CN110456790A - Intelligent network based on adaptive weighting joins electric car queue optimal control method - Google Patents
Intelligent network based on adaptive weighting joins electric car queue optimal control method Download PDFInfo
- Publication number
- CN110456790A CN110456790A CN201910702422.6A CN201910702422A CN110456790A CN 110456790 A CN110456790 A CN 110456790A CN 201910702422 A CN201910702422 A CN 201910702422A CN 110456790 A CN110456790 A CN 110456790A
- Authority
- CN
- China
- Prior art keywords
- queue
- control
- cat
- vehicle
- electric car
- 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
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 33
- 238000000034 method Methods 0.000 title claims abstract description 26
- 241000282326 Felis catus Species 0.000 claims abstract description 71
- 239000011159 matrix material Substances 0.000 claims abstract description 51
- 238000005457 optimization Methods 0.000 claims abstract description 44
- 230000001133 acceleration Effects 0.000 claims description 15
- 238000006243 chemical reaction Methods 0.000 claims description 13
- 230000002452 interceptive effect Effects 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 abstract description 11
- 230000008569 process Effects 0.000 abstract description 7
- 238000013461 design Methods 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 12
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000009415 formwork Methods 0.000 description 2
- 230000010355 oscillation Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 102100021913 Sperm-associated antigen 8 Human genes 0.000 description 1
- 101710098579 Sperm-associated antigen 8 Proteins 0.000 description 1
- 206010044565 Tremor Diseases 0.000 description 1
- SAZUGELZHZOXHB-UHFFFAOYSA-N acecarbromal Chemical compound CCC(Br)(CC)C(=O)NC(=O)NC(C)=O SAZUGELZHZOXHB-UHFFFAOYSA-N 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0223—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving speed control of the vehicle
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0287—Control of position or course in two dimensions specially adapted to land vehicles involving a plurality of land vehicles, e.g. fleet or convoy travelling
- G05D1/0291—Fleet control
- G05D1/0295—Fleet control by at least one leading vehicle of the fleet
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
Abstract
Intelligent network based on adaptive weighting joins electric car queue optimal control method, is related to automobile intelligent safety and automatic Pilot.Linearization process is carried out to intelligent network connection electric car Longitudinal Dynamic Model first, secondly, design the secondary type controller of queue, real-time local optimum and real-time global optimization are carried out to the weight matrix of quadratic form queue controller using tracing mode and search pattern according to the current state of vehicle in queue using cat swarm optimization, it is final to obtain the corresponding optimum control variable of current state and optimum state variable, avoid algorithm when controlling queue, phenomena such as convergence occurred is slow is reduced due to multifarious falling into system chatter caused by locally optimal solution and iteration latter stage, it realizes when under front truck complexity operating condition, rear car fast and stable follows, effectively improve the string stability of intelligent network connection electric car queue control, safety, comfort and accuracy.
Description
Technical field
The present invention relates to automobile intelligent safeties and automatic Pilot, more particularly to a kind of intelligent network based on adaptive weighting
Join electric car queue optimal control method.
Background field
Intelligent network, which joins electric car, has outstanding mobility, control flexibility and driving terseness, it is considered to be improves
Traffic safety one of reduces environmental pollution with the effective way of energy consumption, causes the extensive of national governments and scientific research institution
Concern.
The purpose of intelligent network connection electric car platoon driving is to solve that traffic delay, the workload of driver be big, traffic
The problem of accident rate is higher and energy consumption etc..Intelligent network joins the nonlinear system that electric car is multiple-input and multiple-output variable
System, system have the characteristics that the uncertainty of nonlinearity dynamic characteristic and parameter, design the queue controlling party of high-quality
Safety, string stability and the road passage capability of automobile platoon driving can be improved in method, this joins electronic vapour for China's intelligent network
The development and popularization of vehicle are of great significance.
The main task of intelligent network connection electric car platoon driving is to realize multiple vehicles according to certain following distance and speed
Degree traveling, to improve the traffic capacity of intelligent transportation system.(Wu Yao, Wang Xin, Xia Wei are based on the control of dynamic surface sliding formwork to document [1]
Active Queue Management Algorithm [J] marine electronic engineering of system, 2018,38 (03): 68-71.) using a kind of sliding based on dynamic surface
The Active Queue Management Algorithm of mould control, can overcome the uncertain interference of system, but be only capable of maintaining queue oscillation compared with
Fluctuation, cannot effectively eliminate oscillation in small range.Document [2] (C.Desjardins and B.Chaib-draa, "
Cooperative Adaptive Cruise Control:A Reinforcement Learning Approach,"in
IEEE Transactions on Intelligent Transportation Systems,vol.12,no.4,pp.1248-
1260, Dec.2011.doi:10.1109/TITS.2011.2157145) distributed director is designed using machine learning techniques,
Queue performance can be optimized by improving controller parameter using gradient descent algorithm, although realizing the accurate control of queue,
The speed and acceleration of single vehicle are presented on small range buffeting, are unsatisfactory for the comfort requirement of vehicle platoon control.Intelligent network connection
Electric car queue control system is to contain intelligent traffic signal, nonlinear and time-varying system electrically and mechanically, between vehicle
Speed and distance controlling, vehicle and traffic signals between information transmitting, between vehicle itself and its various components inside
Coordinate, so that intelligent network connection electric car queue control problem becomes extremely complex.
Summary of the invention
It is an object of the invention in view of the above-mentioned problems existing in the prior art, provide a kind of intelligence based on adaptive weighting
It can net connection electric car queue optimal control method.
Intelligent network of the present invention based on adaptive weighting joins electric car queue optimal control method, including following step
It is rapid:
1) intelligent network connection electric car queuing message interactive module by truck traffic system obtained from vehicle and front truck away from
From with the information such as velocity deviation, the speed and acceleration of each intelligent electric automobile in queue are obtained by onboard sensor,
The vehicle-state that will acquire is sent to queue optimal control module;
2) model linearization processor linearizes the non-linear Longitudinal Dynamic Model of electric car, according to vehicle headway
The electric car Longitudinal Dynamic Model after relationship and linearisation between model, vehicle and traffic lights establishes queue control
The state-space expression of system, queue optimal control module include cat swarm optimization optimization, secondary type controller and signal modulus of conversion
Block, cat swarm optimization optimization state according to needed for i-th vehicle of control that information exchange module i is obtained solve adaptive weighting matrix
QiAnd Ri;The adaptive weighting matrix Q that secondary type controller is acquired according to cat swarm optimization optimizationiAnd RiCalculate Optimal state-feedback
Control matrix Ki, and then find out control auxiliary variable ui;Signal conversion module is according to control auxiliary variable uiFind out desired control
Torque Ti, method particularly includes:
(1) model linearization processor is to intelligent network connection electric car Longitudinal Dynamic Model linearisation, in conjunction with following distance
Model establishes the state-space expression that queue controls i-th vehicle;
(2) quadratic form queue controller is designed, initial value is set to weight matrix;
(3) using the initial value of weight matrix parameter as the initial position of each cat in cat swarm optimization optimization, according to combination
Rate (MR indicates ratio of the cat number for executing tracing mode in population) determines that each cat is in search pattern in cat group
Or the flag value of tracing mode selects new position to replace original position according to select probability;Calculate the adaptation of each cat
Angle value, and retain the highest solution of fitness value, cat group is updated according to MR, meets termination condition and jumps out optimization, export current cat
The position of the corresponding cat of group's optimal solution corresponds to adaptive weighting matrix QiAnd RiParameter;
(4) according to the coefficient matrices A of the state-space expression of acquisitioni, control matrix Bi, adaptive weighting matrix QiWith
Ri, utilize differential Riccati equation solution matrix Pi, the Optimal state-feedback control under current state is solved using secondary type controller
Matrix K processedi;According to STATE FEEDBACK CONTROL formula, conversion and control auxiliary variable ui;
(5) signal conversion module is by the control auxiliary variable u of acquisitioniIt is converted into control moment Ti, will be controlled by CAN network
Torque T processediIt is sent to electric motor actuator.
3) signal conversion module is by the control auxiliary variable u of acquisitioniIt is processed into and executes instruction TiIt is sent to the execution of vehicle afterwards
Device, the final queue optimal control realized to intelligent network connection electric car.
The present invention uses sliding formwork control to be easy to appear " buffeting " phenomenon and fit using preset parameter control method for current
The problems such as response is not high, and control system dynamic characteristic is poor proposes a kind of intelligent network connection electric car based on adaptive weighting
Queue optimal control method.The present invention carries out linearization process to intelligent network connection electric car Longitudinal Dynamic Model first,
It is secondary, the secondary type controller of queue is designed, using tracing mode and is searched using cat swarm optimization according to the current state of vehicle in queue
Rope mode carries out real-time local optimum and real-time global optimization to the weight matrix of quadratic form queue controller, final to obtain currently
The corresponding optimum control variable of state and optimum state variable, avoid algorithm when controlling queue, due to falling into local optimum
It phenomena such as system chatter caused by solution and iteration latter stage slow due to the multifarious convergence for reducing appearance, realizes when front truck is multiple
When under miscellaneous operating condition, rear car fast and stable is followed, effectively improve intelligent network connection electric car queue control string stability,
Safety, comfort and accuracy.
Detailed description of the invention
Fig. 1 is that the intelligent network based on adaptive weighting joins electric car queue Optimal Control System structure chart.
Fig. 2 is the weight matrix optimized flow chart of the quadratic form queue controller based on cat swarm optimization.
Fig. 3 is cat swarm optimization search pattern and tracing mode flow chart.
Specific embodiment
Following embodiment will the present invention is further illustrated in conjunction with attached drawing.
The present invention provides a kind of intelligent networks to join electric car queue control method, with intelligent network connection system, fusion
It is shared to realize that vehicle is exchanged with intelligent information such as people, vehicle, road, backstages, has complex environment sense for modern communications and network technology
Know, the functions such as intelligent decision, Collaborative Control and execution are, it can be achieved that safety, comfortable, energy conservation, efficiently traveling.
As shown in Figure 1, the intelligent network connection electric car queue optimization control described in the embodiment of the present invention based on adaptive weighting
System processed, including information exchange module, queue optimal control module and execution module;Information exchange module, queue optimal control
Module, execution module are sequentially connected;Information exchange module be used for by truck traffic system obtained from vehicle at a distance from front truck and
The information such as velocity deviation are obtained the speed and acceleration of each intelligent electric automobile in queue by onboard sensor, will obtained
The vehicle-state taken is sent to queue optimal control module;The queue optimal control module include model linearization processor,
Cat swarm optimization optimization, secondary type controller and signal conversion module;Model linearization processor is used for the non-thread of electric car
Property Longitudinal Dynamic Model linearisation, according to the relationship and linearisation between vehicle headway model, vehicle and traffic lights
Electric car Longitudinal Dynamic Model afterwards establishes the state-space expression of queue control;Cat swarm optimization optimization is used for according to letter
State needed for ceasing i-th vehicle of control that interactive module obtains solves adaptive weighting matrix QiAnd Ri;Secondary type controller according to
The adaptive weighting matrix Q that cat swarm optimization optimization acquiresiAnd RiOptimal state feed-back control matrix is calculated, and then it is auxiliary to find out control
Help variable;Signal conversion module for the control auxiliary variable of acquisition is processed into execute instruction after be sent to execution module, it is real
Now to the queue optimal control of intelligent network connection electric car.
Intelligent network described in the embodiment of the present invention based on adaptive weighting joins electric car queue optimal control method, including
Following steps:
A. information exchange module.
Information exchange module obtains the distance between i-th vehicle and (i-1)-th vehicle by truck traffic system and speed is inclined
The information such as difference obtain the speed and acceleration of each intelligent electric automobile in queue by onboard sensor.Finally it will acquire
Vehicle-state be sent to queue optimal control module i.
First part, intelligent network join electric car and obtain (i-1)-th vehicle speed v by truck traffic systemi-1, (i-1)-th
Vehicle acceleration ai-1With the distance, delta d of i-th vehicle and (i-1)-th vehiclei。
Second part obtains the speed v of i-th vehicle by onboard sensoriWith acceleration aiInformation, will by CAN bus
Signal passes to information exchange module i.
B. queue optimal control module.
Firstly, establish the Longitudinal Dynamic Model of intelligent electric automobile, the linearisation processor that designs a model is by the non-of vehicle
The processing of linear model progressiveization establishes queue control according to the relationship between following distance model, vehicle and traffic lights
State-space expression.Secondly, queue optimal control module obtains the required state of control by information interaction system.Then,
Queue algorithm is divided into cat swarm optimization optimization and liner quadratic regulator device two parts, and cat swarm optimization is to liner quadratic regulator
The weight matrix of device carries out real-time optimization, and liner quadratic regulator device calculates optimal state feed-back control matrix.
The first step initially sets up i-th intelligent network connection electric car Longitudinal Dynamic Model, and design a model linearization process
Device linearizes complicated non-linear Longitudinal Dynamic Model;Secondly, establishing the electric car team knot of N+1 electric car composition
Structure model becomes the acceleration of previous automobile of each electric car (except leader's electric vehicle) as measurable interference
Amount, and send the vehicle to, to more efficiently solve possible coupling condition between front car and rear car, finally obtain
Queue state of a control spatial expression.
Establish the Longitudinal Dynamic Model of i-th vehicle in fleet:
Wherein, miFor the quality of i-th vehicle, Si(i=0,1,2..., n) is coordinate of i-th vehicle in referential, FiIt is vehicle
Driving force, σ represents air mass density, AciIndicate the cross-sectional area of i-th vehicle, cdiIt is resistance coefficient, dmiRepresent i-th
The mechanical resistance of vehicle.
In order to describe the kinetic characteristics of intelligent electric automobile, it is contemplated that in practical situations, the torque of motor, which exports, rings
Should there should be certain delay, in order to make up the non-linear factor of vehicle torque transmitting, therefore the dynamic response is described as one
Rank inertial model, expression formula are as follows:
Wherein, τmiRepresent motor one order inertia time constant, TmiIndicate the torque of the motor of i-th electric car, Tmai
Indicate that the motor of i-th electric car is effectively outputed to the torque of transmission system.
In view of the transmission efficiency problem under actual conditions, there is following equation to set up:
Tdi=ηiRaiTmai (0.3)
Tdi=Fi×reffi (0.4)
Wherein, TdiRepresent the driving moment that i-th vehicle is transmitted to wheel, ηiFor the transmission efficiency of i-th vehicle, RaiIndicate the
The resultant gear ratio of i vehicle, reffiIndicate i-th vehicle radius of wheel;
It is available by formula (1.2) (1.3) (1.4):
By the F of (1.1) formulaiIt is available to bring formula (1.5) into:
Following equation can be obtained by bringing (1.6) formula into the formula after (1.1) formula derivation:
For the nonlinear equation of formula (1.7), linear condition feedback controller is designed, is linearized, as follows:
Wherein, uiIt is control auxiliary variable.
Thus, it is possible to which the linear equation being simplified is as follows:
That is:
Wherein, aiIndicate the acceleration of i-th vehicle.
Fleet is made of the vehicle of N+1 arrangement traveling, mathematics geometry expression formula are as follows:
δi=Si-1-Si-δdi-Li (0.11)
Wherein, S0The position coordinates of leading vehicle, with lower target increase, the position of vehicle more and more rearward, Si(i=
0,1,2..., n) represent coordinate of i-th vehicle in referential, δdiIndicate the expectation following distance between i-th vehicle and (i-1)-th vehicle
From δiIndicate the deviation of expectation vehicle headway and actual range;LiIndicate i-th vehicle vehicle commander.
Using Fixed Time Interval strategy (the Constant-time headway spacing of current relatively broad application
Policy), mathematic(al) representation are as follows:
δdi=τhvi+d0 (0.12)
Wherein, τhFor time interval, i.e. rear car reaches the time required for front truck current location;viIndicate the speed of i-th vehicle
Degree;d0Represent the safe distance after vehicle stops.
It, in order to solve this problem, can be by front truck in view of the case where there may be couplings between front car and rear car
Partial information considers at measurable distracter, herein, using the acceleration of front truck as a measurable interference.
It therefore can be with definition status variable, xi=[δi Δvi ai]T, wi=ai-1
Wherein, wiIt is a measurable interference, that is to say the acceleration a of (i-1)-th vehiclei-1, Δ vi=vi-1-viI.e. i-th
The speed difference of vehicle and the (i-1)-th vehicle, and assume acceleration ai-1I-th can be passed to from the (i-1)-th vehicle by way of wireless network
Vehicle, final queue state of a control spatial expression are as follows:
Wherein each coefficient matrix are as follows:
Second step, using liner quadratic regulator device to adaptive weighting arranged in matrix initial value.
According to Quadratic Optimal Control theory, Qi、RiThe adaptive weighting matrix for respectively indicating state variable, controlling variable,
Wherein, QiFor positive semidefinite matrix, RiFor positive definite matrix.
IfRi=[ri]
Wherein, qdi、qvi、qai、riIt is exactly to i-th vehicle current state and control amount i.e. relative distance, relative velocity, i-th
The state variable and control variable that vehicle actual acceleration and i-th vehicle expectation acceleration are adjusted, therefore available:
Equation Section(Next)
ui=-Ri -1BilT(Pixi+Qi) (1.1)
Wherein, constant value matrix Pi≥0.P simultaneouslyiMatrix also meets Riccati equation:
At the same time, remember:
Qi=[(Ai-Ri -1BilBil TPi)T]-1PiBicwi (1.3)
Therefore the state feedback matrix expression formula of available system are as follows:
K1i=[kpi kvi kai]=- Ri -1Bil TPi (1.4)
K2i=kci=-Ri -1Bil T[(Ai-Ri -1BilBil TPi)T]-1PiBic (1.5)
The closed loop states spatial expression finally obtained are as follows:
It is adjusted by weight matrix of the artificial method to secondary type controller, it is preferable to obtain a control effect
The initial value of the weight matrix of state variable and control variable.
Third step, detailed process using the initial value of the weight matrix of i-th vehicle in queue as cat group as shown in Fig. 2, optimize
The initial position of each cat in algorithm, the speed of the every dimension of every cat of random initializtion are current according to i-th vehicle in queue
State utilizes the adaptive weighting matrix Q of the secondary type controller of cat swarm optimization real-time optimizationi、Ri, detailed process is as follows:
1, N cat is generated, the different parameters of cat swarm optimization, SMP (memory pond), SRD (domain of variation), CDC (variation are initialized
Number), neighbour structure, α, β, C, according to the obtained Q of second stepi、RiValue the position and speed of N cat is initialized.
The four-dimensional array representation of position use of each cat, the position m=1 of the m cat, 2 ... N:
The four-dimensional array representation of speed use of every cat, the speed m=1 of the m cat, 2 ... N:
2, mode locating for every cat is determined according to MR (Percentage bound), wherein
3, flag=1 cat is in search pattern
3.1, every cat j parts are replicated to be put into SMP.General j=SMP, if SPC (self-position judgement) be it is true,
Just using the position having already passed through as the position candidate to be moved, j=SMP-1 at this time, then XikIndicate k-th of i-th cat
The corresponding position of copy, corresponding QikAnd RikIndicate XikI-th secondary type controller of vehicle represented by corresponding parameter it is adaptive
Weight matrix.
3.2, the positional value and velocity amplitude of each copy are updated.
For each copy according to the value of CDC, random adds to the position of current cat or mitigates SRD%.Every cat in this way
Positional value between j copy of duplication is opened with regard to difference, from the point of view of relating to physical significance represented by adaptive weighting matrix,
It is equivalent to and generates a series of state variable and control variable in the current location small area of cat.
3.3, the updated fitness function value of j locations of copies of cat is calculated, calculation formula is as follows: Equation
Section(Next)
Wherein, xikWhat is indicated is that kth copy calculates the i-th vehicle current state of expression substituted into when fitness function
Amount, uikThat indicate is the control amount that kth copy calculates that the expection substituted into when fitness function value is applied on vehicle, Qik
And rikWhat is indicated is state variable corresponding to the position and speed parameter of the kth copy of current i-th cat and control variable
Corresponding weighting matrix.Calculating on the process nature of the value of fitness function is exactly real-time calculating queue state of a control spatial table
Up to the process of the secondary type controller optimum control matrix of formula.
3.4, compare the value of fitness function, record the position of the minimum corresponding cat of fitness function
3.5, cat swarm optimization search pattern terminates
4, flag=0 cat is in tracing mode
4.1, the position and speed of the individual of the cat in tracing mode is updated
The position of i-th cat is updated according to the following formula:
Wherein,Indicate i-th updated position of cat, the corresponding position of each cat is current queue shape
The expected state variable used and control variable parameter under state,To pass
Increase parameter, XbestWhat is indicated is the desired positions that entire cat group lives through, that is, what is indicated is can obtain under current vehicle condition
The corresponding parameter of adaptive weighting matrix of optimal control target is obtained,The speed of preceding i-th cat is updated, the position of cat is more
New is exactly the location updating that subsequent time cat is instructed according to the adaptive weighting matrix of global optimum under current queue state, is drawn
Entering globally optimal solution to instruct the variation of the position of cat is to allow algorithm more quickly to carry out more towards the direction of global optimum
Newly.
The speed of i-th cat is updated according to the following formula:
Wherein,Indicate the i-th updated speed of cat,Expression is passed
The acceleration parameter subtracted, εiIndicate the random vector being evenly distributed between 0-1.
The position and speed of the cat in tracing mode is updated according to two formula above, wherein due to βi(t) and αi
(t) presence of auto-adaptive parameter makes entire cat group be in the shape of dynamic equilibrium in tracing mode and the cat in search pattern
State enhances the global exploring ability of algorithm, improves iteration finally since insufficient caused do not restrain of population diversity is either received
The problems such as slow is held back, is corresponded to practical to the accuracy and stability that will improve control from the point of view of in the control of electric car.
4.2, the fitness function value of updated cat is calculated.
The calculation formula of fitness function is identical as step 3
4.3, compare fitness function value, and retain the position of the smallest cat of fitness function value.
4.4, tracing mode terminates.
5, judge whether to meet termination condition
Meet the position that termination condition will export the minimum corresponding cat of fitness function value, being unsatisfactory for output condition will again
The position and speed of cat is updated, the search pattern for re-starting cat is allocated.
4th step, according to the coefficient matrix for the open loop situations spatial expression that i-th successive vehicles that second step obtains control
Ai, control matrix Bi, according to the adaptive weighting matrix Q that obtains after the optimization of third step cat swarm optimizationi、Ri, utilize differential Li Kati
Equation solution matrix Pi, the optimal state feed-back control matrix under current state is solved according to linear-quadratic optimal control
Ki, control auxiliary variable u is finally acquired according to STATE FEEDBACK CONTROL formulai。
According to the following formula:
Signal conversion module will control auxiliary variable uiBe converted to control moment Ti, and signal is passed into holding for i-th vehicle
Row device.Cat swarm optimization search pattern and tracing mode flow chart are referring to Fig. 3.
C. instruction processing and control execution part
The actuator of i-th vehicle is handled according to executing instruction for acquisition, and final realize joins electric car to intelligent network
Queue control.
In conclusion by the search pattern of optimization cat group and the position and speed update method of tracing mode, according to working as
The state of vehicle in front carries out real-time optimization to the adaptive weighting matrix parameter of secondary type controller, establishes a kind of based on adaptive
The intelligent network of weight joins electric car queue optimal control method, avoids the buffeting occurred when algorithm Control complex systems, calculates
Slowly, algorithm is not restrained so as to cause vehicle control shakiness, dangerous working condition caused by overshoot is excessive, to reach raising intelligence
Accuracy, stability and the purpose of safety of net connection electric car queue control.
The above content is combine optimal technical scheme to the present invention done further description, and it cannot be said that invention
Specific implementation is only limitted to these explanations.For general technical staff of the technical field of the invention, the present invention is not being departed from
Design under the premise of, can also make it is simple deduce and replacement, all should be considered as protection scope of the present invention.
Claims (2)
1. the intelligent network based on adaptive weighting joins electric car queue optimal control method, it is characterised in that including following step
It is rapid:
1) intelligent network connection electric car queuing message interactive module by truck traffic system obtained from vehicle at a distance from front truck and
The information such as velocity deviation are obtained the speed and acceleration of each intelligent electric automobile in queue by onboard sensor, will obtained
The vehicle-state taken is sent to queue optimal control module;
2) model linearization processor linearizes the non-linear Longitudinal Dynamic Model of electric car, according to vehicle headway mould
The electric car Longitudinal Dynamic Model after relationship and linearisation between type, vehicle and traffic lights establishes queue control
State-space expression, queue optimal control module includes cat swarm optimization optimization, secondary type controller and signal conversion module,
Cat swarm optimization optimization state according to needed for i-th vehicle of control that information exchange module i is obtained solves adaptive weighting matrix Qi
And Ri;The adaptive weighting matrix Q that secondary type controller is acquired according to cat swarm optimization optimizationiAnd RiCalculate Optimal state-feedback control
Matrix K processedi, and then find out control auxiliary variable ui;Signal conversion module is according to control auxiliary variable uiFind out desired control force
Square Ti;
3) signal conversion module is by the control auxiliary variable u of acquisitioniIt is processed into and executes instruction TiIt is sent to the actuator of vehicle afterwards,
The final queue optimal control realized to intelligent network connection electric car.
2. the intelligent network based on adaptive weighting joins electric car queue optimal control method, feature as described in claim 1
It is in step 2), it is described to find out desired control moment TiMethod particularly includes:
(1) model linearization processor is to intelligent network connection electric car Longitudinal Dynamic Model linearisation, in conjunction with following distance model,
Establish the state-space expression that queue controls i-th vehicle;
(2) quadratic form queue controller is designed, initial value is set to weight matrix;
(3) using the initial value of weight matrix parameter as the initial position of each cat in cat swarm optimization optimization, according to Percentage bound
(MR indicates ratio of the cat number for executing tracing mode in population) determines that each cat is to be in search pattern also in cat group
It is the flag value of tracing mode, selects new position to replace original position according to select probability;Calculate the fitness of each cat
Value, and retain the highest solution of fitness value, cat group is updated according to MR, meets termination condition and jumps out optimization, exports current cat group
The position of the corresponding cat of optimal solution corresponds to adaptive weighting matrix QiAnd RiParameter;
(4) according to the coefficient matrices A of the state-space expression of acquisitioni, control matrix Bi, adaptive weighting matrix QiAnd Ri, benefit
With differential Riccati equation solution matrix Pi, the optimal state feed-back control square under current state is solved using secondary type controller
Battle array Ki;According to STATE FEEDBACK CONTROL formula, conversion and control auxiliary variable ui;
(5) signal conversion module is by the control auxiliary variable u of acquisitioniIt is converted into control moment Ti。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910702422.6A CN110456790B (en) | 2019-07-31 | 2019-07-31 | Intelligent networking electric automobile queue optimization control method based on adaptive weight |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910702422.6A CN110456790B (en) | 2019-07-31 | 2019-07-31 | Intelligent networking electric automobile queue optimization control method based on adaptive weight |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110456790A true CN110456790A (en) | 2019-11-15 |
CN110456790B CN110456790B (en) | 2020-10-02 |
Family
ID=68484347
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910702422.6A Active CN110456790B (en) | 2019-07-31 | 2019-07-31 | Intelligent networking electric automobile queue optimization control method based on adaptive weight |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110456790B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112820097A (en) * | 2021-01-08 | 2021-05-18 | 重庆邮电大学 | Truck fleet longitudinal hierarchical control method based on 5G-V2X and unmanned aerial vehicle |
CN113219962A (en) * | 2021-02-26 | 2021-08-06 | 北京航空航天大学合肥创新研究院(北京航空航天大学合肥研究生院) | Control method and system for hybrid queue following safety and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2451686B1 (en) * | 2009-07-07 | 2013-05-01 | Volvo Lastvagnar AB | Method and controller for controlling output torque of a propulsion unit. |
CN103956045A (en) * | 2014-05-13 | 2014-07-30 | 中国人民解放军军事交通学院 | Method for achieving collaborative driving of vehicle fleet by means of semi-physical simulation technology |
WO2016154033A1 (en) * | 2015-03-24 | 2016-09-29 | Brownridge Chelsea W | Dog locker/pet locker |
CN106476806A (en) * | 2016-10-26 | 2017-03-08 | 上海理工大学 | Cooperating type self-adaption cruise system algorithm based on transport information |
CN108357596A (en) * | 2018-01-26 | 2018-08-03 | 浙江大学 | A kind of unmanned bicycle of self-balancing that Behavior-based control drives and its control method that behavior drives |
CN109144076A (en) * | 2018-10-31 | 2019-01-04 | 吉林大学 | A kind of more vehicle transverse and longitudinals coupling cooperative control system and control method |
CN109591804A (en) * | 2018-11-22 | 2019-04-09 | 湖南大学 | Consider the vehicle platoon stability control method of communication delay |
-
2019
- 2019-07-31 CN CN201910702422.6A patent/CN110456790B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2451686B1 (en) * | 2009-07-07 | 2013-05-01 | Volvo Lastvagnar AB | Method and controller for controlling output torque of a propulsion unit. |
CN103956045A (en) * | 2014-05-13 | 2014-07-30 | 中国人民解放军军事交通学院 | Method for achieving collaborative driving of vehicle fleet by means of semi-physical simulation technology |
WO2016154033A1 (en) * | 2015-03-24 | 2016-09-29 | Brownridge Chelsea W | Dog locker/pet locker |
CN106476806A (en) * | 2016-10-26 | 2017-03-08 | 上海理工大学 | Cooperating type self-adaption cruise system algorithm based on transport information |
CN108357596A (en) * | 2018-01-26 | 2018-08-03 | 浙江大学 | A kind of unmanned bicycle of self-balancing that Behavior-based control drives and its control method that behavior drives |
CN109144076A (en) * | 2018-10-31 | 2019-01-04 | 吉林大学 | A kind of more vehicle transverse and longitudinals coupling cooperative control system and control method |
CN109591804A (en) * | 2018-11-22 | 2019-04-09 | 湖南大学 | Consider the vehicle platoon stability control method of communication delay |
Non-Patent Citations (3)
Title |
---|
INMO JANG 等: "《Cooperative Control for a Flight Array of UAVs and an Application in Radar Jamming》", 《IFAC PAPERSONLINE》 * |
吴瑶 等: "《基于动态面滑模控制的主动队列管理算法》", 《舰船电子工程》 * |
张云东: "《基于猫群优化算法的线性二次型最优控制器研究》", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112820097A (en) * | 2021-01-08 | 2021-05-18 | 重庆邮电大学 | Truck fleet longitudinal hierarchical control method based on 5G-V2X and unmanned aerial vehicle |
CN113219962A (en) * | 2021-02-26 | 2021-08-06 | 北京航空航天大学合肥创新研究院(北京航空航天大学合肥研究生院) | Control method and system for hybrid queue following safety and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN110456790B (en) | 2020-10-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110969848B (en) | Automatic driving overtaking decision method based on reinforcement learning under opposite double lanes | |
Zhang et al. | Ship motion attitude prediction based on an adaptive dynamic particle swarm optimization algorithm and bidirectional LSTM neural network | |
CN111267831A (en) | Hybrid vehicle intelligent time-domain-variable model prediction energy management method | |
CN109901403A (en) | A kind of face autonomous underwater robot neural network S control method | |
CN111103795B (en) | Hypersonic aircraft reentry section track optimization controller based on intelligent self-adaptive optimization algorithm | |
CN109240091A (en) | A kind of underwater robot control method based on intensified learning and its control method tracked | |
CN110442129B (en) | Control method and system for multi-agent formation | |
CN110456790A (en) | Intelligent network based on adaptive weighting joins electric car queue optimal control method | |
CN112684781B (en) | Multi-agent distributed model prediction control method and system | |
CN113900449B (en) | Multi-unmanned aerial vehicle track planning method and device, unmanned aerial vehicle and storage medium | |
CN109144099B (en) | Fast evaluation method for unmanned aerial vehicle group action scheme based on convolutional neural network | |
CN108394429B (en) | Method for generating automatic driving curve for urban rail train group | |
CN114995119A (en) | Urban traffic signal cooperative control method based on multi-agent deep reinforcement learning | |
CN108919799A (en) | A kind of net connection intelligent vehicle cooperation lane-change method | |
Li | A hierarchical autonomous driving framework combining reinforcement learning and imitation learning | |
Huang et al. | Cooperative collision avoidance method for multi-uav based on kalman filter and model predictive control | |
CN114943168B (en) | Method and system for combining floating bridges on water | |
CN116817909A (en) | Unmanned aerial vehicle relay type navigation method based on deep reinforcement learning | |
CN116176654A (en) | Scene self-adaptive track traffic ATO control system | |
CN114200835B (en) | Heterogeneous fleet stability distributed control method based on Tube MPC | |
CN106373410B (en) | A kind of Optimal Method of Urban Traffic Signal Control | |
Gong et al. | A path planning method based on improved particle swarm optimization algorithm | |
CN116256970A (en) | Data-driven cloud edge cooperative control method and system based on disturbance observer | |
Chen et al. | Decision making for overtaking of unmanned vehicle based on deep Q-learning | |
Li et al. | Prioritized experience replay based reinforcement learning for adaptive tracking control of autonomous underwater vehicle |
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 |