CN105151024A - Vehicle brake control method and device - Google Patents

Abstract

The invention relates to a vehicle brake control method and device and belongs to the field of new energy vehicles. The vehicle brake control method comprises the steps that the control parameters of a vehicle are obtained, wherein the control parameters comprise at least one parameter influencing the brake process of the vehicle; according to the control parameters, the adjustable parameters of a PID controller are obtained through a neural network algorithm; according to the adjustable parameters, the target brake parameters of the vehicle are controlled through the PID controller. The target brake parameters comprise the rotation speed of a motor, regenerative brake torque and hydraulic brake torque. By the adoption of the vehicle brake control method and device, the adjustable parameters of the PID controller are adjusted through the neural network algorithm, and the problems that in relevant technology, accurate PID adjustable parameters are difficult to determine, the brake effect on the vehicle is poor, and the recycling efficiency of regenerative brake kinetic energy is low are solved; and the effects that the accurate PID adjustable parameters can be determined, the brake effect on the vehicle is good, and the recycling efficiency of regenerative brake kinetic energy is high are achieved.

Description

The brake control method of vehicle and device

Technical field

The present invention relates to new-energy automobile field, particularly a kind of brake control method of vehicle and device.

Background technology

Regenerative brake (Regenerativebraking), also known as regenerative braking, is a kind of braking technology be used on elec. vehicle.Kinetic transformation electrical energy for storage when car brakeing vehicle gets up.Regenerative Braking Technology, as a crucial energy-conserving and environment-protective technology of new-energy automobile, more and more receives the concern in vehicle research and development field.

A kind of brake control method of vehicle is had in correlation technique, the method is by PID (proportion integration differentiation, proportion-integration-differentiation) controller comes real-time adjustment regenerative braking moment and hydraulic braking moment, then completes the braking to vehicle.

Contriver is realizing in process of the present invention, find that aforesaid way at least exists following defect: said method is when using PID controller adjustment regenerative braking moment and hydraulic braking moment, be difficult to determine PID adjustable parameter comparatively accurately, cause the braking effect of vehicle poor, the efficiency that regenerative brake reclaims kinetic energy is lower.

Summary of the invention

Be difficult to determine PID adjustable parameter comparatively accurately to solve in correlation technique, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy, embodiments provides a kind of brake control method and device of vehicle.Described technical scheme is as follows:

According to a first aspect of the invention, provide a kind of brake control method of vehicle, described method comprises:

Obtain the controling parameters of vehicle, described controling parameters comprises at least one parameter affecting described car brakeing process;

Neural network algorithm is used to obtain the adjustable parameter of proportion integration differentiation PID controller according to described controling parameters;

Utilize described PID controller to control the object brake parameters of described vehicle according to described adjustable parameter, described object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

Optionally, described adjustable parameter comprises scale parameter, integral parameter and differential parameter,

The described adjustable parameter using neural network algorithm acquisition proportion integration differentiation PID controller according to described controling parameters, comprising:

Obtain the input layer hidden layer weights of current time and the hidden layer output layer weights of current time;

Determine neural network structure, described neural network includes input layer, hidden layer and output layer;

Described controling parameters is inputted to described neural network;

According to described controling parameters and described input layer hidden layer weights determination hidden layer incoming signal;

According to described hidden layer incoming signal determination hidden layer output signal;

The output layer incoming signal of described neural network is determined according to described hidden layer output signal and described hidden layer output layer weights;

The output parameter of described neural network is determined according to the output layer incoming signal of described neural network;

The output parameter of described neural network is defined as described adjustable parameter.

Optionally, the input layer hidden layer weights of described acquisition current time and the hidden layer output layer weights of current time, comprising:

When needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, described n be more than or equal to 2 integer.

Optionally, the described input layer hidden layer weights by adjusting a upper control cycle obtain the input layer hidden layer weights of described current time, comprising:

Obtain the input layer hidden layer weights of a described upper control cycle;

Obtain the input layer hidden layer weights of described current time according to the input layer hidden layer weights of the described upper control cycle of input layer hidden layer weights formula adjustment, described input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\αw}}_{li}^2(k-1),$

Wherein, described in for the input layer hidden layer weights of described current time, described in for described hidden layer incoming signal, described in for described hidden layer output signal, described in for input layer output signal, described α is inertia coefficient, and described η is learning efficiency, described in for parameter preset, described in for the initial value of input layer hidden layer weights, described in for the input layer hidden layer weights of a described upper control cycle;

The described hidden layer output layer weights by adjusting a upper control cycle obtain the hidden layer output layer weights of described current time, comprising:

Obtain the hidden layer output layer weights of a described upper control cycle;

Obtain the hidden layer output layer weights of described current time according to the hidden layer output layer weights of the described upper control cycle of hidden layer output layer weights formula adjustment, described hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right),$

Wherein, described in for the hidden layer output layer weights of described current time, described e (k) is error amount, the actual value of the described object brake parameters that described y (k) is current control period, described y (k-1) is the actual value of the object brake parameters of a described upper control cycle, described in for described output layer incoming signal, described in for output layer output signal, the output valve of the described PID controller that described u (k) is described current control period, described u (k-1) is the output valve of the described PID controller of a described upper control cycle.

Optionally, described when needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, comprising:

When needing the object brake parameters controlling described vehicle for n-th time, obtain error criterion according to error formula, described error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2,$

Wherein, described E (k) is error criterion, and described r (k) is the setting value of described object brake parameters, the actual value of the described object brake parameters that described y (k) is current control period;

Judge whether described error criterion is less than default error threshold;

When described error criterion is not less than described default error threshold, obtained the input layer hidden layer weights of described current time by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle.

According to a second aspect of the invention, provide a kind of braking force control system of vehicle, described device comprises:

Controling parameters acquisition module, be configured to the controling parameters obtaining vehicle, described controling parameters comprises at least one parameter affecting described car brakeing process;

Adjustable parameter acquisition module, is configured to use neural network algorithm to obtain the adjustable parameter of proportion integration differentiation PID controller according to described controling parameters;

Control module, be configured to utilize described PID controller to control the object brake parameters of described vehicle according to described adjustable parameter, described object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

Optionally, described adjustable parameter comprises scale parameter, integral parameter and differential parameter,

Described adjustable parameter acquisition module, comprising:

Weights obtain submodule, are configured to obtain the input layer hidden layer weights of current time and the hidden layer output layer weights of current time;

Structure determination submodule, is configured to determine neural network structure, described neural network includes input layer, hidden layer and output layer;

Parameters input submodule, is configured to input described controling parameters to described neural network;

Hidden layer input submodule, is configured to according to described controling parameters and described input layer hidden layer weights determination hidden layer incoming signal;

Hidden layer output sub-module, is configured to according to described hidden layer incoming signal determination hidden layer output signal;

Output layer input submodule, is configured to the output layer incoming signal determining described neural network according to described hidden layer output signal and described hidden layer output layer weights;

Output sub-module, is configured to the output parameter determining described neural network according to the output layer incoming signal of described neural network;

Parameter determination submodule, is configured to the output parameter of described neural network to be defined as described adjustable parameter.

Optionally, described weights obtain submodule, comprising:

Adjustment unit, be configured to when needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, described n be more than or equal to 2 integer.

Optionally, described adjustment unit, is configured to the input layer hidden layer weights obtaining a described upper control cycle;

Obtain the input layer hidden layer weights of described current time according to the input layer hidden layer weights of the described upper control cycle of input layer hidden layer weights formula adjustment, described input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\αw}}_{li}^2(k-1),$

Wherein, described in for the input layer hidden layer weights of described current time, described in for described hidden layer incoming signal, described in for described hidden layer output signal, described in for input layer output signal, described α is inertia coefficient, and described η is learning efficiency, described in for parameter preset, described in for the initial value of input layer hidden layer weights, described in for the input layer hidden layer weights of a described upper control cycle;

Described adjustment unit, is configured to the hidden layer output layer weights obtaining a described upper control cycle;

Obtain the hidden layer output layer weights of described current time according to the hidden layer output layer weights of the described upper control cycle of hidden layer output layer weights formula adjustment, described hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right),$

Wherein, described in for the hidden layer output layer weights of described current time, described e (k) is error amount, the actual value of the described object brake parameters that described y (k) is current control period, described y (k-1) is the actual value of the object brake parameters of a described upper control cycle, described in for described output layer incoming signal, described in for output layer output signal, the output valve of the described PID controller that described u (k) is described current control period, described u (k-1) is the output valve of the described PID controller of a described upper control cycle.

Optionally, described adjustment unit, be configured to when needing the object brake parameters controlling described vehicle for n-th time, obtain error criterion according to error formula, described error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2,$

Wherein, described E (k) is error criterion, and described r (k) is the setting value of described object brake parameters, the actual value of the described object brake parameters that described y (k) is current control period;

Judge whether described error criterion is less than default error threshold;

When described error criterion is not less than described default error threshold, obtained the input layer hidden layer weights of described current time by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle.

The technical scheme that the embodiment of the present invention provides can comprise following beneficial effect:

Adjusted the adjustable parameter of PID controller by neural network algorithm, solve in correlation technique and be difficult to determine PID adjustable parameter comparatively accurately, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy; Reach and can determine comparatively accurate PID adjustable parameter, the braking effect of vehicle is better, and regenerative brake reclaims the higher effect of the efficiency of kinetic energy.

Should be understood that, it is only exemplary and explanatory that above general description and details hereinafter describe, and can not limit the present invention.

Accompanying drawing explanation

Accompanying drawing to be herein merged in specification sheets and to form the part of this specification sheets, shows embodiment according to the invention, and is used from specification sheets one and explains principle of the present invention.

Fig. 1 is the diagram of circuit of the brake control method of a kind of vehicle that the invention process exemplifies;

Fig. 2-1 is the diagram of circuit of the brake control method of a kind of vehicle that the invention process exemplifies;

Fig. 2-2 obtains input layer hidden layer weights in Fig. 2-1 illustrated embodiment and hidden layer output layer weights obtain diagram of circuit;

Fig. 2-3 is schematic diagrams of neural network in Fig. 2-1 illustrated embodiment;

Fig. 3-1 is the block diagram of the braking force control system of a kind of vehicle that the invention process exemplifies;

Fig. 3-2 is block diagrams of adjustable parameter acquisition module in Fig. 3-1 illustrated embodiment;

Fig. 3-3 is block diagrams that in Fig. 3-1 illustrated embodiment, weights obtain submodule.

By above-mentioned accompanying drawing, illustrate the embodiment that the present invention is clear and definite more detailed description will be had hereinafter.These accompanying drawings and text description be not in order to limited by any mode the present invention design scope, but by reference to specific embodiment for those skilled in the art illustrate concept of the present invention.

Detailed description of the invention

Here will be described exemplary embodiment in detail, its sample table shows in the accompanying drawings.When description below relates to accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawing represents same or analogous key element.Embodiment described in following exemplary embodiment does not represent all embodiments consistent with the present invention.On the contrary, they only with as in appended claims describe in detail, the example of apparatus and method that aspects more of the present invention are consistent.

Fig. 1 is the diagram of circuit of the brake control method of a kind of vehicle that the invention process exemplifies.The brake control method of this vehicle can comprise following several step:

In a step 101, obtain the controling parameters of vehicle, controling parameters comprises at least one parameter affecting car brakeing process.

In a step 102, neural network algorithm is used to obtain the adjustable parameter of proportion integration differentiation PID controller according to controling parameters.

In step 103, utilize PID controller to control the object brake parameters of vehicle according to adjustable parameter, object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

In sum, the brake control method of the vehicle that the embodiment of the present invention provides, the adjustable parameter of PID controller is adjusted by neural network algorithm, solve in correlation technique and be difficult to determine PID adjustable parameter comparatively accurately, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy; Reach and can determine comparatively accurate PID adjustable parameter, the braking effect of vehicle is better, and regenerative brake reclaims the higher effect of the efficiency of kinetic energy.

Fig. 2-1 is the diagram of circuit of the brake control method of a kind of vehicle that the invention process exemplifies.The brake control method of this vehicle can be applied in automobile.The brake control method of this vehicle can comprise following several step:

In step 201, obtain the controling parameters of vehicle, controling parameters comprises at least one parameter affecting car brakeing process.

When the brake control method of the vehicle using the embodiment of the present invention to provide, first the controling parameters of vehicle can be obtained, this controling parameters can comprise at least one parameter affecting car brakeing process, such as Vehicle Speed, the control signal etc. of brake pedal.

In step 202., the input layer hidden layer weights of current time and the hidden layer output layer weights of current time are obtained.

After the controling parameters obtaining vehicle, the input layer hidden layer weights of current time and the hidden layer output layer weights of current time can be obtained, the initial value of input layer hidden layer weights and the initial value of hidden layer output layer weights can be inputted by operating personal, need afterwards to adjust input layer hidden layer weights and hidden layer output layer weights during the object brake parameters controlling vehicle at every turn.Namely the input layer hidden layer weights of current time are the input layer hidden layer weights of current control period, and the hidden layer output layer weights of current time are the hidden layer output layer weights of current control period.

Concrete, as shown in Fig. 2-2, this step can comprise two sub-steps below:

Sub-step 2021, when needing the object brake parameters controlling vehicle for n-th time, obtains the input layer hidden layer weights of current time by the input layer hidden layer weights adjusting a upper control cycle, n be more than or equal to 2 integer.

The process of this step adjustment input layer hidden layer weights can comprise:

1) the input layer hidden layer weights of a upper control cycle are obtained.

2) obtain the input layer hidden layer weights of current time according to the input layer hidden layer weights of the upper control cycle of input layer hidden layer weights formula adjustment, input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\αw}}_{li}^2(k-1).$

Wherein, for the input layer hidden layer weights of current time, for hidden layer incoming signal, for hidden layer output signal, for input layer output signal, α is inertia coefficient, and η is learning efficiency, for parameter preset, for the initial value of input layer hidden layer weights, for the input layer hidden layer weights of a upper control cycle.

Sub-step 2022, when needing the object brake parameters controlling vehicle for n-th time, obtains the hidden layer output layer weights of current time by the hidden layer output layer weights adjusting a upper control cycle.

The process of this step adjustment hidden layer output layer weights can comprise:

1) the hidden layer output layer weights of a upper control cycle are obtained.

2) obtain the hidden layer output layer weights of current time according to the hidden layer output layer weights of the upper control cycle of hidden layer output layer weights formula adjustment, hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right).$

Wherein, for the hidden layer output layer weights of current time, e (k) is error amount, the actual value of the object brake parameters that y (k) is current control period, and y (k-1) is the actual value of the object brake parameters of a upper control cycle, for output layer incoming signal, for output layer output signal, the output valve of the PID controller that u (k) is current control period, u (k-1) is the output valve of the PID controller of a upper control cycle.It should be noted that, sub-step 2021 and sub-step 2022 there is no sequencing when performing.

It should be noted that, this step is when adjusting the hidden layer output layer weights of the input layer hidden layer weights of current time and current time, and first can obtain error criterion according to error formula, error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2;$

Wherein, E (k) is error criterion, the setting value of brake parameters for the purpose of r (k), the actual value of the object brake parameters that y (k) is current control period.

Whether error in judgement index is less than default error threshold afterwards.And when error criterion is not less than default error threshold, the adjustment input layer hidden layer weights of current time and the hidden layer output layer weights of current time.When error criterion is less than default error threshold, can using the hidden layer output layer weights of the input layer hidden layer weights of current time and current time as the input layer hidden layer weights finally determined and hidden layer output layer weights.

In step 203, determine neural network structure, neural network includes input layer, hidden layer and output layer.

After the hidden layer output layer weights of the input layer hidden layer weights and current time that obtain current time, neural network structure can be determined, as Figure 2-3, neural network can include input layer 21, hidden layer 22 and output layer 23, input layer 21 can have m input node, corresponding to the controling parameters of vehicle, output layer 23 can have 3 output nodes, corresponding to 3 adjustable parameter k of PID controller _p, k _iand k _d, wherein k _pfor scale parameter, k _ifor integral parameter, k _dfor differential parameter.

In step 204, neuralward network input control parameter.

After determining the structure of neural network and input layer hidden layer weights and hidden layer output layer weights, can the controling parameters of neuralward network input vehicle.

In step 205, according to controling parameters and input layer hidden layer weights determination hidden layer incoming signal.

In neuralward network after input control parameter, can according to controling parameters and input layer hidden layer weights determination hidden layer incoming signal.Wherein hidden layer incoming signal can input formula to determine according to hidden layer, and hidden layer input formula can be:

${\mathrm{net}}_i^{\left(2\right)}\left(k\right)=\underset{i=0}{\overset{m}{\Σ}}{w}_{ij}^{\left(2\right)}{o}_j^{\left(1\right)}.$

Wherein for hidden layer incoming signal, for input layer hidden layer weights, for input layer output signal.

In step 206, according to hidden layer incoming signal determination hidden layer output signal.

After determining hidden layer incoming signal, can according to hidden layer incoming signal determination hidden layer output signal.Hidden layer output signal can be wherein for hidden layer incoming signal, Sigmoid (S type) function that f (x) is Symmetrical.And wherein e is parameter preset.

In step 207, according to the output layer incoming signal of hidden layer output signal and hidden layer output layer weights determination neural network.

After obtaining hidden layer output signal, can according to the output layer incoming signal of hidden layer output signal and hidden layer output layer weights determination neural network.Wherein output layer incoming signal can be determined according to output layer incoming signal formula, and output layer incoming signal formula can be:

${\mathrm{net}}_i^{\left(3\right)}\left(k\right)=\underset{i=0}{\overset{m}{\Σ}}{w}_{ij}^{\left(3\right)}{o}_j^{\left(2\right)}.$

Wherein for output layer incoming signal, for hidden layer output layer weights, for hidden layer output signal.

In a step 208, according to the output parameter of the output layer incoming signal determination neural network of neural network.

After obtaining output layer incoming signal, can according to the output parameter of the output layer incoming signal determination neural network of neural network.

The output parameter of neural network can be wherein for the output parameter of neural network, for output layer incoming signal, l can be 1 or 2 or 3, g (x) be neuronal activation function, and g (x)=1/2 (1=tanh (x))=e ^x/ e ^x+ e ^-x.

In step 209, the output parameter of neural network is defined as adjustable parameter.

After the output parameter determining neural network, the output parameter of neural network can be defined as adjustable parameter.Namely three adjustable parameters of PID controller can be respectively when l equals 1,2,3, exemplary, $\left\{\begin{array}{c}{o}_1^{\left(3\right)}\left(k\right)=k_p\\ o_2^{\left(3\right)}\left(k\right)=k_i\\ o_3^{\left(3\right)}\left(k\right)=k_d\end{array}\right.,$ K _pfor scale parameter, k _ifor integral parameter, k _dfor differential parameter.

In step 210, PID controller is utilized to control the object brake parameters of vehicle according to adjustable parameter.

After obtaining adjustable parameter, PID controller can be utilized to control the object brake parameters of vehicle according to adjustable parameter, object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

PID dominated formulate can be u (k)=u (k-1)+k _p(e (k-1)+k _ie (k)+k _d(e (k)-2e (k-1)+e (k-2)), wherein u (k) is PID controller output, this output can be the regulation coefficient to object brake parameters, the PID controller that u (k-1) is a upper control cycle exports, the error amount that e (k) is current control period and e (k)=r (k)-y (k), the setting value of brake parameters for the purpose of r (k), the actual value of the object brake parameters that y (k) is current control period, e (k-1) is the error amount of a upper control cycle, e (k-2) is the error amount of a upper control cycle, k _pfor scale parameter, k _ifor integral parameter, k _dfor differential parameter.

It should be added that, the brake control method of the vehicle that the embodiment of the present invention provides, by hidden layer output layer weights and the input layer hidden layer weights of periodic cycle adjustment neural network, reach the effect improving PID controller adjustable parameter accuracy.

In sum, the brake control method of the vehicle that the embodiment of the present invention provides, the adjustable parameter of PID controller is adjusted by neural network algorithm, solve in correlation technique and be difficult to determine PID adjustable parameter comparatively accurately, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy; Reach and can determine comparatively accurate PID adjustable parameter, the braking effect of vehicle is better, and regenerative brake reclaims the higher effect of the efficiency of kinetic energy.

Following is apparatus of the present invention embodiment, may be used for performing the inventive method embodiment.For the details do not disclosed in apparatus of the present invention embodiment, please refer to the inventive method embodiment.

Fig. 3-1 is the block diagram of the braking force control system of a kind of vehicle that the invention process exemplifies.The braking force control system of this vehicle can comprise:

Controling parameters acquisition module 310, be configured to the controling parameters obtaining vehicle, controling parameters comprises at least one parameter affecting car brakeing process.

Adjustable parameter acquisition module 320, is configured to use neural network algorithm to obtain the adjustable parameter of proportion integration differentiation PID controller according to controling parameters.

Control module 330, be configured to utilize PID controller to control the object brake parameters of vehicle according to adjustable parameter, object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

In sum, the braking force control system of the vehicle that the embodiment of the present invention provides, the adjustable parameter of PID controller is adjusted by neural network algorithm, solve in correlation technique and be difficult to determine PID adjustable parameter comparatively accurately, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy; Reach and can determine comparatively accurate PID adjustable parameter, the braking effect of vehicle is better, and regenerative brake reclaims the higher effect of the efficiency of kinetic energy.

Optionally, adjustable parameter comprises scale parameter, integral parameter and differential parameter.

As shown in figure 3-2, adjustable parameter acquisition module 320, comprising:

Weights obtain submodule 321, are configured to obtain the input layer hidden layer weights of current time and the hidden layer output layer weights of current time.

Structure determination submodule 322, is configured to determine neural network structure, neural network includes input layer, hidden layer and output layer.

Parameters input submodule 323, is configured to neuralward network input control parameter.

Hidden layer input submodule 324, is configured to according to controling parameters and input layer hidden layer weights determination hidden layer incoming signal.

Hidden layer output sub-module 325, is configured to according to hidden layer incoming signal determination hidden layer output signal.

Output layer input submodule 326, is configured to the output layer incoming signal according to hidden layer output signal and hidden layer output layer weights determination neural network.

Output sub-module 327, is configured to the output parameter of the output layer incoming signal determination neural network according to neural network.

Parameter determination submodule 328, is configured to the output parameter of neural network to be defined as adjustable parameter.

Optionally, as shown in Fig. 3-3, weights obtain submodule 321, comprising:

Adjustment unit 3211, be configured to when needing the object brake parameters controlling vehicle for n-th time, the input layer hidden layer weights of current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of current time by the hidden layer output layer weights adjusting a upper control cycle, n be more than or equal to 2 integer.

Optionally, adjustment unit 3211, is configured to the input layer hidden layer weights obtaining a upper control cycle.

Obtain the input layer hidden layer weights of current time according to the input layer hidden layer weights of the upper control cycle of input layer hidden layer weights formula adjustment, input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\α\Δw}}_{li}^2(k-1).$

Wherein, for the input layer hidden layer weights of current time, for hidden layer incoming signal, for hidden layer output signal, for input layer output signal, α is inertia coefficient, and η is learning efficiency, for parameter preset, for the initial value of input layer hidden layer weights, for the input layer hidden layer weights of a upper control cycle.

Adjustment unit 3211, is configured to the hidden layer output layer weights obtaining a upper control cycle.

Obtain the hidden layer output layer weights of current time according to the hidden layer output layer weights of the upper control cycle of hidden layer output layer weights formula adjustment, hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right).$

Wherein, for the hidden layer output layer weights of current time, e (k) is error amount, the actual value of the object brake parameters that y (k) is current control period, and y (k-1) is the actual value of the object brake parameters of a upper control cycle, for output layer incoming signal, for output layer output signal, the output valve of the PID controller that u (k) is current control period, u (k-1) is the output valve of the PID controller of a upper control cycle.

Optionally, adjustment unit 3211, be configured to when needing the object brake parameters controlling vehicle for n-th time, obtain error criterion according to error formula, error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2.$

Wherein, E (k) is error criterion, the setting value of brake parameters for the purpose of r (k), the actual value of the object brake parameters that y (k) is current control period.

Whether error in judgement index is less than default error threshold.

When error criterion is not less than default error threshold, obtained the input layer hidden layer weights of current time by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of current time by the hidden layer output layer weights adjusting a upper control cycle.

It should be added that, the braking force control system of the vehicle that the embodiment of the present invention provides, by hidden layer output layer weights and the input layer hidden layer weights of periodic cycle adjustment neural network, reach the effect improving PID controller adjustable parameter accuracy.

In sum, the braking force control system of the vehicle that the embodiment of the present invention provides, the adjustable parameter of PID controller is adjusted by neural network algorithm, solve in correlation technique and be difficult to determine PID adjustable parameter comparatively accurately, the braking effect of vehicle is poor, and regenerative brake reclaims the lower problem of the efficiency of kinetic energy; Reach and can determine comparatively accurate PID adjustable parameter, the braking effect of vehicle is better, and regenerative brake reclaims the higher effect of the efficiency of kinetic energy.

About the device in above-described embodiment, wherein the concrete mode of modules executable operations has been described in detail in about the embodiment of the method, will not elaborate explanation herein.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. a brake control method for vehicle, is characterized in that, described method comprises:

Obtain the controling parameters of vehicle, described controling parameters comprises at least one parameter affecting described car brakeing process;

Neural network algorithm is used to obtain the adjustable parameter of proportion integration differentiation PID controller according to described controling parameters;

Utilize described PID controller to control the object brake parameters of described vehicle according to described adjustable parameter, described object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

2. method according to claim 1, is characterized in that, described adjustable parameter comprises scale parameter, integral parameter and differential parameter,

The described adjustable parameter using neural network algorithm acquisition proportion integration differentiation PID controller according to described controling parameters, comprising:

Obtain the input layer hidden layer weights of current time and the hidden layer output layer weights of current time;

Determine neural network structure, described neural network includes input layer, hidden layer and output layer;

Described controling parameters is inputted to described neural network;

According to described controling parameters and described input layer hidden layer weights determination hidden layer incoming signal;

According to described hidden layer incoming signal determination hidden layer output signal;

The output layer incoming signal of described neural network is determined according to described hidden layer output signal and described hidden layer output layer weights;

The output parameter of described neural network is determined according to the output layer incoming signal of described neural network;

The output parameter of described neural network is defined as described adjustable parameter.

3. method according to claim 2, is characterized in that, the input layer hidden layer weights of described acquisition current time and the hidden layer output layer weights of current time, comprising:

When needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, described n be more than or equal to 2 integer.

4. method according to claim 3, is characterized in that,

The described input layer hidden layer weights by adjusting a upper control cycle obtain the input layer hidden layer weights of described current time, comprising:

Obtain the input layer hidden layer weights of a described upper control cycle;

Obtain the input layer hidden layer weights of described current time according to the input layer hidden layer weights of the described upper control cycle of input layer hidden layer weights formula adjustment, described input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\αw}}_{li}^2(k-1),$

Wherein, described in for the input layer hidden layer weights of described current time, described in for described hidden layer incoming signal, described in for described hidden layer output signal, described in for input layer output signal, described α is inertia coefficient, and described η is learning efficiency, described in for parameter preset, described in for the initial value of input layer hidden layer weights, described in for the input layer hidden layer weights of a described upper control cycle;

The described hidden layer output layer weights by adjusting a upper control cycle obtain the hidden layer output layer weights of described current time, comprising:

Obtain the hidden layer output layer weights of a described upper control cycle;

Obtain the hidden layer output layer weights of described current time according to the hidden layer output layer weights of the described upper control cycle of hidden layer output layer weights formula adjustment, described hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right),$

Wherein, described in for the hidden layer output layer weights of described current time, described e (k) is error amount, the actual value of the described object brake parameters that described y (k) is current control period, described y (k-1) is the actual value of the object brake parameters of a described upper control cycle, described in for described output layer incoming signal, described in for output layer output signal, the output valve of the described PID controller that described u (k) is described current control period, described u (k-1) is the output valve of the described PID controller of a described upper control cycle.

5. the method according to claim 3 or 4, it is characterized in that, described when needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, comprising:

When needing the object brake parameters controlling described vehicle for n-th time, obtain error criterion according to error formula, described error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2,$

Wherein, described E (k) is error criterion, and described r (k) is the setting value of described object brake parameters, the actual value of the described object brake parameters that described y (k) is current control period;

Judge whether described error criterion is less than default error threshold;

When described error criterion is not less than described default error threshold, obtained the input layer hidden layer weights of described current time by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle.

6. a braking force control system for vehicle, is characterized in that, described device comprises:

Controling parameters acquisition module, be configured to the controling parameters obtaining vehicle, described controling parameters comprises at least one parameter affecting described car brakeing process;

Adjustable parameter acquisition module, is configured to use neural network algorithm to obtain the adjustable parameter of proportion integration differentiation PID controller according to described controling parameters;

Control module, be configured to utilize described PID controller to control the object brake parameters of described vehicle according to described adjustable parameter, described object brake parameters comprises motor speed, regenerative braking moment and hydraulic braking moment.

7. device according to claim 6, is characterized in that, described adjustable parameter comprises scale parameter, integral parameter and differential parameter,

Described adjustable parameter acquisition module, comprising:

Weights obtain submodule, are configured to obtain the input layer hidden layer weights of current time and the hidden layer output layer weights of current time;

Structure determination submodule, is configured to determine neural network structure, described neural network includes input layer, hidden layer and output layer;

Parameters input submodule, is configured to input described controling parameters to described neural network;

Hidden layer input submodule, is configured to according to described controling parameters and described input layer hidden layer weights determination hidden layer incoming signal;

Hidden layer output sub-module, is configured to according to described hidden layer incoming signal determination hidden layer output signal;

Output layer input submodule, is configured to the output layer incoming signal determining described neural network according to described hidden layer output signal and described hidden layer output layer weights;

Output sub-module, is configured to the output parameter determining described neural network according to the output layer incoming signal of described neural network;

Parameter determination submodule, is configured to the output parameter of described neural network to be defined as described adjustable parameter.

8. device according to claim 7, is characterized in that,

Described weights obtain submodule, comprising:

Adjustment unit, be configured to when needing the object brake parameters controlling described vehicle for n-th time, the input layer hidden layer weights of described current time are obtained by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle, described n be more than or equal to 2 integer.

9. device according to claim 8, is characterized in that,

Described adjustment unit, is configured to the input layer hidden layer weights obtaining a described upper control cycle;

Obtain the input layer hidden layer weights of described current time according to the input layer hidden layer weights of the described upper control cycle of input layer hidden layer weights formula adjustment, described input layer hidden layer weights formula is:

${\mathrm{\Δw}}_{ij}^{\left(2\right)}\left(k\right)=\mathrm{\η}f\left({\mathrm{net}}_i^2\right(k\left)\right)\underset{l=1}{\overset{3}{\Σ}}{\mathrm{\δ}}_l^{\left(3\right)}{w}_{li}^3\left(k\right)o_j^{\left(1\right)}\left(k\right)+{\mathrm{\αw}}_{li}^2(k-1),$

Wherein, described in for the input layer hidden layer weights of described current time, described in for described hidden layer incoming signal, described in for described hidden layer output signal, described in for input layer output signal, described α is inertia coefficient, and described η is learning efficiency, described in for parameter preset, described in for the initial value of input layer hidden layer weights, described in for the input layer hidden layer weights of a described upper control cycle;

Described adjustment unit, is configured to the hidden layer output layer weights obtaining a described upper control cycle;

Obtain the hidden layer output layer weights of described current time according to the hidden layer output layer weights of the described upper control cycle of hidden layer output layer weights formula adjustment, described hidden layer output layer weights formula is:

${\mathrm{\Δw}}_{li}^{\left(3\right)}=e\left(k\right)sgn\left(\frac{y\left(k\right)-y(k-1)}{u\left(k\right)-u(k-1)}\right)\frac{\∂u\left(k\right)}{\∂o_l^{\left(3\right)}\left(k\right)}{o}_l^{\left(3\right)}\left(k\right),$

Wherein, described in for the hidden layer output layer weights of described current time, described e (k) is error amount, the actual value of the described object brake parameters that described y (k) is current control period, described y (k-1) is the actual value of the object brake parameters of a described upper control cycle, described in for described output layer incoming signal, described in for output layer output signal, the output valve of the described PID controller that described u (k) is described current control period, described u (k-1) is the output valve of the described PID controller of a described upper control cycle.

10. device according to claim 8 or claim 9, is characterized in that,

Described adjustment unit, be configured to when needing the object brake parameters controlling described vehicle for n-th time, obtain error criterion according to error formula, described error formula is:

$E\left(k\right)=\frac{1}{2}{\left(r\right(k)-y(k\left)\right)}^2,$

Wherein, described E (k) is error criterion, and described r (k) is the setting value of described object brake parameters, the actual value of the described object brake parameters that described y (k) is current control period;

Judge whether described error criterion is less than default error threshold;

When described error criterion is not less than described default error threshold, obtained the input layer hidden layer weights of described current time by the input layer hidden layer weights adjusting a upper control cycle, obtained the hidden layer output layer weights of described current time by the hidden layer output layer weights adjusting a upper control cycle.