CN116552520A - High-performance lane keeping control system based on rolling pretreatment - Google Patents

Abstract

The invention discloses a high-performance lane keeping control system based on rolling pretreatment, which is characterized by comprising a guiding controller and a feedback controller, wherein a lane detection module and a steering wheel restraining module are arranged in the feedback controller, the guiding controller comprises a fuzzy module, a PID module and a fusion module, the fuzzy module is used for outputting a fuzzy steering wheel corner, a rolling time domain window and an iteration module are arranged in the PID module, an iteration strategy arranged in the iteration module is used for obtaining an adjusted PID steering wheel corner, a fusion function and a weight coefficient are arranged in the fusion module, and the fusion module is used for fusing the fuzzy steering wheel corner and the PID steering wheel corner to generate a restraining steering wheel corner. The high-performance lane keeping control system based on rolling preprocessing has the advantage that vehicles running at different speeds can be controlled to run without deviating from a specified lane in various road scenes with different curvatures.

Description

High-performance lane keeping control system based on rolling pretreatment

Technical Field

The invention relates to the technical field of vehicle control systems, in particular to a high-performance lane keeping control system based on rolling pretreatment.

Background

With the rise of intelligent driving, advanced driving assistance systems are receiving increasing attention from research institutions and automobile enterprises. The lateral stability control of the vehicle is one of key technologies of a driving auxiliary system, and lateral safety performance of the vehicle under dangerous driving conditions can be effectively improved by adopting a lateral stability control strategy, and the lateral stability control system mainly comprises a lane keeping system, a lane changing system and the like. The control strategy adopted by the lane keeping system can be divided into two main types of a control system with a model and a control system without a model according to the model.

In contrast to the former, the methods based on neural network control, proportional-integral-derivative control, fuzzy control, and the like do not need to consider the influence of model deviation on the system performance. But the neural network control needs to have a good control effect in a large amount of off-line training, and on-line adaptation is difficult in an uncertain environment; the traditional fuzzy control has better performance in transient state, but has larger steady-state error; the proportional-integral-derivative control is simple in structure and easy to realize, but has poor dynamic quality. The cooperative control of the fuzzy control and the proportional-integral-derivative control can achieve improvement of overall performance, but the traditional strategy can only achieve rough adjustment based on fuzzy logic, and cannot achieve maximization of performance. In addition, there is less research on steering wheel shake in current lane keeping systems, and fluctuations in output in high-speed driving conditions will result in deviations from the range of deviations allowed for lane keeping, and thus deviations from the specified lane.

Disclosure of Invention

Aiming at the defects of the prior art, the invention aims to provide a high-performance lane keeping control system based on rolling pretreatment, which has the effect that vehicles running at different speeds can be controlled to run without deviating from a specified lane in various road scenes with different curvatures.

In order to achieve the above purpose, the present invention provides the following technical solutions:

a high-performance lane keeping control system based on rolling preprocessing comprises a guide controller and a feedback controller which are integrated in a processor;

the feedback controller is internally provided with a lane detection module and a steering wheel suppression module, the lane detection module comprises a detection unit and a guiding unit, the detection unit detects lane information in real time according to an ultrasonic detector to generate a lane image, the guiding unit is used for calling the lane image and the lane keeping line displayed on the lane image and displaying the lane image and the lane keeping line on the vehicle-mounted display screen;

the steering controller comprises a fuzzy module, a PID module and a fusion module, wherein an obstacle function is configured in the fuzzy module, the fuzzy module outputs a fuzzy steering wheel corner according to the obstacle function, a feedforward unit is configured in the steering wheel restraining module, a feedforward strategy is configured in the feedforward unit, the feedforward strategy comprises the steps of obtaining the feedforward steering wheel corner according to the wheelbase of a front axle and a rear axle of a vehicle, outputting the steering wheel corner to be determined according to the feedforward steering wheel corner and the fuzzy steering wheel corner, a rolling time domain window designed based on a vehicle dynamics model is configured in the PID module, an iteration module is further configured in the PID module, a corner algorithm and a superposition strategy are arranged in the PID module, the PID module generates the PID steering wheel corner according to the corner algorithm, the superposition strategy further comprises the steps of estimating the running state of the vehicle through the rolling time domain window according to the vehicle dynamics model, obtaining a vehicle overshoot value, performing adaptive optimization according to the rolling time domain window in a view range, obtaining an adjustment interval, and further comprises obtaining an adaptive steering wheel adjustment interval according to the rolling time domain corner and the adaptive iteration optimization value after the rolling time domain window overshoot value;

the fusion module is internally provided with a fusion function and a weight coefficient, the fusion module calls the steering wheel angle to be determined and the PID steering wheel angle to the fusion function, the steering wheel angle is restrained according to the fusion function, the processor sends a suppression steering wheel angle to a steering wheel suppression module, which controls a vehicle steering wheel suppression adjustment to keep the vehicle traveling centered along the lane keep line.

As a further improvement of the invention, the fuzzy module is also internally provided with a rotation angle threshold;

the obstacle function is specifically:

δ _F ＝Δδ _F *l _δ

wherein: delta ^* Characterizing the angular threshold, l _δ Characterizing the barrier function, delta _F Characterizing a fuzzy steering wheel angle.

As a further improvement of the present invention, the feedforward strategy further includes a feedforward algorithm, and the steering wheel angle to be determined is calculated according to the feedforward algorithm, where the feedforward algorithm specifically includes:

δ _out ＝δ _ff +δ _F

δ _ff ＝L*ρ

wherein: delta _out Characterizing steering angle delta _ff The feed-forward steering wheel angle is represented, L represents the wheelbase of the front and rear axles of the vehicle, and ρ represents the curvature of the road.

As a further improvement of the invention, a turning angle algorithm for generating the turning angle of the PID steering wheel is configured in the PID module, and the turning angle algorithm specifically comprises:

δ′ _out ＝δ _ff +δ _PID

wherein: e, e _y Characterization of lateral deviation, de _y Characterization of the rate of change of lateral deviation, K' _p 、k′ _I 、k′ _D Characterizing steering wheel angle coefficient, delta _PID Characterizing PID steering wheel angle.

As a further improvement of the invention, the specific process of obtaining the vehicle overshoot value through the rolling time domain window is as follows:

discretizing a state space equation based on a forward Euler method to obtain a prediction step length N _P The state equation in the range is:

wherein: n (N) _P K is the current time for predicting step length;

analyzing the control performance under the corrected proportional, integral and differential coefficients of the current fuzzy module:

definition setIs { x (k+1), x (k+2),. The. _P ) Sum } and->Is { u (k), u (k+1),. The term, u (k+n _P -1)}；

Defining a set of allowed steady-state fluctuation interval points as follows:

defining a set of points within the allowed steering wheel shake interval as:

the probability of the predicted state quantity and the control quantity being in the stable region output in the predicted rolling time domain window is P _A 、P _B The P is _A 、P _B The method comprises the following steps of:

the overshoot value is M:

wherein:respectively set->And->The set of each element taking absolute value epsilon _x ，ε _u The upper limit value of the stable region is defined respectively.

As a further improvement of the present invention, the iterative process of the adaptive optimization of the iterative strategy is:

while ΔKp > H2o// H is a given minimum inter-zone spacing;

for i=0 to 2n// establish a cyclic search for optimal Kp;

{ kp=kp+Δkp (i-n)/n; adjusting the proportionality coefficient Kp;

if M (i) > temM then// judging the quality of the overshoot according to the formula (10);

for j=0to 2n// establish a cyclic search for the optimal Ki;

{ ki= KiO j/n; adjusting the integral coefficient Ki;

judging whether the stability performance is good or bad according to a formula, wherein lf PA (j) PB (j) > temP then/;

temm=m; temp=pa (j) PB (j); the intermediate variable temM, temP is/are updated;

kp=kp; ki=ki; updating proportional and integral coefficients Kp and Ki;

end if

j++；}

end if

i++；}

Δkp=Δkp/n; the adjustment interval delta Kp of the proportional coefficient is/is updated;

End while。

as a further improvement of the present invention, the weight coefficient is defined as f _P-F The fusion function specifically comprises the following steps:

δ _control ＝f _P-F +(1-f _P-F )δ _F

wherein: p (P) _th Is probability P _A N is the system switching threshold, delta _control And restraining steering wheel rotation angle after fusion.

The invention has the beneficial effects that:

1: the invention provides a high-performance lane keeping control strategy based on rolling pretreatment, which can improve the steady-state performance of a system on the premise of ensuring the transient performance of the control system;

2: the rolling time domain window obtains an overshoot value of the vehicle and carries out self-adaptive optimization iteration according to the current system state, so that the PID steering wheel turning angle after the self-adaptive optimization iteration is obtained, the maximization of the steady state performance of the system is realized, the shaking degree of the steering wheel can be restrained on the premise of keeping the transient performance advantage of the fuzzy module through the obstacle function and the stable region probability, and in various different curvature road scenes, different vehicle speed running can be ensured to be driven along the lane keeping line to keep stable, and the driving of the deviated lane is not easy to occur.

Drawings

FIG. 1 is a control strategy block diagram embodying the present invention;

FIG. 2 is a two degree of freedom vehicle dynamics model;

FIG. 3 is a membership function curve;

FIG. 4 is a plot of the change in steering wheel angle for the fuzzy module control output;

fig. 5 is a graph of fluctuation of steady-state phase deviation versus control variable for traveling on paths of different road curvatures.

Detailed Description

The invention will now be described in further detail with reference to the drawings and examples. Wherein like parts are designated by like reference numerals. It should be noted that the words "front", "back", "left", "right", "upper" and "lower" used in the following description refer to directions in the drawings, and the words "bottom" and "top", "inner" and "outer" refer to directions toward or away from, respectively, the geometric center of a particular component.

Referring to fig. 1 to 5, a specific embodiment of a high performance lane keeping control system based on rolling preprocessing according to the present invention includes a guidance controller and a feedback controller integrated in a processor, in which a two-degree-of-freedom vehicle dynamics model is built, and the building of the vehicle dynamics model is shown in fig. 2.

The lane detection module comprises a detection unit and a guiding unit, the detection unit is used for detecting lane information in real time according to the ultrasonic detector to generate a lane image, forming a lane keeping line positioned at the center of a lane according to the lane image and displaying the lane keeping line in the lane image, and the guiding unit is used for calling the lane image and the lane keeping line displayed on the lane image and displaying the lane image and the lane keeping line on the vehicle-mounted display screen.

The steering controller comprises a fuzzy module, a PID module and a fusion module, wherein an obstacle function is configured in the fuzzy module, the fuzzy module outputs a fuzzy steering wheel corner according to the obstacle function, a feedforward unit is configured in the steering wheel restraining module, a feedforward strategy is configured in the feedforward unit, the feedforward strategy comprises the step of obtaining the feedforward steering wheel corner according to the wheelbase of a front axle and a rear axle of a vehicle, the step of outputting the steering wheel corner to be determined according to the feedforward steering wheel corner and the fuzzy steering wheel corner, a rolling time domain window designed based on a vehicle dynamics model is configured in the PID module, an iteration module is further configured in the PID module, a corner algorithm and a superposition strategy are arranged in the PID module, the PID module generates the PID steering wheel corner according to the corner algorithm, the superposition strategy further comprises the step of estimating the running state of the vehicle through the rolling time domain window according to the vehicle dynamics model, obtaining a vehicle overshoot value, the iteration strategy further comprises the step of carrying out self-adaptive optimization according to the rolling time domain window in a view range, obtaining an adjustment interval, and the iteration strategy further comprises the step of obtaining the adaptive steering wheel adjustment interval according to the rolling time domain window after the rolling time domain window overshoot value.

The fusion module is internally provided with a fusion function and a weight coefficient, the fusion module calls the steering wheel angle to be determined and the PID steering wheel angle to the fusion function, the steering wheel angle is restrained according to the fusion function, the processor sends a suppression steering wheel angle to a steering wheel suppression module, which controls a vehicle steering wheel suppression adjustment to keep the vehicle traveling centered along the lane keep line.

When the vehicle dynamics model is built, the lane line information and the vehicle posture information can be perceived through a camera sensor, an inertial sensor, a global positioning system and other sensors, the lane center line is used as a reference path during lane keeping control, and if the longitudinal speed v of the vehicle is the same as the reference path _x The model of the standard can be approximated as a linear steady system, and the specific state space equation is:

wherein: x is a state quantity, e _y E is the deviation of the vehicle centroid abscissa y (t) from the lane center abscissa yref (t) _y ＝y(t)-yref(t)，e _φ Is the heading angle phi of the vehicle _v Heading angle phi with lane center line _r Deviation of e _φ ＝φ _v -φ _r M is the mass of the whole vehicle, C _αf And C _αr Lateral deflection rigidity of front and rear wheels respectively, l _f And l _r The distances from the mass center to the front and rear axes are respectively I _z Delta is the front wheel angle for the moment of inertia about the vehicle center of mass perpendicular to the ground, which is the Z-axis.

A corner threshold value is also arranged in the fuzzy module;

the obstacle function is specifically:

δ _F ＝Δδ _F *l _δ

wherein: delta ^* Characterizing the angular threshold, l _δ Characterizing the barrier function, delta _F Characterizing a fuzzy steering wheel angle.

The feedforward strategy further comprises a feedforward algorithm, the steering wheel angle to be determined is calculated according to the feedforward algorithm, and the feedforward algorithm specifically comprises the following steps:

δ _out ＝δ _ff +δ _F

δ _ff ＝L*ρ

wherein: delta _out Characterizing steering angle delta _ff The feed-forward steering wheel angle is represented, L represents the wheelbase of the front and rear axles of the vehicle, and ρ represents the curvature of the road.

Fuzzy steering wheel angle delta _F An output control amount for a fuzzy controller based on the obstacle function;

table 1 shows a specific fuzzy logic rule table

Fuzzy rule 1: IF e _y 、de _y The p and v are NB, NB, NB, S and the THEN outputs K 'respectively' _p ，K′ _I ，K′ _D 、Δδ _F 、δ ^* OH, O, O, NB, B respectively;

fuzzy rule 2: IF e _y 、de _y The p and v are NS, NB, NB, S and the THEN outputs K 'respectively' _p ，K′ _I ，K′ _D 、Δδ _F 、δ ^* OH, O, O, NB, B respectively;

fuzzy rule …

Fuzzy rule 375: IF e _y 、de _y The p and v are PB, PB, PB, B and the THEN outputs K 'respectively' _p ，K′ _I ，K′ _D 、Δδ _F 、δ ^* H, H, O, PB, M respectively;

in the case where the vehicle speed is high, fluctuation of the control amount output by the fuzzy controller is small in order to make it whether it is a straight road or a curve.

The PID module is internally provided with a corner algorithm for generating the corner of the PID steering wheel, and the corner algorithm specifically comprises:

δ′ _out ＝δ _ff +δ _PID

wherein: e, e _y Characterization of lateral deviation, de _y Characterization of the rate of change of lateral deviation, K' _p 、k′ _I 、k′ _D Characterizing steering wheel angle coefficient, delta _PID Characterizing PID steering wheel angle.

The specific process for obtaining the vehicle overshoot value through the rolling time domain window comprises the following steps:

discretizing the state space equation based on a forward Euler method to obtain a state equation within a prediction step length NP range, wherein the state equation comprises the following steps:

wherein: n (N) _P K is the current time for predicting step length;

analyzing the control performance under the corrected proportional, integral and differential coefficients of the current fuzzy module:

definition setIs { x (k+1), x (k+2),. The. _P ) Sum } and->Is { u (k), u (k+1),. The term, u (k+n _P -1)}；

Defining a set of allowed steady-state fluctuation interval points as follows:

defining a set of points within the allowed steering wheel shake interval as:

the probability of the predicted state quantity and the control quantity being in the stable region output in the predicted rolling time domain window is P _A 、P _B The P is _A 、P _B The method comprises the following steps of:

the overshoot value is M:

wherein:respectively set->And->The set of each element taking absolute value epsilon _x ，ε _u The upper limit value of the stable region is defined respectively.

The iterative process of the adaptive optimization of the iterative strategy comprises the following steps:

while ΔKp > H2o// H is a given minimum inter-zone spacing;

for i=0 to 2n// establish a cyclic search for optimal Kp;

{ kp=kp+Δkp (i-n)/n; adjusting the proportionality coefficient Kp;

if M (i) > temM then// judging the quality of the overshoot according to the formula (10);

for j=0 to 2n// establish a cyclic search for the optimal Ki;

{ ki= KiO j/n; adjusting the integral coefficient Ki;

judging whether the stability performance is good or bad according to a formula, wherein lf PA (j) PB (j) > temP then/;

temm=m; temp=pa (j) PB (j); the intermediate variable temM, temP is/are updated;

kp=kp; ki=ki; updating proportional and integral coefficients Kp and Ki;

end if

j++；}

end if

i++；}

Δkp=Δkp/n; the adjustment interval delta Kp of the proportional coefficient is/is updated;

End while。

the weight coefficient is defined as f _P-F The fusion function specifically comprises the following steps:

δ _cohtrol ＝f _P-F +(1-f _P-F )δ _F

wherein: p (P) _th Is probability P _A N is the system switching threshold, delta _control And restraining steering wheel rotation angle after fusion.

Working principle and effect:

the invention provides a high-performance lane keeping control strategy based on rolling pretreatment, which can improve the steady-state performance of a control system on the premise of ensuring the transient performance of the control system, a rolling time domain window obtains an overshoot value of a vehicle and carries out self-adaptive optimization iteration according to the current system state, so that the PID steering wheel turning angle after the self-adaptive optimization iteration is obtained, the maximization of the steady-state performance of the system is realized, the jitter degree of the steering wheel can be restrained on the premise of keeping the transient performance advantage of a fuzzy module through an obstacle function and the probability of a stable region, and the stable driving along a lane keeping line can be ensured to be obtained in various different vehicle speed driving in various different curvature road scenes, so that the driving of a deviated lane is not easy to occur.

The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above examples, and all technical solutions belonging to the concept of the present invention belong to the protection scope of the present invention. It should be noted that modifications and adaptations to the present invention may occur to one skilled in the art without departing from the principles of the present invention and are intended to be within the scope of the present invention.

Claims (7)

1. A high-performance lane keeping control system based on rolling pretreatment is characterized in that: the system comprises a guide controller and a feedback controller which are integrated in a processor;

the feedback controller is internally provided with a lane detection module and a steering wheel suppression module, the lane detection module comprises a detection unit and a guiding unit, the detection unit detects lane information in real time according to an ultrasonic detector to generate a lane image, the guiding unit is used for calling the lane image and the lane keeping line displayed on the lane image and displaying the lane image and the lane keeping line on the vehicle-mounted display screen;

the steering controller comprises a fuzzy module, a PID module and a fusion module, wherein an obstacle function is configured in the fuzzy module, the fuzzy module outputs a fuzzy steering wheel corner according to the obstacle function, a feedforward unit is configured in the steering wheel restraining module, a feedforward strategy is configured in the feedforward unit, the feedforward strategy comprises the steps of obtaining the feedforward steering wheel corner according to the wheelbase of a front axle and a rear axle of a vehicle, outputting the steering wheel corner to be determined according to the feedforward steering wheel corner and the fuzzy steering wheel corner, a rolling time domain window designed based on a vehicle dynamics model is configured in the PID module, an iteration module is further configured in the PID module, a corner algorithm and a superposition strategy are arranged in the PID module, the PID module generates the PID steering wheel corner according to the corner algorithm, the superposition strategy further comprises the steps of estimating the running state of the vehicle through the rolling time domain window according to the vehicle dynamics model, obtaining a vehicle overshoot value, performing adaptive optimization according to the rolling time domain window in a view range, obtaining an adjustment interval, and further comprises obtaining an adaptive steering wheel adjustment interval according to the rolling time domain corner and the adaptive iteration optimization value after the rolling time domain window overshoot value;

the fusion module is internally provided with a fusion function and a weight coefficient, the fusion module calls the steering wheel angle to be determined and the PID steering wheel angle to the fusion function, the steering wheel angle is restrained according to the fusion function, the processor sends a suppression steering wheel angle to a steering wheel suppression module, which controls a vehicle steering wheel suppression adjustment to keep the vehicle traveling centered along the lane keep line.

2. A high performance lane keeping control system based on rolling pre-processing according to claim 1, wherein: a corner threshold value is also arranged in the fuzzy module;

the obstacle function is specifically:

δ _F ＝Δδ _F *l _δ

wherein: delta ^* Characterizing the angular threshold, l _δ Characterizing the barrier function, delta _F Characterizing a fuzzy steering wheel angle.

3. A high performance lane keeping control system based on rolling pre-processing according to claim 2, wherein: the feedforward strategy further comprises a feedforward algorithm, the steering wheel angle to be determined is calculated according to the feedforward algorithm, and the feedforward algorithm specifically comprises the following steps:

δ _out ＝δ _ff +δ _F

δ _ff ＝L*ρ

wherein: delta _out Characterizing steering angle delta _ff The feed-forward steering wheel angle is represented, L represents the wheelbase of the front and rear axles of the vehicle, and ρ represents the curvature of the road.

4. A high performance lane keeping control system based on rolling pre-processing according to claim 3, wherein: the PID module is internally provided with a corner algorithm for generating the corner of the PID steering wheel, and the corner algorithm specifically comprises:

δ′ _out ＝δ _ff +δ _PID

wherein: e, e _y Characterization of lateral deviation, de _y Characterization of the rate of change of lateral deviation, K' _p 、k′ _I 、k′ _D Characterizing steering wheel angle coefficient, delta _PID Characterizing PID steering wheel angle.

5. The high performance lane keeping control system based on rolling pre-processing of claim 4, wherein: the specific process for obtaining the vehicle overshoot value through the rolling time domain window comprises the following steps:

discretizing a state space equation based on a forward Euler method to obtain a prediction step length N _P The state equation in the range is:

wherein: n (N) _P K is the current time for predicting step length;

analyzing the control performance under the corrected proportional, integral and differential coefficients of the current fuzzy module:

definition setIs { x (k+1), x (k+2),. The. _P ) Sum } and->Is { u (k), u (k+1),. The term, u (k+n _P -1)}；

Defining a set of allowed steady-state fluctuation interval points as follows:

defining a set of points within the allowed steering wheel shake interval as:

the probability of the predicted state quantity and the control quantity being in the stable region output in the predicted rolling time domain window is P _A 、P _B The P is _A 、P _B The method comprises the following steps of:

the overshoot value is M:

wherein:respectively set->And->The set of each element taking absolute value epsilon _x ，ε _u The upper limit value of the stable region is defined respectively.

6. The high performance lane keeping control system based on rolling pre-processing of claim 5, wherein: the iterative process of the adaptive optimization of the iterative strategy comprises the following steps:

while Δkp > hdo// H is a given minimum inter-interval;

for i=0 to 2n// establish a cyclic search for optimal Kp;

{ kp=kp+Δkp (i-n)/n; adjusting the proportionality coefficient Kp;

if M (i) > temp then// judging the quality of the overshoot according to the formula (10);

for j=0 to 2n// establish a cyclic search for the optimal Ki;

{ ki=ki0×j/n; adjusting the integral coefficient Ki;

if PA (j) PB (j) > temP then// judging whether the stability performance is good or bad according to a formula;

temm=m; temp=pa (j) PB (j); the intermediate variable temM, temP is/are updated;

kp=kp; ki=ki; updating proportional and integral coefficients Kp and Ki;

endif

j++；}

endif

i++；}

Δkp=Δkp/n; the adjustment interval delta Kp of the proportional coefficient is/is updated;

End while。

7. the high performance lane keeping control system based on rolling pre-processing of claim 6, wherein: the weight coefficient is defined as f _P-F The fusion function specifically comprises the following steps:

δ _control ＝f _P-F +(1-f _P-F )δ _F

wherein: p (P) _th Is probability P _A N is the system switching threshold, delta _control And restraining steering wheel rotation angle after fusion.