CN114900085B - Robot joint servo motor model prediction parameter optimization method and device - Google Patents

Abstract

The invention discloses a robot joint servo motor model prediction parameter optimization method and device, and relates to the technical field of robot joint control. Including: modeling the permanent magnet synchronous motor to obtain the permanent magnet synchronous motor mathematical model that meets the robot joint predictive control standard; according to the permanent magnet synchronous motor mathematical model, establish a model predictive controller for the robot joint permanent magnet synchronous motor; The controller obtains the optimal voltage vector of the robot joint motor to be predicted. The invention can optimize the parameter configuration of the value function in the model predictive control method on-line, solve the problem that the parameters of the value function in the model predictive control technology are difficult to configure, and further optimize the control performance of the system.

Description

A robot joint servo motor model prediction parameter optimization method and device

technical field

The invention relates to the technical field of robot joint control, in particular to a method and device for optimizing parameters of a model prediction parameter of a servo motor of a robot joint.

Background technique

Jointed robots, also known as jointed arm robots or jointed mechanical arms, are one of the most common forms of industrial robots in today's industrial fields, and are suitable for mechanical automation operations in many industrial fields. For example, for automatic assembly, painting, handling, welding and other tasks, joint robots are driven by motors, and high-precision permanent magnet synchronous motors are used to achieve high-precision control of robot joints. PMSM (permanent magnet synchronous motor, permanent magnet synchronous motor) has the advantages of small size, small inertia, fast response, and high efficiency. Traditional control methods often require staff to repeatedly debug controller parameters, and the parameter adjustment is often not accurate enough to achieve satisfactory results.

Contents of the invention

The invention proposes the invention aiming at the problem that the weight coefficients are difficult to configure accurately in the traditional model predictive control.

In order to solve the above technical problems, the present invention provides the following technical solutions:

On the one hand, the present invention provides a kind of robot joint servo motor model prediction parameter optimization method, and this method is realized by electronic equipment, and this method comprises:

S1. The permanent magnet synchronous motor is modeled, and the mathematical model of the permanent magnet synchronous motor that meets the robot joint predictive control standard is obtained.

S2. According to the mathematical model of the permanent magnet synchronous motor, a model predictive controller for the permanent magnet synchronous motor of the robot joint is established.

S3. Obtain the optimal voltage vector of the robot joint motor to be predicted according to the model predictive controller.

Optionally, the permanent magnet synchronous motor mathematical model includes permanent magnet synchronous motor d-axis and q-axis voltage and current equations, permanent magnet synchronous motor d-axis and q-axis voltage transformation equations in the rotating coordinate system, flux linkage equations, and permanent magnet synchronous motor electromagnetic torque equation.

Optionally, the model predictive controller includes a determination unit, a current predictive control unit and a parameter optimization unit.

According to the model predictive controller in S3, the optimal voltage vector of the robot joint motor to be predicted includes:

S31. Collect the speed feedback data and current feedback data of the permanent magnet synchronous motor through the determination unit, and judge the motion state of the joint motor of the robot; wherein, the motion state includes a dynamic adjustment phase and a stable adjustment phase.

S32. Obtain a value function according to the current prediction control unit; wherein, the value function includes a dynamic adjustment stage value function and a stable adjustment stage value function.

S33. Perform parameter optimization on the value function by the parameter optimization unit to obtain an optimal voltage vector of the robot joint motor to be predicted.

Optionally, the passing determination unit in S31 collects the speed feedback data and current feedback data of the permanent magnet synchronous motor, and judging the motion state of the robot joint motor includes:

The speed feedback data and current feedback data of the permanent magnet synchronous motor with set conditions are collected through the determination unit; wherein the set conditions include a set cycle period and a set sampling time.

According to the speed feedback data and the current feedback data, it is determined that the motion state of the robot joint motor is in the dynamic adjustment stage or the stable adjustment stage.

Optionally, according to the current prediction control unit in S32, the value function obtained includes:

Construct current prediction mathematical model, electromagnetic torque prediction mathematical model and inverter switching frequency prediction model.

According to the current prediction mathematical model, the electromagnetic torque prediction mathematical model and the inverter switching frequency prediction model, the value function is obtained.

Optionally, the value function is shown in the following formula (1):

g _m ＝λ[|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-βT _e ^k+1 +γS _w ^k+1 (1)

Among them, λ represents the weight coefficient of the current error; id ^k+1 and i _q ^k+1 are the _d and q axis currents at time k+1 after compensation prediction; _id ^k and i _q ^k are the currents after compensation prediction d and q axis currents at time k+1 and time k; β represents the weight coefficient of electromagnetic torque; T _e ^{k +1} represents the predicted value of electromagnetic torque at time k+1; γ represents the weight coefficient of inverter switching frequency; S _w ^k+1 represents the number of switching frequency changes of the inverter at time k+1.

Optionally, optimizing the parameters of the value function through the parameter optimization unit in S33 includes:

The parameter optimization of the weight coefficient of the value function in the dynamic adjustment stage is performed online through the optimization algorithm of the parameter optimization unit.

And through the optimization algorithm of the parameter optimization unit, the parameters of the weight coefficients of the value function in the stable adjustment stage are optimized online.

Optionally, online parameter optimization of the weight coefficient of the value function in the dynamic adjustment stage through the optimization algorithm of the parameter optimization unit includes:

Determine whether the rise time of the dynamic adjustment stage exceeds the preset threshold, and if so, reduce the weight coefficient of the current error in the value function of the dynamic adjustment stage, and increase the weight coefficient of the electromagnetic torque in the value function of the dynamic adjustment stage.

Optionally, online parameter optimization of the weight coefficient of the value function in the stable adjustment stage through the optimization algorithm of the parameter optimization unit includes:

Determine whether the static error in the stable adjustment stage exceeds the preset threshold, and if so, increase the weight coefficient of the current error in the value function of the stable adjustment stage, and decrease the weight coefficient of the electromagnetic torque in the value function of the stable adjustment stage.

In another aspect, the present invention provides a robot joint servo motor model prediction parameter optimization device, which is applied to realize a robot joint servo motor model prediction parameter optimization method, and the device includes:

The mathematical model building block is used to model the permanent magnet synchronous motor to obtain a permanent magnet synchronous motor mathematical model that meets the robot joint predictive control standard.

The controller building block is used to establish a model predictive controller of the permanent magnet synchronous motor of the robot joint according to the mathematical model of the permanent magnet synchronous motor.

The output module is used to predict the controller according to the model to obtain the optimal voltage vector of the robot joint motor to be predicted.

Optionally, the permanent magnet synchronous motor mathematical model includes permanent magnet synchronous motor d-axis and q-axis voltage and current equations, permanent magnet synchronous motor d-axis and q-axis voltage transformation equations in the rotating coordinate system, flux linkage equations, and permanent magnet synchronous motor electromagnetic torque equation.

Optionally, the model predictive controller includes a determination unit, a current predictive control unit and a parameter optimization unit.

Optionally, output modules, further used to:

S31. Collect the speed feedback data and current feedback data of the permanent magnet synchronous motor through the determination unit, and judge the motion state of the joint motor of the robot; wherein, the motion state includes a dynamic adjustment phase and a stable adjustment phase.

S32. Obtain a value function according to the current prediction control unit; wherein, the value function includes a dynamic adjustment stage value function and a stable adjustment stage value function.

S33. Perform parameter optimization on the value function by the parameter optimization unit to obtain an optimal voltage vector of the robot joint motor to be predicted.

Optionally, output modules, further used to:

The speed feedback data and current feedback data of the permanent magnet synchronous motor with set conditions are collected through the determination unit; wherein the set conditions include a set cycle period and a set sampling time.

According to the speed feedback data and the current feedback data, it is determined that the motion state of the robot joint motor is in the dynamic adjustment stage or the stable adjustment stage.

Optionally, output modules, further used to:

Construct current prediction mathematical model, electromagnetic torque prediction mathematical model and inverter switching frequency prediction model.

According to the current prediction mathematical model, the electromagnetic torque prediction mathematical model and the inverter switching frequency prediction model, the value function is obtained.

Optionally, the value function is shown in the following formula (1):

g _m ＝λ[|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-βT _e ^k+1 +γS _w ^k+1 (1)

Among them, λ represents the weight coefficient of the current error; id ^k+1 and i _q ^k+1 are the _d and q axis currents at time k+1 after compensation prediction; _id ^k and i _q ^k are the currents after compensation prediction d and q axis currents at time k+1 and time k; β represents the weight coefficient of electromagnetic torque; T _e ^{k +1} represents the predicted value of electromagnetic torque at time k+1; γ represents the weight coefficient of inverter switching frequency; S _w ^k+1 represents the number of switching frequency changes of the inverter at time k+1.

Optionally, output modules, further used to:

The parameter optimization of the weight coefficient of the value function in the dynamic adjustment stage is performed online through the optimization algorithm of the parameter optimization unit.

And through the optimization algorithm of the parameter optimization unit, the parameters of the weight coefficients of the value function in the stable adjustment stage are optimized online.

Optionally, output modules, further used to:

Determine whether the rise time of the dynamic adjustment stage exceeds the preset threshold, and if so, reduce the weight coefficient of the current error in the value function of the dynamic adjustment stage, and increase the weight coefficient of the electromagnetic torque in the value function of the dynamic adjustment stage.

Optionally, output modules, further used to:

Determine whether the static error in the stable adjustment stage exceeds the preset threshold, and if so, increase the weight coefficient of the current error in the value function of the stable adjustment stage, and decrease the weight coefficient of the electromagnetic torque in the value function of the stable adjustment stage.

In one aspect, an electronic device is provided, the electronic device includes a processor and a memory, and at least one instruction is stored in the memory, and the at least one instruction is loaded and executed by the processor to realize the robot joint servo motor described above Model prediction parameter optimization method.

In one aspect, a computer-readable storage medium is provided, wherein at least one instruction is stored in the storage medium, and the at least one instruction is loaded and executed by a processor to implement the above method for optimizing parameters of robot joint servo motor model prediction.

The beneficial effects brought by the technical solutions provided by the embodiments of the present invention at least include:

In the above scheme, a method for optimizing parameters of robot joint servo motor model prediction is provided, and a method of parameter tuning is used in traditional model prediction. This method solves the problem that the weight coefficients of traditional model predictive control methods are difficult to configure accurately. Through The algorithm optimizes the weight coefficient ratio of the value function online, which can improve the dynamic performance and stability of the system.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

Fig. 1 is a schematic flow chart of a method for optimizing a model prediction parameter of a robot joint servo motor provided by an embodiment of the present invention;

Fig. 2 is a control block diagram of the method for optimizing the prediction parameters of the robot joint servo motor model provided by the embodiment of the present invention;

Fig. 3 is a schematic flowchart of an optimization algorithm based on a model predictive control method provided by an embodiment of the present invention;

4 is a block diagram of a robot joint servo motor model prediction parameter optimization device provided by an embodiment of the present invention;

Fig. 5 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention.

Detailed ways

In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will describe in detail with reference to the drawings and specific embodiments.

As shown in FIG. 1 , an embodiment of the present invention provides a method for optimizing parameters of robot joint servo motor model prediction, which can be implemented by electronic equipment. As shown in Figure 1, the flow chart of the method for optimizing the prediction parameters of the robot joint servo motor model, the processing flow of the method may include the following steps:

S1. The permanent magnet synchronous motor is modeled, and the mathematical model of the permanent magnet synchronous motor that meets the robot joint predictive control standard is obtained.

Optionally, the permanent magnet synchronous motor mathematical model includes permanent magnet synchronous motor d-axis and q-axis voltage and current equations, permanent magnet synchronous motor d-axis and q-axis voltage transformation equations in the rotating coordinate system, flux linkage equations, and permanent magnet synchronous motor electromagnetic torque equation.

Among them, the d-axis and q-axis voltage and current equations of the permanent magnet synchronous motor are shown in the following formula (1):

Among them, u _d and u _q represent the d-axis voltage and q-axis voltage respectively, R _s represents the stator equivalent resistance, id and i _q represent the _d -axis current and q-axis current respectively, ω _t represents the angular velocity of the motor at time t, L Indicates the direct axis and quadrature axis inductance, ψ _f is the total flux vector of the rotor.

The d-axis and q-axis voltage transformation equations of the permanent magnet synchronous motor in the rotating coordinate system are shown in the following formula (2):

Among them, ω represents the angular velocity of the motor at time _t , and u _a , _ub and uc represent the voltages in the natural coordinate system.

The flux linkage equation is shown in the following formula (3):

Among them, ψ _d , ψ _q are the d-axis and q-axis components of the motor stator flux linkage, respectively.

The electromagnetic torque equation of the permanent magnet synchronous motor is shown in the following formula (4):

Among them, p is the differential operator, id and i _q represent the _d -axis current and q-axis current respectively, L _d and L _q represent the equivalent inductance of the d-axis and q-axis stators respectively, ψ _f is the total flux vector of the rotor, ψ _d , ψ _q are the d-axis and q-axis components of the motor stator flux linkage, J represents the moment of inertia of the motor rotor, T _e represents the electromagnetic torque, and p _n is the number of pole pairs of the motor.

S2. According to the mathematical model of the permanent magnet synchronous motor, a model predictive controller for the permanent magnet synchronous motor of the robot joint is established.

S3. Obtain the optimal voltage vector of the robot joint motor to be predicted according to the model predictive controller.

Optionally, the model predictive controller includes a determination unit, a current predictive control unit and a parameter optimization unit.

According to the model predictive controller in S3, the optimal voltage vector of the robot joint motor to be predicted includes:

S31. Collect the speed feedback data and current feedback data of the permanent magnet synchronous motor through the determination unit to determine the motion state of the joint motor of the robot.

Among them, the motion state includes a dynamic adjustment phase and a stable adjustment phase.

Optionally, the passing determination unit in S31 collects the speed feedback data and current feedback data of the permanent magnet synchronous motor, and judging the motion state of the robot joint motor includes:

The speed feedback data and current feedback data of the permanent magnet synchronous motor with set conditions are collected through the determination unit.

Wherein, the setting condition includes a set cycle period and a set sampling time.

According to the speed feedback data and the current feedback data, it is determined that the motion state of the robot joint motor is in the dynamic adjustment stage or the stable adjustment stage.

In a feasible implementation, determining that the motion state of the robot joint motor is a dynamic adjustment phase or a stable adjustment phase can be to set a cycle and sampling time to obtain the speed data of the permanent magnet synchronous motor, and the data acquisition range can be set to 1000 After the collection is completed, judge whether the rising time is long (the data within 10% of the static difference is less than 3/4 of the total data); judge whether the static error meets the system requirements (the average value of the last 1/5 data static difference is less than the given 0.5% of the fixed value).

分别表示转速给定值与反馈值，I_a、I_b、I_c表示电机自然坐标系下的三相电流，I_α、I_β表示静止坐标系下的电流，I_d、I_q表示同步旋转坐标系下的d-q轴电流。I_q ^*、I_d ^*表示预测控制模型的电流参考值，I_q ^*由PI控制器输出，I_d ^*设定I_d ^*＝0，将转速参考值与转速检测器检测到的电机角速度作差得到差值，通过速度环PI控制器输出参考电流，通过模型预测控制的判定单元采集转速数据，判断电机的状态及性能。As shown in Figure 2, where,

Represent the speed given value and feedback value respectively, I _a , I _b , I _c represent the three-phase current in the motor natural coordinate system, I _α , I _β represent the current in the stationary coordinate system, I _d , I _q represent synchronous rotation The dq-axis current in the coordinate system. I _q ^* and I _d ^* represent the current reference value of the predictive control model, I _q ^* is output by the PI controller, I _d ^* is set to I _d ^* = 0, and the rotational speed reference value and the motor angular velocity detected by the rotational speed detector are made The difference is obtained by the difference, the reference current is output through the speed loop PI controller, and the speed data is collected through the judgment unit of the model predictive control to judge the state and performance of the motor.

S32. Obtain a value function according to the current prediction control unit.

Wherein, the value function includes a dynamic adjustment stage value function and a stable adjustment stage value function.

Optionally, according to the current prediction control unit in S32, the value function obtained includes:

Construct current prediction mathematical model, electromagnetic torque prediction mathematical model and inverter switching frequency prediction model.

Among them, the current prediction mathematical model is expressed as the following formula (5) (6):

Among them, id ^k and i _q ^k are _d and q axis currents at time k; id ^k+1 and i _q ^k+1 are _d and q axis currents at time k+1 after compensation prediction; λ _d ^k , λ _q ^k is the compensation factor; , L _d , and L _q represent the equivalent inductance of the d-axis and q-axis stators respectively; ψ _f is the total flux vector of the rotor; Δt is the sampling period.

The mathematical model of electromagnetic torque prediction is shown in the following formula (7):

Among them, T _e ^k+1 represents the predicted value of electromagnetic torque at the next moment; id ^k+1 and i _q ^k+1 are the _d and q axis currents at time k+1 after compensation prediction; L _d , L _q Represent the equivalent inductance of the d-axis and q-axis stators respectively; ψ _f is the total flux vector of the rotor.

The model prediction unit outputs 8 alternative basic voltage vectors for the optimal voltage vector, including:

U ₀ (000), U ₁ (001), U ₂ (010), U ₃ (011), U ₄ (100), U ₅ (101), U ₆ (110), U ₇ (111), corresponding to 8 different voltage vectors for a three-phase voltage source inverter.

The inverter switching frequency prediction model is expressed as the following formula (8):

Among them, U ^k+1 and U ^k represent the optimal voltage vector at the next moment and the current moment respectively, S _w ^k+1 represents the amount of voltage vector change at the moment k+1, which can be expressed as the frequency of inverter switching, U ₁ ^′ _i represents the variation of the i-th dimension of the inverter switch at the next moment.

Specifically, U' _1i represents the 1st row and the ith column of the 1-row 3-column matrix U'.

According to the current prediction mathematical model, the electromagnetic torque prediction mathematical model and the inverter switching frequency prediction model, the value function is obtained.

Optionally, the value function is shown in the following formula (9):

g _m ＝λ[|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-βT _e ^k+1 +γS _w ^k+1 (9)

Among them, λ represents the current error weight coefficient; id ^k+1 and i _q ^k+1 are the _d and q axis currents at time k+1 after compensation prediction; _id ^k and i _q ^k are k after compensation prediction +1 time and k time d, q axis current; β represents the electromagnetic torque weight coefficient; T _e ^k+1 represents the electromagnetic torque prediction value at k+1 time; γ represents the inverter switching frequency weight coefficient; S _w ^{k +1} represents the number of inverter switching frequency changes at time k+1.

S33. Perform parameter optimization on the value function by the parameter optimization unit to obtain an optimal voltage vector of the robot joint motor to be predicted.

Optionally, optimizing the parameters of the value function through the parameter optimization unit in S33 includes:

S331. Perform online parameter optimization on the weight coefficients of the value function in the dynamic adjustment stage through the optimization algorithm of the parameter optimization unit.

Optionally, online parameter optimization of the weight coefficient of the value function in the dynamic adjustment stage through the optimization algorithm of the parameter optimization unit includes:

Determine whether the rise time of the dynamic adjustment stage exceeds the preset threshold, and if so, reduce the weight coefficient of the current error in the value function of the dynamic adjustment stage, and increase the weight coefficient of the electromagnetic torque in the value function of the dynamic adjustment stage.

In a feasible implementation, when the speed reaches a given value for the first time, it is recorded as the rise time t _r , when the speed error is stable at ±5%, it is recorded as the adjustment time t _s , and the value function g is selected within the time of t _{s m} ＝λ ₁ [|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-β ₁ T _e ^k+1 +γ ₁ S _w ^k+1 , initial value λ ₁ =β ₁ , γ ₁ =0.

After t _s , select g _m ＝λ ₂ [|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-β ₂ T _e ^k+1 +γ ₂ S _w ^{k+ 1} , where λ ₂ =β ₂ , γ ₂ =0.

S332. Perform online parameter optimization on the weight coefficients of the value function in the stable adjustment stage by using the optimization algorithm of the parameter optimization unit.

Optionally, online parameter optimization of the weight coefficient of the value function in the stable adjustment stage through the optimization algorithm of the parameter optimization unit includes:

Determine whether the static error in the stable adjustment stage exceeds the preset threshold, and if so, increase the weight coefficient of the current error in the value function of the stable adjustment stage, and decrease the weight coefficient of the electromagnetic torque in the value function of the stable adjustment stage.

In a feasible implementation, as shown in Figure 3, according to the obtained feedback data, the online calculation of the rise time and static error of the system meets the system requirements, and if the requirements are not met, the value of each stage is optimized online through the parameter optimization unit optimization algorithm The weight coefficient of the function, if the rising time is long, reduce the weight coefficient of the current error in the value function of the rising stage, increase the weight coefficient of the virtual torque, and decrease or increase by 1% each time, that is, λ ₁ ′=λ ₁ *99% , β ₁ ′=β ₁ *101%; if the static error is large, increase the weight coefficient of the current error in the value function in the stable stage, reduce the weight coefficient of the virtual torque, and decrease or increase by 1% each time, that is, λ ₂ ′=λ ₂ *101%, β ₂ ′=β ₂ *99%, until the dynamic performance and stability of the system meet the requirements of the system, then introduce the inverter switching frequency value function γ ₁ =10*λ ₁ , γ ₂ =10*λ ₂ .

In a feasible implementation mode, the robot joint servo motor model prediction parameter optimization method of the embodiment of the present invention establishes a permanent magnet synchronous motor mathematical model that meets the robot joint model predictive control standard; according to the established permanent magnet synchronous motor mathematical model, establishes Speed loop PI control unit, current prediction control unit, and parameter optimization unit; according to the current prediction control unit, the speed feedback data is obtained to judge the state of the robot joint motor. The operating state includes the dynamic adjustment stage and the stable adjustment stage, and the actual current value of the motor is obtained. , compared with the given value output by the PI controller, set a stage-by-stage value function to control the system; according to the parameter optimization unit, the online calculation system whether the rise time and static error meet the system requirements, if not, optimize the parameters The unit optimization algorithm optimizes the weight coefficients of the value function of each stage online. By adopting the invention, the parameters of the value function in the model predictive control method can be optimized online, and the dynamic performance, steady-state performance and switching frequency of the inverter are comprehensively considered to further optimize the system performance.

In the embodiment of the present invention, a method for optimizing parameters of robot joint servo motor model prediction is provided. A method of parameter tuning is used in traditional model prediction. This method solves the problem that the weight coefficients of traditional model predictive control methods are difficult to accurately configure. , through the online optimization of the value function weight coefficient ratio by this algorithm, the dynamic performance and stability performance of the system can be well improved.

As shown in Figure 4, the embodiment of the present invention provides a robot joint servo motor model prediction parameter optimization device 400, the robot joint servo motor model prediction parameter optimization device 400 is applied to realize the robot joint servo motor model prediction parameter optimization method, the Device 400 includes:

The mathematical model building module 410 is used to model the permanent magnet synchronous motor to obtain a permanent magnet synchronous motor mathematical model that meets the robot joint predictive control standard.

The controller construction module 420 is used to establish a model predictive controller of the permanent magnet synchronous motor of the robot joint according to the mathematical model of the permanent magnet synchronous motor.

The output module 430 is configured to obtain the optimal voltage vector of the robot joint motor to be predicted according to the model predictive controller.

Optionally, the permanent magnet synchronous motor mathematical model includes permanent magnet synchronous motor d-axis and q-axis voltage and current equations, permanent magnet synchronous motor d-axis and q-axis voltage transformation equations in the rotating coordinate system, flux linkage equations, and permanent magnet synchronous motor electromagnetic torque equation.

Optionally, the model predictive controller includes a determination unit, a current predictive control unit and a parameter optimization unit.

Optionally, the output module 430 is further used for:

S31. Collect the speed feedback data and current feedback data of the permanent magnet synchronous motor through the determination unit, and judge the motion state of the joint motor of the robot; wherein, the motion state includes a dynamic adjustment phase and a stable adjustment phase.

S32. Obtain a value function according to the current prediction control unit; wherein, the value function includes a dynamic adjustment stage value function and a stable adjustment stage value function.

S33. Perform parameter optimization on the value function by the parameter optimization unit to obtain an optimal voltage vector of the robot joint motor to be predicted.

Optionally, the output module 430 is further used for:

The speed feedback data and current feedback data of the permanent magnet synchronous motor with set conditions are collected through the determination unit; wherein the set conditions include a set cycle period and a set sampling time.

According to the speed feedback data and the current feedback data, it is determined that the motion state of the robot joint motor is in the dynamic adjustment stage or the stable adjustment stage.

Optionally, the output module 430 is further used for:

Construct current prediction mathematical model, electromagnetic torque prediction mathematical model and inverter switching frequency prediction model.

According to the current prediction mathematical model, the electromagnetic torque prediction mathematical model and the inverter switching frequency prediction model, the value function is obtained.

Optionally, the value function is shown in the following formula (1):

g _m ＝λ[|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-βT _e ^k+1 +γS _w ^k+1 (1)

Among them, λ represents the weight coefficient of the current error; id ^k+1 and i _q ^k+1 are the _d and q axis currents at time k+1 after compensation prediction; _id ^k and i _q ^k are the currents after compensation prediction d and q axis currents at time k+1 and time k; β represents the weight coefficient of electromagnetic torque; T _e ^{k +1} represents the predicted value of electromagnetic torque at time k+1; γ represents the weight coefficient of inverter switching frequency; S _w ^k+1 represents the number of switching frequency changes of the inverter at time k+1.

Optionally, the output module 430 is further used for:

The parameter optimization of the weight coefficient of the value function in the dynamic adjustment stage is performed online through the optimization algorithm of the parameter optimization unit.

And through the optimization algorithm of the parameter optimization unit, the parameters of the weight coefficients of the value function in the stable adjustment stage are optimized online.

Optionally, the output module 430 is further used for:

Determine whether the rise time of the dynamic adjustment stage exceeds the preset threshold, and if so, reduce the weight coefficient of the current error in the value function of the dynamic adjustment stage, and increase the weight coefficient of the electromagnetic torque in the value function of the dynamic adjustment stage.

Optionally, the output module 430 is further used for:

Determine whether the static error in the stable adjustment stage exceeds the preset threshold, and if so, increase the weight coefficient of the current error in the value function of the stable adjustment stage, and decrease the weight coefficient of the electromagnetic torque in the value function of the stable adjustment stage.

In the embodiment of the present invention, a method for optimizing parameters of robot joint servo motor model prediction is provided. A method of parameter tuning is used in traditional model prediction. This method solves the problem that the weight coefficients of traditional model predictive control methods are difficult to accurately configure. , through the online optimization of the value function weight coefficient ratio by this algorithm, the dynamic performance and stability performance of the system can be well improved.

5 is a schematic structural diagram of an electronic device 500 provided by an embodiment of the present invention. The electronic device 500 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 501 And one or more memory 502, wherein at least one instruction is stored in the memory 502, at least one instruction is loaded and executed by the processor 501 to realize the following robot joint servo motor model prediction parameter optimization method:

S1. The permanent magnet synchronous motor is modeled, and the mathematical model of the permanent magnet synchronous motor that meets the robot joint predictive control standard is obtained.

S2. According to the mathematical model of the permanent magnet synchronous motor, a model predictive controller for the permanent magnet synchronous motor of the robot joint is established.

S3. Obtain the optimal voltage vector of the robot joint motor to be predicted according to the model predictive controller.

In an exemplary embodiment, there is also provided a computer-readable storage medium, such as a memory including instructions, which can be executed by a processor in a terminal to implement the above method for optimizing parameters of robot joint servo motor model prediction. For example, the computer readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, among others.

Those of ordinary skill in the art can understand that all or part of the steps for implementing the above embodiments can be completed by hardware, and can also be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium. The above-mentioned The storage medium mentioned may be a read-only memory, a magnetic disk or an optical disk, and the like.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (9)

1. A robot joint servo motor model prediction parameter optimization method is characterized by comprising the following steps:

s1, modeling is carried out on a permanent magnet synchronous motor to obtain a permanent magnet synchronous motor mathematical model meeting a robot joint prediction control standard;

s2, establishing a model prediction controller of the robot joint permanent magnet synchronous motor according to the permanent magnet synchronous motor mathematical model;

s3, obtaining an optimal voltage vector of the robot joint motor to be predicted according to the model prediction controller;

the model predictive controller includes a determination unit, a current predictive control unit, and a parameter optimization unit;

in the S3, obtaining the optimal voltage vector of the robot joint motor to be predicted according to the model prediction controller includes:

s31, acquiring speed feedback data and current feedback data of the permanent magnet synchronous motor through the judging unit, and judging the motion state of the robot joint motor; wherein the motion state comprises a dynamic regulation phase and a stable regulation phase;

s32, obtaining a value function according to the current prediction control unit; wherein the cost function comprises a dynamic adjustment stage cost function and a stable adjustment stage cost function;

and S33, performing parameter optimization on the cost function through the parameter optimization unit to obtain the optimal voltage vector of the robot joint motor to be predicted.

2. The method of claim 1, wherein the PMSM mathematical model comprises a PMSM d-axis and q-axis voltage-current equation, a PMSM d-axis and q-axis voltage-to-rotating coordinate system transformation equation, a flux linkage equation, and a PMSM electromagnetic torque equation.

3. The method according to claim 1, wherein the step S31 of collecting speed feedback data and current feedback data of the permanent magnet synchronous motor by the determination unit and determining the motion state of the robot joint motor comprises:

acquiring speed feedback data and current feedback data of the permanent magnet synchronous motor under set conditions through the judging unit; wherein the set condition comprises a set cycle period and a set sampling time;

and judging the motion state of the robot joint motor to be a dynamic regulation stage or a stable regulation stage according to the speed feedback data and the current feedback data.

4. The method of claim 1, wherein the step of obtaining a cost function according to the current prediction control unit in S32 comprises:

constructing a current prediction mathematical model, an electromagnetic torque prediction mathematical model and an inverter switching frequency prediction model;

and obtaining a value function according to the current prediction mathematical model, the electromagnetic torque prediction mathematical model and the inverter switching frequency prediction model.

5. The method of claim 1, wherein the cost function is represented by the following equation (1):

g _m ＝λ[|i _q ^k+1 -i _q ^k |+|i _d ^k+1 -i _d ^k |]-βT _e ^k+1 +γS _w ^k+1 (1)

wherein λ represents a weighting coefficient of the current error; i.e. i _d ^k+1 、i _q ^k+1 D-axis current and q-axis current at the k +1 moment after compensation prediction; i.e. i _d ^k 、i _q ^k D and q axis currents at the k moment; β represents a weight coefficient of the electromagnetic torque; t is _e ^k+1 Representing the predicted value of the electromagnetic torque at the moment k + 1; gamma represents a weight coefficient of the switching frequency of the inverter; s. the _w ^k+1 Indicating the number of inverter switching frequency changes at time k + 1.

6. The method according to claim 1, wherein the parameter optimization of the cost function by the parameter optimization unit in S33 comprises:

the weight coefficient of the dynamic adjustment stage cost function is optimized on line through the optimization algorithm of the parameter optimization unit,

and carrying out parameter optimization on the weight coefficient of the value function in the stable adjusting stage on line through the optimization algorithm of the parameter optimization unit.

7. The method of claim 6, wherein the parameter optimizing weight coefficients of the dynamic adjustment phase cost function on-line by an optimization algorithm of the parameter optimization unit comprises:

and judging whether the rising time of the dynamic regulation stage exceeds a preset threshold value, if so, reducing the weight coefficient of the current error in the value function of the dynamic regulation stage, and increasing the weight coefficient of the electromagnetic torque in the value function of the dynamic regulation stage.

8. The method of claim 6, wherein the parameter optimizing the weight coefficients of the stability adjustment phase cost function on-line by the optimization algorithm of the parameter optimization unit comprises:

and judging whether the static error of the stable regulation stage exceeds a preset threshold value, if so, increasing the weight coefficient of the current error in the cost function of the stable regulation stage, and reducing the weight coefficient of the electromagnetic torque in the cost function of the stable regulation stage.

9. A robot joint servo motor model prediction parameter optimization device is characterized by comprising:

the mathematical model construction module is used for modeling the permanent magnet synchronous motor to obtain a permanent magnet synchronous motor mathematical model which accords with the robot joint prediction control standard;

the controller building module is used for building a model prediction controller of the robot joint permanent magnet synchronous motor according to the permanent magnet synchronous motor mathematical model;

the output module is used for predicting the controller according to the model to obtain the optimal voltage vector of the robot joint motor to be predicted;

the model predictive controller includes a determination unit, a current predictive control unit, and a parameter optimization unit;

the obtaining of the optimal voltage vector of the robot joint motor to be predicted according to the model prediction controller comprises:

s31, acquiring speed feedback data and current feedback data of the permanent magnet synchronous motor through the judging unit, and judging the motion state of the robot joint motor; wherein the motion state comprises a dynamic regulation stage and a stable regulation stage;

s32, obtaining a value function according to the current prediction control unit; wherein the cost function comprises a dynamic adjustment stage cost function and a stable adjustment stage cost function;

and S33, performing parameter optimization on the value function through the parameter optimization unit to obtain the optimal voltage vector of the robot joint motor to be predicted.

Images