CN112660124A - Collaborative adaptive cruise control method for lane change scene - Google Patents

Abstract

The invention relates to a cooperative self-adaptive cruise control method with a lane change auxiliary function, which mainly comprises the following steps: the method comprises the following steps: the method comprises the following steps that a piloted vehicle establishes a track changing track function with obstacle avoidance capacity according to the states before and after the vehicle changes the track and the critical collision condition; step two: the pilot vehicle calculates the optimal track change track of the vehicle under the critical collision condition according to the speeds of the own vehicle and the front vehicle, the geometric parameters of the vehicle body, the road parameters and the like; step three: the fleet tracks the optimal trajectory in accordance with the method of the present invention. The invention mainly has the following three innovation points: (1) providing an optimal track changing track calculation method with a collision avoidance function by combining a polynomial method and an optimization method; (2) the nonlinear queue model is established for the heterogeneous vehicles of the fleet, and compared with the traditional method, the method has the advantages that the fleet is simplified into the linear model and is closer to the reality; (3) the designed method can simultaneously ensure the stability of the transverse and longitudinal queues when the motorcade changes lanes.

Description

Collaborative adaptive cruise control method for lane change scene

Technical Field

The invention relates to a design of an advanced assistant driving system (ADAS), in particular to a design of a cooperative adaptive cruise control system control method (CACC) with a lane change assistant function.

Background

Nowadays, the holding capacity of vehicles is getting larger and larger, and a series of problems such as traffic jam, environmental pollution, frequent traffic accidents and the like are brought along with the holding capacity of vehicles. With the development of communication technology, the V2X technology is receiving more and more attention, wherein the collaborative adaptive cruise control has considerable development potential because it can form vehicles to run cooperatively, thereby greatly reducing energy consumption and accident rate during the running process of the vehicles.

The driving of the vehicle formation needs to meet various requirements, firstly, the safety is required, namely, the vehicles are required not to be collided; comfort, i.e. vehicle acceleration etc. cannot be too great; queue stability, i.e., errors upstream of the fleet do not scale up as they propagate downstream. According to the development of communication technology, vehicles in a fleet can obtain various information of a front vehicle, such as a desired track, a desired speed, an expected acceleration and the like of the front vehicle, and the information is fully utilized, so that the vehicles in the fleet can cooperatively advance, and the high efficiency and the safety of a traffic system are ensured.

At present, many researches on CACC systems are focused on the aspect of single-lane longitudinal driving, for example, summer, sunset and the like take traffic information as an available condition, a model prediction control method is adopted, riding comfort, fuel economy, safety and following performance are taken as targets to be optimized, and a cooperative adaptive cruise system method is designed by adding the restriction of the self capacity of a vehicle.

Compared with the situation of longitudinal movement, the situation of the motorcade is more complex when moving transversely, such as lane changing, overtaking and the like, and the transverse dynamics, the selection of lane changing safe routes, the transverse stability of the single vehicle, the queue stability and the like need to be considered. There is relatively little research currently being done to address this problem. Danruina et al, the Qinghua university, has designed a vehicle multi-target cooperative lane change assisted adaptive cruise control method, which can guarantee the tracking, comfort and safety requirements of the vehicle during lane change under the lane change condition, but only aims at the problem of single-vehicle lane change cruise and is not suitable for the motorcade lane change cruise scene (Chinese patent: CN201410033746.2, a vehicle multi-target cooperative lane change assisted adaptive cruise control method).

For a motorcade, firstly, a smooth lane change track which meets the requirements of safety and comfort needs to be planned in a lane change scene, then, motorcade vehicles track the track to run, and the vehicles in the motorcade can always track the expected track in the running process and always meet the requirements of safety, comfort, asymptotic stability and queue stability in the process.

Disclosure of Invention

The invention mainly solves the problem of team cooperative lane change under the condition of obstacles, and mainly comprises the following points:

firstly, designing a lane change track function with an obstacle avoidance function according to predefined states of a lane change starting part and a lane change ending part aiming at an obstacle lane change scene;

secondly, planning an optimal lane changing track of the pilot vehicle by using an optimization method under the condition of ensuring comfort and safety;

and thirdly, coordinating the actual running state of the motorcade vehicles by using a distributed model predictive control method, wherein the requirements of traceability, comfort, safety, asymptotic stability and queue stability are met.

In order to solve the technical problems, the invention adopts the following technical scheme:

a cooperative self-adaptive cruise control method with a lane change assisting function comprises the following steps:

the method comprises the following steps: the method comprises the following steps that a fleet firstly runs at a constant speed, a pilot vehicle determines the optimal lane changing time according to the longitudinal state of a lane changing starting point and a lane changing finishing point predefined by the vehicle and the road condition, and a lane changing track function with an obstacle avoiding function is designed:

defining the starting time of lane change as t_iThe piloted vehicle state is

The termination time is t_fThe piloted vehicle state is

Critical collision time t_cThe piloted vehicle coordinate is (X)_c,Y_c) And establishing the following track changing track function with the obstacle avoiding function according to a polynomial method:

step two: the method comprises the following steps that a pilot vehicle calculates the optimal lane changing track of the vehicle under a critical collision condition according to the speeds of a self vehicle and a front vehicle, the geometric parameters of the vehicle body, the optimal lane changing time and the like, and an optimization method is adopted to obtain the optimal lane changing track by establishing a cost function with the minimum longitudinal comfort, the minimum transverse comfort and the minimum longitudinal distance during lane changing as an optimization target:

firstly, the optimal track-changing end time t needs to be determined_fAnd the distance dist between the vehicle and the front vehicle when changing the lane:

subject to

Eq.(1)

and then obtaining the optimal track changing tracks X (t) and Y (t) according to the obtained optimal track changing ending time and the optimal vehicle distance.

Step three: a distributed model prediction control method is designed to ensure the driving safety, tracking performance, comfort, gradual stability and queue stability of a motorcade;

the motorcade tracks the optimal track according to a given method, and designs a cooperative adaptive cruise control method with a lane change auxiliary function, which mainly comprises the following steps:

firstly, establishing a local optimal control method of a single vehicle, and ensuring safety, traceability and comfort:

(1): establishing a nonlinear state space model aiming at a single vehicle i in a fleet:

(2): designing a local optimal control method to ensure safety, traceability and comfort:

For i∈{1,2,...,N}at time t

subject to

(3): design distributed model predictive control method

Each vehicle simultaneously solves the corresponding local optimal control problem, and additional constraints are added for other vehicles except the pilot vehicle in the fleet

σ ∈ (x, y) to guarantee queue stability.

The scheme of the invention has the following beneficial effects:

(1) the invention designs an optimal lane changing track aiming at the obstacle lane changing scene, so that the lane changing process can be smooth and comfortable and the safety is ensured;

(2) the local optimal method is established for the single vehicle in the motorcade, so that the tracking performance, the safety and the comfort of the single vehicle in the running process are ensured;

(3) the invention expands the bicycle control method into a distributed model predictive control method suitable for the motorcade, and ensures the asymptotic stability and the queue stability in the driving process of the motorcade.

Drawings

Fig. 1 is a schematic diagram of a lane change in an unobstructed scene.

Fig. 2 is a schematic diagram of lane change in an obstructed scene.

Fig. 3 is a diagram of a fleet topology.

FIG. 4 is a single track bicycle model.

Fig. 5 is a flow chart of method execution.

FIG. 6a is the running track of the fleet of vehicles at the initial speed of 15 m/s;

FIG. 6b is the running track of the fleet at the initial speed of 25 m/s.

FIG. 7a is the longitudinal speed variation of the fleet at the initial lane change speed of 15 m/s;

FIG. 7b is the longitudinal speed variation of the platoon at a lane change initial speed of 25 m/s.

FIG. 8a is the lateral velocity variation of the fleet at lane-change initial speeds of 15m/s (a) and 25m/s (b), respectively;

FIG. 8b shows the lateral velocity variation of the fleet at the lane-change initial speed of 15m/s (a) and 25m/s (b), respectively.

FIG. 9a is the longitudinal position deviation of the fleet at the initial lane change speed of 15 m/s;

FIG. 9b shows the longitudinal position deviation of the fleet at a lane-change initial speed of 25m/s (b).

FIG. 10a shows the lateral position deviation of the fleet at the lane-change initial speed of 15 m/s;

FIG. 10b shows the lateral position deviations of the fleet at initial lane change speeds of 25 m/s.

FIG. 11a shows absolute values of longitudinal position deviations of fleets at a lane change initial speed of 15m/s, respectively;

FIG. 11b shows absolute values of longitudinal position deviations of the fleet at initial lane change speeds of 25 m/s.

FIG. 12a shows the absolute values of lateral position deviations of the fleet at a lane-change initial speed of 15 m/s;

FIG. 12b shows the absolute values of the lateral position deviations of the fleet at a lane change initial speed of 25 m/s.

Detailed Description

The invention provides a cooperative self-adaptive cruise control method with a lane change auxiliary function, which comprises the following steps of: the method comprises the following steps: the method comprises the following steps that a piloted vehicle establishes a track changing track function with obstacle avoidance capacity according to the states before and after the vehicle changes the track and the critical collision condition; step two: the pilot vehicle calculates the optimal track change track of the vehicle under the critical collision condition according to the speeds of the own vehicle and the front vehicle, the geometric parameters of the vehicle body, the road parameters and the like; step three: and designing a distributed model prediction control method to control the fleet to track the planned optimal trajectory for running.

1) Establishing a track-changing track function with obstacle avoidance capability

FIG. 1 is a schematic view of a vehicle lane change without obstacles

L_iAnd L_fRespectively defining the states of the starting time and the ending time of the lane change of the piloting vehicles of the fleet as t_iThe position coordinate of which is (X)_i,Y_i) The termination time is t_fThe position coordinate of which is (X)_f,Y_f). The path planning in the course of changing the path is to find a path from L_iTo L_fThe constraint condition to be satisfied by the curve is the moving state parameters of the vehicle at the lane change starting position and the lane change ending position, namely the lane change initial state

Lane change termination state

The lane change initial state and the lane change termination state have 12 parameters, and the solution of 12 unknowns can be satisfied, so that a fifth-order polynomial can be used to represent the transverse position x (t) and the longitudinal position y (t) of the vehicle, namely:

and substituting the state values before and after the lane change of the vehicle into the formula to determine the lane change track.

Fig. 2 is a schematic view of a lane-changing scene with obstacles. The obstacle in the figure refers to a vehicle with a relatively slow speed in front of the vehicle fleet.

The states of the vehicles with slower speeds in front of the motorcade at the beginning and the end of the corresponding lane change in the definition map are respectively P_iAnd P_f. To avoid a collision and to ensure safety, a critical collision state during lane change is considered, namely at t_cAt the moment, the front right angle of the piloting vehicle of the fleet is the same as the longitudinal position of the rear left angle of the vehicle in front of the fleet, the transverse displacement of the piloting vehicle is the width W of the vehicle in front of the fleet, and the transverse displacement and the longitudinal displacement are monotonically increased when the vehicles change lanes, so that the piloting vehicle only needs to be used at t_cNo collision occurs at all times and no collision occurs later. L is_cAnd P_cRespectively, states of a fleet pilot vehicle and a fleet front vehicle at the critical collision moment. The critical state satisfies the following equation:

where x is_cAnd y_cRespectively carrying out longitudinal displacement and transverse displacement from the initial lane change moment of the pilot vehicle to the critical collision moment;

longitudinal displacement of a vehicle in front of a fleet from an initial lane changing moment to a critical collision moment; w_pAnd L_pRespectively the width and length of the front vehicle of the fleet; dist is the distance between two vehicles at the time of lane change;

for piloting the vehicle at t_cA yaw angle of a moment; h is the half length of the diagonal line of the piloted vehicle; alpha is the included angle between the diagonal line of the piloting vehicle and the longitudinal center line of the vehicle body.

In the actual track change, the transverse velocity is generally much smaller than the longitudinal velocity, so psi_c< α, where the critical collision equation can be rewritten as:

in consideration of the critical collision condition required to be met during vehicle trajectory planning, a certain parameter to be determined needs to be added on the basis of collision-free lane change trajectory planning, and in order to take account of both lane change timeliness and comfort, only the high-order term of the longitudinal motion equation is added. Therefore, the adjustment equation (8) is:

2) calculating optimal track change track of vehicle

The text sets the fleet first

At constant speed for a period of time, at t_iAt time t, the lane change is started_fThe time of lane change is over, and then

And (5) driving at a constant speed for a period of time. The state of the vehicle at the beginning and end of lane change position can be represented as

And

the parameters of the body of the piloted vehicle and of the vehicles ahead of the fleet and the predefined states during the lane change are shown in table 1, where the parameters are known, it is determined that the lane change is to be madeTime of track t_fAnd the inter-vehicle distance dist between the pilot vehicle and the front vehicle, and the following optimal problems are established by taking the two variables as free variables:

subject to

Eq.(12)

where w is₁,w₂,w₃,w₄,w₅Is a weight coefficient, penalty function J (dist)_,t_f) The device is used for ensuring the comfort of the track changing process in the transverse direction and the longitudinal direction, and simultaneously requires that the distance in the longitudinal direction of the track changing process is as small as possible.

Determining the optimal track-changing time t at different speeds_fAnd the optimal distance dist, the optimal track changing tracks X (t) and Y (t) of the vehicle are determined. Table 1 shows the relevant parameters used in the planning process, respectively

And

two sets of comparison experiments are carried out to plan the optimal path and the transverse and longitudinal speedsThe changes are shown in the relevant curves in fig. 6, 7 and 8 respectively.

3) Design distributed model predictive control method

(1) System model

As shown in FIG. 3, in order to ensure the real-time performance of the method, the present invention adopts a relatively common single-track bicycle model for a single vehicle i in a fleet.

The state of each vehicle is defined here as:

the output is as follows:

control amount: u. of_i(t)＝[u_a,i,u_δ,i]^TAmount of interference

Here, the

And

respectively, the longitudinal and transverse speeds of the vehicle in the world coordinate system, then we can obtain the state space form of each vehicle as:

here, the

And

are all twice continuously differentiable and can be used,

m in the above formula_iAs vehicle mass, v_x,iAnd v_y,iThe velocities in the direction of the vehicle body and perpendicular thereto, respectively_iIs the yaw angle of the vehicle, /)_f,iAnd l_r,iRespectively, the distance, omega, between the center of mass of the vehicle and the centers of the front and rear wheels_iIs angular velocity, I_iIs the moment of inertia of the vehicle, C_fAnd C_rCornering stiffness, alpha, of front and rear wheel tyres respectively_f,iAnd alpha_r,iSlip angles, delta, of front and rear wheels, respectively_iAs a steering angle of the vehicle, a_iAnd u_a,iThe actual acceleration of the vehicle i in the direction of the vehicle body and the control variable u given by the controller_δ,iIs the desired angle of rotation, T, given by the controller_a,iAnd T_δ,iThe delay times of the engine and the steering system actuators, respectively.

As shown in fig. 4, the first vehicle provides a reference track, and following vehicles in the fleet of vehicles can receive information of the first vehicle and the vehicles in the vicinity of the first vehicle. The fleet's state, output, input function and interference function are defined as:

the discrete dynamics equation of the fleet at this time is as follows:

here, the

Υ＝diag{γ₁,...,γ_N}，Z＝diag{ζ₁...ζ_N}，

(2) Control target

Safety: safety is the central importance of the vehicle during driving and is used in the literature [3]"triangle area criterion" in (1) to avoid collision, as shown in FIG. 2, the ith vehicleAre respectively provided with four corners of { A_i,B_i,C_i,D_iAccording to this rule, the constraint condition for avoiding collision can be expressed as follows:

where i, k — p indicates a vehicle with a low vehicle speed ahead of the vehicle group.

Traceability: here, it is mainly meant that the lateral and longitudinal position errors approach zero, i.e.

Queue stability: queue stability means that the position error and the like of upstream vehicles of a fleet caused by interference are smaller and smaller when the upstream vehicles are propagated downstream, namely

Where α is_X,i≤1,α_Y,i≤1。

(3) Method design

A distributed nonlinear model predictive control method is employed herein. The method can solve the problem of multi-constraint control, and is suitable for the control problems with various requirements such as safety, stability, comfort, economy and the like of vehicle movement.

Defining the expected output of the ith vehicle in the fleet as

The desired control amount is u_des,i(t)＝[u_ades,i,u_δdes,i]^T。

In the running process of the vehicles, a fixed vehicle distance strategy is adopted, namely the expected vehicle distance between the vehicles is a fixed value d, and the expected output and the expected control quantity of each vehicle in the fleet can be obtained by combining the planned track.

Where X is_des(t),Y_des(t) are respectively the planned desired lateral and longitudinal positions,

is the planned desired transverse longitudinal speed, a_Xdes(t) is the projected lateral expected acceleration,

is the planned expected path curvature, and in order to meet the actual lane change requirement of the vehicle in the experiment, the rotation angle value is specified to be positive anticlockwise and negative clockwise.

(4) Local optimal control problem

An optimal control problem for a single vehicle within the prediction horizon [ t, t + p-1] is first analyzed.

Here, it is defined that:

the output and the control amount used in prediction are respectively.

The assumed output and control amount are actually the optimum amount calculated in one step in the preceding vehicle.

And solving the optimal quantity obtained by the optimization problem.

Optimizing the problem:

For i∈{1,2,...,N}at time t

subject to

Eq.(15)

the constraints on the control amount here include both acceleration limit for ensuring driver comfort and turning angle limit depending on the physical limitations of the vehicle tires.

Defining a loss function:

here Q_i≥0,R_i≥0,F_i≥0,G_iNot less than 0 is a weight coefficient, D ═ D,0,0,0]^T。

(5) Distributed model prediction control method

The previous subsection introduces the local optimal control problem of a single vehicle, and the section expands the local optimal control problem to the whole fleet to establish a distributed model predictive control method.

1 deg. initialization.

At the moment t is 0, each vehicle in the fleet is defined to run at an initial expected speed at a constant speed, and the expected distance d between the vehicles is maintained, so that the initial state x of each vehicle can be obtained_i(0) And initial input u_i(0)＝[0,0]^T. Then we can get the assumed control quantity and output that vehicle i transmits to other vehicles at the initial time as:

where y is_i^p(k |0) can be obtained from the following equation:

2 degree DMPC method iteration

For any time t > 0, vehicles i in the fleet perform the following steps:

step one, a vehicle i is based on the current state x_i(t), currently assuming output

And the received assumed output of vehicle i-1

To solve its local optimal control problem and obtain optimal control sequence

The following two points need to be noted in the solving process:

A. for vehicle i-1, the assumed output of the preceding vehicle received by the vehicle is the expected output planned within time [ t, t + p-1]

[y_des(t),y_des(t+1)...y_des(t+p-1)]。

B. For vehicle i-2, 3n, additional constraint conditions need to be satisfied in the solution process compared to the first vehicle i-1

Here σ ∈ (x, y).

And step two, solving the optimal state and outputting by using the optimal control sequence obtained in the previous step:

step three, calculating assumed output

First calculate the hypothetical input

The hypothetical control output is then calculated:

taking the first item of the optimal control sequence as a control input:

step five, returning to 1) iterative calculation until t is t +1_stop

(6) Simulation experiment results

In order to verify the effectiveness of the method of the present invention, simulation result graphs of the method are shown in FIGS. 6-12, wherein (a) and (b) represent driving conditions of the vehicle team under the two conditions of lane changing initial speed of 15m/s and 25m/s, respectively. In the figure 6, the operation track of the piloted vehicles of the fleet is almost completely overlapped with the planned optimal track changing track, the subsequent vehicles also completely track the respective expected tracks, and the track in the whole tracking process is smooth and has no collision phenomenon; FIG. 7 is a diagram of longitudinal speed variation during lane change, and it can be seen from the diagram that the absolute value of the speed deviating from the expected speed in two speed cases is about 0.004m/s at most, which is almost ignored relative to the speed per se, and the whole longitudinal driving process is very stable; FIG. 8 is a comparison graph of lateral velocity changes of the fleets at two initial velocities, and it can be seen that the lateral velocity tracking effect is very good; fig. 9 and fig. 10 are actual values of the longitudinal and lateral position deviation between the fleet vehicle and the adjacent front vehicle in two speed situations, respectively, and it can be seen from the figures that the inter-vehicle distance error between the second vehicle and the lead vehicle in the fleet is opposite to the change situation of the lateral and longitudinal tracking error value between the vehicle behind the fleet vehicle and the adjacent front vehicle, because each vehicle must not only track the adjacent front vehicle when solving its local optimum function, but also consider the tracking expected trajectory, and therefore the following vehicle will generate the "compensation" effect on the premise of ensuring safety when the second vehicle tracks the lead vehicle and generates the position deviation; fig. 11 and 12 show absolute values of the lateral and longitudinal tracking errors of the fleet at two speeds, which are used for verifying whether the fleet meets the stability of the fleet during the running process, and it can be seen from the figures that the absolute values of the tracking errors in both the longitudinal and the lateral directions at the two speeds are smaller and smaller along with the increase of the serial number of the vehicle, which proves that the fleet meets the stability of the fleet during the running process.

TABLE 1 vehicle-related vehicle parameters in a fleet of vehicles

	mi(kg)	lf,i(m)	lr,i(m)	Ii(kg·m2)	τa,i(s)	τδ,i(s)
							1	1751	1.22	1.73	2583.3	0.73	0.73
2	1506	1.15	1.65	2111.1	0.65	0.65
							3	1891	1.27	1.76	2869.2	0.77	0.77
4	1699	1.21	1.71	2480.0	0.71	0.71
							5	1959	1.29	1.79	3012.4	0.79	0.79
6	1255	1.08	1.58	1664.8	0.58	0.58

TABLE 2 parameters used in the trajectory planning procedure

Claims (1)

Translated fromChinese

1.一种具有换道辅助功能的协同自适应巡航控制方法，包括步骤如下：1. A collaborative adaptive cruise control method with a lane change assist function, comprising the following steps:步骤一：车队首先匀速行驶，领航车辆根据自车预定义换道起点和终点纵向状态和道路情况确定最优换道时间，设计具有避障功能的换道轨迹函数：Step 1: The team first drives at a constant speed, and the leading vehicle determines the optimal lane-changing time according to the longitudinal state of the pre-defined lane-changing starting point and ending point and road conditions, and designs a lane-changing trajectory function with obstacle avoidance function:定义换道开始时刻为t_i，领航车辆状态为

终止时刻为t_f，领航车辆状态为

临界碰撞时刻为t_c，领航车辆坐标为(X_c,Y_c)，根据多项式法建立如下具有避障功能的换道轨迹函数：Define the start time of lane change as t_i , and the state of the leading vehicle as

The termination time is t_f , and the state of the leading vehicle is

The critical collision moment is t_c , the coordinates of the leading vehicle are (X_c , Y_c ), and the following lane-changing trajectory function with obstacle avoidance function is established according to the polynomial method:

步骤二：领航车辆根据自车和前车车速、车身几何参数以及最优换道时间等在临界碰撞条件下计算车辆最优换道轨迹，计算车辆最优换道轨迹；Step 2: The leading vehicle calculates the optimal lane-changing trajectory of the vehicle under critical collision conditions according to the speed of the vehicle and the preceding vehicle, the geometric parameters of the vehicle body, and the optimal lane-changing time, and calculates the optimal lane-changing trajectory of the vehicle;首先需要确定最优换道结束时刻t_f以及换道时与前车间距dist：First, it is necessary to determine the optimal lane change end time t_f and the distance dist from the preceding vehicle when changing lanes:

subject tosubject toEq.(1)Eq.(1)

而后再根据得到的最优换道结束时间以及最优车间距可以得到最优换道轨迹X(t)和Y(t)；Then, according to the obtained optimal lane-changing end time and optimal vehicle spacing, the optimal lane-changing trajectories X(t) and Y(t) can be obtained;步骤三：车队依据给出的方法跟踪该最优轨迹，设计具有换道辅助功能的协同自适应巡航控制方法，主要包含以下步骤：Step 3: The team tracks the optimal trajectory according to the given method, and designs a collaborative adaptive cruise control method with lane-changing assist function, which mainly includes the following steps:(1)：针对车队中单个车辆i建立其非线性状态空间模型：(1): Establish its nonlinear state space model for a single vehicle i in the fleet:

(2)：设计局部最优控制问题(2): Designing local optimal control problems

(3)：设计分布式模型预测控制方法(3): Design a distributed model predictive control method每辆车同时求解其对应的局部最优控制问题，并针对车队中除领航车辆的其他车辆添加额外约束

来保证队列稳定性。Each vehicle solves its corresponding local optimal control problem at the same time, and adds additional constraints to other vehicles in the fleet except the lead vehicle

to ensure queue stability.

Images