CN-122008898-A - Fuzzy self-adaptive PI control method for vehicle driving motor

CN122008898ACN 122008898 ACN122008898 ACN 122008898ACN-122008898-A

Abstract

The application relates to the technical field of vehicle control, in particular to a fuzzy self-adaptive PI control method for a vehicle driving motor, which comprises the following steps of collecting and processing accelerator pedal signals to obtain pedal change rate; collecting and processing a motor rotating speed signal, comparing the motor rotating speed signal with a target rotating speed to obtain a rotating speed loop error, taking the pedal change rate and the rotating speed loop error as cooperative input quantities, feeding the cooperative input quantities into a fuzzy logic decision-making device, calculating an on-line adjustment quantity of a proportional coefficient and an integral coefficient of a PI controller in real time based on the cooperative input quantities by the fuzzy logic decision-making device according to a preset rule mapping relation, dynamically refreshing parameters of the PI controller by using the on-line adjustment quantity, and carrying out rotating speed closed-loop control on a driving motor by adopting the PI controller after the parameters are refreshed. The application combines the real-time operation intention of the driver with the system state, and dynamically adjusts the parameters of the PI controller on line so as to comprehensively improve the dynamic response quality and driving experience of the vehicle driving system.

Inventors

WANG ZHIJUN
GUO GAOSHANG
LI FEI
YAO XIN

Assignees

河南嘉晨智能控制股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260213

Claims (10)

1. A fuzzy adaptive PI control method for a vehicle drive motor, comprising the steps of: Collecting and processing accelerator pedal signals to obtain a pedal change rate representing real-time operation intention of a driver; Collecting and processing a motor rotating speed signal, and comparing the motor rotating speed signal with a target rotating speed to obtain a rotating speed loop error representing the tracking deviation of the system; Taking the pedal change rate and the rotating speed ring error as cooperative input quantities, and feeding the cooperative input quantities into a fuzzy logic decision-making device; the fuzzy logic decision-making device calculates the online adjustment quantity of the proportional coefficient and the integral coefficient of the PI controller in real time based on the collaborative input quantity according to a preset rule mapping relation; Dynamically refreshing parameters of the PI controller by using the online adjustment quantity; and carrying out rotating speed closed-loop control on the driving motor by adopting the PI controller after refreshing the parameters.
2. The method of claim 1, wherein the fuzzy logic decider fuzzifies the collaborative input quantity prior to performing the real-time resolution; The fuzzification processing comprises the step of respectively converting the pedal change rate and the accurate value of the rotating speed ring error into a plurality of corresponding fuzzy language variables and membership degrees of the fuzzy language variables by utilizing a preset membership degree function.
3. The method of claim 2, wherein the step of the fuzzy logic decider performing the real-time resolution comprises: inputting a plurality of fuzzy language variables and membership degrees thereof output by the fuzzification processing into the preset rule mapping relation; the rule mapping relation is configured to calculate and output on-line adjustment amounts of the proportional coefficient and the integral coefficient in parallel based on a combination of the fuzzy linguistic variable state corresponding to the pedal change rate and the fuzzy linguistic variable state corresponding to the rotational speed ring error.
4. A method according to claim 3, wherein said rule mapping relationship comprises a first rule sub-library for generating said on-line adjustment to said scaling factor; The rules of the first rule sub-library are configured to trigger generation of the online adjustment amount for performing enhancement adjustment on the proportional coefficient when the fuzzy linguistic variable state obtained by blurring the pedal change rate is determined to represent the sudden acceleration intention and the fuzzy linguistic variable state obtained by blurring the rotational speed ring error is determined to represent the rotational speed tracking serious hysteresis.
5. The method of claim 3, wherein the rule mapping relationship further comprises a second rule sub-library for generating the online adjustment to the integral coefficient; The rules of the second rule sub-library are configured to trigger generation of the online adjustment amount for steady-state optimization adjustment of the integral coefficient when the fuzzy linguistic variable state obtained by blurring the pedal change rate is determined to represent acceleration intention and the fuzzy linguistic variable state obtained by blurring the rotational speed ring error is determined to represent system tracking well.
6. A method according to claim 3, wherein the fuzzy logic decider further performs a defuzzification process after the parallel solutions are completed; The defuzzification processing is configured to execute weighted fusion calculation on at least one fuzzy adjustment instruction corresponding to each of the proportional coefficient and the integral coefficient according to the corresponding membership degree to generate each quantization adjustment amount; and combining the quantization adjustment quantity with a preset PI controller parameter basic value to form a parameter update value which is finally used for dynamically refreshing the PI controller parameter.
7. The method of claim 6, wherein the step of combining the quantization adjustment with a preset PI controller parameter base value to form the parameter update value uses different calculation methods according to the sign or magnitude of the quantization adjustment: When the absolute value of the quantization adjustment quantity is smaller than a preset adjustment amplitude threshold value, combining the quantization adjustment quantity with a corresponding parameter base value in a linear superposition mode; When the absolute value of the quantization adjustment quantity is larger than or equal to the adjustment amplitude threshold value, multiplying the quantization adjustment quantity by a preset weight coefficient, and then performing weighted linear combination with a corresponding parameter base value to form the parameter updating value.
8. The method of claim 1, wherein the dynamically refreshing parameters of the PI controller using the online adjustment amount comprises; Converting the online adjustment quantity into a scaling factor and an integral scaling factor according to a preset scaling mapping relation; The integral coefficient is updated by multiplying the scaling factor by a preset scaling coefficient base value and multiplying the integral scaling factor by a preset integral coefficient base value.
9. The method of claim 1, further comprising a compound control mode; in the compound control mode, the proportional coefficient and the integral coefficient of the PI controller are set and maintained to be fixed values; The fuzzy logic decision device calculates the online adjustment quantity in real time according to the collaborative input quantity and takes the online adjustment quantity as a feedforward compensation signal; the PI controller calculates a feedback control signal based on the rotational speed loop error; The feedforward compensation signal is superimposed with the feedback control signal to generate a final control command applied to the drive motor.
10. The method of claim 1, wherein the fuzzy logic decider and the PI controller are configured to operate cooperatively at different operating frequencies; The fuzzy logic decision maker continuously performs the real-time resolving and the dynamic refreshing steps at a first control frequency; the PI controller executes the rotating speed closed-loop control by using the latest refreshed parameters at a second control frequency lower than the first control frequency; The output of the fuzzy logic decision maker completes at least one parameter update between two adjacent control periods of the PI controller.

Description

Fuzzy self-adaptive PI control method for vehicle driving motor Technical Field The application relates to the technical field of vehicle control, in particular to a fuzzy self-adaptive PI control method for a vehicle driving motor. Background With the popularization of electric vehicles, a driving motor control technology has become a core for improving the driving quality and energy efficiency of vehicles. The proportional-integral (PI) controller is widely used for closed-loop regulation of motor rotation speed due to simple structure and strong robustness. However, the actual running conditions of the vehicle are complex and changeable, and from a congested road section frequently started and stopped to a overtaking scene needing quick response, the severe requirements are set for the dynamic adaptability of the control system. The traditional PI controller with fixed parameters is difficult to give consideration to quick response and stability and comfort under all working conditions, so how to enable a control system to actively adapt to different driving intentions and real-time working conditions becomes a key technical challenge for improving the driving experience of an electric automobile. In order to solve the limitation of fixed parameter PI control, the prior art explores a method for combining an intelligent algorithm with traditional control. For example, patent document CN110834544B discloses a fuzzy-PI compound control system for regenerative braking of a pure electric vehicle. The key point of the scheme is that the fuzzy controller and the PI controller are switched according to the absolute value of the deviation between the target current and the actual output current, wherein the fuzzy controller is started to pursue rapid adjustment when the deviation is large, and the PI controller is switched to maintain stability when the deviation is small. Another idea is presented in patent document CN117424503a, which proposes a fuzzy adaptive sliding mode control scheme for brushless dc motors. According to the scheme, a fuzzy reasoning mechanism is introduced into a speed loop, parameters of a PID controller are dynamically adjusted on line according to two system state quantities, namely a rotating speed error and a change rate of the rotating speed error, and the aim of improving tracking performance and disturbance rejection capability of the rotating speed of a motor is achieved. Although the above technique enhances the adaptability of the system by introducing fuzzy logic, the control decision is completely dependent on hysteresis state information of "error generated between the actual output of the system and the target value", and the "dynamic characteristic of the operation intention of the driver" cannot be taken as a feedforward or parallel decision basis. This results in parameter tuning or mode switching of the system being essentially a post-remediation passive response rather than a pre-matched active adaptation. In particular, in a vehicle acceleration scenario, the degree of intensity of the driver's acceleration intention is first and directly reflected in the rate at which the accelerator pedal is depressed, i.e. the pedal change rate, which is a dynamic signal rich in predictive information that produces a significant tracking error prior to the motor speed. However, the prior art solutions have to wait for a sufficiently large deviation between the motor speed and the target value before triggering the corresponding control action. This mechanism introduces unavoidable perceived and response delays at the instants of severe driving intent changes. This defect may cause a perceived degradation of experience in typical driving scenarios. Taking the expressway overtaking as an example, drivers often step on the accelerator pedal to express a strong immediate power demand for safety and efficiency. At this time, an ideal drive system should be able to establish peak torque in a very short time, achieving rapid acceleration. With prior art systems, however, at the initial instant of sudden pedal depression, the control parameters may remain at a moderate level suitable for cruise due to the rotational speed error not yet being sufficiently established. The system needs to go through a short waiting error accumulation before passively switching or adjusting to aggressive control modes, which causes a hysteresis window in the power output. On the contrary, when the driver slowly steps on the pedal during the urban slow driving and the vehicle is driven, the driver expects to obtain smooth and linear starting, and at this time, if the system is adjusted only according to a tiny rotating speed error, the dynamic characteristic of the controller may not be optimized finely, or unnecessary parameter jumping is caused by tiny fluctuation of the error, so that the vehicle generates a channeling feeling, and the comfort is affected. Therefore, the core problem of the prior art is tha