CN108829110A - A kind of pilot model modeling method of cross/longitudinal movement Unified frame - Google Patents

Abstract

A driver model modeling method with a unified framework for lateral/longitudinal motion, comprising the following steps: using a hyperbolic tangent function to establish an analytical formula for a vehicle's lateral single-lane change trajectory, collecting information on the road, the vehicle itself, and its driving state, and based on the road state information Obtain the constraints on the lateral displacement of the vehicle, establish an index function according to the driver's characteristics and safety requirements, obtain the optimized lateral displacement parameters, calculate the lateral force constraints on the vehicle combined with the vehicle's driving state and environmental factors, and use the lateral force constraints to obtain the smoothness of the vehicle trajectory. , combined with the driver's characteristics to establish an index function to obtain the optimized value of the smoothness parameter, and finally to obtain the optimized trajectory expectation. Taking lateral expectation and longitudinal speed as expected values, a driver model describing the unified framework of lateral/vertical motion is established to achieve the purpose of trajectory tracking.

Description

A Driver Model Modeling Method Based on the Unified Framework of Horizontal/Longitudinal Motion

technical field

The present invention relates to the technical field of intelligent vehicle control, and in particular to a driver model design method based on the analytical expression of lane-changing trajectories based on driving safety and driver characteristics, parameter constraint calculation and optimization, and horizontal/vertical motion.

Background technique

Vehicle trajectory planning and driver model modeling occupy an important research position in the field of intelligent driving. Trajectory planning, as the reference output source of driver modeling, is the first guarantee for the safe operation of the vehicle in the entire intelligent driving process. However, most of the current trajectory planning usually considers the smoothness of the route itself as a safety indicator. This approach is too conservative in trajectory planning, and the planned trajectory that ignores the driving characteristics of the driver cannot guarantee the comfort of the driver when driving.

Driver modeling is a means to realize intelligent driving of vehicles. Many researchers divide the driver model into two models according to the direction of motion: lateral motion driver model and longitudinal motion driver model. When designing the driver model based on the controller, the model is also designed separately. However, considering most of the working conditions and the characteristics of the controlled object of the driver model—vehicle, the movement of the vehicle in two directions has coupling characteristics, and considering the movement state of a single direction to design the driver model separately will result in neglecting the movement state of the other direction. The deviation makes its control effect have a large error. Therefore, in order to make the controller design more accurate, the design of the driver model needs to consider the characteristics of the mutual influence of its motion, and use the state quantity in the other direction as a reference factor for the controller design to ensure that the desired control effect is achieved.

Contents of the invention

The technology of the present invention solves the problem: for the intelligent driving of vehicles, a method for analytical representation, parameter constraints and optimization of a lane-changing reference trajectory based on the combination of driving safety and driver characteristics is provided. At the same time, a horizontal and vertical driver model based on state compensation PID controller is given, so that vehicle motion can realize trajectory tracking and longitudinal speed tracking.

A driver model modeling method of horizontal/longitudinal motion unified framework, it is characterized in that, described driver model modeling method comprises the steps:

Step 1: According to the vehicle lane change trajectory, establish a road coordinate system, and use the hyperbolic tangent function with parameters to analyze and express the lateral lane change trajectory;

Step 2: Collect vehicle movement status and road environment information in real time

In real time, the on-board sensor collects the vehicle state during the driving process, obtains information such as the vehicle's lateral speed, longitudinal speed, and heading angle, and then collects road information through external sensors such as on-board cameras and radars to obtain road width, road surface friction coefficient, and vehicle body width. , frontal area, vehicle lateral position and other effective information;

Step 3: Optimizing the lateral displacement parameters in the analytical formula of lane change trajectory

According to the driver's driving intention analysis, the driver's longitudinal driving expectation and the longitudinal distance expectation for completing the lateral movement are obtained, and the path parameters of the trajectory analysis formula in step 1 are obtained, combined with the vehicle lateral position, road width, and body width collected in step 2 to obtain The position parameter and lateral displacement parameter constraint field of the middle trajectory analysis formula, the optimization index function of the lateral displacement parameter is established in combination with the safety requirements and driving conditions, and the optimized lateral displacement parameter is obtained by calculating and combining the lateral displacement parameter constraint domain;

Step 4: Optimize the flatness parameter in the analytical formula of lane change trajectory

According to the longitudinal driving expectation, the vehicle frontal area, the current road friction coefficient and the lateral displacement obtained in step 3, the smoothness parameter constraint domain in the trajectory analysis formula is obtained, and the optimization index function of the flatness parameter is established in combination with safety requirements and driving conditions. Optimizing the smoothness parameter value in the constraint domain, so as to obtain the desired trajectory expression described in step 1;

Step 5: Obtain the expected lateral velocity of the vehicle according to the lateral displacement

Taking the expected trajectory obtained in step 4 as the reference of lateral displacement, the lateral velocity satisfying the lateral trajectory tracking in the earth coordinate system is obtained by designing the PID controller, and the lateral velocity is converted into the vehicle coordinate system according to the heading angle information and the driving longitudinal expectation Lateral/vertical velocity expectations.

Step 6: Determine the driver model output

Taking the horizontal/longitudinal desired movement velocity obtained in step 5 as a reference value, use the PID controller to design the upper controller of the driver model combined with the horizontal/longitudinal movement, and obtain the horizontal trajectory tracking and longitudinal velocity acting on the vehicle coordinate system The planned values ΣF _x , ΣF _y and ΣM _z of the tracked resultant longitudinal force, resultant lateral force and yaw moment.

When performing lane change planning as described in step 1, the vehicle lane change trajectory is mathematically expressed by establishing a road coordinate system and using the hyperbolic tangent function. The form of the selected hyperbolic tangent function is:

y＝k·tanh[a·(x-b)]+h

Among them, y is the lateral displacement change of the vehicle, b is the distance parameter of the trajectory, which means 1/2 of the expected longitudinal travel distance of the vehicle performing the lane change operation, and k is the lateral displacement parameter of the trajectory, indicating the lateral expected distance of the vehicle performing the lane change operation. 1/2 of the total displacement, h is the initial lateral position of the vehicle on the earth coordinate system, x is the current longitudinal distance of the trajectory, and a is the smoothness parameter of the trajectory, indicating the smoothness of the trajectory.

In Step 3, considering the relationship between the vehicle’s driving trajectory and the vehicle’s stability and safety when the vehicle is driving, and according to the relationship between the lane-changing behavior and the road, the constraint domain of the lateral displacement parameter k is obtained, and then according to the safety requirements and driving conditions Build a performance indicator function with respect to k:

Among them, J _k1 represents the safety index, J _k2 represents the driver’s characteristic index, w ₁ and w ₂ are their respective weight coefficients, w ₂ ≠ 0, that is, w ₂ can be positive or negative, which is used to represent the driver’s expectation of lateral displacement A positive number indicates that the driver's expectation of the lateral displacement is as small as possible, and a negative number indicates that the driver's expectation of the lateral displacement is as large as possible. By designing different w ₁ and w ₂ to reflect the difference in the amount of lateral displacement expected by different drivers expected value of k.

As described in step 4, examine the relationship between the trajectory of the vehicle and the stability and safety of the vehicle when the vehicle is running. According to Newton’s second law and the friction circle constraint, the constraint domain of the flatness parameter a in the trajectory analysis formula is obtained. According to the flatness of the trajectory and the Safety requirements and driving conditions establish a performance index function on a:

w ₃ >0, w ₄ ≠0, w ₅ ≠0

Among them, J _a1 represents the safety index, J _a2 and J _a3 represent the performance indexes of the driver on the lateral velocity and acceleration of the vehicle, w ₃ , w ₄ , and w ₅ are their respective weight coefficients, and w ₄ ≠ 0, that is, w ₄ can be a positive number or It can be a negative number, which is used to represent the characteristics of the driver’s expectation on the lateral speed. A positive number indicates that the driver’s expectation on the lateral speed is as small as possible, that is, the smoother the driving process, the better. A negative number indicates that the driver’s expectation on the lateral speed is greater. , that is, the steering process is as fast as possible. In the same way for w ₅ , the expected value a of the lateral displacement acceleration expected by different drivers is reflected by designing different w ₃ , w ₄ , and w ₅ .

Described based on state compensation PID controller design combines the driver model controller of horizontal/longitudinal motion, its controlled object state space equation is:

Among them, u _x and u _y are the longitudinal and transverse velocities in the geodetic coordinate system, respectively, is the heading angle of the vehicle, ω _r is the yaw rate of the vehicle, v _x and v _y are the longitudinal and lateral velocities in the body coordinate system respectively; the PID controller in step 5 and step 6 is shown in the following formula:

u _y = k _py e _y + k _sy ∫ e _y + k _dy de _y

u _x1 ＝k _pvx e _vx +k _svx ∫e _vx +k _dvx de _vx , ∑F _x ＝u _x1 -mω _r v _y

u _vy ＝k _pvy e _vy +k _svy ∫e _vy +k _dvy de _vy , ∑F _y ＝u _vy +mω _r v _x

∑M _z ＝k _pwvx ·e _vx +k _swvx ·∫e _vy +k _dwvx ·de _vx

Among them, u _x1 and u _vy are the intermediate control variables of the state compensation of the longitudinal velocity and the transverse velocity, e _y , ∫ey _y , de _y are the lateral displacement deviation and its integral and differential items respectively, k _py , k _sy , k _dy are the _proportional , _integral and differential _parameters in the _{lateral displacement PID} controller; Proportional, integral, and differential parameters in the controller; e _vy , _{∫evy, and e vy are} the lateral velocity deviation and its integral and differential items, respectively, and k _pvy , k _svy , and k _dvy are the lateral velocity tracking PID of ∑F _y Proportional, integral, and differential parameters in the controller; k _pwvy , k _swvy , and k _dwvy are the proportional, integral, and differential parameters in the lateral velocity tracking PID controller of ∑M _z , respectively.

Beneficial effects: the present invention combines the external state information collected by the vehicle sensor and the state information of the vehicle itself, uses the hyperbolic tangent function to analyze and express the vehicle lane change trajectory, and at the same time analyzes and obtains the analytical parameter constraint domain, and then obtains the feasible trajectory cluster and optimizes it. The optimized trajectory was obtained by analysis, and the driver's horizontal/vertical combined controller was designed at the same time to realize the simultaneous tracking of longitudinal velocity and lateral trajectory displacement, and realize the unification of the horizontal and vertical motion control of the driver model, which has a good theory for the intelligent driving and realization of vehicles Guiding significance and application prospects.

Description of drawings

Fig. 1 is the driver's model work flowchart that the present invention proposes;

Fig. 2 is a road environment figure for the research of the present invention;

Fig. 3 is the road coordinate system established based on the road environment map;

Fig. 4 is the value range image of smoothness parameter a changing with k;

Fig. 5 is the reference trajectory image for experiment;

Figure 6 shows the longitudinal speed tracking effect of working condition 1;

Fig. 7 is working condition 1 longitudinal velocity tracking deviation;

Fig. 8 is the tracking effect of lateral displacement in working condition 1;

Figure 9 shows the lateral displacement deviation of working condition 1;

Figure 10 shows the response effect of longitudinal acceleration control in working condition 2;

Figure 11 shows the longitudinal velocity deviation of working condition 2;

Figure 12 is the tracking effect of uniform acceleration in working condition 3;

Figure 13 shows the deviation of the uniform acceleration tracking effect in working condition 3;

Figure 14 is the tracking of working condition 4 sudden acceleration working condition;

Figure 15 shows the tracking deviation of sudden acceleration in working condition 4;

Figure 16 is the tracking of working condition 4 sudden deceleration working condition;

Figure 17 shows the tracking deviation of sudden deceleration in working condition 4;

Figure 18 is the tracking of working condition 5 variable acceleration working condition;

Figure 19 shows the variable acceleration tracking error of working condition 5;

Fig. 20 is the tracking effect of lateral displacement in working condition 6;

Figure 21 shows the lateral displacement tracking deviation of working condition 6;

Figure 22 shows the longitudinal speed tracking effect of working condition 6;

Fig. 23 shows the longitudinal speed tracking deviation of working condition 6.

Detailed ways

The proposed design scheme will be further elaborated and illustrated below in conjunction with the accompanying drawings.

The present invention proposes a driver model design method that integrates vehicle lane change trajectory planning and lateral/longitudinal motion for vehicle driving safety, and the workflow layer diagram of the driver model involved is shown in Figure 1.

Module ① represents the horizontal and vertical expected quantity acquisition module of the driver model. Through the perception of the external environment and the parameters of the vehicle itself, the lateral displacement parameters and the constraint domains of the flatness parameters after the trajectory analysis of the vehicle when the vehicle performs lateral lane change are obtained, and then through the environment and The driving safety requirements and the driver's own characteristics optimize the trajectory parameters to obtain the optimized parameters, and then obtain the complete reference trajectory analysis formula, which is used as the lateral expected quantity of the driver model.

Module ② represents the driver model controller of the unified framework of horizontal and vertical motion, and the expected value required by the driver model controller is obtained in the expected value acquisition module: longitudinal velocity and lateral displacement. Among them, the lateral displacement is converted into the expected amount of lateral velocity through steering planning, and then the obtained expected amount of lateral velocity and longitudinal velocity is converted into the lateral resultant force, longitudinal resultant force, and yaw resultant moment on the body coordinate system by using the tracking controller , and then convert the resultant force and resultant moment into the input quantity conforming to the input form of the driver to the vehicle through the output conversion, so as to complete the driver model design.

Module ③ represents the controlled object of the driver model controller: the vehicle-road model, which operates the vehicle through the driver model, so that the vehicle obtains the corresponding longitudinal speed, lateral speed, and yaw rate, and then produces longitudinal displacement on the road, Lateral displacement, heading angle, so as to obtain the error with the expected amount in module ①, and then control it through the driver model controller.

The present invention proposes a driver model design method that is aimed at vehicle lane change trajectory planning and horizontal/longitudinal motion for vehicle safety, and is implemented in the following steps:

1) Establish a road coordinate system based on the vehicle lane-changing trajectory, and use the hyperbolic tangent function with parameters to analyze and express the lateral lane-changing trajectory.

First observe the road environment as shown in Figure 2:

In Figure 2, the vehicle is subdivided into mass points. Considering the collision avoidance safety requirements of the vehicle’s own width and the minimum safety distance, the gray dotted line area at the upper and lower ends is the safe area where the vehicle’s center of mass exists after the vehicle completes the lane change operation. By reading relevant driver’s lateral model literature and own actual observation, when the vehicle performs lane change behavior, considering that the analytical form expression method of the trajectory has a reference effect on the trajectory constraint expression, in this study, the trajectory of the single lane change operation is passed through the hyperbolic The tangent function is approximated, and the form of the selected hyperbolic tangent function is:

y=k tanh[a (x-b)]+h (1)

Among them, y is the lateral displacement change of the vehicle, b is the distance parameter of the trajectory, which represents 1/2 of the expected longitudinal travel distance of the vehicle performing the lane change operation, and k is the lateral displacement parameter of the trajectory, representing the distance generated by the vehicle performing the lane change operation. 1/2 of the total expected lateral displacement, h is the initial lateral position of the vehicle on the earth coordinate system, x is the current longitudinal distance of the trajectory, and a is the smoothness parameter of the trajectory, indicating the smoothness of the trajectory.

2) Real-time collection of vehicle movement status and road environment information

In real time, the on-board sensor collects the vehicle state during the driving process, obtains information such as the vehicle's lateral speed, longitudinal speed, and heading angle, and then collects road information through external sensors such as on-board cameras and radars to obtain road width, road surface friction coefficient, and vehicle body width. , frontal area, vehicle lateral position and other effective information;

3) Optimize the lateral displacement parameters in the analytical formula of lane change trajectory

Through the analysis of the driver's driving intention, the driver's longitudinal driving expectation and the longitudinal distance expectation for completing the lateral movement are obtained, and the constraint domain of the position parameter and lateral displacement parameter in the trajectory analysis formula is obtained, and the optimized lateral direction is obtained by calculating and combining the lateral displacement parameter constraint domain displacement parameter.

According to the hyperbolic tangent function formula image, the road coordinate system is established. Take the boundary line between the initial road and the target road as the X-axis, and the horizontal direction of the starting position of the road as the Y-axis, as shown in Figure 3 below.

Among them, D is the width of the road, w is the width of the vehicle itself, and d _buff is the minimum safe distance for vehicle collision avoidance. Here it is assumed that the longitudinal road distance is known, that is, b is a known constant. The constraint on the initial position of the vehicle can be obtained through the safe distance between the vehicle and the sideline. From the road coordinate system, the constraint on the initial lateral position h of the vehicle is:

In Figure 3, the area between the upper and lower dotted lines of the X-axis is the feasible area of the center of mass before and after the vehicle lane change operation. Through the two feasible areas and the initial position of the vehicle, the constraints on the lateral displacement of the vehicle's center of mass before and after the lane change can be obtained, namely Constraints on k:

After obtaining the constraint of k, according to the safety requirements of the vehicle and the driving characteristics of the driver, the performance index function of k is established:

Among them, J _k1 represents the safety index, J _k2 represents the driver’s characteristic index, w ₁ and w ₂ are their respective weight coefficients, w ₂ ≠ 0, that is, w ₂ can be positive or negative, which is used to represent the driver’s expectation of lateral displacement Positive numbers indicate that the driver's expectations for lateral displacement are as small as possible, and negative numbers indicate that the driver's expectations for lateral displacement are as large as possible. Different w ₁ and w ₂ are designed to reflect the difference in the amount of lateral displacement expected by different drivers. expected value of k.

The mapping between the k value and different weight coefficients w ₁ and w ₂ is obtained by analysis:

4) Optimize the smoothness parameter in the analytical formula of lane change trajectory

After the constraints of h and k are determined, only a remains unprocessed for the entire curve parameter, so the constraint set of the entire trajectory can be obtained by obtaining the constraint of a according to the actual driving safety constraints.

Combining the longitudinal driving expectation and the obtained vehicle frontal area information, according to the current road friction coefficient and the trajectory parameters, the smoothness parameter constraint domain in the trajectory analysis formula is obtained, and the optimization index function of the smoothness parameter is established by combining the safety requirements and driver characteristics. Calculate and combine the smoothness parameter constraint domain to obtain the optimized smoothness parameter.

According to the actual requirements of the trajectory, the lateral displacement parameters, the force analysis of the vehicle, Newton's second theorem, and the friction circle principle, the constraints of the smoothness parameter a are obtained:

a∈[a _min ,a _max ] (6)

Among them, a _min is a smoothness parameter that shows the slope of the trajectory close to 0 when the trajectory is expressed by the hyperbolic tangent function.

where a _ymax is the upper limit of the lateral acceleration of the vehicle, expressed as follows:

In the formula, m is the total mass of the vehicle, g is the acceleration of gravity, μ is the friction coefficient of the road, ρ is the air density, CD is the air resistance coefficient, _A is the windward area of the vehicle, and f is the rolling resistance coefficient of the vehicle.

After the constraint of a is obtained, the performance index function of a is established according to the smoothness of the trajectory, safety requirements and driver characteristics:

w ₃ >0, w ₄ ≠0, w ₅ ≠0 (9)

Among them, J _a1 represents the safety index, J _a2 and J _a3 represent the performance indexes of the driver on the lateral velocity and acceleration of the vehicle, w ₃ , w ₄ , and w ₅ are their respective weight coefficients, and w ₄ ≠ 0, that is, w ₄ can be a positive number or It can be a negative number, which is used to represent the characteristics of the driver’s expectation on the lateral speed. A positive number indicates that the driver’s expectation on the lateral speed is as small as possible, that is, the smoother the driving process, the better. A negative number indicates that the driver’s expectation on the lateral speed is greater. , that is, the steering process is as fast as possible. In the same way for w ₅ , the expected value a of the lateral displacement acceleration expected by different drivers is reflected by designing different w ₃ , w ₄ , and w ₅ .

The mapping between a value and different weight coefficients w ₃ , w ₄ , and w ₅ is obtained through analysis:

Figure 4 shows the upper and lower limit curves of a under the condition that the lateral displacement parameter k is between 1.25 and 2.25, where the vehicle mass m is taken as 1359.8kg, the acceleration of gravity g is taken as 9.8m/s ² , the road surface friction coefficient μ is taken as 1, and the windward area Take A as 2.2m ² , air resistance coefficient C _D as 0.30, rolling resistance coefficient f as 0.01, and v _x as 30m/s.

5) Obtain the expected lateral velocity of the vehicle according to the lateral displacement.

Taking the expected trajectory obtained in step 4 as the reference of lateral displacement, by designing the lateral velocity of the linear PID controller to satisfy the lateral trajectory tracking in the earth coordinate system, according to the heading angle information obtained in step 2 and the expectation obtained in step 3 Longitudinal motion translates this lateral velocity into a desired lateral/longitudinal velocity on the vehicle coordinate system.

Establish the three-degree-of-freedom model of the design vehicle longitudinal velocity, lateral velocity and yaw rate, the expression of the model is shown in the balance equation (11)

Among them, u _x and u _y are the longitudinal and transverse velocities in the geodetic coordinate system, respectively, is the heading angle of the vehicle, ω _r is the yaw rate of the vehicle, v _x , v _y are the longitudinal and lateral velocities in the body coordinate system, respectively.

According to the relationship between the lateral displacement and the lateral velocity of the earth, the PID controller is designed as shown in formula (12):

u _y = k _py e _y + k _sy ∫ e _y + k _dy de _y (12)

Among them, e _y , ∫ey _y , de _y are the lateral displacement deviation and its integral and differential items respectively, and k _py , k _sy , k _dy are the proportional, integral and differential parameters in the lateral displacement PID controller, respectively.

The lateral velocity in the geodetic coordinates obtained by the above formula can be converted to obtain the lateral expectation v _y in the tracking controller. The conversion formula is as in formula (13):

6) Determine the driver model output

Taking the horizontal/longitudinal desired motion obtained from formula (13) as a reference value, using the state feedback PID controller to design the upper layer controller of the driver model combined with the horizontal/longitudinal motion, obtain the lateral trajectory tracking and The planning values ∑F _x , ∑F _y and ∑M _z of the resultant longitudinal force, resultant lateral force and yaw moment of longitudinal velocity tracking.

According to the vehicle longitudinal velocity balance equation, the following PID controller is designed:

u _x1 ＝k _pvx e _vx +k _svx ∫e _vx +k _dvx de _vx (14)

∑F _x ＝u _x -mω _r V _y (15)

Among them, u _x1 is the longitudinal resultant force input of the PID including state compensation in the longitudinal velocity, _evx , _∫evx , _evx are the longitudinal velocity deviation and its integral and differential items respectively, k _pvx , k _svx , k _dvx are the longitudinal tracking PID Proportional, integral, derivative parameters in the controller.

According to the balance equation of vehicle lateral velocity, the following PID controller is designed:

u _vy ＝k _pvy e _vy +k _svy ∫e _vy +k _dvy de _vy (16)

∑F _y ＝u _vy +mω _r v _x (17)

Among them, u _vy is the lateral resultant force input of the PID including state compensation in the lateral velocity, e _vy , ∫ev _{y , e vy are} the lateral velocity deviation and its integral and differential items respectively, k _pvy , k _svy , k _dvy are the lateral tracking PID respectively Proportional, integral, derivative parameters in the controller.

According to the vehicle yaw rate balance equation, the following PID controller is designed:

∑M _z ＝k _pwvx ·e _vx +k _swvx ·∫e _vy +k _dwvx ·de _vx

Among them, k _pwvy , k _swvy , and k _dwvy are proportional, integral, and differential parameters in the horizontal tracking PID controller, respectively.

By adjusting nine PID parameters, a complete PID controller that realizes desired tracking can be obtained.

The simulation experiment data of the technical solutions provided by the present invention are given below.

A trajectory in the trajectory cluster satisfying the constraints is used as a reference trajectory, the function is y=1.75[tanh[0.08(v _x x-45)]+1]-1.5, where v _x =30m/s, and the trajectory is described as: The vehicle has a lateral displacement of 3.5m while passing 90m, and the trajectory image is shown in Figure 5.

Working condition 1: The vehicle maintains a longitudinal velocity of v _x = 30m/s to track the trajectory. Figures 6 to 9 are the simulation results. Among them, Figures 6 and 8 are the comparison between the expected actual value and the expected value of the longitudinal velocity and lateral displacement , Figure 7 and Figure 9 are the respective deviations.

Working condition 2: The vehicle travels straight, the longitudinal speed starts from 0, and it is required to accelerate rapidly to 30m/s. At this time, the lateral displacement is expected to be 0, so it is only necessary to observe the longitudinal speed response effect. Figure 10 and Figure 11 are the simulation results.

Working condition 3: The vehicle travels in a straight line, the longitudinal velocity starts from 0, and uniform acceleration is required at an acceleration of 5m/s2. Figure ¹² and Figure 13 are the simulation results.

Working condition 4: The vehicle runs smoothly at 20m/s, suddenly accelerates/decelerates at a certain moment, and the variation is ±10m/s. The tracking effect of the designed controller is shown in Figure 14-17.

Working condition 5: The vehicle travels straight, and the longitudinal speed changes in a sinusoidal manner, with a period of 3s and a range from 0 to 30. Figures 18 and 19 show the simulation results.

Working condition 6: The vehicle travels longitudinally at the speed in working condition 5, and tracks the expected trajectory laterally. Figure 20-Figure 23 shows the simulation results.

From the tracking effect of longitudinal velocity and lateral displacement in working condition 1, it can be seen that the driver model proposed in the present invention can realize lateral trajectory tracking when the vehicle is running at a constant speed, and has a good tracking effect, which shows that the design is effective in longitudinal and uniform lateral tracking. It has better tracking performance.

As can be seen from the tracking and corresponding effects of the longitudinal speed in working condition 2-working condition 5, the driver model proposed by the present invention can keep good tracking and response characteristics in various longitudinal driving conditions, indicating that the present invention can be well It realizes various longitudinal operations that the driver needs to perform while driving.

From the tracking effect of longitudinal velocity and lateral displacement in working condition 6, it can be seen that the driver model algorithm proposed by the present invention can realize lateral track tracking when the longitudinal acceleration changes, and has a good tracking effect, which shows that the design can be used when the longitudinal acceleration changes. It has good adaptability in tracking. It shows that the driver model can realize the combination of horizontal and vertical, and unify the control of the two directions of motion.

The present invention combines the external state information collected by the vehicle sensor and the state information of the vehicle itself, uses the hyperbolic tangent function to analyze the vehicle lane change trajectory, and simultaneously analyzes and obtains the analytical parameter constraint domain, and then obtains feasible trajectory clusters and optimizes them through optimization analysis Trajectory, at the same time, the driver's horizontal/vertical combined controller is designed to realize the simultaneous tracking of longitudinal velocity and lateral trajectory displacement, and realize the unification of the horizontal and vertical motion control of the driver model, which has good theoretical guiding significance and realization for the intelligent driving of vehicles. Application prospect.

Claims (5)

1. a kind of cross/longitudinal movement Unified frame pilot model modeling method, which is characterized in that the pilot model is built Mould method includes the following steps：

Step 1：According to vehicle lane change track, path coordinate system is established, using the hyperbolic tangent function with parameter to lateral lane change Track carries out analytic representation；

Step 2：Acquisition state of motion of vehicle and road environment information in real time

In real time by vehicle-state in onboard sensor acquisition vehicle travel process, obtain the lateral velocity of vehicle, longitudinal velocity, The information such as course angle, then road information is acquired by external sensors such as vehicle-mounted camera, radars, obtain road width, road surface The effective informations such as coefficient of friction, body width, front face area, lateral direction of car position；

Step 3：Optimize the lateral displacement parameter in the analytic expression of lane change track

It analyzes to obtain the expectation of driver's longitudinal drive and complete longitudinal distance phase of transverse movement according to driver's driving intention Hope, obtain step 1 in track analytic expression distance parameter, in conjunction with acquired in step 2 lateral direction of car position, road width, Body width obtain step 1 in the location parameter of track analytic expression and the constrained domain of lateral displacement parameter, in conjunction with demand for security with Driving cycles establish the optimizing index function of lateral displacement parameter, excellent by calculating and combining lateral displacement restriction on the parameters domain to obtain The lateral displacement parameter of change；

Step 4：Optimize the gentle degree parameter in the analytic expression of lane change track

The lateral displacement obtained in foundation longitudinal drive expectation, vehicle front face area, present road coefficient of friction and step 3, The gentle degree restriction on the parameters domain in the analytic expression of track is obtained, the optimization of gentle degree parameter is established in conjunction with demand for security and driving cycles Target function optimizes gentle degree parameter value, to obtain desired trajectory expression formula described in step 1 in constrained domain；

Step 5：Vehicle expectation lateral velocity is obtained according to lateral displacement

It is referred to using desired trajectory obtained in step 4 as lateral displacement, obtains earth coordinates by designing PID controller Under the lateral velocity for meeting transverse path tracking, according to course angle information and drive longitudinal expectation and convert the lateral velocity to The transverse direction that vehicle coordinate is fastened/longitudinal velocity expectation.

Step 6：Determine pilot model output quantity

Be used as reference value using laterally/longitudinal desired motion speed obtained in step 5, using PID controller design combine it is horizontal/ The pilot model upper controller of longitudinal movement acquires the tracking of realization transverse path and indulge for acting on that vehicle coordinate is fastened To the planning value ∑ F of the longitudinal resultant force of synthesis of speed tracing, the lateral resultant force of synthesis and yaw moment_x、∑F_yAnd ∑ M_z。

2. a kind of pilot model modeling method of cross/longitudinal movement Unified frame according to claim 1, feature exist In：When carrying out lane change planning described in step 1, by establishing path coordinate system using hyperbolic tangent function to vehicle lane change track Mathematical analysis expression is carried out, selected hyperbolic tangent function form is：

Y=ktanh [a (x-b)]+h

Wherein, y is the lateral displacement variable quantity of vehicle, and the distance parameter of the track b indicates that vehicle executes lane change and operates its lengthwise rows 1/2, the k for sailing desired distance is the lateral displacement parameter of track, indicates that vehicle executes lane change and operates its transverse direction expectation total displacement 1/2, h is vehicle in the initial lateral position that geodetic coordinates is fastened, and x is current longitudinal distance of track, and a is the gentle degree of track Parameter indicates the gradual degree of track.

3. a kind of pilot model modeling method of cross/longitudinal movement Unified frame according to claim 1, feature exist In：Relationship when vehicle driving between its driving trace and vehicle stabilization safety is considered described in step 3, according to lane change row Relationship between road obtains the constrained domain of lateral displacement parameter k, further according to demand for security and driving cycles establish about The performance index function of k：

Wherein, J_k1Indicate safety index, J_k2Indicate driver characteristics index, w₁、w₂For respective weight coefficient, w₂≠ 0 i.e. w₂It can be with It is positive number can also be negative, to indicate driver to the desired characteristic of lateral displacement, positive number indicates driver to lateral position Shifting expectation is the smaller the better, and negative number representation driver is the bigger the better to the expectation of lateral displacement, passes through and designs different w₁, w₂To embody The expectation k value of the different desired lateral displacement amounts of driver.

4. a kind of pilot model modeling method of cross/longitudinal movement Unified frame according to claim 1, feature exist In：Relationship when examining vehicle driving described in step 4 between its driving trace and vehicle stabilization safety, it is fixed according to newton second Rule is constrained with friction circle, obtains the constrained domain that parameter a is gently spent in the analytic expression of track, is gently being spent according to track and demand for security The performance index function about a is established with driving cycles：

w₃> 0, w₄≠ 0, w₅≠0

Wherein, J_a1Indicate safety index, J_a2、J_a3Indicate performance indicator of the driver to vehicle lateral speed and acceleration, w₃、 w₄、w₅For respective weight coefficient, w₄≠ 0 i.e. w₄Can be positive number can also be negative, to indicate driver to laterally expectation speed The characteristic of degree, it is the smaller the better that positive number indicates that driver it is expected lateral velocity, i.e., the more flat driving procedure the better, and negative number representation is driven The person of sailing is the bigger the better to the expectation of lateral velocity, i.e., steering procedure is as quick as possible.w₅Similarly, by designing different w₃、w₄、w₅ To embody the expectation a value of the different desired lateral displacement acceleration of driver.

5. a kind of pilot model modeling method of cross/longitudinal movement Unified frame according to claim 1, feature exist In：It is described based on state compensation PID controller design combination cross/longitudinal movement pilot model controller, controlled device State space equation is：

Wherein, u_x、u_yVertical and horizontal speed respectively under earth coordinates,For the course angle of vehicle, ω_rFor the cross of vehicle Pivot angle speed, v_x、v_yVertical and horizontal speed respectively under vehicle body coordinate system；Step 5: PID controller is as follows in step 6 Shown in formula：

u_y=k_py·e_y+k_sy·∫e_y+k_dy·de_y

u_x1=k_pvx·e_vx+k_svx·∫e_vx+k_dvx·de_vx, ∑ F_x=u_x1-mω_rv_y

u_vy=k_pvy·e_vy+k_svy·∫e_vy+k_dvy·de_vy, ∑ F_y=u_vy+mω_rv_x

∑M_z=k_pwvx·e_vx+k_swvx·∫e_vy+k_dwvx·de_vx

Wherein, u_x1、u_vyFor the intermediate control amount of the state compensation of longitudinal velocity and lateral velocity, e_y、∫e_y、de_yIt is laterally respectively Offset deviation and its integral and differential term, k_py、k_sy、k_dyRatio, integral, differential respectively in lateral displacement PID controller Parameter；e_vx、∫e_vx、e_vxIt is longitudinal velocity deviation and its integral and differential term, k respectively_pvx、k_svx、k_dvxRespectively longitudinal tracking Ratio, integral in PID controller, differential parameter；e_vy、∫e_vy、e_vyIt is lateral velocity deviation and its integral and differential respectively , k_pvy、k_svy、k_dvyRespectively ∑ F_yLateral velocity tracking PID controller in ratio, integral, differential parameter；k_pwvy、 k_swvy、k_dwvyRespectively ∑ M_zLateral velocity tracking PID controller in ratio, integral, differential parameter.