CN107168104A  Pure electric intelligent automobile longitudinal method for controlling driving speed based on observer  Google Patents
Pure electric intelligent automobile longitudinal method for controlling driving speed based on observer Download PDFInfo
 Publication number
 CN107168104A CN107168104A CN201710483937.2A CN201710483937A CN107168104A CN 107168104 A CN107168104 A CN 107168104A CN 201710483937 A CN201710483937 A CN 201710483937A CN 107168104 A CN107168104 A CN 107168104A
 Authority
 CN
 China
 Prior art keywords
 speed
 torque
 formula
 moment
 drive shaft
 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
 238000005457 optimization Methods 0.000 claims abstract description 41
 238000004088 simulation Methods 0.000 claims abstract description 29
 238000005096 rolling process Methods 0.000 claims abstract description 17
 238000004891 communication Methods 0.000 claims abstract description 7
 238000009826 distribution Methods 0.000 claims abstract description 7
 230000000717 retained Effects 0.000 claims abstract description 4
 238000000034 methods Methods 0.000 claims description 15
 230000005540 biological transmission Effects 0.000 claims description 10
 230000001133 acceleration Effects 0.000 claims description 8
 238000005070 sampling Methods 0.000 claims description 7
 238000009795 derivation Methods 0.000 claims description 6
 230000005484 gravity Effects 0.000 claims description 4
 229940004296 Formula 21 Drugs 0.000 claims description 3
 239000011159 matrix materials Substances 0.000 claims description 3
 230000000306 recurrent Effects 0.000 claims description 3
 239000003570 air Substances 0.000 claims 7
 239000000203 mixtures Substances 0.000 claims 2
 280000312338 Pavement companies 0.000 claims 1
 230000000694 effects Effects 0.000 description 11
 238000010586 diagram Methods 0.000 description 8
 239000007787 solid Substances 0.000 description 8
 238000005516 engineering process Methods 0.000 description 5
 238000004458 analytical method Methods 0.000 description 4
 238000004422 calculation algorithm Methods 0.000 description 4
 230000005611 electricity Effects 0.000 description 4
 230000035852 Tmax Effects 0.000 description 3
 238000011068 load Methods 0.000 description 3
 238000004364 calculation method Methods 0.000 description 2
 229910052744 lithium Inorganic materials 0.000 description 2
 206010039203 Road traffic accident Diseases 0.000 description 1
 230000037396 body weight Effects 0.000 description 1
 238000006243 chemical reaction Methods 0.000 description 1
 238000002485 combustion reaction Methods 0.000 description 1
 235000013399 edible fruits Nutrition 0.000 description 1
 238000003912 environmental pollution Methods 0.000 description 1
 239000002828 fuel tank Substances 0.000 description 1
 230000002045 lasting Effects 0.000 description 1
 150000002641 lithium Chemical group 0.000 description 1
 WHXSMMKQMYFTQSUHFFFAOYSAN lithium Chemical compound data:image/svg+xml;base64,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 data:image/svg+xml;base64,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 [Li] WHXSMMKQMYFTQSUHFFFAOYSAN 0.000 description 1
 238000004519 manufacturing process Methods 0.000 description 1
 238000003825 pressing Methods 0.000 description 1
 230000000452 restraining Effects 0.000 description 1
 230000001629 suppression Effects 0.000 description 1
 230000001360 synchronised Effects 0.000 description 1
 230000001131 transforming Effects 0.000 description 1
Classifications

 G—PHYSICS
 G05—CONTROLLING; REGULATING
 G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
 G05B17/00—Systems involving the use of models or simulators of said systems
 G05B17/02—Systems involving the use of models or simulators of said systems electric
Abstract
A kind of pure electric intelligent automobile longitudinal method for controlling driving speed based on observer, belongs to technical field of automobile control.The purpose of the present invention is to design controller using the rolling time horizon system optimizing control based on observer, go out the torque of operator demand by controller optimization, then it is driven and braking torque distribution, so as to realize the pure electric intelligent automobile longitudinal method for controlling driving speed based on observer that longitudinal speed is effectively controlled.The present invention realizes Matlab/Simulink and AMESim associative simulation, the interface module communicated with simulink is added in AMESim interfaces, after System build, model information in AMESim is retained in Simulink in the form of S function, so as to realize both associative simulation and communication.Present invention is generally directed to pure electric intelligent automobile longitudinal speed control problem, observer is designed for the important parameter of system, rolling time horizon system optimizing control can be good at completing online optimization solution, while processing constraint that can be dominant.
Description
Technical field
The invention belongs to technical field of automobile control.
Background technology
In order to reduce the generation of traffic accident, reduction internalcombustion engines vehicle is adjoint to energy resource consumption and the influence of environmental pollution
The development of internet, information, electronics and intellectual technology, the intellectuality of automobile and motorized technology, which have become, solves abovementioned ask
The effective way of topic.In recent years, the famous internet such as the wellknown automobile manufacturing enterprise such as masses, BMW, Audi and Baidu, Google
Enterprise, all in continuous increase to the manpower of intelligent driving automotive field, the input of financial resources, to seize the forward position skill of intelligent driving
Art.The development of intelligent driving technology will necessarily lead the major transformation of an automobile industry new round.Longitudinal speed control is used as pure electricity
The bottom control algorithm of dynamic intelligent automobile, its control effect directly affects the property such as security and ride comfort of intelligent automobile
Energy.For pure electric intelligent automobile, due to the fast response time of motor, the moment of torsion and rotating speed of motor are easily obtained, and this is pure electricity
Longitudinal speed control of dynamic intelligent automobile provides good basic condition.It is directed to centralized pure electric intelligent automobile longitudinal car
Speed control, mainly there is problems with：
1. the control of intelligent automobile longitudinal direction speed, i.e., reasonably produce drive demand torque by designing controller（Including driving
Torque and braking moment）, so as to realize the tracing control of longitudinal speed.
2. pure electric intelligent automobile longitudinal vehicle speed control system has nonlinear.Meanwhile, the output of controller, which will be met, holds
The actual maximum output torque of the firm constraints of row device motor and brake, i.e. driving and braking moment signal no more than motor
With the maximum braking moment of brake.
3. electric automobile needs electrical source of power, existing frequentlyused is lithium battery group, to be powered to motor, the power supply electricity of motor
Pressure also contributes to the maximum output torque of motor, therefore also must be considered that battery pack is defeated when considering the maximum output torque of motor
It is a change constraint to go out the actual maximum output torque of the influence of voltage, i.e. motor.
4. systematic parameter can not be surveyed, complete vehicle quality is as the important parameter of influence system model, and it is with rider's
How many and body weight and change, and it can not simultaneously be surveyed.
The content of the invention
The purpose of the present invention is to design controller using the rolling time horizon system optimizing control based on observer, passes through control
The torque of device optimization operator demand processed, is then driven and braking torque distribution, so as to realize that longitudinal speed is effectively controlled
The pure electric intelligent automobile longitudinal method for controlling driving speed based on observer of system.
The present invention realizes Matlab/Simulink and AMESim associative simulation,
1. the environmental variance of PC computers is must be provided with, makes both interrelated；
2. the interface module communicated with simulink is added in AMESim interfaces, by between Matlab/Simulink and AMESim
The variable of communication is needed to be connected to this module；
3. pass through System build after, the model information in AMESim to be retained in the form of Sfunction in Simulink,
So as to realize both associative simulation and communication.
The present invention step be：
First, centralized electric vehicle simulation model buildings：
Electric vehicle simulation model includes Electric drive module, transmission module, tire module and longitudinal direction of car dynamics, vehicle mould
Shape parameter such as table one
The electric automobile parameter list of table one
2nd, the rolling time horizon optimal controller based on observer：
2.1 Controllerorienteds, which design a model, to be built
2.1.1 longitudinal vehicle dynamic model
Travel longitudinal stress in the case where not considering cross force, on vehicle sloping route has according to Newton's second law：
（1）
Wherein：For driving quality,For driving force,For running resistance；Including air drag, pavement friction resistance
Power, grade resistanceAnd mechanical braking force；
Car weightWith driving qualityRelation is represented with formula following formula：
（2）
WhereinFor the inertia of a wheel,For radius of wheel；
The automobile travelled on the road of slope is by grade resistanceFor：
（3）
WhereinFor acceleration of gravity；
The air drag that the automobile travelled on road surface is subject toFor：
（4）
WhereinFor air stickiness density,For air resistance coefficient,For the front face area of vehicle,For wind speed,For car
Speed；
Ignore automobile air speed influence, therefore air drag is expressed as：
（5）
Frictional resistanceIt is the frictional force between road and tire, passes through following formula：
（6）
WhereinFor surface friction coefficient,For viscous friction coefficient；
Mechanical braking force,For braking moment；
Obtaining the running resistance that vehicle is subject to is：
（7）；
2.1.2 power train is modeled
2.1.2.1 clutch
By rigidity it is assumed that the torque of its transmission is：
（8）
WhereinFor motor output torque,For clutch output torque,For motor output speeds,Exported for clutch
Rotating speed；
2.1.2.2 speed changer
Gearbox output torqueModeling such as following formula：
（9）
WhereinTo reverse damped coefficient,For output speed,For gearratio,For gear gearratio,For
Base ratio；
2.1.2.3 drive shaft
（10）
WhereinExported for drive shaft,For drive shaft output speed；
Formula 8 and formula 9 are substituted into formula 10 to arrange：
（11）
WithDriving force, is usedRadius of wheel is represented, then the relation between power and torque, while speed, therefore convolution 11：
（12）
Wherein radius of wheelTried to achieve by following formula, in formulaFor hub radius,For tire aspect ratio,For tyre width；
（13）；
2.2 joint observation devices：
2.2.1 recurrent least square method quality is recognized
Convolution 1 and formula 7, obtain following equation：
（14）
In combination with formula（2）、（3）、（4）、（5）、（6）It is organized into least square form, drive shaft torque estimate, conversion
For driving force, obtain
（15）
WhereinLongitudinal direction of car acceleration is represented,Equivalent rotary quality, its value, whereinFor a wheel
Inertia,For radius of wheel；Represent to include drive shaft estimator system output quantity,Retrievable data vector is represented,For amount to be identified,For the process white noise of system；
According to principle of least square method, the complete vehicle quality that K1, K moment System Discrimination obtains is defined respectively is、, then quality identification model is obtained：
（16）
In formula,For the forgetting factor at K moment；
Forgetting factorRule is：
（17）；
2.2.2 drive shaft torque observer
If drive shaft two ends rotating speed can be surveyed, then drive shaft torque generation model is built：
（18）
WhereinDrive shaft torque is calculated for open loop,、For drive shaft two ends rotating speed,For gearbox output speed,
For vehicle wheel rotational speed,For drive shaft equivalent stiffness coefficients,For drive shaft Equivalent damping coefficient；
Obtain equivalent wheel power model：
（19）
WhereinFor vehicle wheel rotational speed estimate,For driving moment,For the moment of resistance,For the rotation of jack shaft end
Inertia,
（20）
Driving moment passes through resistanceMathematical modeling is tried to achieve, i.e.,, convolution 7
（21）
WhereinGrade resistance square is represented,Air drag square is represented,Represent frictional resistance moment,For mechanical braking torque,Represent car resistance square；Used during drive shaft torque estimatorInstead of；
Define deviation, i.e. vehicle wheel rotational speed estimate subtracts actual value, to deviationDerivation, in combination with formula 19 and formula
21：
（22）
Take control inputFor following form：
（23）
Wherein,Inputted for virtual controlling, utilize feedback linearization method, the linearised form of formula 22
（24）
By virtual controlling Input design into PI forms, i.e.,：
（25）
Wherein,For proportionality coefficient,For integral coefficient；
Association type 23 and 25, obtains controller：
（26）
Define Lyapunov functions such as formula（27）：
（27）
Both sides derivation is obtained：
（28）
By formula（25）Bring above formula into and arrange：
（29）
Therefore haveWhen, so；
Therefore drive shaft torque is：
（30）.
Longitudinal speed controller of the invention based on observer：
Numerically converted between motor torque and mechanical braking torque by gearratio, i.e.,：
（31）
That is braking moment and driving moment is collectively expressed as,
Arranged by formula (1), (7) and (12) and obtain following relational expression：
（32）
It is quantity of state wherein to choose longitudinal speed, i.e.,, it is control variable to choose drive demand, i.e.,, same choosing
Longitudinal speed is taken as output quantity, i.e.,；Discretization is carried out to state space equation by Euler's method, usedExpression is adopted
Sample steplength, then existAt the moment, it is as follows that discretization obtains system discrete model：
（33）
For output factor matrix, the prediction time domain for defining system is, control the time domain to be, it is necessary to meet,
So existThe prediction output sequence at moment is expressed as：
（34）
Meanwhile, the optimal control list entries at k momentIt is expressed as：
（35）
In sampling instant, the value of quantity of state is, derive quantity of state and the prediction process such as formula (19) of output quantity
(20) shown in：
（36）
（37）
Meanwhile, reference input is desired speed, therefore obtain reference input sequence：
（38）
Need to consider following constraint in controller design：
（39）
Simultaneously in order to ensure the tracking of longitudinal speed, while the comfortableness taken is improved, therefore controller chooses performance indications mesh
It is designated as：
（40）
Wherein；
And then it is described as the optimization problem of (41) formula, even if object functionValue is minimum：
（41）
In formula (41),Reality output speed and the deviation of desired speed are reflected,The power of drive demand is reflected,WithThe respectively power of output signal sequence and control signal sequence
Repeated factor；The optimization problem of 41 formulas, the control input sequence of optimization system, then by first in sequence are solved using NAG
ElementAct on system；Subsequent time, repeats abovementioned Optimization Solution process, the i.e. closed loop to the longitudinal speed of autonomous driving
Optimal control,
（42）；
Need that the part for being more than 0 in driver's torque demand is turned into driving moment to the torque progress Torque distribution of optimization,
Part less than or equal to 0 turns to braking moment, i.e.,
（43）
WhereinFor the driving torque demand of optimization,Expect torque for motor,For braking moment, by brake force
Square is converted into mechanical braking signal by following formula：
（44）.
Present invention is generally directed to pure electric intelligent automobile longitudinal speed control problem, for the important parameter design view of system
Device is surveyed, rolling time horizon system optimizing control can be good at completing online optimization solution, while processing constraint that can be dominant.Tool
The mechanism model said by setting up longitudinal vehicle speed control system of body, obtains driving the predictive equation of torque demand, then constructs
Cost function, while taking into full account constraints, Optimization Solution obtains optimal driver's torque demand.
The beneficial effects of the invention are as follows：
1. in longitudinal speed control aspect, traditional control algolithm majority is all not based on model, and up in real road
The vehicle working condition sailed is complicated and changeable, it is difficult to which finding a group controller parameter meets all operating modes.Estimator is added to control simultaneously
In device design process processed, it is suppressed that the influence that this parameter of vehicle mass is controlled automobile longitudinal speed, the rolling based on observer
Dynamic Optimization of Time Domain control algolithm mechanism model based on system in controller design, by some vehicles such as vehicle mass of observation
Operating condition information is introduced directly into mechanism model, so that the control of longitudinal direction of car speed is more accurate.
2. the longitudinal vehicle speed control system designed in the present invention is a nonlinear system, and it is mechanical, electrical to take into account electricity
The actuator firm constraints of pond group and brake, traditional control algolithm can not effective processing system constraint, and roll
Optimization of Time Domain control algolithm can effectively handle the control problem of belt restraining, directly be compiled into constraints when solving
Line solver in S_function in simulink.
Brief description of the drawings
Fig. 1 is to implement of the present invention based on the longitudinal speed control block diagram of rolling time horizon optimization；
Fig. 2 is to implement centralized electric automobile AMESim whole vehicle models of the present invention；
Fig. 3 is traveling longitudinal stress schematic diagram on vehicle sloping route of the present invention；
Fig. 4 is drive shaft torque estimation scheme of the present invention；
Fig. 5 is motor maximum output torque MAP of the present invention；
Fig. 6 is longitudinal speed controller design flow diagram of the design torsion based on rolling optimization algorithm in the present invention；
Fig. 7 is the desired speed that the present invention is controlled the complicated city operating mode used during device compliance test result, and abscissa is the time,
Unit s, ordinate is longitudinal speed, and unit is m/s;
Fig. 8 is the simulation result of the present invention under level road operating mode, and quality identification correlation curve is followed successively by from top to bottom, is driven
Moving axis torque estimates correlation curve and speed tracking curve.Wherein solid line is actual vehicle matter in quality identification correlation curve
Amount, dotted line is identification quality.Dotted line is estimation torque in torque correlation curve, and solid line is actual torque.Actual vehicle speed and expectation
In speed comparison diagram, dotted line is actual vehicle speed, and solid line is desired speed；
Fig. 9 is the simulation result of the present invention under constant heavy grade operating mode, and quality identification contrast is followed successively by from top to bottom
Curve, drive shaft torque estimation correlation curve and speed tracking curve.Wherein solid line is actual in quality identification correlation curve
Vehicle mass, dotted line is identification quality.Dotted line is estimation torque in torque correlation curve, and solid line is actual torque.Actual vehicle speed
In desired speed comparison diagram, dotted line is actual vehicle speed, and solid line is desired speed；
Change in road slope situation when Figure 10 is change gradient emulation of the present invention；
Figure 11 is of the present invention in the simulation result become close to true road surface under gradient operating mode, and quality is followed successively by from top to bottom
Recognize correlation curve, drive shaft torque estimation correlation curve and speed tracking curve.It is wherein real in quality identification correlation curve
Line is actual vehicle quality, and dotted line is identification quality.Dotted line is estimation torque in torque correlation curve, and solid line is actual torque.
In actual vehicle speed and desired speed comparison diagram, dotted line is actual vehicle speed, and solid line is desired speed.
Embodiment
Control block diagram such as Fig. 1 that the electric automobile torque optimization method based on datadriven PREDICTIVE CONTROL is implemented in the present invention
Shown, speed optimal controller is built in Simulink in figure, and the input of controller is desired speed, and actual vehicle speed is made
For that can survey signal, quality and the gradient are as measurable disturbances Realtime Feedback to controller, and Tmax is motor maximum drive torque, and it is
Together decided on by the mechanical property of motor and the output voltage of battery, both embodied the actuator firm constraints bar of motor in itself
Part, has been embodied as battery discharge time increases influence of the voltage reduction to vehicle performance again.The drive that entire car controller is obtained
Kinetic moment is necessarily less than equal to Tmax, therefore Tmax is to be given to controller as constraint.Centralized pure electric automobile mould in Fig. 2
Type is built in AMESim, for simulating the operation of real vehicle.Controller optimization drive demand torque is by Torque distribution
Driving and braking moment signal are given to motor and brake module respectively, control the operation of vehicle, and the actual vehicle speed conduct of vehicle
Feedback signal back is to controller.
The control targe of the present invention is the actual vehicle speed and desired speed that longitudinal speed controller is returned according to Realtime Feedback
Signal contrast, under the premise of constraints is met, optimization obtains drive demand torque, then by Torque distribution, so as to obtain
Driving moment and braking moment signal, and the motor in whole vehicle model and brake module are given to, the operation of vehicle is controlled, is finally allowed
The upper desired speed of actual vehicle speed tracking.
The invention provides a set of based on the device for operating above principle and running.It is offline electronic i.e. based on PC
Vehicle torsional moment optimization design test platform.Build and running is as follows：
Software is selected
The controlled device of the control system and the simulation model of controller pass through software Matlab/Simulink and AMESim respectively
Built, software version is respectively Matlab R2009a and AMESim R10, solver is respectively chosen as ode3 and Euler.
Simulation step length is fixed step size, and steplength selection is 0.01s.
The present invention will realize Matlab/Simulink and AMESim associative simulation.
1. the environmental variance of PC computers must be as requested set first, make both interrelated.
2. the interface module communicated with simulink is added and then in AMESim interfaces, by Matlab/Simulink with
The variable of communication is needed to be connected to this module between AMESim；
3. eventually pass after System build, the model information in AMESim is retained in Simulink in the form of Sfunction
In, so as to realize both associative simulation and communication.
When running Simulink simulation models, AMESim models are also calculated and solved at the same time.In simulation process
The exchange of data is constantly carried out between the two.If be modified to the model structure or parameter setting in AMESim,
Need to recompilate.It is worth noting that, both simulation step lengths must be consistent.
Step of the present invention is：
First, centralized electric vehicle simulation model buildings：
As shown in Fig. 2 whole electric vehicle simulation model includes Electric drive module, transmission module, tire module and vehicle and indulged
To several parts such as dynamics, whole vehicle model parameter is as shown in Table 1.
The electric automobile parameter list of table one
。
Power drive system includes battery compartment and motor part, and the battery pack of pure electric automobile is lithium battery group, by multiple
Cell connection in seriesparallel is constituted.The terminal voltage that battery pack is externally exported is single battery output voltage and battery system output
Terminal voltage is the voltage that battery pack is supplied to motor；Permagnetic synchronous motor is used in the present invention.
Transmission system includes speed changer, three parts of differential mechanism and drive shaft.The power of motor output is logical via speed changer
Cross different tooth radius to produce not at the same speed than increasing square so as to slow down, the present invention have ignored the lateral dynamic of vehicle, then
The output speed of differential mechanism both sides is identical, i.e., differential mechanism does not work, and differential mechanism output speed is the input speed of drive shaft,
And the output speed of drive shaft is equal with vehicle wheel rotational speed, the torque transmitted on axle is calculated by drive shaft two ends speed discrepancy.Motor
The torque of output is produced not at the same speed than increasing square so as to slow down via speed changer by different tooth radius, and this model master subtracts
Speed is than being 2.2786, and gear speed reducing ratio is that 3.9431, i.e. gearratio are 8.6847.
Longitudinal direction of car kinetic part, power, brake force and traveling are driven it considers vehicle in the process of moving
The effect of resistance, wherein running resistance include air drag, rolling resistance and frictional resistance.Car can be set in this module
The parameter such as total quality, the gradient, wind speed.
2nd, the rolling time horizon optimal controller based on observer：
2.1 Controllerorienteds, which design a model, to be built
2.1.1 longitudinal vehicle dynamic model
In order to realize longitudinal speed control and the research of vehicle mass parameter Estimation, it is necessary to set up longitudinal vehicle dynamic model.
Longitudinal stress schematic diagram is travelled in the case where not considering cross force, on vehicle sloping route as shown in Figure 3.In Fig. 3For road slope
Degree,For driving quality, then gravity be,For driving force,For air drag with,For pavement friction resistance.According to
Newton's second law has：
（1）
Wherein：For driving quality,For driving force,For running resistance；Including air drag, pavement friction resistance
Power, grade resistanceAnd mechanical braking force.
It is to be noted that car weightWith driving qualityAnd it is unequal, it is used to when calculating driving quality by travel direction
The influence of property effect, its relation can be represented approximately with formula following formula：
（2）
WhereinFor the inertia of a wheel,For radius of wheel；Driving force/brake forceIt is described in detail in next trifle,
Running resistance to being subject in vehicle travel process belowCarry out analysis introduction.
The automobile travelled on the road of slope is by grade resistanceFor：
（3）
WhereinFor acceleration of gravity；For car weight.In view of the gradient in AMESim modelsFor the road percentagewised
The gradient, the unification of gradient form when being tested for subsequent simulation, we are unified here calculates the gradient by hundredmark system, then need to turn
Change。
The air drag that the automobile travelled on road surface is subject toFor：
（4）
WhereinFor air stickiness density,For air resistance coefficient,For the front face area of vehicle,For wind speed,For car
Speed；Here because air drag is smaller relative to grade resistance and frictional resistance numerical value, the automobile wind then travelled in city
Speed is again smaller relative to speed, and the influence of wind speed is have ignored in controller design, therefore air drag is expressed as：
（5）
Frictional resistanceIt is the frictional force between road and tire, passes through following formula：
（6）
WhereinFor surface friction coefficient,For viscous friction coefficient；
Mechanical braking force,For braking moment；
Obtaining the running resistance that vehicle is subject to is：
（7）；
2.1.2 power train is modeled
Rigidity is carried out to clutch, power transmission shaft, drive shaft during modeling it is assumed that have ignored simultaneously main reducing gear and speed transmission it
Between transmit torque loss.
2.1.2.1 clutch
By rigidity it is assumed that the torque of its transmission is：
（8）
WhereinFor motor output torque,For clutch output torque,For motor output speeds,Turn for clutch output
Speed.
2.1.2.2 speed changer
Here because we have ignored the torque transmitted between main reducing gear and speed transmission loss, therefore we are right in modeling
It is unified, gearbox output torqueModeling such as following formula：
（9）
WhereinTo reverse damped coefficient,For output speed, reversing the product of damped coefficient and output speed is used to approximately rub
Torque loss is wiped,For gearratio,For gear gearratio,For base ratio.
2.1.2.3 drive shaft
（10）
WhereinExported for drive shaft,For drive shaft output speed；As vehicle wheel rotational speed.
Formula 8 and formula 9 are substituted into formula 10 to arrange：
（11）
WithDriving force, is usedRadius of wheel is represented, then the relation between power and torque, while speed,
Therefore convolution 11：
（12）
Wherein radius of wheelTried to achieve by following formula, in formulaFor hub radius,For tire aspect ratio,For tyre width；
（13）；
2.2 joint observation devices：
In order to accurately estimate vehicle mass, our designing qualities and drive shaft torque joint observation device.It is contemplated that to quality and
There is coupled relation between drive shaft torque, that is, it is to need to use drive shaft moment information to estimate quality, and estimation torque is desirable
Use vehicle mass information.Thus it is determined that during estimation scheme, analyze two respective characteristics of amount to be observed, the quality of automobile
Change mainly in number of passengers, fuel tank it is oily number and loading goods number cause, quality is in vehicle travel process
In stablize relatively, thus vehicle mass is slow variable, can be recognized with the method for identification, while considering this characteristic with ought
The quality identification result at preceding moment is estimated the torque of subsequent time to produce torque estimated result influence very little, so may be used
To improve the efficiency of algorithm, while also solving coupled relation therebetween.
2.2.1 recurrent least square method quality is recognized
The kinetic model that analysis vehicle is travelled on the road of slope, convolution 1 and formula 7, obtain following equation：
（14）
In combination with formula（2）、（3）、（4）、（5）、（6）Least square form is organized into, in combination with analysis, is used during estimation quality
To drive shaft torque estimate, it is converted into driving force, obtain
（15）
WhereinLongitudinal direction of car acceleration is represented,Equivalent rotary quality, its value, whereinFor wheel
Inertia,For radius of wheel；Represent to include drive shaft estimator system output quantity,Retrievable data vector is represented,
For amount to be identified,For the process white noise of system；It is to be noted that we are by drive shaft during identification vehicle mass
Torque is as measurable, then according to the relation between power and torque, driving force is also measurable, while thinking longitudinal acceleration
It is that can survey parameter in real time with mechanical braking torque.
According to previously described principle of least square method, the complete vehicle quality that K1, K moment System Discrimination are obtained is defined respectively
For、, then quality identification model is obtained：
（16）
In formula,For the forgetting factor at K moment.
Vehicle mass is slow variable, because the quality initial value and actual mass of setting there may be larger deviation, therefore
Identification is just when starting, it is necessary to set a larger degree of belief decay, i.e., the forgetting factor value now chosen is smaller, with distinguishing
The continuous progress known, identification result can be converged near actual vehicle quality, then be accomplished by a larger forgetting factor, with
Less degree of belief is obtained to decay, therefore the forgetting factor chosen hereinRule is：
（17）；
2.2.2 drive shaft torque observer
This technology devises closeloop driven axle torque observer, by torque generation model open loop result of calculation, then passes through car
Speed deviation divided ring result of calculation is corrected, so that drive shaft torque observation problem is converted into rotatingspeed tracking problem.Estimation
Scheme is as shown in Figure 4.In view of the nonlinear feature of loading moment, with reference to the advantage of feedback linearization method, therefore use anti
Linearization method carries out torque observer design.
If drive shaft two ends rotating speed can be surveyed, then drive shaft torque generation model is built：
（18）
WhereinDrive shaft torque is calculated for open loop,、For drive shaft two ends rotating speed,For gearbox output speed,
For vehicle wheel rotational speed,For drive shaft equivalent stiffness coefficients,For drive shaft Equivalent damping coefficient.
According to Lagrange's dynamical equations, and assume that wheel does pure rolling positive motion in running car, therefore obtain
To equivalent wheel power model：
（19）
WhereinFor vehicle wheel rotational speed estimate,For driving moment,For the moment of resistance,It is used to for the rotation of jack shaft end
Amount, is approximately calculated with formula 20 and obtained
（20）
The relation between power and torque is considered simultaneously, and driving moment passes through resistanceMathematical modeling is tried to achieve, i.e.,,
Convolution 7
（21）
WhereinGrade resistance square is represented,Air drag square is represented,Represent frictional resistance moment,For mechanical braking torque,Represent car resistance square；While the expression in view of using last moment quality when being driven axle power moments estimation
As a result, that is, calculatingWhen use, therefore used in design drive torque estimatorInstead of。
Define deviation, i.e. vehicle wheel rotational speed estimate subtracts actual value, to deviationDerivation, in combination with formula
19 and formula 21：
（22）
Take control inputFor following form：
（23）
Wherein,Inputted for virtual controlling, utilize feedback linearization method, the linearised form of formula 22
（24）
By virtual controlling Input design into PI forms, i.e.,：
（25）
Wherein,For proportionality coefficient,For integral coefficient；
Association type 23 and 25, obtains controller：
（26）
Define Lyapunov functions such as formula（27）：
（27）
Both sides derivation is obtained：
（28）
By formula（25）Bring above formula into and arrange：
（29）
Therefore haveWhen, so；
In summary, estimate that obtained drive shaft torque is：
（30）.
Longitudinal speed controller of the invention based on observer：
This paper control targe is to realize the tracing control of longitudinal speed during autonomous driving, under different operating modes, passes through control
The desired drive demand of device optimization processed, then realizes that actual vehicle speed tracking is last by executing agencies such as motor and mechanical brakings
Hope speed.
Controller design optimizes for convenience, and we uniformly optimize mechanical braking torque and motor braking torque,
I.e. in the case where considering that motor is desired motor, optimization reaches operator demand during desired speed, then passing through
Certain control strategy is classified as motor torque demand and mechanical force moment demand.
Closed to realize unified optimization driver's torque demand, it is necessary to set up between mechanical braking torque and motor torque
It is that we ignore the inertia loss of power train here, then numerically pass through between motor torque and mechanical braking torque and pass
Move than being converted, i.e.,：
（31）
That is braking moment and driving moment is collectively expressed as, the system model that to sum up analysis set is above set up, while in order to
Speed control is more accurately realized, we use the quality that joint estimator is estimated by the complete vehicle quality when designing controller.
Arranged by formula (1), (7) and (12) and obtain following relational expression：
（32）
It is quantity of state wherein to choose longitudinal speed, i.e.,, it is control variable to choose drive demand, i.e.,, equally
Longitudinal speed is chosen as output quantity, i.e.,；Discretization is carried out to state space equation by Euler's method, usedTable
Show sampling step length, then existAt the moment, it is as follows that discretization obtains system discrete model：
（33）
For output factor matrix, according to Model Predictive Control Theory, the prediction time domain for defining system is, control the time domain to be, it is necessary to meet, thenThe prediction output sequence at moment is expressed as：
（34）
Meanwhile, the optimal control list entries at k momentIt is expressed as：
（35）.
In sampling instant, the value of quantity of state is, according to the basic principle and correlation theory of Model Predictive Control,
Derive shown in quantity of state and the prediction process such as formula (19) of output quantity and (20)：
（36）
（37）.
The state variable value and last moment system input value at current time are calculated by analyzing, Optimization Solution goes out controlled quentity controlled variable
Sequence, and first amount of the controlled quentity controlled variable sequence for only going out Optimization Solution acts on system.And in next sampling instant, electronic vapour
Vehicle model can feed back to new input variable and quantity of state, and controller reoptimization solves control problem.
Meanwhile, reference input is desired speed, therefore obtain reference input sequence：
（38）
During autonomous driving, in order to ensure driving safety, it is necessary to enter row constraint to quantity of state, while in view of motor in itself
Characteristic, its output torque has constraint, while in view of the limitation of mechanical structure, having the constraint of maximum braking moment again, to sum up
It is described, need to consider following constraint in controller design：
（39）
The maximum moment that wherein motor is provided is tabled lookup by the MAP in Fig. 5 to be obtained.
Simultaneously in order to ensure the tracking of longitudinal speed, while improving the comfortableness taken（Accelerate to control with braking procedure
Braking is made as small as possible）, therefore controller choose performance indications target be：
（40）
Wherein；
And then it is described as the optimization problem of (41) formula, even if object functionValue is minimum：
（41）
In formula (41),Reality output speed and the deviation of desired speed are reflected,The power of drive demand is reflected,WithRespectively output signal sequence and control signal sequence
Weight factor.Size reflect speed tracing precision requirement,Bigger, the deviation of speed tracing is closer to zero.Then
Reflect the requirement to control action,Bigger, control action is smaller, and ride comfort is more preferable.Use a kind of NAG (rolling time horizons
Optimized algorithm MATLAB solve tool box) solve 41 formulas optimization problem, the control input sequence of optimization system, then by sequence
First element in rowAct on system；Subsequent time, repeats abovementioned Optimization Solution process, i.e., to autonomous driving longitudinal direction
The closed optimized control of speed,
（42）；
The flow of rolling time horizon Optimal Controller Design is as shown in Figure 6：Rolling time horizon optimal controller optimization driver's torque
Demand, but the control signal needed in control is motor torque demand and mechanical braking signal, therefore need to optimization
Torque carry out Torque distribution by driver's torque demand be more than 0 part turn to driving moment, the part less than or equal to 0
Braking moment is turned to, i.e.,
（43）
WhereinFor the driving torque demand of optimization,Expect torque for motor,For braking moment, by brake force
Square is converted into mechanical braking signal by following formula：
（44）.
Experimental verification
Control parameter is adjusted repeatedly, and output signal sequence and the weight factor Γ of control signal sequence are chosen respectivelyy =100, Γu =2, the sampling time is 0.01s, we have selected acceleration and deceleration frequently city operating mode, desired speed such as Fig. 7.Set vehicle matter
Measure as 1500kg, controller is carried out under level road operating mode, constant heavy grade and the change gradient operating mode for pressing close to true road surface respectively
Checking.
1）Level road operating mode simulating, verifying
We select level road to verify first, set road grade as 0, and in emulation, wind speed setting is 0, and vehicle mass is
1500 Kg, i.e. vehicle noload running, simulation result are followed successively by quality identification contrast song from top to bottom as shown in figure 8, being provided in figure
Line, drive shaft torque estimation correlation curve and speed tracking curve, as can be seen from the figure estimator and controller all have good
Good effect.
2）Constant heavy grade operating mode simulating, verifying
Under simulated environment, we set road grade as 10%, i.e., on constant larger slope road, verify long lasting for larger
Whether the control effect of slope road travel controller is stablized, and simulation result is as shown in Figure 9.Provided in figure and be followed successively by quality from top to bottom
Recognize correlation curve, drive shaft torque estimation correlation curve and speed tracking curve.Simulation result shows, up on larger slope road
Sail, very well, actual vehicle speed can track desired speed in the most of the time to estimator estimation effect, but left in 200 ~ 300s
The right side, actual vehicle speed does not track desired speed, but speed maintains 20 m/s, this be due to motor expect torque by
The effect of our motor maximum motor torque constraints during controller design, by Fig. 5 motor maximum output torques map can
Know, motor now expects that torque is just equal to motor maximum output torque.Illustrate that controller's effect is good, simultaneity factor constraint
Play a good role.
3）Become gradient operating mode simulating, verifying
In actual vehicle running environment, road grade can't keep constant, thus we set closer actual motion work
The change gradient of condition（Road grade such as Figure 10）Verified.Simulation result can be seen that as shown in Figure 10 mono from simulation result
Joint estimator estimation effect is good, while longitudinal direction of car speed tracks desired speed well in the case where becoming gradient operating mode.
The present invention is directed to longitudinal speed controller of the pure electric intelligent Automobile Design based on rolling time horizon optimization method, this
The method of kind realizes online optimization well, while dominant processing constraint.In order to verify having for longitudinal speed optimal controller
Effect property, has built centralized electric automobile model, and in complicated urban road, exist respectively in the senior simulation softwares of AMESim
The performance of controller is demonstrated under the change gradient operating mode of level road operating mode, constant heavy grade operating mode and closing to reality road surface.Emulation knot
Fruit shows that the longitudinal speed controller of rolling time horizon optimization is under different driving cycles, with good control performance.
Claims (3)
1. a kind of pure electric intelligent automobile longitudinal method for controlling driving speed based on observer, it is characterised in that：Realize Matlab/
Simulink and AMESim associative simulation,
1. the environmental variance of PC computers is must be provided with, makes both interrelated；
2. the interface module communicated with simulink is added in AMESim interfaces, by between Matlab/Simulink and AMESim
The variable of communication is needed to be connected to this module；
3. pass through System build after, the model information in AMESim to be retained in the form of Sfunction in Simulink,
So as to realize both associative simulation and communication.
2. the pure electric intelligent automobile longitudinal method for controlling driving speed based on observer described in claim 1, it is characterised in that：
First, centralized electric vehicle simulation model buildings：
Electric vehicle simulation model includes Electric drive module, transmission module, tire module and longitudinal direction of car dynamics, vehicle mould
Shape parameter such as table one
The electric automobile parameter list of table one
2nd, the rolling time horizon optimal controller based on observer：
2.1 Controllerorienteds, which design a model, to be built
2.1.1 longitudinal vehicle dynamic model
Travel longitudinal stress in the case where not considering cross force, on vehicle sloping route has according to Newton's second law：
（1）
Wherein：For driving quality,For driving force,For running resistance；Including air drag, pavement friction resistance
Power, grade resistanceAnd mechanical braking force；
Car weightWith driving qualityRelation is represented with formula following formula：
（2）
WhereinFor the inertia of a wheel,For radius of wheel；
The automobile travelled on the road of slope is by grade resistanceFor：
（3）
WhereinFor acceleration of gravity；
The air drag that the automobile travelled on road surface is subject toFor：
（4）
WhereinFor air stickiness density,For air resistance coefficient,For the front face area of vehicle,For wind speed,For car
Speed；
Ignore automobile air speed influence, therefore air drag is expressed as：
（5）
Frictional resistanceIt is the frictional force between road and tire, passes through following formula：
（6）
WhereinFor surface friction coefficient,For viscous friction coefficient；
Mechanical braking force,For braking moment；
Obtaining the running resistance that vehicle is subject to is：
（7）；
2.1.2 power train is modeled
2.1.2.1 clutch
By rigidity it is assumed that the torque of its transmission is：
（8）
WhereinFor motor output torque,For clutch output torque,For motor output speeds,Turn for clutch output
Speed；
2.1.2.2 speed changer
Gearbox output torqueModeling such as following formula：
（9）
WhereinTo reverse damped coefficient,For output speed,For gearratio,For gear gearratio,For
Base ratio；
2.1.2.3 drive shaft
（10）
WhereinExported for drive shaft,For drive shaft output speed；
Formula 8 and formula 9 are substituted into formula 10 to arrange：
（11）
WithDriving force, is usedRadius of wheel is represented, then the relation between power and torque, while speed, therefore convolution 11：
（12）
Wherein radius of wheelTried to achieve by following formula, in formulaFor hub radius,For tire aspect ratio,For tyre width；
（13）；
2.2 joint observation devices：
2.2.1 recurrent least square method quality is recognized
Convolution 1 and formula 7, obtain following equation：
（14）
In combination with formula（2）、（3）、（4）、（5）、（6）It is organized into least square form, drive shaft torque estimate, it is converted into
Driving force, obtain
（15）
WhereinLongitudinal direction of car acceleration is represented,Equivalent rotary quality, its value, whereinFor a wheel
Inertia,For radius of wheel；Represent to include drive shaft estimator system output quantity,Retrievable data vector is represented,For amount to be identified,For the process white noise of system；
According to principle of least square method, the complete vehicle quality that K1, K moment System Discrimination obtains is defined respectively is、, then quality identification model is obtained：
（16）
In formula,For the forgetting factor at K moment；
Forgetting factorRule is：
（17）；
2.2.2 drive shaft torque observer
If drive shaft two ends rotating speed can be surveyed, then drive shaft torque generation model is built：
（18）
WhereinDrive shaft torque is calculated for open loop,、For drive shaft two ends rotating speed,For gearbox output speed,
For vehicle wheel rotational speed,For drive shaft equivalent stiffness coefficients,For drive shaft Equivalent damping coefficient；
Obtain equivalent wheel power model：
（19）
WhereinFor vehicle wheel rotational speed estimate,For driving moment,For the moment of resistance,It is used to for the rotation of jack shaft end
Amount,
（20）
Driving moment passes through resistanceMathematical modeling is tried to achieve, i.e.,, convolution 7
（21）
WhereinGrade resistance square is represented,Air drag square is represented,Represent frictional resistance moment,For mechanical braking torque,Represent car resistance square；Used during drive shaft torque estimatorInstead of；
Define deviation, i.e. vehicle wheel rotational speed estimate subtracts actual value, to deviationDerivation, in combination with the He of formula 19
Formula 21：
（22）
Take control inputFor following form：
（23）
Wherein,Inputted for virtual controlling, utilize feedback linearization method, the linearised form of formula 22
（24）
By virtual controlling Input design into PI forms, i.e.,：
（25）
Wherein,For proportionality coefficient,For integral coefficient；
Association type 23 and 25, obtains controller：
（26）
Define Lyapunov functions such as formula（27）：
（27）
Both sides derivation is obtained：
（28）
By formula（25）Bring above formula into and arrange：
（29）
Therefore haveWhen, so；
Therefore drive shaft torque is：
（30）.
3. the pure electric intelligent automobile longitudinal method for controlling driving speed according to claim 1 or 2 based on observer, its feature
It is：Longitudinal speed controller based on observer：
Numerically converted between motor torque and mechanical braking torque by gearratio, i.e.,：
（31）
That is braking moment and driving moment is collectively expressed as,
Arranged by formula (1), (7) and (12) and obtain following relational expression：
（32）
It is quantity of state wherein to choose longitudinal speed, i.e.,, it is control variable to choose drive demand, i.e.,, equally
Longitudinal speed is chosen as output quantity, i.e.,；Discretization is carried out to state space equation by Euler's method, usedTable
Show sampling step length, then existAt the moment, it is as follows that discretization obtains system discrete model：
（33）
For output factor matrix, the prediction time domain for defining system is, control the time domain to be, it is necessary to meet,
So existThe prediction output sequence at moment is expressed as：
（34）
Meanwhile, the optimal control list entries at k momentIt is expressed as：
（35）
In sampling instant, the value of quantity of state is, derive quantity of state and output quantity prediction process such as formula (19) and
(20) shown in：
（36）
（37）
Meanwhile, reference input is desired speed, therefore obtain reference input sequence：
（38）
Need to consider following constraint in controller design：
（39）
Simultaneously in order to ensure the tracking of longitudinal speed, while the comfortableness taken is improved, therefore controller chooses performance indications mesh
It is designated as：
（40）
Wherein；
And then it is described as the optimization problem of (41) formula, even if object functionValue is minimum：
（41）
In formula (41),Reflect the inclined of reality output speed and desired speed
Difference,The power of drive demand is reflected,WithRespectively output signal sequence and control signal sequence
The weight factor of row；The optimization problem of 41 formulas, the control input sequence of optimization system, then by sequence are solved using NAG
First elementAct on system；Subsequent time, repeats abovementioned Optimization Solution process, i.e., to the longitudinal speed of autonomous driving
Closed optimized control,
（42）；
Need that the part for being more than 0 in driver's torque demand is turned into driving moment to the torque progress Torque distribution of optimization,
Part less than or equal to 0 turns to braking moment, i.e.,
（43）
WhereinFor the driving torque demand of optimization,Expect torque for motor,For braking moment, by brake force
Square is converted into mechanical braking signal by following formula：
（44）.
Priority Applications (1)
Application Number  Priority Date  Filing Date  Title 

CN201710483937.2A CN107168104B (en)  20170623  20170623  Observerbased longitudinal speed control method for pure electric intelligent automobile 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

CN201710483937.2A CN107168104B (en)  20170623  20170623  Observerbased longitudinal speed control method for pure electric intelligent automobile 
Publications (2)
Publication Number  Publication Date 

CN107168104A true CN107168104A (en)  20170915 
CN107168104B CN107168104B (en)  20200616 
Family
ID=59820095
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

CN201710483937.2A Active CN107168104B (en)  20170623  20170623  Observerbased longitudinal speed control method for pure electric intelligent automobile 
Country Status (1)
Country  Link 

CN (1)  CN107168104B (en) 
Cited By (6)
Publication number  Priority date  Publication date  Assignee  Title 

CN108897928A (en) *  20180613  20181127  吉林大学  A kind of intelligent vehicle slope road energy conservation speed optimization method based on the search of nested Monte Carlo tree 
CN109268159A (en) *  20180918  20190125  吉林大学  LeanBurn Gasoline Engine fuel air ratio system control method 
CN109521674A (en) *  20181126  20190326  东南大学  A kind of electric vehicle drive robot controller parameter selflearning method 
CN110727994A (en) *  20191028  20200124  吉林大学  Parameter decoupling electric automobile mass and gradient estimation method 
CN107168104B (en) *  20170623  20200616  吉林大学  Observerbased longitudinal speed control method for pure electric intelligent automobile 
CN111976736A (en) *  20200827  20201124  浙江吉利新能源商用车集团有限公司  Automatic driving control system and method for vehicle 
Citations (6)
Publication number  Priority date  Publication date  Assignee  Title 

CN102529976A (en) *  20111215  20120704  东南大学  Vehicle running state nonlinear robust estimation method based on sliding mode observer 
EP2160308B1 (en) *  20070621  20130731  Renault S.A.S.  Regenerative braking control method for hybrid vehicles and/or fourwheeldrive vehicles and arrangement for vehicle using said method 
CN103308325A (en) *  20130626  20130918  东莞中山大学研究院  Driving system semiphysical simulation platform of electric automobile 
CN103921786A (en) *  20140411  20140716  北京工业大学  Nonlinear model prediction control method of regenerative braking of electric vehicle 
CN104175891A (en) *  20140807  20141203  吉林大学  Pure electric vehicle energy recovery and regeneration brake control method 
CN104401232A (en) *  20141221  20150311  吉林大学  Electric vehicle torque optimization method based on data driven predictive control 
Family Cites Families (1)
Publication number  Priority date  Publication date  Assignee  Title 

CN107168104B (en) *  20170623  20200616  吉林大学  Observerbased longitudinal speed control method for pure electric intelligent automobile 

2017
 20170623 CN CN201710483937.2A patent/CN107168104B/en active Active
Patent Citations (6)
Publication number  Priority date  Publication date  Assignee  Title 

EP2160308B1 (en) *  20070621  20130731  Renault S.A.S.  Regenerative braking control method for hybrid vehicles and/or fourwheeldrive vehicles and arrangement for vehicle using said method 
CN102529976A (en) *  20111215  20120704  东南大学  Vehicle running state nonlinear robust estimation method based on sliding mode observer 
CN103308325A (en) *  20130626  20130918  东莞中山大学研究院  Driving system semiphysical simulation platform of electric automobile 
CN103921786A (en) *  20140411  20140716  北京工业大学  Nonlinear model prediction control method of regenerative braking of electric vehicle 
CN104175891A (en) *  20140807  20141203  吉林大学  Pure electric vehicle energy recovery and regeneration brake control method 
CN104401232A (en) *  20141221  20150311  吉林大学  Electric vehicle torque optimization method based on data driven predictive control 
NonPatent Citations (3)
Title 

SUWATKUNTANAPREEDA: "《Supertwisting slidingmode tractio ncontrol of vehicle swith tractive force observer》", 《CONTROLENGINEERINGPRACTICE》 * 
余卓平 等: "《车辆行驶过程中的状态估计问题综述》", 《机械工程学报》 * 
胡云峰 等: "《基于观测器的输出反馈电子节气门控制器设计》", 《自动化学报》 * 
Cited By (8)
Publication number  Priority date  Publication date  Assignee  Title 

CN107168104B (en) *  20170623  20200616  吉林大学  Observerbased longitudinal speed control method for pure electric intelligent automobile 
CN108897928A (en) *  20180613  20181127  吉林大学  A kind of intelligent vehicle slope road energy conservation speed optimization method based on the search of nested Monte Carlo tree 
CN108897928B (en) *  20180613  20200421  吉林大学  Intelligent vehicle slope energysaving vehicle speed optimization method based on nested Monte Carlo tree search 
CN109268159A (en) *  20180918  20190125  吉林大学  LeanBurn Gasoline Engine fuel air ratio system control method 
CN109268159B (en) *  20180918  20200818  吉林大学  Control method of fuelair ratio system of leanburn gasoline engine 
CN109521674A (en) *  20181126  20190326  东南大学  A kind of electric vehicle drive robot controller parameter selflearning method 
CN110727994A (en) *  20191028  20200124  吉林大学  Parameter decoupling electric automobile mass and gradient estimation method 
CN111976736A (en) *  20200827  20201124  浙江吉利新能源商用车集团有限公司  Automatic driving control system and method for vehicle 
Also Published As
Publication number  Publication date 

CN107168104B (en)  20200616 
Similar Documents
Publication  Publication Date  Title 

CN107168104A (en)  Pure electric intelligent automobile longitudinal method for controlling driving speed based on observer  
CN110228462B (en)  Yaw stability control method for fourwheel hub motordriven electric automobile  
CN105857309A (en)  Automotive adaptive cruise control method taking multiple targets into consideration  
Xu et al.  Torque optimization control for electric vehicles with four inwheel motors equipped with regenerative braking system  
CN107719372A (en)  Fourdrive electric car dynamics multi objective control system based on dynamic control allocation  
CN104175891B (en)  Pure electric automobile energy regenerating regenerating brake control method  
De Castro et al.  Minimumtime manoeuvring in electric vehicles with four wheelindividualmotors  
CN104554251A (en)  Hybrid car energysaving prediction control method based on road gradient information  
CN104590247A (en)  Hybrid electric vehicle energy conservation predictive control method based on traffic signal lamp information  
CN104401232A (en)  Electric vehicle torque optimization method based on data driven predictive control  
Saini et al.  Genetic algorithm based gear shift optimization for electric vehicles  
Yu et al.  Optimal design and control of 4IWD electric vehicles based on a 14DOF vehicle model  
Phan et al.  Intelligent energy management system for conventional autonomous vehicles  
Lv et al.  Cyberphysical system based optimization framework for intelligent powertrain control  
Goodarzi et al.  Stability enhancement and fuel economy of the 4wheeldrive hybrid electric vehicles by optimal tyre force distribution  
Wu  Optimization of steering control to improve the energy consumption of internal combustion engine vehicles  
Amiri et al.  Feedforward modelling and fuzzy logic based control strategy for powertrain efficiency improvement in a parallel hybrid electric vehicle  
Anderson et al.  Modelling minimumtime manoeuvering with global optimisation of local receding horizon control  
CN110717218B (en)  Electric drive vehicle distributed power drive system reconstruction control method and vehicle  
Holdstock et al.  Linear and nonlinear methods to analyse the drivability of a throughtheroad parallel hybrid electric vehicle  
Ruan et al.  Realtime energy management strategy based on driveractionimpact MPC for series hybrid electric vehicles  
Zhao et al.  Distributed Electric Powertrain Test Bench With Dynamic Load Controlled by Neuron PI SpeedTracking Method  
Mohebbi et al.  Adaptive neuro control of parallel hybrid electric vehicles  
Ganji et al.  Backward modelling and lookahead fuzzy energy management controller for a parallel hybrid vehicle  
Zhu et al.  Realtime cooptimization of vehicle route and speed using generic algorithm for improved fuel economy 
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 