CN106882080A - A kind of differential steering system and its adaptive neural network fault tolerant control method - Google Patents

Abstract

The invention discloses a kind of differential steering system and its adaptive neural network fault tolerant control method, the system includes steering wheel angle sensor, steering wheel, steering column, rack and pinion steering gear, four wheels and wheel hub motor, front axle, finished vehicle electronic control unit, batteries, yaw-rate sensor, vehicle speed sensor, rear axle and motor control unit.In the process of moving, finished vehicle electronic control unit Real-time Collection steering wheel angle, yaw velocity and GES, calculate the difference of preferable yaw velocity and actual yaw velocity, wheel hub motor output torque is recalculated by the neural network control device for designing, and this dtc signal is delivered to electric machine controller, and current signal is sent from electric machine controller to four wheel hub motors, complete wheel hub motor and normally controlled with the steering stability under disabled status.The present invention can improve the stability and security when the reliability and running car of differential steering system.

Description

A kind of differential steering system and its adaptive neural network fault tolerant control method

Technical field

The present invention relates to four-wheel steering technical field, more particularly to a kind of differential steering system and its adaptive neural network Fault tolerant control method.

Background technology

For conventional truck, clutch, speed changer, power transmission shaft, differential mechanism or even transfer gear be all it is essential, And these parts not only weight weight, vehicle structure are complicated, while there is also the problem for needing periodic maintenance and fault rate.But wheel Hub motor just solves this problem well.In addition to structure is more simple, use the vehicle of In-wheel motor driving can be with More preferable space availability ratio is obtained, while transmission efficiency.

Because wheel hub motor possesses the characteristic that single wheel independently drives, therefore it is easily achieved forerunner, rear-guard or four Drive drive form.Wheel hub motor can adjust left and right wheelses torque or rotating speed realizes differential steering simultaneously, greatly reduce vehicle Radius of turn, pivot stud can be almost realized under special circumstances.

But wheel hub motor there may be failure conditions, there is problem in reliability.How under wheel hub motor failure conditions, Ensure that the problem of stability of automobile is badly in need of solving.

The content of the invention

The technical problems to be solved by the invention are directed to involved defect in background technology, there is provided a kind of differential turns To system and its adaptive neural network fault tolerant control method.

The present invention uses following technical scheme to solve above-mentioned technical problem：

A kind of differential steering system, including steering wheel angle sensor, steering wheel, steering column, rack and pinion steering Device, first to fourth wheel, first to fourth wheel hub motor, front axle, finished vehicle electronic control unit, batteries, speed sensing Device, yaw-rate sensor, rear axle and motor control unit；

Described steering column one end and steering wheel are fixedly linked, and the other end passes through rack and pinion steering gear and the front axle It is connected；

The steering wheel angle sensor is arranged on steering column, for obtaining steering wheel angle；

The vehicle speed sensor and yaw-rate sensor are arranged on automobile, are respectively used to obtain the speed of automobile And yaw velocity；

First wheel, the second wheel are separately positioned on the two ends of the front axle, the 3rd wheel, the 4th wheel point The two ends of the rear axle are not arranged on；

First to fourth wheel hub motor is correspondingly arranged on first to fourth wheel respectively, for driving first To the 4th wheel；

The batteries are set on automobile, for powering；

The finished vehicle electronic control unit is sensed with steering wheel angle sensor, vehicle speed sensor, yaw velocity respectively Device, electric machine controller, batteries are electrically connected, for according to steering wheel angle sensor, vehicle speed sensor and yaw angle speed The data that degree sensor is measured calculate four torques of wheel hub motor and produce corresponding current signal to pass to the motor control Device processed；

The electric machine controller is electrically connected with four wheel hub motors, batteries respectively, for according to the electric current for receiving Signal controls four wheel hub motor work.

The invention also discloses a kind of adaptive neural network fault tolerant control method based on the differential steering system, including Following steps：

Step 1), calculate the relation of preferable yaw velocity and steering wheel angle；

Step 2), set up the state-space model of differential steering system；

Step 3), the state-space model based on differential steering system sets up its adaptive neural network fault-tolerant control system State-space model, and the adaptive neural network fault-tolerant control system based on differential steering system state-space model, The state for setting up adaptive neural network fault-tolerant control system of the differential steering system in the case of wheel hub motor breaks down is empty Between model；

Step 4), set up the reference model and inversion model of adaptive neural network fault-tolerant control system；

Step 5), reference model, inversion model and preferable yaw angle based on adaptive neural network fault-tolerant control system Speed sets up the Network Compensator of adaptive neural network fault-tolerant control system with the relation of steering wheel angle；

Step 6), based on the Network Compensator of adaptive neural network fault-tolerant control system, set up adaptive neural network The neural network control device of network fault tolerance control system；

Step 7), the self adaptation under being broken down to reference model and wheel hub motor based on neural network control device Error between the output of neutral net fault-tolerant control system carries out self-adaptative adjustment.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 1) described in preferable yaw velocity ω_r ^*With steering wheel angle θ_swRelation is：

In formula：a₀=k_fk_r(a+b)²+(k_rb-k_fa)mu²；b₀=k_fk_r(a+b)u；L is antero posterior axis axle Away from；U is car speed；K_sIt is default yaw velocity adjusting parameter；k_f、k_rRespectively front and back wheel cornering stiffness；A is barycenter To front shaft away from；B is barycenter to rear axle wheelbase；M is complete vehicle quality.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 2) described in the state-space model of differential steering system be：

In formula,

δ_fIt is front wheel angle；β is side slip angle；ω_rIt is yaw velocity；D is half wheelbase；J_sFor equivalent turn of steering wheel Dynamic inertia；G is rack and pinion steering gear gearratio；I is vehicle around z-axis rotary inertia；B_sIt is steering wheel equivalent damping, R is wheel Tire radius；d₂For tire drags square；d₁It is stub lateral shift square；T_swThe torque of steering wheel is acted on for driver；T_fl、T_fr、 T_rl、T_rrBefore respectively left front, right, left back, off hind wheel hub motor output torque.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 3) described in the state-space model of adaptive neural network fault-tolerant control system of differential steering system be：

In formula, f (x (t))=Ax (t)；G (x (t))=λ B；H (x (t))=Cx (t)；T is the time；λ₁、λ₂、λ₃、λ₄Before respectively left front, right, the probability that breaks down of left back, off hind wheel hub motor.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 3) described in adaptive neural network faults-tolerant control of the differential steering system in the case of wheel hub motor breaks down The state-space model of system is：

In formula, σ (x (t), u (t), w (t)) is the disturbance input function under wheel hub motor failure.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 4) described in the reference model of adaptive neural network fault-tolerant control system be：

In formula：x_mT () is the state vector of reference model；u_mT () is the input dominant vector of reference model, y_mT () is ginseng Examine the output vector of model；A_m=A；B_m=λ B；C_m=C.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 4) described in the inversion model of adaptive neural network fault-tolerant control system be：

U (t)=g^-1(t)[v(t)-f(x)]

In formula：V (t) is given tracking response.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 5) described in Network Compensator be：

In formula：Δ is inversion model error；y_sIt is the s layers of output of neutral net；w_isIt is i-th neuron to s layers of god Through the weight of unit；g_iX () is i-th neuron output value；I is the natural number less than or equal to n more than or equal to 1, and n is neuron Number, s is the Current Situation of Neural Network number of plies.

As a kind of further side of optimization of adaptive neural network fault tolerant control method of differential steering system of the invention Case, step 6) described in neural network control device be：

In formula：u_eerT () is the compensation error of inner ring system；K_pIt is parameter matrix；u_NNIt is neural network control device Output.

The present invention uses above technical scheme compared with prior art, with following technique effect：

1., according to electronic control unit Real-time Collection steering wheel angle, yaw velocity and GES, calculate preferable Yaw velocity and the difference of actual yaw velocity, wheel hub electricity is recalculated by the neural network control device for designing Machine output torque, and current signal is sent from ECU to wheel hub motor, complete wheel hub motor normally steady with steering under disabled status Qualitative contrlol；

2. carried control method is easy to be reliable, while effectively overcoming the inverse mould that steering failure causes using neutral net The influence of type error and non-linear factor, so as to realize the real-time model- following control to differential steering model；

3. the method need not be known a priori by position and the size of failure generation, also need not carry out parameter identification to system, just Differential steering system accurate trace model output in case of a fault can be ensured, so as to reach preferable dynamic property.

Brief description of the drawings

Fig. 1 is the structural representation of differential steering system in the present invention；

Fig. 2 is the schematic flow sheet of adaptive neural network fault tolerant control method in the present invention.

In figure, 1- steering wheel angle sensors, 2- steering wheels, 3- steering columns, 4- the near front wheels and wheel hub motor, 5- gears Tooth bar steering gear, 6- off-front wheels and wheel hub motor, 7- front axles, 8- finished vehicle electronic control units, 9- batteries, 10- left rear wheels And wheel hub motor, 11- off hind wheels and wheel hub motor, 12- vehicle speed sensor, 13- yaw-rate sensors, 14- rear axles, 15- Electric machine controller.

Specific embodiment

Technical scheme is described in further detail below in conjunction with the accompanying drawings：

As shown in figure 1, the present invention develops a kind of differential steering system, including steering wheel angle sensor 1, steering wheel 2, Steering column 3, rack and pinion steering gear 5, first to fourth wheel, first to fourth wheel hub motor, front axle 7, finished vehicle electronic Control unit 8, batteries 9, vehicle speed sensor 12, yaw-rate sensor 13, rear axle 14 and motor control unit 15；

The one end of the steering column 3 and steering wheel 2 are fixedly linked, and the other end is by rack and pinion steering gear 5 and described Front axle 7 is connected；

The steering wheel angle sensor 1 is arranged on steering column 3, for obtaining steering wheel angle；

The vehicle speed sensor 12 and yaw-rate sensor 13 are arranged on automobile, are respectively used to obtain automobile Speed and yaw velocity；

First wheel, the second wheel are separately positioned on the two ends of the front axle 7, the 3rd wheel, the 4th wheel It is separately positioned on the two ends of the rear axle 14；

First to fourth wheel hub motor is correspondingly arranged on first to fourth wheel respectively, for driving first To the 4th wheel；

The batteries 9 are arranged on automobile, for powering；

The finished vehicle electronic control unit 8 respectively with steering wheel angle sensor 1, vehicle speed sensor 12, yaw velocity Sensor 13, electric machine controller 15, batteries 9 are electrically connected, for according to steering wheel angle sensor 1, vehicle speed sensor 12 and the data that measure of yaw-rate sensor 13 calculate four torques of wheel hub motor and produce corresponding current signal to pass Pass the electric machine controller 15；

The electric machine controller 15 is electrically connected with four wheel hub motors, battery 9 respectively, for according to the electricity for receiving Stream signal controls four wheel hub motor work.

As shown in Fig. 2 the present invention also disclosed a kind of fault-tolerant control of the adaptive neural network based on the differential steering system Method processed, it is characterised in that comprise the following steps：

Step 1), calculate preferable yaw velocity ω_r ^*With steering wheel angle θ_swRelation：

In formula：a₀=k₁k₂(a+b)²+(k₂b-k₁a)mu²；b₀=k₁k₂(a+b)u；L is antero posterior axis Wheelbase；U is car speed；K_sIt is default yaw velocity adjusting parameter, its scope can like according to driver to be chosen, preferentially It is 0.12-0.37；k₁、k₂Respectively front and back wheel cornering stiffness；A be barycenter to front shaft away from；B is barycenter to rear axle wheelbase；M is Complete vehicle quality.

Step 2), set up differential steering system state-space model：

Differential steering system state-space model is：

In formula,

δ_fIt is front wheel angle；β is side slip angle；ω_rIt is yaw velocity；D is half wheelbase；J_sFor equivalent turn of steering wheel Dynamic inertia；G is rack and pinion steering gear gearratio；I is vehicle around z-axis rotary inertia；B_sIt is steering wheel equivalent damping, R is wheel Tire radius；d₂For tire drags square；d₁It is stub lateral shift square；T_swThe torque of steering wheel is acted on for driver；T_fl、T_fr、 T_rl、T_rrBefore respectively left front, right, left back, off hind wheel hub motor output torque.

Step 3), the state-space model based on differential steering system sets up its adaptive neural network fault-tolerant control system State-space model, and the adaptive neural network fault-tolerant control system based on differential steering system state-space model, The state for setting up adaptive neural network fault-tolerant control system of the differential steering system in the case of wheel hub motor breaks down is empty Between model.

Initially set up the adaptive neural network fault-tolerant control system state-space model of differential steering system：

In formula, f (x (t))=Ax (t)；G (x (t))=λ B；H (x (t))=Cx (t)；T is the time；λ₁、λ₂、λ₃、λ₄Before respectively left front, right, the probability that breaks down of left back, off hind wheel hub motor.

Adaptive neural network fault-tolerant control system state-space model based on above-mentioned differential steering system, when differential system During system failure, the differential steering system state-space model of adaptive neural network fault tolerance is：

In formula, σ (x (t), u (t), w (t)) is the disturbance input function in the case of failure；

Step 4), set up the reference model and inversion model of adaptive neural network fault-tolerant control system；

Initially set up the reference model of adaptive neural network fault-tolerant control system：

In formula：x_mT () is the state vector of reference model；u_mT () is the input dominant vector of reference model, y_mT () is ginseng Examine the output vector of model；A_m=A；B_m=λ B；C_m=C.

Next sets up the inversion model of adaptive neural network fault-tolerant control system：

U (t)=g^-1(t)[v(t)-f(x)]

In formula：V (t) is given tracking response.

Step 5), reference model, inversion model and preferable yaw based on above-mentioned adaptive neural network fault-tolerant control system The relation of angular speed and steering wheel angle, the Network Compensator for setting up adaptive neural network fault-tolerant control system can be stated For：

In formula：Δ is inversion model error；y_sIt is the s layers of output of neutral net；w_isIt is i-th neuron to s layers of god Through the weight of unit；g_iX () is i-th neuron output value；I is the natural number less than or equal to n more than or equal to 1, and n is neuron Number, s is the Current Situation of Neural Network number of plies.

Step 6), based on the Network Compensator of adaptive neural network fault-tolerant control system, set up adaptive neural network The neural network control device of network fault tolerance control system is：

In formula：u_eerT () is the compensation error of inner ring system；K_pIt is parameter matrix；u_NNIt is neural network control device Output.

Step 7), based on adaptive neural network adjuster to the adaptive neural network under reference model and wheel hub motor failure Error between the output of network fault tolerance control system carries out self-adaptative adjustment.

As can be seen that inversion model error delta and inversion model from Fig. 2 neural network control device structure charts of the present invention Input and system output relation be：

The performance indications being defined as follows：

In formula：y_jsIt is j-th neuron output；e_jIt is j-th neuron error.

Compensation error u first with off-line training in the case of various_eerSo that the output of network approaches u_eer, so that complete Into the effect of feedback compensation.

On the basis of off-line learning, data when Real-time Collection differential steering system is run, using online adaptive Algorithm undated parameter is practised, in order to improve the stability of adaptive neural network fault-tolerant control system, is entered for neural network weight Row adjustment, using following online adaptive learning algorithm：

In formula：W () is weights；T, P are positive definite matrix；Q is RBF.

In the process of moving, electronic control unit Real-time Collection steering wheel angle, yaw velocity and GES, meter The difference of preferable yaw velocity and actual yaw velocity is calculated, is recalculated by the neural network control device for designing Wheel hub motor output torque, and from motor control unit to wheel hub motor send current signal, complete wheel hub motor normally with mistake Steering stability control under effect situation, it is achieved thereby that a kind of four rotation with adaptive neural network faults-tolerant control function To system and its control method.

Those skilled in the art of the present technique it is understood that unless otherwise defined, all terms used herein (including skill Art term and scientific terminology) have with art of the present invention in those of ordinary skill general understanding identical meaning.Also It should be understood that those terms defined in such as general dictionary should be understood that with the context of prior art in The consistent meaning of meaning, and unless defined as here, will not be explained with idealization or excessively formal implication.

Above-described specific embodiment, has been carried out further to the purpose of the present invention, technical scheme and beneficial effect Describe in detail, should be understood that and the foregoing is only specific embodiment of the invention, be not limited to this hair Bright, all any modification, equivalent substitution and improvements within the spirit and principles in the present invention, done etc. should be included in the present invention Protection domain within.

Claims (10)

1. a kind of differential steering system, it is characterised in that including steering wheel angle sensor (1), steering wheel (2), steering column (3), rack and pinion steering gear (5), first to fourth wheel, first to fourth wheel hub motor, front axle (7), finished vehicle electronic control Unit (8) processed, batteries (9), vehicle speed sensor (12), yaw-rate sensor (13), rear axle (14) and motor control Unit (15)；

Described steering column (3) one end and steering wheel (2) are fixedly linked, and the other end passes through rack and pinion steering gear (5) and institute Front axle (7) is stated to be connected；

The steering wheel angle sensor (1) is arranged on steering column (3), for obtaining steering wheel angle；

The vehicle speed sensor (12) and yaw-rate sensor (13) are arranged on automobile, are respectively used to obtain automobile Speed and yaw velocity；

First wheel, the second wheel are separately positioned on the two ends of the front axle (7), the 3rd wheel, the 4th wheel point The two ends of the rear axle (14) are not arranged on；

First to fourth wheel hub motor is correspondingly arranged on first to fourth wheel respectively, for driving first to the Four wheels；

The batteries (9) are arranged on automobile, for powering；

The finished vehicle electronic control unit (8) is fast with steering wheel angle sensor (1), vehicle speed sensor (12), yaw angle respectively Degree sensor (13), electric machine controller (15), batteries (9) are electrically connected, for according to steering wheel angle sensor (1), The data that vehicle speed sensor (12) and yaw-rate sensor (13) are measured calculate four torques of wheel hub motor and produce phase The current signal answered passes to the electric machine controller (15)；

The electric machine controller (15) is electrically connected with four wheel hub motors, battery (9) respectively, for according to the electricity for receiving Stream signal controls four wheel hub motor work.

2. the adaptive neural network fault tolerant control method of the differential steering system being based on described in claim 1, it is characterised in that Comprise the following steps：

Step 1), calculate the relation of preferable yaw velocity and steering wheel angle；

Step 2), set up the state-space model of differential steering system；

Step 3), the state-space model based on differential steering system sets up the shape of its adaptive neural network fault-tolerant control system State space model, and the adaptive neural network fault-tolerant control system based on differential steering system state-space model, set up The state space mould of adaptive neural network fault-tolerant control system of the differential steering system in the case of wheel hub motor breaks down Type；

Step 4), set up the reference model and inversion model of adaptive neural network fault-tolerant control system；

Step 5), reference model, inversion model and preferable yaw velocity based on adaptive neural network fault-tolerant control system With the Network Compensator that the relation of steering wheel angle sets up adaptive neural network fault-tolerant control system；

Step 6), based on the Network Compensator of adaptive neural network fault-tolerant control system, set up adaptive neural network The neural network control device of fault-tolerant control system；

Step 7), the adaptive neural network under being broken down to reference model and wheel hub motor based on neural network control device Error between the output of network fault tolerance control system carries out self-adaptative adjustment.

3. the adaptive neural network fault tolerant control method of the differential steering system according to claims 2, its feature exists In step 1) described in preferable yaw velocity ω_r ^*With steering wheel angle θ_swRelation is：

$\frac{{\mathrm{\θ}}_{sw}}{{{\mathrm{\ω}}_r}^{*}}=\frac{a_{0}u}{b_0{K}_s(L+K_u{u}^2)}$

In formula：a₀=k_fk_r(a+b)²+(k_rb-k_fa)mu²；b₀=k_fk_r(a+b)u；L is antero posterior axis wheelbase；u It is car speed；K_sIt is default yaw velocity adjusting parameter；k_f、k_rRespectively front and back wheel cornering stiffness；A is barycenter to preceding Axle wheelbase；B is barycenter to rear axle wheelbase；M is complete vehicle quality.

4. the adaptive neural network fault tolerant control method of the differential steering system according to claims 3, its feature exists In step 2) described in the state-space model of differential steering system be：

$\left\{\begin{array}{c}\stackrel{\·}{x}\left(t\right)=Ax\left(t\right)+Bu\left(t\right)+Ew\left(t\right)\\ y\left(t\right)=Cx\left(t\right)\end{array}\right.$

In formula,

C=[0 00 1]；

U (t)=[T_fl T_fr T_rl T_rr]^T；W (t)=[T_sw]^T；Y (t)=[ω_r]^T；

δ_fIt is front wheel angle；β is side slip angle；ω_rIt is yaw velocity；D is half wheelbase；J_sFor steering wheel Equivalent Rotational is used Amount；G is rack and pinion steering gear gearratio；I is vehicle around z-axis rotary inertia；B_sIt is steering wheel equivalent damping, R is tire half Footpath；d₂For tire drags square；d₁It is stub lateral shift square；T_swThe torque of steering wheel is acted on for driver；T_fl、T_fr、T_rl、T_rr Before respectively left front, right, left back, off hind wheel hub motor output torque.

5. the adaptive neural network fault tolerant control method of the differential steering system according to claims 4, its feature exists In step 3) described in the state-space model of adaptive neural network fault-tolerant control system of differential steering system be：

$\left\{\begin{array}{c}\stackrel{\·}{x}\left(t\right)=f\left(x\right(t\left)\right)+g\left(x\right(t\left)\right)u\left(t\right)+w\left(t\right)\\ y\left(t\right)=h\left(x\right(t\left)\right)\end{array}\right.$

In formula, f (x (t))=Ax (t)；G (x (t))=λ B；H (x (t))=Cx (t)；T is the time； λ₁、λ₂、λ₃、λ₄Before respectively left front, right, the probability that breaks down of left back, off hind wheel hub motor.

6. the adaptive neural network fault tolerant control method of the differential steering system according to claims 5, its feature exists In step 3) described in adaptive neural network faults-tolerant control of the differential steering system in the case of wheel hub motor breaks down The state-space model of system is：

$\left\{\begin{array}{c}\stackrel{\·}{x}\left(t\right)=f\left(x\right(t\left)\right)+g\left(x\right(t\left)\right)u\left(t\right)+\mathrm{\σ}\left(x\right(t),u(t),w(t\left)\right)\\ y\left(t\right)=h\left(x\right(t\left)\right)\end{array}\right.$

In formula, σ (x (t), u (t), w (t)) is the disturbance input function under wheel hub motor failure.

7. the adaptive neural network fault tolerant control method of the differential steering system according to claims 6, its feature exists In step 4) described in the reference model of adaptive neural network fault-tolerant control system be：

$\left\{\begin{array}{c}{\stackrel{\·}{x}}_m\left(t\right)=A_m{x}_m\left(t\right)+B_m{u}_m\left(t\right)\\ y_m\left(t\right)=C_m{x}_m\left(t\right)\end{array}\right.$

In formula：x_mT () is the state vector of reference model；u_mT () is the input dominant vector of reference model, y_mT () is to refer to mould The output vector of type；A_m=A；B_m=λ B；C_m=C.

8. the adaptive neural network fault tolerant control method of the differential steering system according to claims 7, its feature exists In step 4) described in the inversion model of adaptive neural network fault-tolerant control system be：

U (t)=g^-1(t)[v(t)-f(x)]

In formula：V (t) is given tracking response.

9. the adaptive neural network fault tolerant control method of the differential steering system according to claims 8, its feature exists In step 5) described in Network Compensator be：

$\mathrm{\Δ}=y_s=\underset{i=1}{\overset{n}{\Σ}}{w}_{is}{g}_i\left(x\right)$

In formula：Δ is inversion model error；y_sIt is the s layers of output of neutral net；w_isIt is i-th neuron to s layers of neuron Weight；g_iX () is i-th neuron output value；I is the natural number less than or equal to n more than or equal to 1, and n is neuron number, s It is the Current Situation of Neural Network number of plies.

10. the adaptive neural network fault tolerant control method of the differential steering system according to claims 9, its feature exists In step 6) described in neural network control device be：

$\stackrel{\·}{x}\left(t\right)=u\left(t\right)+u_{eer}\left(t\right)=K_p\left(x\right(t)-{\stackrel{\·}{x}}_m(t\left)\right)+x_m\left(t\right)-u_{NN}+u_{eer}$

In formula：u_eerT () is the compensation error of inner ring system；K_pIt is parameter matrix；u_NNIt is defeated for neural network control device Go out.