098] According to (Formula s10), S=42.365 can be calculated.
[0099] Therefore, braking is initiated when the vehicle is 42.365m from the station, at a speed of -0.5m / s². 2 The first stage of deceleration is achieved by continuous braking for 5.03s at a deceleration rate of -0.8m / s. The braking distance is 28.615m. 2 The deceleration rate is continuously braked for 3.02s to achieve the second stage of deceleration, and the braking distance is 9.724m; at -1.0m / s 2 The vehicle continued braking for 2.01s to achieve the third stage of deceleration, with a braking distance of 2.026m; and finally stopped at the station.
[0100] From this we can get the acceleration curve, such as Figure 6 shown.
[0101] According to (Equation 7) and (Equation 8), each deceleration stage is discretized to obtain the speed discrete points, as shown in Figure 7 shown.
[0102] The velocity discrete points are fitted using a 5th-order polynomial, such as... Figure 8 shown.
[0103] The speed curve polynomial coefficients are calculated as:
[0104] -1.2482e-05, 0.00087, -0.01618, 0.07897, -0.61574, 6.97698;
[0105] That is, the speed curve is:
[0106] V=-1.2482e -5 *t 5 +0.00087*t^ 4 -0.01618*t^ 3 +0.07897*t^ 2 -0.61574*t+6.97698;
[0107] The acceleration curve equation can be calculated by taking the derivative of the velocity curve polynomial:
[0108] A = -6.2412e -5 *t 4 +0.003504*t 3 -0.04854*t 2 +0.15794*t-0.61574;
[0109] By discretely sampling the above velocity curve and acceleration curve and fusing the information with the planned path, the complete planned trajectory can be obtained, as shown in Figure 9 As shown, the planned trajectory contains several discrete trajectory points, each containing velocity and acceleration values, which are obtained by discretely sampling the velocity and acceleration curves at the end of the calculation process.
[0110] Due to the smooth deceleration curve or the unstable speed tracking of the subsequent control module, excessive braking may occur (the distance between the vehicle and the end point or obstacle when it stops is much greater than the safety distance S min ), insufficient braking (the distance between the vehicle and the end point or obstacle when it stops is much less than the safe distance S min ) and other issues, which may affect the driving experience and even cause collision risks. Therefore, the safe parking judgment is set as follows:
[0111] a. When the distance between the vehicle and the destination or obstacle is less than the safety distance S min When the vehicle speed is greater than the set safe speed V safe Then, with the set maximum deceleration A max Apply brakes;
[0112] b. When the vehicle speed is reduced to the safe speed V safe When the distance between the vehicle and the destination or obstacle is greater than S min Buffer distance S from the set safety distance buff The sum of , reduces the deceleration.
[0113] It can stop the vehicle safely at the end point or behind an obstacle and maintain a safe distance.
[0114] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All modifications made according to the spirit and essence of the main technical solution of the present invention should be covered within the scope of protection of the present invention.
Claims
1. A speed planning method for an autonomous driving vehicle during braking, characterized in that: include: Set n braking stages, n ≥ 2, and set the deceleration value A1, A2...A in each braking stage n , and the duration of each braking phase accounts for P1, P2...P n ; When the vehicle is about to reach the destination or senses an obstacle ahead, the autonomous vehicle plans the speed during braking, based on the safety distance S between the final stop position and the destination or obstacle. min , and the current speed V of the autonomous vehicle, calculate the duration of each deceleration stage, and the distance S between the autonomous vehicle and the end point or obstacle when braking begins; The method for calculating the duration of each deceleration stage includes: S01, assuming that the duration of each deceleration stage is T1, T2...T n ; P1+P2+...+P n =1 Formula (1.1) From formula (1.1), we can get: T2=T1*(P2 / P1) Formula (1.2) ... T n =T1*(P n / P1) Formula (1.n) S02, assuming that the initial speed of each deceleration stage is V1, V2...V n ; V1=V Formula (2.1) V2=V1+A1*T1 Formula (2.2) ... V n =V n-1 +A n-1 *T n-1 Equation (2.n) V n =-A n *T n Formula (2.n+1) Combining equation (2.1) to equation (2.n+1), we get: V + A1*T1 + A2*T2 +... + A n-1 *T n-1 = -A n *T n Equation (2.n + 2) Combining equation (1.2) to equation (1.n) and equation (2.n+2), we can get: V + A1*T1 + A2*T1*(P2 / P1)+...+ A n *T1*(P n / P1) = 0 Equation (2.n + 3) S03, calculate the duration of each deceleration stage T1, T2...T n ; The value of T1 is calculated by formula (2.n+3): T1=-V / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) Formula (3.1) The duration of each deceleration stage is calculated from equation (3.1) and equation (1.2) to equation (1.n): T2=-V*(P2 / P1) / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) Formula (3.2) ... T n =-V*(P n / P1) / (A1+A2*(P2 / P1)+...+A n *(P n / P1)) Formula (3.n); The method for calculating the distance S between the autonomous driving vehicle and the end point or obstacle when braking begins includes: S04, calculate the initial speed V1, V2...V of each deceleration stage n ; From formula (2.2), we get V1=V Formula (4.1) Combining equations (2.2), (3.1) and (4.1), we can get: V2 = V1 - A1*V / (A1 + A2*(P2 / P1)+...+A n *(P n / P1)) Equation (4.2) ... Combining equation (2.n+1), equation (3.n), and equation (4.1), we can get: V n = A n * V * (P n / P1) / (A1 + A2 * (P2 / P1) +... + A n * (P n / P1)) Equation (4.n) S05, calculate the braking distance S1, S2...S of each deceleration stage n ; S1=V1*T1+0.5*A1*T1 2 Formula (5.1) S2 = V2 * T2 + 0.5 * A2 * T2 2 Equation (5.2) ... S n =V n *T n +0.5*A n *T n 2 Equation (5.n) S06, calculating the distance S between the vehicle and the end point or obstacle when braking begins; S1+S2+...+S n =SS min Formula (6.1) Combining equations (3.1) to (3.n), (4.1) to (4.n), (5.1) to (5.n) and (6.1) to calculate the value of S: S = S min + V1 * T1 + 0.5 * A1 * T1 2 + V2 * T2 + 0.5 * A2 * T2 2 +... + V n * T n + 0.5 * A n * T n 2 Equation (6.2) When the mileage of the vehicle to the end point or obstacle is less than or equal to the S value calculated by formula (6.2), braking begins; After calculating the duration of each deceleration stage and the starting deceleration distance, the deceleration curve is also smoothed: Taking ∆t as the unit time, the deceleration process is discretized and the speed values of all speed discrete points are calculated: T m =m*∆t Formula (7) V m =V m-1 +A x *∆t Equation (8) Combine equations (7) and (8), and take m=1, 2, 3, ...; Among them, when T m <=T1, A x =A1;When T1 <T m <=T2, A x =A2,..., when T n-1 <T m <=T n When A x =A n ; Calculate the speed values of all speed discrete points; Perform curve fitting on discrete speed points to make the first and second order deceleration curves derivativeable, that is, the acceleration and its rate of change are continuous; When the distance between the vehicle and the destination or obstacle is less than the safety distance S min When the vehicle speed is greater than the set safe speed V safe , then the maximum deceleration A is set max Apply brakes; When the vehicle speed is reduced to a safe speed V safe When the distance between the vehicle and the destination or obstacle is greater than S min Buffer distance S from the set safety distance buff The sum of the two reduces the deceleration.
2. The speed planning method during braking of an autonomous driving vehicle according to claim 1, characterized in that: The curve fitting for the velocity discrete points includes one of polynomial fitting, B-spline curve fitting and Bezier curve fitting.
Citation Information
Patent Citations
Automatic driving deceleration parking method, electronic equipment and storage medium
CN111552283A
Speed planning method and device, electronic equipment and storage medium
CN112526999A
Automatic driving speed planning method and device
CN112965501A
Speed planning guidance line for mild slow down
CN113060118A
Public transport interval vehicle speed guiding method
CN111223300A
