CN108859648A - A kind of suspension damper damping control neural network based switching weighting coefficient determines method - Google Patents
A kind of suspension damper damping control neural network based switching weighting coefficient determines method Download PDFInfo
- Publication number
- CN108859648A CN108859648A CN201810933191.5A CN201810933191A CN108859648A CN 108859648 A CN108859648 A CN 108859648A CN 201810933191 A CN201810933191 A CN 201810933191A CN 108859648 A CN108859648 A CN 108859648A
- Authority
- CN
- China
- Prior art keywords
- suspension
- operating condition
- vehicle body
- damping force
- wheel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60G—VEHICLE SUSPENSION ARRANGEMENTS
- B60G17/00—Resilient suspensions having means for adjusting the spring or vibration-damper characteristics, for regulating the distance between a supporting surface and a sprung part of vehicle or for locking suspension during use to meet varying vehicular or surface conditions, e.g. due to speed or load
- B60G17/015—Resilient suspensions having means for adjusting the spring or vibration-damper characteristics, for regulating the distance between a supporting surface and a sprung part of vehicle or for locking suspension during use to meet varying vehicular or surface conditions, e.g. due to speed or load the regulating means comprising electric or electronic elements
- B60G17/0152—Resilient suspensions having means for adjusting the spring or vibration-damper characteristics, for regulating the distance between a supporting surface and a sprung part of vehicle or for locking suspension during use to meet varying vehicular or surface conditions, e.g. due to speed or load the regulating means comprising electric or electronic elements characterised by the action on a particular type of suspension unit
- B60G17/0157—Resilient suspensions having means for adjusting the spring or vibration-damper characteristics, for regulating the distance between a supporting surface and a sprung part of vehicle or for locking suspension during use to meet varying vehicular or surface conditions, e.g. due to speed or load the regulating means comprising electric or electronic elements characterised by the action on a particular type of suspension unit non-fluid unit, e.g. electric motor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
Abstract
The invention discloses suspension damper damping control neural network based switching weighting coefficients to determine method, including：Automobile monitoring data are acquired, there is the seven freedom differential equation of motion of linear character based on the vertical deviation comprising four vehicle body vertical deviation, vehicle body pitch angle, angle of heel and front and back wheels, and are translated into the sample data of neural network；Classified according to road conditions classification to the sample data, obtains the corresponding sample data sets of each road surface operating condition；Road conditions identification is carried out according to automobile Real-time Monitoring Data, and it is exported the governing equation comprising the handoff features weighting coefficient as control strategy, the present invention controls suspension damper damping force for different operating conditions, improves safety and the comfort of automobile, realizes steady switching.
Description
Technical field
The present invention relates to the multi-state switching control fields of vehicle active suspension, more particularly to one kind to be based on neural network
Switching control characteristic weighing coefficient determine method.
Background technique
Suspension system is the important component of vehicle.With the development of science and technology and control technology is constantly progressive, tradition
Passive type suspension system due to its parameter it is immutable, constrain suspension system effect performance, and active suspension system because
For Parameter adjustable, it can be achieved that best performance under each operating condition, and meet future automobile low-carbon, lightweight, intelligence, a
Property development trend, thus become research hot topic and trend.In order to improve comfort and the safety of automobile, Hyundai Motor
The increasingly developed of technology keeps suspension system more and more intelligent.Actively, semi-active suspension system makes vehicle by real-time control
The optimal of ride performance and control stability can be reached under various road surface operating conditions, have traditional suspension incomparable
Advantage.
In vehicle travel process, driving cycle constantly changes, and different control modes pair is taken for different operating conditions
Active suspension control, and improves comfort and the safety of automobile.When road surface, operating condition changes, switching control power
It will appear Caton, phenomena such as mutation, so the research of Active suspension multi-state switching control smoothly switched seems especially heavy
It wants.
Summary of the invention
The present invention has designed and developed suspension damper damping control switching neural network based weighting coefficient determination side
Method establishes neural network model for different operating conditions, is obtained under different operating conditions according to the vehicle condition data real-time monitored
The switching control coefficient of automotive suspension, and suspension damper damping force is controlled accordingly, improve the safety of automobile and is relaxed
Adaptive realizes steady switching.
Technical solution provided by the invention is：
A kind of suspension damper damping control neural network based switching weighting coefficient determines method, including：
Acquire automobile monitoring data, establish seven freedom differential equation of motion, by vehicle body vertical deviation, vehicle body pitch angle,
The vertical deviation of four wheels of angle of heel and front and back is converted into the sample data of neural network；
Classified according to road conditions classification to the sample data, obtains the corresponding sample data set of each road surface operating condition
It closes；
Neural network model is established respectively according to the corresponding sample data sets of each road surface operating condition, including：
Using the vehicle body vertical deviation, vehicle body pitch angle, four wheels of angle of heel and front and back vertical deviation as input
Layer vector constructs neural network, parses in neural network to input layer vector characteristics, and obtaining indicates the road surface operating condition pair
The vector group for the switching weighting coefficient answered；
All neural network models are permeated neural network；And
Road conditions identification is carried out according to automobile Real-time Monitoring Data, and the governing equation for switching weighting coefficient will be included
It is exported as control strategy；
Wherein, the road conditions classification includes：Straight road operating condition, gradient road surface operating condition and continuous deceleration strip.
Preferably, the differential equation of motion is：
The near front wheel catenary motion equation：
Wherein, mu1For the near front wheel nonspring carried mass, zflFor the near front wheel vertical deviation,For the near front wheel vertical velocity,
For the near front wheel vertical acceleration, ktFor tire stiffness, kfFor suspension rate, q1For the vertical deviation of the near front wheel road excitation, Fk1
For left front suspension shock-absorber damping force；
Off-front wheel catenary motion equation：
Wherein, mu2For off-front wheel nonspring carried mass, zfrFor off-front wheel vertical deviation,For off-front wheel vertical velocity,
For off-front wheel vertical acceleration, q2For the vertical deviation of off-front wheel road excitation, Fk2For right front suspension absorber damping force；
Left rear wheel catenary motion equation：
Wherein, mu3For left rear wheel nonspring carried mass, zrlFor left rear wheel vertical deviation,For left rear wheel vertical velocity,
For left rear wheel vertical acceleration, q3For the vertical deviation of left rear wheel road excitation, Fk3Left rear suspension absorber damping force；
Off hind wheel catenary motion equation：
Wherein, mu4For off hind wheel nonspring carried mass, zrrFor off hind wheel vertical deviation,For off hind wheel vertical velocity,
For off hind wheel vertical acceleration, q4For the vertical deviation of off hind wheel road excitation, Fk4Right rear suspension absorber damping force；
zs1For the vertical deviation of the near front wheel vehicle body and suspension junction, Zs1=Zs-Lfθ+φB/2；
zs2For the vertical deviation of off-front wheel vehicle body and suspension junction,
zs3For the vertical deviation of left rear wheel vehicle body and suspension junction,
zs4For the vertical deviation of off hind wheel vehicle body and suspension junction,
zsFor vehicle body vertical deviation, LfIt is mass center away from front suspension distance, LrIt is mass center away from the right hinge joint distance of suspension, φ is
Vehicle roll angle, θ be vehicle body pitch angle, B be left and right wheels away from；
Vehicle body catenary motion equation：
Vehicle body pitching movement equation:
The body roll equation of motion：
Wherein,For the vertical velocity of the near front wheel vehicle body and suspension junction,It is connect for off-front wheel vehicle body with suspension
The vertical velocity at place,For the vertical velocity of left rear wheel vehicle body and suspension junction,It is connect for off hind wheel vehicle body with suspension
The vertical velocity at place,For vehicle body rate of pitch,For body roll angular speed, IxFor the rotation on body roll direction
Inertia, IyFor the rotary inertia in vehicle body pitch orientation, CfSuspension front damper damped coefficient, CrFor the damping of suspension rear shock absorber
Coefficient, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension resistance of shock absorber
Power, Fk4Right rear suspension absorber damping force.
Preferably, the neural network is three-layer neural network model, is successively formatted to input layer vector,
Determine the input layer vector of three-layer neural networkThe input layer DUAL PROBLEMS OF VECTOR MAPPING
To hidden layer, the hidden layer vector is Y={ y1,y2,y3,y4···ym, m is node number, and output layer vector is
Wherein, zsFor vehicle body vertical deviation vector,For vehicle body vertical velocity, θ is vehicle body pitch angle,For vehicle body pitching
Angular speed, φ are vehicle roll angle,For body roll angular speed, zflFor front-wheel vertical deviation,For front-wheel vertical velocity,
zrlFor rear-wheel vertical deviation,For rear-wheel vertical velocity；ladma1Switch weighting coefficient matrix, ladma for straight road conditions2For
Gradient road conditions switch weighting coefficient matrix, ladma3Switch weighting coefficient matrix for continuous deceleration strip road conditions,For corresponding operating condition
The near front wheel suspension damper handoff features weighting coefficient,Off-front wheel suspension damper handoff features for corresponding operating condition add
Weight coefficient,Left rear wheel suspension damper for corresponding operating condition switches weighting coefficient,Off hind wheel suspension for corresponding operating condition subtracts
It shakes device and switches weighting coefficient.
Preferably, input layer parameter is formatted using following formula：
Wherein, xiFor index coefficient after formatting, TiFor input layer parameter, TimaxMaximum value is corresponded to for input layer parameter,
TiminMinimum value is corresponded to for input layer parameter.
Preferably, the node in hidden layer is 13.
Preferably, the governing equation of the switching weighting coefficient is：
Wherein, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension subtracts
Vibration device damping force, Fk4Right rear suspension absorber damping force,For the near front wheel suspension shock-absorber damping force under straight road operating condition,For the near front wheel suspension shock-absorber damping force under the operating condition of gradient road surface,Subtract for the near front wheel suspension under continuous deceleration strip operating condition
Vibration device damping force,For straight road operating condition lower right front wheel suspension absorber damping force,For off-front wheel under the operating condition of gradient road surface
Suspension shock-absorber damping force,For the near front wheel suspension shock-absorber damping force under continuous deceleration strip operating condition；For straight road operating condition
Lower left rear wheel suspension shock-absorber damping force,For left rear wheel suspension shock-absorber damping force under the operating condition of gradient road surface,It is continuous
Left rear wheel suspension shock-absorber damping force under deceleration strip operating condition；For straight road operating condition lower right rear wheel suspension absorber damping force,For gradient road surface operating condition lower right rear wheel suspension absorber damping force,For left rear wheel suspension vibration damping under continuous deceleration strip operating condition
Device damping force.
Beneficial effect of the present invention
The switching control characteristic weighing coefficient that the present invention has designed and developed a kind of neural network determines that method, the present invention are set
The switching control characteristic weighing coefficient that meter develops a kind of neural network determines method, establishes neural network for different operating conditions
Model obtains the control coefrficient of the automotive suspension under different operating conditions according to the vehicle condition data real-time monitored, and hangs accordingly
Frame damping force of vibration damper is controlled, and safety and the comfort of automobile are improved, and realizes steady switching, when road surface operating condition occurs
When variation, phenomena such as switching control power will appear Caton, mutation, the control of Active suspension multi-state is realized using neural network algorithm
System smoothly switches.
Detailed description of the invention
Fig. 1 is seven freedom Full Vehicle Dynamics illustraton of model of the present invention.
Fig. 2 is the schematic diagram of neural network model of the present invention.
Fig. 3 is multi-state road surface of the present invention simulation result diagram.
Fig. 4 is straight road conditions handoff features weighting coefficient of the present invention.
Fig. 5 is gradient road conditions handoff features weighting coefficient of the present invention.
Fig. 6 is continuous deceleration strip road conditions handoff features weighting coefficient of the present invention.
Specific embodiment
Present invention will be described in further detail below with reference to the accompanying drawings, to enable those skilled in the art referring to specification
Text can be implemented accordingly.
As shown in Figure 1, suspension damper damping control handoff features weighting neural network based system provided by the invention
Number determines method, including：Establish the vertical position comprising vehicle body vertical deviation, vehicle body pitch angle, angle of heel and the wheel of front and back four
Move the seven freedom differential equation of motion with nonlinear characteristic：
The near front wheel catenary motion equation：
Wherein, mu1For the near front wheel nonspring carried mass, zflFor the near front wheel vertical deviation,For the near front wheel vertical velocity,
For the near front wheel vertical acceleration, ktFor tire stiffness, kfFor suspension rate, q1For the vertical deviation of the near front wheel road excitation, Fk1
For left front suspension shock-absorber damping force；
Off-front wheel catenary motion equation：
Wherein, mu2For off-front wheel nonspring carried mass, zfrFor off-front wheel vertical deviation,For off-front wheel vertical velocity,
For off-front wheel vertical acceleration, q2For the vertical deviation of off-front wheel road excitation, Fk2For right front suspension absorber damping force；
Left rear wheel catenary motion equation：
Wherein, mu3For left rear wheel nonspring carried mass, zrlFor left rear wheel vertical deviation,For left rear wheel vertical velocity,
For left rear wheel vertical acceleration, q3For the vertical deviation of left rear wheel road excitation, Fk3Left rear suspension absorber damping force；
Off hind wheel catenary motion equation：
Wherein, mu4For off hind wheel nonspring carried mass, zrrFor off hind wheel vertical deviation,For off hind wheel vertical velocity,
For off hind wheel vertical acceleration, q4For the vertical deviation of off hind wheel road excitation, Fk4Right rear suspension absorber damping force；
zs1For the vertical deviation of the near front wheel vehicle body and suspension junction, Zs1=Zs-Lfθ+φB/2；
zs2For the vertical deviation of off-front wheel vehicle body and suspension junction,
zs3For the vertical deviation of left rear wheel vehicle body and suspension junction,
zs4For the vertical deviation of off hind wheel vehicle body and suspension junction,
zsFor vehicle body vertical deviation, LfIt is mass center away from front suspension distance, LrIt is mass center away from the right hinge joint distance of suspension, φ is
Vehicle roll angle, θ be vehicle body pitch angle, B be left and right wheels away from；
Vehicle body catenary motion equation：
Vehicle body pitching movement equation:
The body roll equation of motion：
Wherein,For the vertical velocity of the near front wheel vehicle body and suspension junction,It is connect for off-front wheel vehicle body with suspension
The vertical velocity at place,For the vertical velocity of left rear wheel vehicle body and suspension junction,It is connect for off hind wheel vehicle body with suspension
The vertical velocity at place,For vehicle body rate of pitch,For body roll angular speed, IxIt is used for the rotation on body roll direction
Amount, IyFor the rotary inertia in vehicle body pitch orientation, CfSuspension front damper damped coefficient, CrIt is damped for suspension rear shock absorber and is
Number, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension resistance of shock absorber
Power, Fk4Right rear suspension absorber damping force.
Acquisition automobile monitoring data simultaneously classify to sample data according to road conditions classification, obtain each road surface operating condition pair
The sample data sets answered；Neural network model, road conditions are established respectively according to the corresponding sample data sets of each road surface operating condition
Classification includes at least：Three straight road operating condition, gradient road surface operating condition and continuous deceleration strip classifications, establish at least three nerve nets
The training process of one of neural network model is given below in network model, by taking straight road conditions neural network model as an example：
As shown in Fig. 2, neural network module uses BP neural network network, the state of seven freedom whole vehicle model is rung
Answer vectorIt is inputted as training sample from input layer, output signal isIt is reality output vector；
zsFor vehicle body vertical deviation vector,For vehicle body vertical velocity, θ is vehicle body pitch angle,For vehicle body pitch angle speed
Degree, φ is vehicle roll angle,For body roll angular speed, zflFor front-wheel vertical deviation,For front-wheel vertical velocity, zrlFor
Rear-wheel vertical deviation,For rear-wheel vertical velocity.
Input layer parameter is formatted using following formula：
Wherein, xiFor index coefficient after formatting, TiFor input layer parameter, inputting parameter includes：
TimaxFor input layer parameterCorresponding maximum value, TiminFor input layer parameterCorresponding minimum value.zsFor vehicle body vertical deviation vector,For vehicle body vertical velocity, θ
For vehicle body pitch angle,For vehicle body rate of pitch, φ is vehicle roll angle,For body roll angular speed, zflIt hangs down for front-wheel
To displacement,For front-wheel vertical velocity, zrlFor rear-wheel vertical deviation,For rear-wheel vertical velocity
Carry out the training of BP neural network：After establishing BP neural network nodal analysis method, BP neural network can be carried out
Training.The sample of training, and the connection between given input node i and hidden layer node j are obtained according to the empirical data of product
Weight wij, hidden node j and output node layer k between connection weight wjk, the threshold θ of hidden node jj, export node layer k
Threshold value wij、wjk、θj、θkIt is the random number between -1 to 1.
In the training process, w is constantly correctedijAnd wjkValue, until systematic error be less than or equal to anticipation error when, complete
The training process of neural network.
Wherein, Δ δ is training error,For output layer vector, n=1,2,3,4,For sample coefficient.
The learning rules of system are that the error sum of squares of desired output and reality output is allowed to reach a certain setting value, with this
To adjust connection weight and threshold vector.After error is reduced to setting value, system stops study, weight and threshold value quilt at this time
It is retained in internal system, becomes internal system knowledge.
Node in hidden layer is 13, and activation primitive is to use symmetrical saturated linear transmission function, and output layer number of nodes is
3, activation primitive is log-S type transmission function, and training method is selected as self study applicability algorithm.
For hidden layer, activation primitive f [] uses symmetrical saturated linear transmission function (satlins)
For output layer, activation primitive f [] uses log-S type transmission function,
IW { 1,1 } and LW { 2,1 } is connection weight between layers, and b { 1,1 } and b { 2,1 } are weight.
As shown in table 1, given the value of each node in one group of training sample and training process.
Each nodal value of 1 training process of table
Referring to above-mentioned training method, the mind of three straight road operating condition, gradient road surface operating condition and continuous deceleration strip classifications is obtained
Through network model, three neural network models are integrated into a neural network model
Output layer vector is
Wherein, ladma1Switch weighting coefficient matrix, ladma for straight road conditions2Switch weighting coefficient square for gradient road conditions
Battle array, ladma3Switch weighting coefficient matrix for continuous deceleration strip road conditions,To switch weighting coefficient to the near front wheel of operating condition,Off-front wheel for corresponding operating condition switches weighting coefficient,Left rear wheel for corresponding operating condition switches weighting coefficient,For corresponding work
The off hind wheel of condition switches weighting coefficient.
It obtains comprising lamda1、lamda2、lamda3It is the governing equation of characteristic weighing coefficient described in reality output vector
Including：
Wherein, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension subtracts
Vibration device damping force, Fk4Right rear suspension absorber damping force, F10For left front suspension shock-absorber damping force under straight road operating condition, F20For
Left front suspension shock-absorber damping force, F under the operating condition of gradient road surface30For left front suspension shock-absorber damping force under continuous deceleration strip operating condition.
Experimental example
The multi-state road surface of the straight road of machine, slope road, continuous deceleration strip is agreed to emulate mould as shown in figure 3, initially setting up
Type, time 0-15s.Wherein 0-5s is random straight road operating condition, 5-10s is gradient road surface operating condition, 10-15s is continuously to accelerate
Band operating condition.
The random straight road operating condition of 0-5s is described using filtering white noise road surface model, and the equation of motion is：
Wherein, G0=6.4e-6 is road roughness coefficient, and w (t) is the white Gaussian noise input that t moment mean value is zero,
V=20m/s is speed, f0=0.2Hz is lower limiting frequency.
It is 3% mould that the gradient, which is added, in the gradient road surface operating condition of 5-10s on the basis of filtering the random straight road that white noise is established
Block.
The elimination gradient module that the gradient is -3% is added at the end 10s in the continuous deceleration strip operating condition of 10-15s, and the
Pulse module is added in the end 11s, this pulse module describes high 0.04m wide 0.3m, is spaced 2 meters between two deceleration strips.
Automobile monitoring data, the motor racing equation based on 7 degree of freedom are acquired, and artificial road condition classification is to sample number
According to classifying, the corresponding sample data sets of each road surface operating condition are obtained
It is inputted as training sample from the input layer of neural network, the learning rules of system are to allow the mistake of desired output and reality output
Poor quadratic sum reaches a certain setting value, adjusts connection weight and threshold vector with this.After error is reduced to setting value, system
Stop study, weight and threshold value at this time is retained in internal system, becomes internal system knowledge.As a preference, when instruction
Practice errorWhen, training terminates；Wherein, Δ δ is training error,For output layer vector, n=1,
2,3,4,It is threshold value, lamda for corresponding sample coefficient K1、lamda2、lamda3It is reality output vector.
Output layer vector is
As Figure 4-Figure 6, using neural network algorithm, the corresponding three characteristic weighing coefficients of three kinds of road surface operating conditions are obtained
Mean filter is carried out to three weighting coefficients respectively, retains the variation that road surface operating condition changes the moment, the value of remaining period carries out
Filtering processing, so filtering since the 2nd second, is filtered lamda1 since preceding two second data is distorted, and retains 7 seconds front and backs
Variation in 0.05 second；Lamda2 is filtered, the variation in 0.05 second of 7 seconds and 12 seconds front and backs is retained；Lamda3 is filtered
Wave retains the variation in 0.05 second of the 12nd second front and back.Three weighting coefficients are obtained after filtering.
It obtains comprising lamda1、lamda2、lamda3It is that the governing equation of weighting coefficient described in reality output vector includes：
Wherein, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension subtracts
Vibration device damping force, Fk4Right rear suspension absorber damping force,For the near front wheel bearing spring vertical load power under straight road operating condition,For the near front wheel suspension shock-absorber damping force under the operating condition of gradient road surface,For the near front wheel suspension vibration damping under continuous deceleration strip operating condition
Device damping force,For straight road operating condition lower right front wheel suspension absorber damping force,It is outstanding for off-front wheel under the operating condition of gradient road surface
Frame absorber damping force,For the near front wheel suspension shock-absorber damping force under continuous deceleration strip operating condition；For under straight road operating condition
Left rear wheel suspension shock-absorber damping force,For left rear wheel suspension shock-absorber damping force under the operating condition of gradient road surface,Continuously to subtract
Left rear wheel suspension shock-absorber damping force under speed belt operating condition；For straight road operating condition lower right rear wheel suspension absorber damping force,
For gradient road surface operating condition lower right rear wheel suspension absorber damping force,For left rear wheel suspension shock-absorber under continuous deceleration strip operating condition
Damping force.
As it can be seen that taking progress of the different control modes to Active suspension for different operating conditions using neural network algorithm
Control, establishes neural network model for different operating conditions, is obtained under different operating conditions according to the vehicle condition data real-time monitored
Automotive suspension control coefrficient, and suspension damper damping force is controlled accordingly, improves the safety of automobile and comfortable
Property, steady switching is realized when road surface operating condition changes, phenomena such as switching control power will appear Caton, mutation, using nerve
Network algorithm realizes smoothly switching for Active suspension multi-state control.
Although the embodiments of the present invention have been disclosed as above, but its institute not only in the description and the implementation
Column use, it can be fully applied to various fields suitable for the present invention, for those skilled in the art, can hold
It changes places and realizes other modification, therefore without departing from the general concept defined in the claims and the equivalent scope, the present invention is simultaneously
It is not limited to specific details and legend shown and described herein.
Claims (6)
1. a kind of suspension damper damping control switching weighting coefficient neural network based determines method, which is characterized in that packet
It includes：
Automobile monitoring data are acquired, seven freedom differential equation of motion is established, by vehicle body vertical deviation, vehicle body pitch angle, are rolled
The vertical deviation of four wheels in angle and front and back is converted into the sample data of neural network；
Classified according to road conditions classification to the sample data, obtains the corresponding sample data sets of each road surface operating condition；
Neural network model is established respectively according to the corresponding sample data sets of each road surface operating condition, including：
Using the vehicle body vertical deviation, vehicle body pitch angle, four wheels of angle of heel and front and back vertical deviation as input layer to
Amount building neural network, parses input layer vector characteristics in neural network, and obtaining indicates that the road surface operating condition is corresponding
Switch the vector group of weighting coefficient；
All neural network models are permeated neural network；And
Road conditions identification is carried out according to automobile Real-time Monitoring Data, and using the governing equation comprising the switching weighting coefficient as control
System strategy output；
Wherein, the road conditions classification includes：Straight road operating condition, gradient road surface operating condition and continuous deceleration strip operating condition.
2. suspension damper damping control neural network based according to claim 1 switches weighting coefficient determination side
Method, which is characterized in that the differential equation of motion is：
The near front wheel catenary motion equation：
Wherein, mu1For the near front wheel nonspring carried mass, zflFor the near front wheel vertical deviation,For the near front wheel vertical velocity,It is left front
Take turns vertical acceleration, ktFor tire stiffness, kfFor suspension rate, q1For the vertical deviation of the near front wheel road excitation, Fk1It is left front outstanding
Frame absorber damping force；
Off-front wheel catenary motion equation：
Wherein, mu2For off-front wheel nonspring carried mass, zfrFor off-front wheel vertical deviation,For off-front wheel vertical velocity,Before the right side
Take turns vertical acceleration, q2For the vertical deviation of off-front wheel road excitation, Fk2For right front suspension absorber damping force；
Left rear wheel catenary motion equation：
Wherein, mu3For left rear wheel nonspring carried mass, zrlFor left rear wheel vertical deviation,For left rear wheel vertical velocity,It is left back
Take turns vertical acceleration, q3For the vertical deviation of left rear wheel road excitation, Fk3Left rear suspension absorber damping force；
Off hind wheel catenary motion equation：
Wherein, mu4For off hind wheel nonspring carried mass, zrrFor off hind wheel vertical deviation,For off hind wheel vertical velocity,Behind the right side
Take turns vertical acceleration, q4For the vertical deviation of off hind wheel road excitation, Fk4Right rear suspension absorber damping force；
zs1For the vertical deviation of the near front wheel vehicle body and suspension junction, Zs1=Zs-Lfθ+φB/2；
zs2For the vertical deviation of off-front wheel vehicle body and suspension junction,
zs3For the vertical deviation of left rear wheel vehicle body and suspension junction,
zs4For the vertical deviation of off hind wheel vehicle body and suspension junction,
zsFor vehicle body vertical deviation, LfIt is mass center away from front suspension distance, LrIt is mass center away from the right hinge joint distance of suspension, φ is vehicle body
Angle of heel, θ be vehicle body pitch angle, B be left and right wheels away from；
Vehicle body catenary motion equation：
Vehicle body pitching movement equation:
The body roll equation of motion：
Wherein,For the vertical velocity of the near front wheel vehicle body and suspension junction,For hanging down for off-front wheel vehicle body and suspension junction
To speed,For the vertical velocity of left rear wheel vehicle body and suspension junction,For the vertical of off hind wheel vehicle body and suspension junction
Speed,For vehicle body rate of pitch,For body roll angular speed, IxFor the rotary inertia on body roll direction, IyFor vehicle
Rotary inertia in body pitch orientation, CfSuspension front damper damped coefficient, CrFor suspension rear shock absorber damped coefficient, Fk1For a left side
Front suspension absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension absorber damping force, Fk4Right rear overhang
Frame absorber damping force.
3. suspension damper damping control neural network based according to claim 2 switches weighting coefficient determination side
Method, which is characterized in that the neural network is three-layer neural network model, is successively formatted to input layer vector, is determined
The input layer vector of three-layer neural networkThe input layer DUAL PROBLEMS OF VECTOR MAPPING is to hidden
Containing layer, the hidden layer vector is Y={ y1,y2,y3,y4…ym, m is node number, and output layer vector isI=1,2,3；
Wherein, zsFor vehicle body vertical deviation vector,For vehicle body vertical velocity, θ is vehicle body pitch angle,For vehicle body pitch angle speed
Degree, φ is vehicle roll angle,For body roll angular speed, zflFor front-wheel vertical deviation,For front-wheel vertical velocity, zrlIt is rear
Vertical deviation is taken turns,For rear-wheel vertical velocity；ladma1Switch weighting coefficient matrix, ladma for straight road conditions2For gradient road conditions
Switch weighting coefficient matrix, ladma3Switch weighting coefficient matrix for continuous deceleration strip road conditions,The near front wheel for corresponding operating condition is outstanding
Frame damper handoff features weighting coefficient,For the off-front wheel suspension damper handoff features weighting coefficient of corresponding operating condition,For
The left rear wheel suspension damper of corresponding operating condition switches weighting coefficient,Off hind wheel suspension damper for corresponding operating condition switches weighting
Coefficient.
4. suspension damper damping control neural network based according to claim 3 switches weighting coefficient determination side
Method, which is characterized in that input layer parameter is formatted using following formula：
Wherein, xiFor index coefficient after formatting, TiFor input layer parameter, TimaxMaximum value, T are corresponded to for input layer parameteriminFor
Input layer parameter corresponds to minimum value.
5. suspension damper damping control neural network based according to claim 3 switches weighting coefficient determination side
Method, which is characterized in that the node in hidden layer is 13.
6. suspension damper damping control neural network based according to claim 5 switches weighting coefficient determination side
Method, which is characterized in that it is described switching weighting coefficient governing equation be：
Wherein, Fk1For left front suspension shock-absorber damping force, Fk2For right front suspension absorber damping force, Fk3Left rear suspension damper
Damping force, Fk4Right rear suspension absorber damping force,For the near front wheel suspension shock-absorber damping force under straight road operating condition,For slope
The near front wheel suspension shock-absorber damping force under the operating condition of road surface is spent,It is damped for the near front wheel suspension shock-absorber under continuous deceleration strip operating condition
Power,For straight road operating condition lower right front wheel suspension absorber damping force,For gradient road surface operating condition lower right front wheel suspension damper
Damping force,For the near front wheel suspension shock-absorber damping force under continuous deceleration strip operating condition；For left rear wheel suspension under straight road operating condition
Absorber damping force,For left rear wheel suspension shock-absorber damping force under the operating condition of gradient road surface,It is left under continuous deceleration strip operating condition
Rear wheel suspension absorber damping force；For straight road operating condition lower right rear wheel suspension absorber damping force,For gradient road surface operating condition
Lower right rear wheel suspension absorber damping force,For left rear wheel suspension shock-absorber damping force under continuous deceleration strip operating condition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810933191.5A CN108859648B (en) | 2018-08-16 | 2018-08-16 | Suspension shock absorber damping control switching weighting coefficient determination method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810933191.5A CN108859648B (en) | 2018-08-16 | 2018-08-16 | Suspension shock absorber damping control switching weighting coefficient determination method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108859648A true CN108859648A (en) | 2018-11-23 |
| CN108859648B CN108859648B (en) | 2020-02-18 |
Family
ID=64319227
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810933191.5A Active CN108859648B (en) | 2018-08-16 | 2018-08-16 | Suspension shock absorber damping control switching weighting coefficient determination method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108859648B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110020499A (en) * | 2019-04-19 | 2019-07-16 | 南京航空航天大学 | A kind of road surface fluctuating disturbance identification method based on inverse Wilson method |
| CN113183707A (en) * | 2021-05-21 | 2021-07-30 | 辽宁工业大学 | Design method and system of whole-vehicle active suspension self-adaptive fixed-period sampling controller |
| CN113183709A (en) * | 2021-06-04 | 2021-07-30 | 合肥工业大学 | Automobile electric control suspension pre-aiming control method |
| CN113570057A (en) * | 2021-09-27 | 2021-10-29 | 岚图汽车科技有限公司 | Vehicle wheel center vertical displacement measuring method and device based on model training |
| CN113752773A (en) * | 2020-06-02 | 2021-12-07 | 丰田自动车株式会社 | Vibration damping control device and vibration damping control method |
| CN113752770A (en) * | 2020-06-03 | 2021-12-07 | 观致汽车有限公司 | Semi-active suspension control system and method, computer storage medium, and electronic device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0870132A1 (en) * | 1995-07-26 | 1998-10-14 | Siemens Aktiengesellschaft | Circuitry for controlling a running or driving system of a motor vehicle |
| US6212466B1 (en) * | 2000-01-18 | 2001-04-03 | Yamaha Hatsudoki Kabushiki Kaisha | Optimization control method for shock absorber |
| CN101269618A (en) * | 2008-04-25 | 2008-09-24 | 江苏大学 | Control method for damping value of electronic control air spring damping apparatus with three-gear |
| CN101367324A (en) * | 2008-10-15 | 2009-02-18 | 江苏大学 | Pavement grade prediction technique based on electronic control air spring vehicle altimetric sensor |
| CN108274969A (en) * | 2018-02-27 | 2018-07-13 | 江苏大学 | A kind of feed energy suspension and its damping method for handover control based on speed subregion |
-
2018
- 2018-08-16 CN CN201810933191.5A patent/CN108859648B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0870132A1 (en) * | 1995-07-26 | 1998-10-14 | Siemens Aktiengesellschaft | Circuitry for controlling a running or driving system of a motor vehicle |
| US6212466B1 (en) * | 2000-01-18 | 2001-04-03 | Yamaha Hatsudoki Kabushiki Kaisha | Optimization control method for shock absorber |
| CN101269618A (en) * | 2008-04-25 | 2008-09-24 | 江苏大学 | Control method for damping value of electronic control air spring damping apparatus with three-gear |
| CN101367324A (en) * | 2008-10-15 | 2009-02-18 | 江苏大学 | Pavement grade prediction technique based on electronic control air spring vehicle altimetric sensor |
| CN108274969A (en) * | 2018-02-27 | 2018-07-13 | 江苏大学 | A kind of feed energy suspension and its damping method for handover control based on speed subregion |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110020499A (en) * | 2019-04-19 | 2019-07-16 | 南京航空航天大学 | A kind of road surface fluctuating disturbance identification method based on inverse Wilson method |
| CN113752773A (en) * | 2020-06-02 | 2021-12-07 | 丰田自动车株式会社 | Vibration damping control device and vibration damping control method |
| CN113752770A (en) * | 2020-06-03 | 2021-12-07 | 观致汽车有限公司 | Semi-active suspension control system and method, computer storage medium, and electronic device |
| CN113752770B (en) * | 2020-06-03 | 2023-09-15 | 观致汽车有限公司 | Semi-active suspension control system and method, computer storage medium, and vehicle |
| CN113183707A (en) * | 2021-05-21 | 2021-07-30 | 辽宁工业大学 | Design method and system of whole-vehicle active suspension self-adaptive fixed-period sampling controller |
| CN113183707B (en) * | 2021-05-21 | 2024-04-02 | 辽宁工业大学 | Design method and system of self-adaptive fixed-period sampling controller of full-vehicle active suspension |
| CN113183709A (en) * | 2021-06-04 | 2021-07-30 | 合肥工业大学 | Automobile electric control suspension pre-aiming control method |
| CN113183709B (en) * | 2021-06-04 | 2022-09-27 | 合肥工业大学 | Preview control method for automobile electric control suspension |
| CN113570057A (en) * | 2021-09-27 | 2021-10-29 | 岚图汽车科技有限公司 | Vehicle wheel center vertical displacement measuring method and device based on model training |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108859648B (en) | 2020-02-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108859648A (en) | A kind of suspension damper damping control neural network based switching weighting coefficient determines method | |
| CN104105628B (en) | The control device of vehicle and the control method of vehicle | |
| CN104080627B (en) | The control device of vehicle | |
| CN106985627B (en) | Vehicle road surface recognition system and suspension mode switching method | |
| CN104080671B (en) | The control device of vehicle and the control method of vehicle | |
| CN104203609B (en) | The control device of vehicle and the control method of vehicle | |
| CN104080672B (en) | The control device of vehicle and the control method of vehicle | |
| Yoshimura et al. | Steering and suspension system of a full car model using fuzzy reasoning based on single input rule modules | |
| CN108891221A (en) | A kind of active suspension system and its working method based on mode energy distribution method | |
| Sharma | Ride, eigenvalue and stability analysis of three-wheel vehicle using Lagrangian dynamics | |
| Caponetto et al. | A soft computing approach to fuzzy sky-hook control of semiactive suspension | |
| CN105539052B (en) | A kind of controllable suspension sliding formwork tracking controller using vehicle plateau as reference | |
| CN109927501A (en) | A kind of intelligent control method of Vehicle Semi-active Suspension System | |
| CN102501737B (en) | Intelligent particle swarm fuzzy hybrid control method for automotive semi-active suspension systems | |
| CN102975587B (en) | Vehicle semiactive suspension based on double controllable dampers and control method thereof | |
| CN107054353A (en) | The driving-force control apparatus of vehicle | |
| Fang et al. | Semi-active suspension of a full-vehicle model based on double-loop control | |
| CN108345218A (en) | Vehicle active suspension PID controller design method based on teaching optimization algorithm | |
| CN113591360B (en) | Magneto-rheological damper structural parameter optimization method based on whole vehicle dynamics model | |
| Wang et al. | Study of semi-active suspension control strategy based on driving behaviour characteristics | |
| Liu et al. | Study of ride comfort of active suspension based on model reference neural network control system | |
| Gao et al. | Optimization of electrically interconnected suspension for vibration control | |
| CN110765554A (en) | Intelligent control method of automobile semi-active suspension system based on TS model | |
| Chen et al. | Study on control of an two hub-motor electric vehicle | |
| Mohan et al. | Comfort Parameters Tuning Analysis for Vehicle Suspension Pitch Performance |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |