CN115432009A

CN115432009A - Automatic driving vehicle trajectory tracking control system

Info

Publication number: CN115432009A
Application number: CN202211227880.7A
Authority: CN
Inventors: 陈振斌; 杨峥; 欧阳颖; 李培新; 赖佳琴; 张天虎
Original assignee: Hainan University
Current assignee: Hainan University
Priority date: 2022-10-09
Filing date: 2022-10-09
Publication date: 2022-12-06
Anticipated expiration: 2042-10-09
Also published as: CN115432009B

Abstract

The embodiment of the invention discloses a trajectory tracking control system of an automatic driving vehicle, which comprises: the system comprises a signal processing subsystem, a parameter adaptation module, a system model library, an optimization solver, an emergency braking module and a system control module. The system can generate a corresponding prediction time domain according to the current vehicle state information and the road information, solve the optimal corner and acceleration by using an optimization function under the condition of meeting the system constraint condition, and convert the optimal corner and acceleration into control signals through a system control module, so that the automatic driving vehicle runs according to the expected track and the expected speed, and the speed change control of the automatic driving vehicle is realized. When the obstacle suddenly appearing in the front is detected, the system can automatically generate corresponding acceleration according to the current speed and the distance to the obstacle, so that emergency braking and obstacle avoidance are realized, and the safety of drivers and passengers is guaranteed.

Description

Automatic driving vehicle trajectory tracking control system

Technical Field

The invention relates to the field of automatic driving, in particular to the field of a track tracking control system of an automatic driving vehicle.

Background

The key technologies of the automatic driving vehicle mainly comprise: environmental awareness, behavioral decision-making, path planning, and vehicle motion control. The motion control is in the last link and is also a very important link. The studies on motion control are mainly divided into two main categories, lateral and longitudinal control: the transverse control mainly controls the front wheel turning angle of the automobile so as to complete the tracking of an expected track and ensure the running stability of the automobile; longitudinal control is to track the speed of the vehicle to ensure that the vehicle adapts to different road environments by varying the speed. The common vehicle needs to cooperate with the transverse control and the longitudinal control to achieve the purpose of running in different states. Although many scholars have studied lateral control, longitudinal control, and integrated control of vehicles, there are still shortcomings:

many researches adopt a model prediction control method to carry out track tracking control, and the method can limit various constraints through an objective function, so that the accuracy and the stability of vehicle track tracking are ensured. The prediction time domain is a key parameter in model prediction control, and the size of the prediction time domain directly influences the control effect of trajectory tracking.

However, many researches use the prediction time domain in the model prediction control method as a fixed value, which causes that the method cannot adapt to the working condition with large speed variation, and when the prediction time domain is not well selected, a large track deviation is generated, which causes large tracking lag, so that the accuracy of vehicle track tracking is greatly reduced.

In addition, many embodiments couple the transverse direction and the longitudinal direction, and the design complexity is high, so most scholars design controllers only considering transverse control, namely control over the front wheel steering angle, and set the vehicle speed to be a constant speed. However, the vehicle speed needs to be changed according to the surrounding environment, track tracking is realized only by controlling the front wheel steering angle, so that the vehicle tires are not properly rubbed on the ground, dangerous working conditions such as sideslip or drift of the rear axle are easy to occur, and the generation of hidden dangers is greatly increased.

In summary, in the research of many model predictive control methods at the present stage, the prediction time domain is not changed, and when a vehicle runs on an actual road, the curvature of the road and the road environment change constantly, the speed may change at any time, and the vehicle cannot keep moving forward at a constant speed constantly, which results in poor control accuracy of the traditional model predictive control method, and danger is easy to occur during speed change, and the traditional model predictive control method cannot cope with an emergency.

Disclosure of Invention

The embodiment of the invention provides a trajectory tracking control system of an automatic driving vehicle, which comprises:

a signal processing subsystem for determining curvature of a road ahead and status information of the vehicle;

the parameter adaptation module comprises a prediction time domain neural network and an adapter, wherein the prediction time domain neural network is used for processing according to the current vehicle speed, the expected vehicle speed and the curvature of the road ahead to obtain a corner prediction time domain parameter and an acceleration prediction time domain parameter; the adapter is used for selecting a corresponding corner prediction time domain and an acceleration prediction time domain in the self-defined corner interval group and acceleration interval group according to the sizes of the corner prediction time domain parameters and the acceleration prediction time domain parameters;

the system model base is used for obtaining corner prediction model output parameters and acceleration prediction model output parameters according to the state information of the vehicle, the corner prediction time domain and the acceleration prediction time domain;

the optimization solver is used for obtaining the corner control quantity and the acceleration control quantity according to the expected track, the expected vehicle speed, the corner prediction model output parameters, the corner prediction time domain, the acceleration prediction model output parameters and the acceleration prediction time domain;

and the system control module is used for receiving the steering angle control quantity and the acceleration control quantity, generating a corresponding control command and controlling the vehicle to execute corresponding deflection, acceleration and deceleration operations.

The embodiment of the invention also provides a trajectory tracking control method of the automatic driving vehicle, which comprises the following steps:

detecting the environment of a road in front of the vehicle according to an object detection module in a signal processing subsystem, and judging whether an obstacle exists on the road in front;

generating corresponding corner prediction time domain parameters and acceleration prediction time domain parameters based on the current vehicle speed, the expected vehicle speed and the road curvature through a prediction time domain neural network in a parameter adaptation module;

then selecting a corner prediction time domain and an acceleration prediction time domain corresponding to the corner prediction time domain parameter and the acceleration prediction time domain parameter in the range by the adapter based on the magnitudes of the corner prediction time domain parameter and the acceleration prediction time domain parameter;

a corner prediction model in the system model library obtains a corner prediction model output parameter according to the vehicle state information obtained by the state estimation module and a corner prediction time domain;

an acceleration prediction model in a system model library obtains an acceleration prediction model output parameter according to the vehicle state information and an acceleration prediction time domain;

a corner optimization function in the optimization solver is based on the expected track, and corner control quantity is obtained according to corner prediction model output parameters and a corner prediction time domain;

an acceleration optimization function in the optimization solver is based on the expected vehicle speed, and an acceleration control quantity is obtained according to the output parameters of the speed prediction model and the acceleration prediction time domain;

and after receiving the steering angle control quantity and the acceleration control quantity, a logic converter in the system control module generates a corresponding control command through a command generator in the system control module to control the vehicle to execute corresponding deflection, acceleration and deceleration operations.

Compared with the prior art, the embodiment of the invention provides the trajectory tracking control system of the automatic driving vehicle, which realizes the longitudinal and transverse comprehensive control of the automatic driving vehicle by depending on the corner prediction model and the acceleration prediction model, so that the steering angle and the driving speed can be automatically adjusted according to the change of the road curvature and the road environment when the vehicle runs on an actual road; in addition, the system can automatically generate a corresponding prediction time domain according to the running speed and the road information, so that the accuracy of track tracking can be met when the vehicle runs at a variable speed, the running stability can be ensured, and the dynamic control performance and the anti-interference performance of the automatic driving vehicle are improved; finally, the system can detect the road ahead through the object detection module. When the obstacle suddenly appears in front, the corresponding acceleration can be automatically generated according to the current speed and the distance to the obstacle, so that the emergency braking and obstacle avoidance are realized, and the driving safety of drivers and passengers is guaranteed.

Drawings

In order to make the technical solutions and advantages of the embodiments of the present application more clearly understood, the following description of the exemplary embodiments of the present application with reference to the attached drawings makes clear that the drawings in the following description are only embodiments disclosed in the present specification, and that other drawings may be obtained by those skilled in the art without inventive efforts.

FIG. 1 is a schematic diagram of the module details of a vehicle trajectory tracking system;

FIG. 2 is a schematic flow chart of the operation of the vehicle trajectory tracking system;

FIG. 3 is a schematic diagram of a vehicle corner control model built in conjunction with a tire model;

FIG. 4 is a schematic flow chart of the logic converter in the operating mode 1;

FIG. 5 is a schematic flow chart of the logic converter in the operation mode 2;

FIG. 6 is a flow chart diagram of a vehicle trajectory tracking method.

Detailed Description

The scheme provided by the embodiment of the specification is described below with reference to the accompanying drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. The described embodiments are only some of the embodiments of the present specification, and not all of the embodiments. All other embodiments obtained by a person skilled in the art based on the embodiments in the present specification without any inventive step are within the scope of the present application.

The embodiment of the invention discloses a trajectory tracking control system of an automatic driving vehicle, which consists of a signal processing subsystem, a system model library, a parameter adaptation module, an optimization solver and a system control module, wherein the signal processing subsystem comprises a state estimation module and an object detection module. Under the condition that an expected track and an expected vehicle speed are known, the system can generate a corresponding prediction time domain according to current vehicle state information and road information, solve an optimal corner and acceleration by using an optimization function under the condition of meeting system constraint conditions, and convert the optimal corner and acceleration into control signals through a system control module, so that the automatic driving vehicle runs according to the expected track and the expected vehicle speed, and the speed change control of the automatic driving vehicle is realized. And continuously updating the current vehicle state information and road information by using the data of the vehicle-mounted sensor assembly through the state estimation module in the driving process, repeating the process and finally finishing the track tracking control of the automatic driving vehicle. When the obstacle suddenly appearing in the front is detected, the system can automatically generate corresponding acceleration according to the current speed and the distance to the obstacle, so that emergency braking and obstacle avoidance are realized, and the safety of drivers and passengers is guaranteed.

The specific components of the modules of the vehicle trajectory tracking system are shown in fig. 1, and the specific module embodiment is as follows:

s110: signal processing subsystem

The signal processing subsystem is composed of a state estimation module and an object detection module, and comprises:

s111: state estimation module

The state estimation module may perform estimation operations based on measurement data acquired by the in-vehicle sensor assembly to obtain state information of the autonomous vehicle.

S112: object detection module

The object detection module comprises an external binocular camera and a laser radar, can detect the driving environment in front of the vehicle, and obtains the road curvature and the obstacle information of the road in front, including the distance to the obstacle.

S120: parameter adaptation module

The parameter adaptation module may generate a corresponding prediction time domain according to the vehicle state information and the road curvature, where the prediction time domain includes:

s121: predictive time domain neural network

The prediction time domain neural network takes the current vehicle speed, the expected vehicle speed and the road curvature as input, and outputs a corner prediction time domain parameter and an acceleration prediction time domain parameter.

S122: adapter

The adapter outputs parameter P according to the predicted time domain neural network ₁ and P₂ Is selected to be within the corresponding corner prediction time domain N _P1 Sum acceleration prediction time domain N _P2 。

S130: system model library

The system model base can predict the time domain N according to the state information of the vehicle _P1 、N _P2 Output parameter Y of the steering angle prediction model ₁ And an output parameter Y of the acceleration prediction model ₂ Performing real-time updating, including:

s131: acceleration prediction model

The acceleration prediction model adopts an acceleration prediction time domain N according to the state information of the vehicle _P2 Calculating the output parameter Y of the acceleration prediction model ₂ 。

S132: corner prediction model

The corner prediction model adopts a corner prediction time domain N according to the state information of the vehicle _P1 Calculate the turning angleOutput parameter Y of prediction model ₁ 。

S140: optimization solver

The optimization solver can output parameters Y according to the expected track, the expected vehicle speed and the model ₁ and Y₂ And predicting the time domain N _P1 and N_P2 Angle of rotation control u ₂ And acceleration control amount u ₃ Comprises the following steps:

s141: acceleration optimization function

The acceleration optimization function is based on the expected vehicle speed and the output parameter Y of the acceleration prediction model ₂ And predicting the time domain N _P2 The current acceleration control amount u can be obtained ₃ 。

S142: corner optimization function

The corner optimization function is based on the output parameters Y of the expected track and the corner prediction model ₁ And predicting the time domain N _P1 The current steering angle control amount u can be obtained ₂ 。

S150: emergency brake module

The emergency brake is composed of an acceleration neural network and an alarm, wherein the acceleration neural network comprises the following components:

s151: acceleration neural network

The acceleration neural network can generate corresponding emergency braking acceleration control quantity u according to the current vehicle speed and the distance of the current obstacle ₁ And controlling the vehicle to brake emergently.

S152: alarm device

The alarm can give out sound and light alarms to prompt drivers and passengers in the vehicle according to the current speed and the distance between the current vehicle and the obstacle in front of the vehicle.

S160: system control module

The system control module is composed of a logic converter and an instruction generator, and comprises:

s161: logic converter

The logic converter can convert the acceleration control quantity into an accelerator/brake control quantity;

s162: instruction generator

The command generator can generate corresponding control commands for the steering angle control quantity, the accelerator control quantity and the brake control quantity and send the control commands to the current automatic driving vehicle.

The specific working flow of the trajectory tracking control system of the automatic driving vehicle disclosed by the embodiment of the invention is shown in fig. 2, and the specific implementation mode of the system is described in the following with reference to the attached drawings:

firstly, processing signals, under the condition that an expected track and an expected vehicle speed are known, simultaneously carrying out vehicle state estimation and obstacle detection by a signal processing subsystem, and acquiring state information, road curvature and obstacle information of a vehicle, wherein the processing comprises the following steps:

1) Estimating state information of a currently running vehicle using a state estimation module

Relevant measurement data is acquired by an onboard sensor assembly of the autonomous vehicle and input to a state estimation module. The state estimation module can perform estimation operation according to the measurement data so as to obtain the current state information of the automatic driving vehicle and the road information of the environment where the automatic driving vehicle is located, and the information is transmitted to the parameter adaptation module and the system model base.

2) Detecting obstacle information in a forward driving road using an object detection module

The vehicle external binocular camera is used for acquiring first image data and second image data in a vehicle running environment, wherein the first image data is road curvature of a road in front of a vehicle; the second image data is road environment data in the forward direction of the vehicle, and is used for judging whether an obstacle exists in front of the road.

The laser radar collects road data in a vehicle running environment, is used for judging whether an obstacle exists in front of a road or not, and calculates the distance a to the obstacle.

And the object detection module is combined with the second image data of the binocular camera and the road data of the laser radar to generate the obstacle information.

When the system judges that an obstacle exists in front, the emergency braking module is started to give an alarm and control the vehicle to carry out emergency braking and obstacle avoidance. The emergency braking module comprises an alarm and an acceleration neural network, wherein the emergency braking module comprises:

1) Acceleration generation

The system utilizes the trained acceleration neural network to obtain the corresponding emergency braking acceleration control quantity u ₁ Then u is added ₁ And sending the data to a system control module.

The acceleration neural network input layer is provided with 2 nodes which are respectively the current speed and the distance to the obstacle; the output layer has 1 node which is the acceleration control quantity u ₁ . The current speed and the distance to the obstacle are used as input to obtain a corresponding output value, and the acceleration control quantity u is obtained through inverse normalization processing ₁ 。

The logic converter in the system control module receives the acceleration control quantity u ₁ Then, the working mode 1 is started, and after a corresponding control command is generated by a command generator, the vehicle can be controlled according to the braking acceleration u ₁ And carrying out emergency braking. Due to u ₁ Is generated on line in real time by using a neural network, so that different u can be produced according to different vehicle speeds and distances to obstacles ₁ The emergency braking device can adapt to different working conditions, can reduce the pause and frustration in the emergency braking process while ensuring the driving safety of drivers and passengers, and improves the smoothness of vehicles.

2) Alarm device

The alarm comprises an alarm circuit consisting of a voice broadcaster and an indicator light, and can alarm according to different conditions. The voice broadcaster broadcasts the distance to the barrier with a certain frequency.

When the distance a to the obstacle is greater than a defined safety distance a ₀ When the vehicle is in a collision state, the voice broadcaster broadcasts that 'an obstacle exists in the front and please pay attention to avoidance'; the orange warning light flickers, and the red warning light is not on.

When the distance a to the obstacle is less than or equal to the defined safe distance a ₀ In time, the voice broadcaster broadcasts 'dangerous driving, collision risk in front'; orange warning light not bright, redThe color warning light flashes.

And when the system judges that no obstacle exists in front, starting a parameter adaptation module, and generating a corresponding prediction time domain according to the current vehicle speed, the expected vehicle speed and the road curvature. The parameter adaptation module comprises a prediction time domain neural network and an adapter, wherein the parameter adaptation module comprises:

1) Predictive temporal parameter generation

The system utilizes the trained prediction time domain neural network to obtain a corresponding corner prediction time domain parameter P ₁ And acceleration prediction time domain parameter P ₂ Then P is added ₁ and P₂ And sending to the adapter.

The input layer of the prediction time domain neural network is provided with 3 nodes which are respectively the current vehicle speed, the expected vehicle speed and the road curvature; the output layer has 2 nodes, and the time domain parameters P are respectively predicted for the corners ₁ And acceleration prediction time domain parameter P ₂ . The current vehicle speed, the expected vehicle speed and the road curvature are used as input to obtain corresponding output values, and the corresponding corner prediction time domain parameter P in the range of 0-1 can be obtained through inverse normalization processing ₁ And acceleration prediction time domain parameter P ₂ 。

2) Parameter adaptation

Adapter with i ₁ For generating spacing within 0-1

Individual intervals, combined into an interval group, wherein i ₁ Is the interval parameter of the current interval. Each interval in the interval group has a corresponding prediction time domain. The adapter has two interval groups, a corner interval group and an acceleration interval group.

When the neural network outputs the parameter P ₁ and P₂ The adapter can then follow the parameter P ₁ and P₂ Finding out corresponding corner prediction time domain N in respective interval group _P1 And acceleration prediction time domain N _P2 Therefore, the online generation of the prediction time domain is realized, and the accuracy of the track tracking control is improved.

Then, the parameter adaptation module obtains N _P1 and N_P2 Send to the system modelA pattern library and an optimization solver.

The system model library comprises a corner prediction model and an acceleration prediction model, and can predict a time domain N according to the state information of the vehicle _P1 、N _P2 Output parameter Y of the steering angle prediction model ₁ And an output parameter Y of the acceleration prediction model ₂ Updating and updating the updated Y ₁ and Y₂ Sending the data to an optimization solver, wherein:

1) Corner prediction model

Based on a vehicle dynamics model and a tire model, a linear time-varying prediction model based on dynamics is designed by applying a model prediction control principle. The specific principle is as follows:

first, vehicle dynamics modeling is performed. Because the vehicle system is complex, the difficulty coefficient of establishing an accurate model is high, and some reasonable assumptions are required before all models are established. It is assumed that fig. 3 is a vehicle rotation angle control model established by combining a tire model, and a specific embodiment of a rotation angle prediction model is described below by combining the following drawings:

in this model, the state quantities are

The turning angle control amount is u ₂ ＝δ _f (ii) a An output of

Wherein m is the vehicle mass; a. b is the distance from the center of mass to the front and rear axes respectively;

is the centroid yaw angle;

is the centroid yaw rate;

is the centroid yaw angular acceleration;

and

vehicle longitudinal and lateral speeds, respectively;

and

longitudinal acceleration and lateral acceleration, respectively; I.C. A _z Is the moment of inertia of the vehicle about the z-axis; delta. For the preparation of a coating _f Is the corner of the front wheel; c _cf and C_cr The cornering stiffness of the front and rear wheels, respectively; c _lf and C_lr Longitudinal stiffness of the front and rear wheels, respectively; s is _f and s_r Respectively the slip rates of the front and rear wheels; x and Y are the lateral and longitudinal displacements of the vehicle in the inertial frame, respectively. m, a, b, I _z 、C _cf 、C _cr 、C _lf 、C _lr 、s _f 、s _r Are all known values.

And then, carrying out linearization and discretization by applying a model predictive control principle. Firstly, a first-order expansion is carried out by utilizing a Taylor formula, and the above model can be linearized and simplified to obtain:

wherein ,

wherein ,

and then discretizing the model by using a forward Euler method to obtain a discrete state space expression:

ξ ₁ (k+1)＝A ₁ (k)ξ ₁ (k)+B ₁ (k)u ₂ (k)

wherein ,A₁ (k)＝I+TA ₁ (t)；B ₁ (k)＝TB ₁ (t); k is the current sampling moment, and k +1 is the next sampling moment; t is the sampling period.

Selecting a corner increment delta u ₂ As a control quantity. Solving to obtain the control increment delta u at the current moment ₂ (k) Then, after adding the control quantity known at the previous moment, the control quantity u at the current moment can be obtained ₂ (k) In that respect Setting:

from this a new state space expression is obtained:

ξ(k+1|t)＝A ₂ ξ(k|t)+B ₂ Δu ₂ (k|t)

let the model output be:

η(k|t)＝C ₁ ξ(k|t)

setting the prediction time domain of the model to be N _P1 (ii) a Controlling the time domain to be N _C1 Is known, and N _C1 <N _P1 . Then the future N is available _P1 The output at that time is:

the current state quantity can be measured by a sensor or estimated by a state, so ξ (k | t) is known and the control increment Δ U in the control time domain ₂ (t) can be obtained by calculation, so that the output quantity in the prediction time domain can be obtained.

Finally, according to the model and the state information of the vehicle, the corner prediction time domain N is adopted _P1 Calculating the output parameter Y of the corner prediction model ₁

2) Acceleration prediction model

And analyzing the vehicle longitudinal dynamics model, and solving the expected acceleration by using a model prediction control principle. First, using a first order inertial system to express longitudinal control of the vehicle, we can obtain:

wherein K is the system gain; tau. _d Is a time constant; a is the current acceleration of the vehicle; a is _des Is the desired acceleration.

The above model is converted into a state space expression:

wherein ,

the state quantity is x = [ va = [ [ va ]] ^T (ii) a The acceleration control amount is u ₃ ＝a _des (ii) a The velocity v is taken as the system output.

x(k+1)＝A ₄ x(k)+B ₄ u ₃ (k)

wherein ,

the model output is then:

y(k|t)＝C ₂ x(k|t)

finally, the acceleration is adopted to predict the time domain N according to the model and the state information of the vehicle _P2 Calculating an output parameter Y of the acceleration prediction model ₂

The optimization solver comprises an acceleration optimization function and a corner optimization function, and can output parameters Y according to an expected track, an expected vehicle speed and a model of a system model library ₁ and Y₂ And the predicted time domain N of the adapter output _P1 and N_P2 Control the turning angle u ₂ And an acceleration control amount u ₃ And solving for u ₂ and u₃ Sending the data to a system control module, wherein the system control module comprises:

1) Solution of turning angle control quantity

The optimization solver can adopt the corner to predict the time domain N according to the expected track and the corner optimization function _P1 And solving the optimal control corner under the constraint condition.

According to the principle of model predictive control, the corner optimization function can be obtained as follows:

matrix Q ₁ Is a weight matrix of tracking deviations; matrix R ₁ Is a weight matrix that controls the magnitude of the increments.

Referring to the expected track, according to the output parameter Y of the corner prediction model ₁ And predicting the time domain N _P1 A series of optimal rotation angle increment delta U under the system constraint condition can be solved ₂ (t) taking the first angular increment of rotation of the series Δ u ₂ (k | t) and the current steering angle control amount u is obtained by adding the steering angle control amount at the previous time ₂ 。

2) Acceleration control quantity solving

The optimization solver can predict the time domain N by adopting the acceleration according to the expected vehicle speed and the acceleration optimization function _P2 And solving the optimal control acceleration under the constraint condition.

According to the principle of model predictive control, an acceleration optimization function can be obtained as

Referring to the expected vehicle speed, the output parameter Y of the model is predicted according to the acceleration ₂ And predicting the time domain N _P2 A series of optimal acceleration increments delta U under the system constraint condition can be solved ₃ (t) taking the first angular increment of the series, Δ u ₃ (k | t), and the current acceleration control amount u can be obtained by adding the acceleration control amount at the previous time ₃ 。

The system control module receives the rotation angle control quantity u ₂ And an acceleration control amount u ₃ Then, the acceleration control amount u is not directly used for vehicle control, and the acceleration control amount u is required ₃ The accelerator/brake control amount is converted to control the autonomous vehicle.

The system control module comprises a logic converter and can convert the acceleration signal into an accelerator/brake signal, and then the rotation angle signal and the accelerator/brake signal generate a corresponding control instruction through an instruction generator, so that the automatic driving vehicle is controlled to run according to an expected track and an expected vehicle speed or perform emergency braking. The system control module comprises two working modes and has the following specific principles:

1) Mode of operation 1

The logic converter receives the emergency braking acceleration control quantity u ₁ Then, the operation mode 1 is started, and the operation flow is shown in fig. 4:

s410: firstly, whether the driver takes over the operation or not is judged after the alarm gives an alarm. When the driver takes over the operation, the logic converter has no output and does not perform any operation. When the driver does not take over the operation, u is added ₁ And a maximum limit value r ₀ And an obstacle distance a and a safety distance a ₀ Carrying out comparison;

s420: when u is ₁ <-r ₀ Or a<a ₀ When the braking force is applied, the emergency braking acceleration control amount u is described ₁ Having exceeded the maximum limit, or the vehicle being too close to the obstacle, the maximum limit r is followed ₀ Performing braking operation and outputting braking control quantity k ₁ r ₀ ；

S430: otherwise, according to the acceleration control quantity u ₁ Performs a braking operation to output a braking control amount k ₁ u ₁ 。k ₁ Is the braking coefficient.

The working mode 1 can avoid the dangerous condition caused by overlarge braking acceleration and can also avoid the condition that the vehicle cannot be braked and stopped because of insufficient braking force when the distance is too short.

2) Mode of operation 2

The logic converter receives the acceleration control quantity u ₃ Then, the operation mode 2 is turned on, and the operation flow is shown in fig. 5. :

s510: controlling the acceleration u ₃ And controlling the regulation coefficient r ₁ Carrying out comparison;

s520: when u is ₃ <-r ₁ When the brake is operated, the brake control amount k is outputted ₁ u ₃ ；

S530: when u is ₃ >r ₁ When the throttle is closed, the driving operation is performed and the throttle control amount k is outputted ₂ u ₃ ；

wherein k₂ Is a driving coefficient;

s540: when-r ₁ ＜＜u ₃ ＜＜r ₁ When the operation is finished, the logic converter has no output and does not perform any operation.

The working mode 2 can avoid frequent switching of an accelerator pedal/a brake pedal as much as possible, so that the riding comfort can be improved, and the loss of parts can be reduced; on the other hand, the accelerator pedal and the brake pedal can be prevented from being operated simultaneously, and the driving safety is improved.

After the command generator receives the parameters output by the emergency braking module and the optimization solver, the turning angle control quantity u can be obtained ₂ Throttle control k ₂ u ₃ And a braking control amount k ₁ r ₀ Or k ₁ u ₃ And generating a corresponding control command and sending the control command to the automatic driving vehicle.

After receiving the control signal, the automatic driving vehicle executes corresponding deflection and acceleration/deceleration operations, so that the vehicle runs according to an expected track and an expected vehicle speed, track tracking control is realized, or emergency braking risk avoidance is carried out. And then, acquiring relevant measurement data in real time through the vehicle-mounted sensor assembly, and inputting the data to the state estimation module. And the control is repeated in a circulating way, and finally, the longitudinal and transverse gear shifting control of the automatic driving vehicle is realized.

The embodiment of the invention also discloses a method for tracking and controlling the track of the automatic driving vehicle, the specific flow of the method is shown in fig. 6, and the method is implemented as follows:

s610: vehicle state acquisition and obstacle determination

Under the condition that the signal processing subsystem obtains the expected speed and the expected track of the current automatic driving vehicle, the state estimation module estimates according to the obtained measurement data to obtain the state information of the vehicle and detects the running environment in front to obtain the curvature and the obstacle information of the road in front; if an obstacle exists in the front, the emergency braking module is started, and if no obstacle exists in the front, the parameter adapting module is started;

s620: emergency braking alert

Generating an output value in real time based on the current speed and the distance a between the vehicle and the obstacle by using the trained acceleration neural network, and then reversing the output valueNormalization processing is carried out to obtain real-time control quantity u of corresponding acceleration ₁ And sending the data to a system control module, wherein the alarm judges a and a ₀ Alarm the magnitude of the value, wherein a ₀ The current safe distance between the vehicle and the obstacle is obtained;

s630: generating a predicted time domain

Generating an output value in real time based on the current vehicle speed, the expected vehicle speed and the curvature by using the trained prediction time domain neural network, and then carrying out inverse normalization processing on the output value to obtain a corresponding corner prediction time domain parameter P ₁ And acceleration prediction time domain parameter P ₂ And sent to the adapter, by the adapter, based on P ₁ and P₂ Is selected to be within the corresponding corner prediction time domain N _P1 Sum acceleration prediction time domain N _P2 And sending the data to a system model library and an optimization solver;

s640: updating prediction model parameters in real time

According to the state information and N, the acceleration prediction model and the rotation angle prediction model _P1 and N_P2 Updating output parameter Y of corner prediction model in real time ₁ And an output parameter Y of the acceleration prediction model ₂ ；

S650: optimizing vehicle travel track in real time

According to the expected track, the expected vehicle speed and Y, the acceleration optimization function and the rotation angle optimization function ₁ 、Y ₂ and N_P1 、N _P2 Calculating to obtain a rotation angle control quantity u ₂ And an acceleration control amount u ₃ And will u ₂ and u₃ Sending the data to a system control module;

s660: converting the control parameters into control parameters and sending the control parameters to the automatic driving vehicle

The logic converter receives u ₁ 、u ₂ and u₃ Generating a corresponding control instruction by an instruction generator, and controlling the current vehicle to be in accordance with u ₁ 、u ₂ and u₃ And executing corresponding deflection and acceleration and deceleration operations.

The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only examples of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made on the basis of the technical solutions of the present invention should be included in the scope of the present invention.

Claims

1. An autonomous vehicle trajectory tracking control system, the system comprising:

the parameter adaptation module comprises a prediction time domain neural network and an adapter, wherein the prediction time domain neural network is used for processing according to the current vehicle speed, the expected vehicle speed and the curvature of the road ahead to obtain a corner prediction time domain parameter and an acceleration prediction time domain parameter; the adapter is used for selecting a corresponding corner prediction time domain and an acceleration prediction time domain in a self-defined corner interval group and an acceleration interval group according to the sizes of corner prediction time domain parameters and acceleration prediction time domain parameters;

2. The system of claim 1, further comprising an emergency braking module comprising:

the acceleration neural network is used for generating corresponding emergency braking acceleration control quantity according to the current speed and the distance between the vehicle and the obstacle, which is obtained by the object detection module in the signal processing subsystem, and sending the control quantity to the system control module; then a system control module receives the emergency braking acceleration control quantity and generates a corresponding control instruction to control the emergency braking of the vehicle;

and the alarm gives an alarm by comparing the distance from the vehicle to the obstacle and a preset safety distance.

3. The system of claim 1, wherein the library of system models includes a corner prediction model having a mathematical expression of:

in the expression, the state quantity is

The turning angle control amount is u ₂ ＝δ _f (ii) a An output of

Wherein m is the overall vehicle mass of the vehicle; a. b is the distance from the vehicle's center of mass to the front and rear axles, respectively; phi is the centroid yaw angle; phi is a ^· Is the centroid yaw rate; phi is a ^·· Is the centroid yaw angular acceleration; x is the number of ^· and y^· The vehicle longitudinal speed and lateral speed, respectively; x is a radical of a fluorine atom ^·· and y^·· Longitudinal acceleration and lateral acceleration, respectively; i is _z Is the moment of inertia of the vehicle about the z-axis; delta. For the preparation of a coating _f Is the corner of the front wheel; c _cf and C_cr The cornering stiffness of the front and rear wheels, respectively; c _lf and C_lr Longitudinal stiffness of the front and rear wheels, respectively; s is _f and s_r The slip rates of the front wheel and the rear wheel are respectively; x and Y are the lateral and longitudinal displacements of the vehicle in the inertial frame, respectively; wherein m, a, b, I _z 、C _cf 、C _cr 、C _lf 、C _lr 、s _f 、s _r Are all known values.

4. The system of claim 3, wherein the mathematical expression of the corner prediction model is linearized and discretized using a model predictive control principle to obtain a discrete state space expression:

ξ ₁ (k+1)＝A ₁ (k)ξ ₁ (k)+B ₁ (k)u ₂ (k)

wherein ,A₁ (k)＝I+TA ₁ (t)；B ₁ (k)＝TB ₁ (t); k is the current sampling moment, and k +1 is the next sampling moment; t is the sampling period;

selecting a corner increment delta u ₂ As a control quantity, solving to obtain a control increment delta u at the current moment ₂ (k) Then adding the control quantity known at the previous moment to obtain the control quantity u at the current moment ₂ (k) (ii) a Setting:

this results in a new state space expression:

ξ(k+1|t)＝A ₂ ξ(k|t)+B ₂ Δu ₂ (k|t)

let the model output be:

η(k|t)＝C ₁ ξ(k|t)

setting the prediction time domain of the model as a corner prediction time domain N _P1 (ii) a Controlling the time domain to be N _C1 Is known, and N is _C1 <N _P1 (ii) a Then the future N is obtained _P1 The output at that time is:

predicting a time domain N by the model and according to the state information of the vehicle using a corner _P1 Using the formula:

obtaining a corner prediction model output parameter Y of the corner prediction model after solving ₁ 。

5. The system of claim 1, wherein the library of system models includes an acceleration prediction model having a mathematical expression of:

wherein K is the system gain; tau. _d Is a time constant; a is the current acceleration of the vehicle; a is _des Is the desired acceleration;

the conversion of the mathematical expression into a spatial expression is:

wherein ,

the state quantity is x = [ va = [ [ va ]] ^T (ii) a The acceleration control amount is u ₃ ＝a _des (ii) a Velocity v as the system output;

discretizing the model by using a forward Euler method to obtain a discrete state space expression:

x(k+1)＝A ₄ x(k)+B ₄ u ₃ (k)

wherein ,

the model output is then:

y(k|t)＝C ₂ x(k|t)

analyzing the vehicle longitudinal dynamics model, and adopting acceleration to predict a time domain N according to the state information of the vehicle _P2 Using the formula:

obtaining an acceleration prediction model output parameter Y after solving ₂ 。

6. The system of claim 1, wherein the steering angle control amount optimization function is:

wherein, the matrix Q ₁ Is a weight matrix, matrix R, of the tracking deviations ₁ Is a weight matrix that controls the incremental magnitude;

outputting parameter Y by expected track and corner prediction model ₁ Sum-rotation angle prediction time domain N _P1 Solving a series of optimal rotation angle increment delta U under the system constraint condition ₂ (t) taking the first angular increment of rotation of the series Δ u ₂ (k | t) and the steering angle control amount at the previous time to obtain the current steering angle control amount u ₂ 。

7. The system of claim 1, wherein the optimization function of the acceleration control amount is:

the output parameter Y of the model is predicted by the expected speed and the acceleration ₂ Sum acceleration prediction time domain N _P2 Solving a series of optimal acceleration increment delta U under the constraint condition of the system ₃ (t) taking the first angular increment of rotation of the series Δ u ₃ (k | t), and adding the acceleration control amount at the previous time to obtain the current acceleration control amount u ₃ 。

8. The system of claim 1, wherein the system control module comprises a logic converter having two modes of operation:

the logic converter starts the mode 1 after receiving the emergency braking acceleration control quantity, and according to the fact that whether a driver takes over or not after the alarm of the alarm exists, the driver takes over, and the logic converter does not have output and does not perform any operation; the driver is not taken over, and the following steps are executed:

when u is ₁ <-r ₀ Or a<a ₀ According to r ₀ Performs a braking operation to output a braking control amount k ₁ r ₀ ；

Otherwise, the braking operation is carried out according to the emergency braking acceleration control quantity u1, and the braking control quantity k is output ₁ u ₁； wherein ,r₀ Is the maximum limit value of the vehicle acceleration, k ₁ Is the braking coefficient;

the logic converter receives the acceleration control quantity u ₃ Rear opening mode 2, when u ₃ <-r ₁ When the brake is operated, the brake control amount k is outputted ₂ u ₃ (ii) a When-r ₁ ＜＜u ₃ ＜＜r ₁ When the operation is finished, the logic converter has no output and does not perform any operation;

when u is ₃ >r ₁ When the throttle is closed, the driving operation is performed to output the throttle control amount k ₂ u ₃； wherein ,r₁ To control the adjustment coefficient, k ₂ Is the drive factor.

9. A method for trajectory tracking control of an autonomous vehicle, the method comprising:

then the adapter selects a corner prediction time domain and an acceleration prediction time domain corresponding to the corner prediction time domain parameter and the acceleration prediction time domain parameter in the range based on the sizes of the corner prediction time domain parameter and the acceleration prediction time domain parameter;

an acceleration prediction model in a system model library obtains an acceleration prediction model output parameter according to the vehicle state information and the acceleration prediction time domain;