CN106527119B - Derivative-precedence PID system based on fuzzy control - Google Patents

Abstract

The invention relates to a differential advance PID control system based on fuzzy control, comprising a differential advance link, a fuzzy controller, a proportional link, an integral link and a feedforward controller; the fuzzy controller takes speed deviation and the rate of change of the speed deviation as inputs , the output control quantity is the proportional coefficient increment of the proportional link; the differential leading link is set on the feedback loop, and the speed measurement value and its change speed value are input into the proportional link and the integral link as the measurement value; the The output of the proportional link and the integral link is used as the control signal of the PID control system; the feedforward controller only controls for a period of time when the speed setting value changes, that is, suppresses the output, and then promotes the change of the output. The invention can effectively overcome the overshoot, improve the dynamic performance and enhance the anti-disturbance capability and robustness of the system.

Description

A Differential Look-ahead PID Control System Based on Fuzzy Control

technical field

The invention relates to the technical field of automatic control, in particular to a differential advance PID control system based on fuzzy control.

Background technique

Smart car refers to a kind of competition car body with automatic driving, automatic transmission, and even automatic road recognition function controlled by its own program. In the self-designed smart car, speed control is the core of the entire smart car system control. In order to complete the race faster and more stably on the specified track, the smart car should decelerate in advance when it encounters a curve to prevent it from rushing out of the track, and should accelerate when it exits the curve and enters the straight. This variable route results in frequent changes in speed setpoints. The differential link in the traditional PID algorithm is to differentiate the set value and the output value at the same time. When it is applied to the system, it will bring different degrees of oscillation and cause the intelligent vehicle system to lose control. Even if the traditional PID can make the system reach the speed setpoint smoothly, the response of the system will be slow and the dynamic performance will be poor. In addition, many uncertain factors such as external disturbance and motor parameters also cause some problems in traditional PID control.

In recent years, the motor control technology of smart cars has developed rapidly, and more and more control algorithms have been applied in this field. For example, in order to improve the anti-interference ability of the smart car, the fuzzy PID controller that adjusts the control amount of the system is introduced; in order to reduce the overshoot of the system, the differential link in the PID of the smart car motor is modified, and the PID controller of some differential links is introduced. In terms of control methods, the above methods have improved the control indicators more or less, but there are still some defects.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a differential advance PID control system based on fuzzy control, which can effectively overcome overshoot, improve dynamic performance and enhance the anti-disturbance capability and robustness of the system.

The technical scheme adopted by the present invention to solve the technical problem is: to provide a differential advance PID control system based on fuzzy control, including a differential advance link, a fuzzy controller, a proportional link, an integral link and a feedforward controller; the fuzzy control The controller takes the speed deviation and the rate of change of the speed deviation as the input, and the output control amount is the proportional coefficient increment of the proportional link; the differential leading link is set on the feedback loop, and the speed measurement value and its change speed value are used as the measurement value. Input into the proportional link and the integral link; the output of the proportional link and the integral link is used as the control signal of the PID control system; the feedforward controller only performs control for a period of time when the speed setting value changes, that is, the output Quantity is suppressed, and then the change of output quantity is promoted.

The differential look-ahead link only differentiates the output measurement value.

The fuzzy controller is a two-input single-output fuzzy controller, which takes the speed deviation and the rate of change of the speed deviation as inputs, wherein the speed deviation has 8 levels and the rate of change of the speed variation has 7 levels, using the center of gravity method. The increment of the deblurred output scale factor.

The control action of the feedforward controller is determined by the current deviation level, the current action time and the proportional coefficient of the output of the feedforward controller.

beneficial effect

Compared with the prior art, the present invention has the following advantages and positive effects due to the adoption of the above-mentioned technical scheme: the combination of the fuzzy controller of the present invention and the differential advance link avoids the online adjustment of the three parameters by the traditional fuzzy controller. The complexity and uncertainty of the time, but also bring excellent control effect. The use of fuzzy controller can improve the anti-interference ability and robustness of the system. The fuzzy control table of the fuzzy controller can be calculated in advance and directly stored in the memory of the single-chip microcomputer. When using, it can be directly retrieved from the memory and queried. , greatly reducing the computational complexity of the microcontroller. The combination of the feedforward controller and the fuzzy controller greatly exerts the adjustment effect of the fuzzy controller when the speed setting value changes, and further optimizes the dynamic performance of the system. The feedforward controller only plays a role when the speed setting value changes. So as to ensure the stability of the system.

Description of drawings

Fig. 1 is the control block diagram of the present invention;

Fig. 2 is a block diagram of the differential look-ahead link algorithm;

Fig. 3 is the block diagram of fuzzy controller;

Fig. 4 is the membership function diagram of velocity deviation;

Fig. 5 is the membership function diagram of velocity deviation rate of change;

Fig. 6 is the membership function diagram of proportional coefficient increment;

Fig. 7 is a fuzzy output surface graph;

Fig. 8 is the response comparison curve diagram of the PID control system of the present invention and the conventional PID control system;

FIG. 9 is a response comparison curve diagram of the PID control system of the present invention and the fuzzy differential PID control method without a feedforward controller.

Detailed ways

The present invention will be further described below in conjunction with specific embodiments. It should be understood that these examples are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of the present application.

Embodiments of the present invention relate to a differential-look-ahead PID control system based on fuzzy control, as shown in FIG. 1 , including a differential-look-ahead link, a fuzzy controller, a proportional link, an integral link, and a feedforward controller; the fuzzy controller is composed of a The speed deviation and the rate of change of the speed deviation are used as input, and the output control quantity is the proportional coefficient increment of the proportional link; the differential leading link is set on the feedback loop, and the speed measurement value and its change speed value are input to the measurement value as the measurement value. In the proportional link and the integral link; the output of the proportional link and the integral link is used as the control signal of the PID control system; the feedforward controller only performs control for a period of time when the speed setting value changes, that is, the output is controlled. Inhibit, and then promote the change in output.

The present invention introduces the differential leading link into the PID control system; adds a fuzzy controller before the proportional link to adjust the proportional coefficient in real time; and sets the "first suppression first then upward" type feedforward control compensation at the output end of the controller.

The essence of the differential first link is to perform differential operation on the output quantity in advance, and reflect the changing state of the output quantity into the deviation. The algorithm frame is shown in Figure 2. The model analysis of the differential first link is as follows: As can be seen from Figure 2, E(S)=R(S) _-UD ( _S ), where R(S) is the input of the PID control system, and UD(S) is the differential The output of the leading link, E(S) is the error between the input of the PID control system and the output of the differential leading link; the transfer function of the differential leading link is: In the formula, H(S) is the input of the differential advance link, γ represents the time coefficient, γ<1, T _D is the differential time constant, S represents the Laplace transform factor, and 1/(γT _D S+1) is a first-order The inertia link is equivalent to a low-pass filter. The above transfer function is converted to time domain form as: Among them, u _D (t) represents the output of the differential preceding link at time k, and h(t) is the output measurement value of the controlled object. For microcontroller processing, convert to differential form: Among them, u _D (k) represents the output of the differential leading link at time k, h(k) represents the output measurement value of the controlled object (the measurement value of the encoder), and T is a main program cycle of the microcontroller. The above formula has been sorted out. have to: It is not difficult to find that the introduced differential leading link only differentiates the output measured value h(t), and does not differentiate the speed set value.

The beneficial effect of the differential leading link is: the output signal of the differential link includes the measured value h(t) and its change speed value h'(t), when they are input into the proportional link and the integral link as the measured value, it can effectively suppress the The overshoot of the motor speed can overcome the adverse effect of the changing speed setting value on the system, and has a better control effect.

The output of the differential link includes the output of the previous moment, and this quantity will be attenuated according to a certain proportion, which will bring a certain hysteresis effect to the system control, which will ease the change of the controlled quantity. And in actual operation, the control process mechanism of the motor is complex, and has the characteristics of nonlinearity, time-varying uncertainty and lag. Under the influence of factors such as noise and load disturbance, the process parameters and even the model structure of the motor will change with time and working environment. This requires that the parameters of the PID can be adjusted online to meet the requirements of real-time control.

The present invention combines the feedforward control and adds a fuzzy controller before the proportional integral link. When there is external disturbance or the set value changes, the system can adjust the proportional parameters of the PID optimally according to different situations, so as to make up for the lag effect of the leading differential link and strengthen the robustness and anti-interference ability of the system.

The fuzzy controller is designed as follows:

In the PID control system, if the proportional coefficient K _p is increased, the error coefficient of the system will increase, thereby reducing the steady-state error; on the other hand, the increase of the proportional coefficient will cause the phase margin of the system to decrease, and the system will appear Large overshoot, becomes unstable. Therefore, in the initial stage of control, a larger proportional coefficient K _p can be set to speed up the response speed and reduce the rise time; in the middle stage of control, appropriately reduce the proportional coefficient K _p to reduce overshoot; in the later stage of control, appropriately increase Large proportional coefficient K _p to improve control accuracy. The integral link is used to eliminate the static error, but excessive integral action will cause the system to oscillate. The differential link is to improve the dynamic performance of the system.

The fuzzy controller mainly optimizes the control of the system in terms of speed, and the differential link has been set in the feedback network, so the proportional coefficient K _p is used as the adjustment object of the fuzzy gain, and the integral coefficient K _i is set to a moderate fixed value.

The fuzzy controller of the present invention is a fuzzy controller with two inputs and one output. Taking the deviation e and the deviation change rate ec as the input of the controller, the increment of the proportional coefficient is calculated by using fuzzy rules and fuzzy reasoning to output ΔK _p . The process of the fuzzy self-tuning is to find out the fuzzy relationship between ΔK _p and the deviation e and the deviation change rate ec. During the running process, the fuzzy controller regularly detects the deviation e and the deviation change rate ec, and performs on-line modification of ΔK _p after fuzzy processing, fuzzy reasoning, clustering output and de-fuzzification to meet different deviation e and deviation changes. The rate ec has different requirements for control parameters, so that the controlled object can achieve good dynamic and static performance.

The design of the feedforward controller is as follows:

During the driving process of the smart car, the fuzzy controller only plays a major role in the transition between the curve and the straight road, and the fuzzy controller can only output a large proportional coefficient increment at the beginning of the transition. , with the decrease of the deviation e and the deviation change rate ec, the increment ΔK _p of the proportional coefficient has begun to slowly decrease or even less than zero. In a sense, this is not taking advantage of the maximum tuning effect of the fuzzy controller. It is generally expected that the fuzzy controller can output a larger increment of the proportional coefficient at the beginning of the transition, and then rely on a larger attenuation of the proportional coefficient to balance the overshoot in the later stage, so as to achieve better dynamic performance. Therefore, some compensation method needs to be adopted to optimize the fuzzy controller in the transition process.

The feedforward controller of the present invention requires the single-chip microcomputer to set some flag bits after detecting the type of the runway according to the road deviation, and use these flag bits to perform software compensation for a specific period of time in the PID control.

The control method of the feed-forward controller is the type of "first suppression and then raising", that is, the output of the PID controller is appropriately "reduced", and then it is appropriately "increased". Here, "decrease" means to appropriately suppress the change of the output of the PID control system, and "increase" means to appropriately promote the change of the output of the PID control system. The "minus" effect can make the fuzzy controller output a longer ΔK _p through the feedback network. In order to prevent overshoot, the "minus" effect should gradually become smaller until it becomes a "plus" effect, so as to obtain a relative effect in the middle and later stages. A stronger effect of attenuating K _p . When the compensation is over, the fuzzy controller has the characteristics of strong continuity in the later stage of the transformation process, thus ensuring the smooth adjustment in the later stage.

The present invention will be further described below through specific smart car embodiments.

In this embodiment, the fuzzy controller can be divided into 6 parts.

1. Determine the level division of the input and output of the fuzzy controller. As shown in FIG. 3 , the control algorithm sets the deviation e of the speed value, the deviation change rate ec and the proportional coefficient increment ΔK _p as 8, 7 and 7 levels respectively. That is, the level of the fuzzy set of deviation e is {-3, -2, -1, -0, +0, 1, 2, 3}, and the language expression is {negative large (NB), negative medium (NM), negative Small (NS), Negative Zero (NZ), Positive Zero (PZ), Positive Small (PS), Positive Medium (PM), Positive Large (PB)}; the rank of the fuzzy set for the deviation change rate ec and the proportional coefficient increment ΔK _p All are {-3, -2, -1, 0, 1, 2, 3}, and the language expressions are {negative large (NB), negative medium (NM), negative small (NS), zero (ZO), positive Small (PS), Medium (PM), Chia (PB)}.

2. Compile fuzzy rules. The smart car system can restore the speed set value in time when it is disturbed, and it is necessary to increase the control amount of the motor; when it recovers or is about to reach the set value, it is necessary to reduce the control amount of the motor as much as possible, so that the system can smoothly reach the set value. constant value without speed oscillation. For this purpose, the fuzzy control rule table shown in Table 1 is designed.

Table 1 Fuzzy control rule table

Part of the rule explanation of the fuzzy control rule table: when E is NB and EC is PB, the speed deviation is negative, and the rate of change of the speed deviation is positive, indicating that although the deviation is large now, the deviation is at a fast speed. is shrinking, so the proportional coefficient increment at this time is set to ZO, and the proportional coefficient basically remains unchanged; when E is PM and EC is ZO, the speed deviation is in the middle, and the change rate of the speed deviation at this time is ZO, indicating that this The time deviation is relatively large, but the deviation change speed is basically unchanged. At this time, the proportional coefficient (PS) can be slightly increased to make the deviation smaller at a faster speed. The interpretations of other rules are basically similar and will not be repeated here.

3. Determine the fuzzy universe and scale factor of each input and output variable. If the actual domain of the speed deviation value is determined by the actual maximum value _{em of the deviation e, it is [-e m ,+e m ] .} The fuzzy domain of the specified deviation e is [-3,3], then the scale factor of the deviation e is In the same way, the fuzzy domain of the specified deviation change rate ec is [-3, 3], then the scale factor of the deviation change rate ec is determined by its actual maximum value ec _m as The _{em and ec m can} be obtained by returning the speed deviation and the rate of change of the speed deviation after the complete run to the host computer through the wireless communication between the smart car and the host computer. Similarly, the fuzzy domain of the specified proportional coefficient increment ΔK _p is [-3, 3], and the determination of the proportional factor can be performed by MATLAB simulation using the mathematical transfer function of the motor to obtain the PID under different proportional coefficients. The simulation curve is used to determine an approximate value, which is then determined by the host computer and actual adjustment.

4. Determine the membership function of each input and output quantity under different fuzzy levels. The membership functions that set the input and output quantities have small intersections in the PB, PM, NB, and NM intervals, and are slightly denser in other intervals. Fig. 4 is the membership function diagram of the speed deviation; Fig. 5 is the membership function diagram of the rate of change of the speed deviation; Fig. 6 is the membership function diagram of the proportional coefficient increment.

The membership functions of the variables are shown in Table 2.

Table 2 Membership function table of each variable

5. Defuzzification. Directly use the Fuzzy toolbox of MATLAB to import the characteristics, rules, membership functions, etc. of input variables and output variables into the Fuzzy toolbox, and compile them to obtain a set of output quantities corresponding to different input quantities. In practice, the shape and interval of the membership function can be modified according to the fuzzy rule observer and the output surface observer to obtain a satisfactory control effect. Figure 7 is a blurred output surface graph.

6. Discretization of input quantities. The purpose of discretizing the input is to discretize the output. For example, for a discrete interval A of the deviation and a discrete interval B of the deviation rate of change, any two input quantities that fall within the interval A and B can be de-fuzzified with a specific output value. In this way, a fuzzy control table can be obtained by using the "readfis" and "evalfis" functions of MATLAB, and the table can be directly looked up in actual operation, thereby reducing the operation of the single-chip microcomputer, which is beneficial to the processing of the single-chip microcomputer. In order to reduce the memory burden of the microcontroller and obtain a more stable dynamic performance, the interval between the deviation and the deviation change rate near 0 can be divided into a smaller value, and a larger value can be divided into a larger value. The fuzzy control table is shown in Table 3.

Table 3 Fuzzy Control Table

The discrete interval of speed deviation e is [-3,2.5], [-2.5,-2], [-2,-1.5], [-1.5,-1], [-1,-0.75], [-0.75, -0.5],[-0.25,-0.125],[-0.125,0.125],[0.125,0.25],[0.25,0.5],[0.5,0.75],[0.75,1],[1,1.5],[ 1.5,2], [2,2.5], [2.5,3]. The corresponding table lookup values for each interval are -2.75, -2.25, -1.75, -1.25, -1, -0.75, -0.5, -0.25, 0, 0.25, 0.5, 0.75, 1, 1.25, 1.75, 2.25, 2.75.

The discrete interval of the speed deviation rate of change ec is [-3, 2.5], [-2.5, -2], [-2, -1.5], [-1.5, -1], [-1, -0.75] ,[-0.75,-0.5],[-0.25,-0.125],[-0.125,0.125],[0.125,0.25],[0.25,0.5],[0.5,0.75],[0.75,1],[1 , 1.5], [1.5, 2], [2, 2.5], [2.5, 3]. The corresponding table lookup values for each interval are -2.75, -2.25, -1.75, -1.25, -1, -0.75, -0.5, -0.25, 0, 0.25, 0.5, 0.75, 1, 1.25, 1.75, 2.25, 2.75.

In the process of self-adaptive adjustment of the proportional coefficient, the single-chip microcomputer matches the detected deviation and the deviation change rate to the quantized interval and looks up the table, thereby obtaining the corresponding proportional coefficient increment.

In this embodiment, the feedforward controller can be divided into three parts.

1. Design of parameters related to feedforward compensation. When the single-chip microcomputer judges the sign bit of the transition and the sign bit of the transition direction according to the change of the road deviation, it initializes the values of four variables, which are respectively km, k, _s , and k ₁ . Among them, k _m is the number of the entire compensation action; k is the time mark of the current compensation action, which is used to indicate the number of times the compensation action has been performed; s is the change in the nature of the control compensation action (from "subtraction" action to "addition" action) The threshold of ; k ₁ is used to adjust the size of the compensation output. In compensation control, in order to enhance the robustness of the compensation effect, the single-chip microcomputer needs to control the compensation effect according to the level of the deviation. Therefore, the deviation level parameter nn is also indispensable, and this parameter can be determined according to the level of the quantization interval in which the deviation e in the fuzzy control table is located.

2. Design of "minus" and "plus" effects. When doing control, the compensation effect needs to be controlled by multiple variables. The expression of the overall control variable is: U _dd =K ₁ ·nn·(ks), where k ₁ is the proportional coefficient; k is at the beginning of the compensation effect. 0, the value increases by one after each compensation action ends, until it is greater than km, the compensation action ends; _s is an empirical parameter between (0, _km ), which controls the moment when the nature of the compensation action changes. The change process of U _dd can be seen intuitively from the above formula: the compensation effect is the largest at the initial time, and as the deviation decreases, the overall compensation effect decreases; the nature of the compensation effect changes at time s.

3. Parameter setting. k ₁ can be determined according to the PID output value returned by the smart car to the upper computer, while other parameters need to be continuously debugged to obtain the optimal control effect.

As shown in Figure 8 and Figure 9, the speed response curve of the PID control system of the present invention is the solid line part, the speed response curve of the conventional PID control system is the dotted line part, and the speed response of the fuzzy differential leading PID is not provided with the feedforward controller. The curve is the dot-dash line part.

It can be seen from Figure 8 that when the speed setting value is 2m/s and the disturbance of 100ms is set at 1200ms, the beneficial effects of the present invention relative to the conventional PID are:

1. The overshoot of the PID is significantly reduced, the adjustment time is less, and the dynamic performance is better.

2. The anti-interference ability of PID is stronger.

It can be seen from FIG. 9 that when the speed setting value is 2m/s, the beneficial effects of the present invention relative to the fuzzy differential advance PID without the feedforward controller are: the adjustment time of the PID is shorter and the dynamic performance is good.

After actual testing, the PID control system of the present invention not only enables the intelligent vehicle system to have good dynamic performance and response speed, but also enhances the anti-interference ability of the system, so that the intelligent vehicle can smoothly adjust the speed of the running type. achieve the optimal path. Experiments show that: the control system proposed by the present invention effectively improves the performance of the smart car. When running on the same runway, the speed of the smart car using the present invention is higher than that of the smart car using the feedback control of the traditional PID control system. The time to run a lap is reduced by an average of 2.72 seconds, and the algorithm also improves the stability of the smart car's operation and improves the smart car's adaptability to the runway.

Claims (4)

1. a differential leading PID control system based on fuzzy control, is characterized in that, comprises differential leading link, fuzzy controller, proportional link, integral link and feedforward controller; Described fuzzy controller changes with speed deviation and speed deviation The control value of the output is the proportional coefficient increment of the proportional link; the differential leading link is set on the feedback loop, and the speed measurement value and its change speed value are input into the proportional link and the integral link as the measurement value. ; The output of the proportional link and the integral link is used as the control signal of the PID control system; the feedforward controller only controls the output for a period of time when the speed setting value changes, that is, suppresses the output, and then promotes the output The change. 2 . The differential look-ahead PID control system based on fuzzy control according to claim 1 , wherein the differential look-ahead link only differentiates the output measurement value. 3 . 3. the differential advance PID control system based on fuzzy control according to claim 1, is characterized in that, described fuzzy controller is the fuzzy controller of two inputs and single output, with speed deviation and speed deviation rate of change as input, wherein , there are 8 levels of speed deviation and 7 levels of rate of change of speed variation, and the increment of proportional coefficient of output is solved by gravity center method. 4 . The differential advance PID control system based on fuzzy control according to claim 1 , wherein the control action of the feedforward controller is determined by the current deviation level, the current action time and the output of the feedforward controller. 5 . is determined by the proportionality factor.