Precise force position control method for servo electric actuator

Abstract

The invention provides a precise force and position control method for a servo electric actuator. The precise force and position control method comprises the following steps: S1, constructing and training a two-input and three-output force control fuzzy neural network model and a position control fuzzy neural network control model; s2, constructing and training a two-input two-output segmented neural network model; s3, acquiring a current force value, and calculating a force controller parameter adjustment amount; s4, acquiring a current position value and a position controller parameter adjustment amount; s5, calculating a weighting coefficient of the force controller parameter adjustment amount and a weighting coefficient of the position controller parameter adjustment amount, and adjusting the force controller parameter adjustment amount and the position controller parameter adjustment amount; and S6, adjusting parameters of a force controller and a position controller, and respectively calculating a force control quantity and a position control quantity. The force-position hybrid control method is used for force-position hybrid control of the servo electric actuator so as to improve the tail end control precision.

Description

Technical Field

[0001] The present invention relates to the technical field of servo electric actuator control, and in particular to a precise force and position control method for a servo electric actuator. Background Art

[0002] Servo electric actuators are an indispensable component of modern industrial automation control. Traditionally, position control has been used to control servo electric actuators, ensuring they follow a predetermined trajectory. However, for some applications requiring force control, position control alone is insufficient. Therefore, hybrid force-position control technology for servo electric actuators has emerged.

[0003] Currently, the application areas of servo electric actuator force-position hybrid control technology are very broad. For example, in the field of robotics, taking assembly robots as an example, CN106402080B discloses an integrated miniature force-position hybrid servo hydraulic cylinder. The displacement sensor, force sensor and hydraulic cylinder are integrated to achieve modularization of force-position servo hydraulic control, improve control accuracy, reduce, and easily adapt to various hydraulic drive robot systems. However, in some high-demand applications, the problem of poor end-point accuracy still exists in the process of servo electric actuator force-position hybrid control.

[0004] In view of this, the present invention is proposed. Summary of the Invention

[0005] The object of the present invention is to provide a precise force-position control method for a servo electric actuator, which is used for force-position hybrid control of the servo electric actuator to improve the terminal control accuracy.

[0006] To solve the above technical problems, the present invention provides a method for precise force and position control of a servo electric actuator, comprising the following steps: S1: Construct and train a two-input and three-output force control fuzzy neural network model and a position control fuzzy neural network control model. The input of the force control fuzzy neural network model is the force deviation and the force deviation change rate, and the output is the parameter adjustment amount of the force controller. The input of the position control fuzzy neural network model is the position deviation and the position deviation change rate, and the output is the parameter adjustment amount of the position controller. S2: Construct and train a two-input and two-output segmented neural network model. The input of the segmented neural network model is the force deviation and the position deviation, and the output is the weighted coefficient of the force control fuzzy neural network model and the weighted coefficient of the position control fuzzy neural network model. S3: Obtain the current force value and calculate the deviation and deviation change rate between the current force value and the preset force target value, namely the force deviation and force deviation change rate; input the force deviation and force deviation change rate into the force control fuzzy neural network model trained in S1 to calculate the force controller parameter adjustment amount; S4: Obtain the current position value and calculate the deviation and deviation change rate between the current position value and the preset position target value, i.e., the position deviation and position deviation change rate; input the position deviation and position deviation change rate into the position control fuzzy neural network model trained in S1 to calculate the position controller parameter adjustment amount; S5: Input the force deviation calculated by S3 and the position deviation calculated by S4 into the segmented neural network model trained by S2, calculate the weighted coefficients of the force controller parameter adjustment amount and the position controller parameter adjustment amount, and adjust the force controller parameter adjustment amount and the position controller parameter adjustment amount; S6: adjusting the parameters of the force controller and the position controller respectively according to the force controller parameter adjustment amount and the position controller parameter adjustment amount adjusted in S5, and calculating the force control amount and the position control amount respectively.

[0007] Furthermore, the fuzzy neural network control model in S1 has four layers, namely input layer, fuzzification layer, fuzzy inference layer and output layer; The input layer consists of two nodes, corresponding to the deviation e and the deviation change rate ec respectively; The fuzzification layer consists of 2*n nodes. The node of the output layer deviation e is connected to the 1~n nodes of the fuzzification layer. The node of the output layer deviation change rate ec is connected to the n+1~2n nodes of the fuzzification layer. n is the number of quantization levels of deviation e and deviation change rate ec. The fuzzy reasoning layer consists of n*n nodes, and the 1~n nodes and n+1~2n nodes of the fuzzification layer are connected to each node of the fuzzy reasoning layer in pairs; The output layer consists of three nodes, corresponding to △Kp, △Ki, and △Kd respectively. Each node in the output layer is connected to each node in the fuzzy inference layer.

[0008] Furthermore, the calculation formula of each node in the input layer is as follows: ; Among them, x1 is e, x2 is ec; The calculation formula of each node in the fuzzy layer is as follows: ; Among them, m 1b 、l 1b , b=1,2,...,n are the center and width of the membership function of each language value of e, m 1b 、l 1b , b=n+1,n+2,...,2n are the center and width of the membership function of each language value of ec respectively; The calculation formula for each node in the reasoning layer is as follows: ; Wherein, b1 and b2 are two nodes in the fuzzification layer connected to the inference layer node c; The calculation formula of each node in the output layer is as follows: ; Among them, w dc is the weight connecting each node of the inference layer with the node d of the output layer, and f4(1), f4(2), and f4(3) are △Kp, △Ki, and △Kd respectively.

[0009] Furthermore, when training the fuzzy neural network control model, the loss value is calculated by the following formula: ; Among them, g4(1), g4(2), and g4(3) are the expected outputs of △Kp, △Ki, and △Kd respectively; The fuzzy neural network control model is trained by updating m through the following formula ab 、l ab 、w dc : ; ; ; Where n is the number of iterations, 0<λ<1, 0<Y<1.

[0010] Furthermore, the segmented neural network model in S2 has three layers, namely, an input layer, a hidden layer, and an output layer; The input layer consists of two nodes, corresponding to stress deviation e1 and position deviation e2 respectively; The hidden layer consists of m nodes, and each node in the input layer is connected to each node in the hidden layer, 3≤m≤12; The output layer consists of two nodes, which are the weighted coefficients α output by the stress control fuzzy neural network model and β output by the position control fuzzy neural network model respectively. Each node in the hidden layer is connected to each node in the output layer; 0<α<1, 0<β<1.

[0011] Furthermore, when training the segmented neural network model, m is set to 3, 4, ..., and 12 for training; an evaluation function is set, and the value of the evaluation function is calculated when m is 3, 4, ..., and 12 for training, and m corresponding to the maximum value of the evaluation function is set as the number of nodes in the hidden layer of the segmented neural network model. The form of the evaluation function is as follows: ; Among them, j is the number of evaluation indicators, θ iThe sum of is 1, F i is the value of the preset evaluation index i, F i,min is the minimum value of the evaluation index i preset during training for each m value, F i,max It is the maximum value of the evaluation index i preset during training for each m value.

[0012] Furthermore, before training the segmented neural network model, the following steps are performed to obtain θ i : A1: Construct a two-dimensional expert evaluation table, where each row is θ i , each column is an importance level, and each importance level is assigned a value from high to low; A2: Obtain the expert evaluation form and calculate the total score of each indicator H(θ i ), and calculate θ by the following formula i : .

[0013] The present invention provides a precise force and position control method for a servo electric actuator, which has the following beneficial effects: a force control fuzzy neural network model and a position control fuzzy neural network control model are constructed to calculate the parameter adjustment amount of the force controller and the parameter adjustment amount of the position controller; in addition, a segmented neural network model is constructed to comprehensively calculate the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount based on the force deviation and the position deviation; and the final force controller parameters and position controller parameters are determined based on the parameter adjustment amount of the force controller and the parameter adjustment amount of the position controller, as well as the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount, so as to determine the force control amount and the position control amount, thereby improving the end control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The present invention is a control flow chart of a method for precise force and position control of a servo electric actuator according to an embodiment. DETAILED DESCRIPTION

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0016] Example 1: A method for precise force and position control of a servo electric actuator comprises the following steps: S1: Construct and train a two-input and three-output force control fuzzy neural network model and a position control fuzzy neural network control model. The input of the force control fuzzy neural network model is the force deviation and the force deviation change rate, and the output is the parameter adjustment amount of the force controller. The input of the position control fuzzy neural network model is the position deviation and the position deviation change rate, and the output is the parameter adjustment amount of the position controller.

[0017] S2: Construct and train a two-input and two-output segmented neural network model. The input of the segmented neural network model is the force deviation and the position deviation, and the output is the weighted coefficient of the force control fuzzy neural network model and the weighted coefficient of the position control fuzzy neural network model.

[0018] S3: Obtain the current force value and calculate the deviation and deviation change rate between the current force value and the preset force target value, namely the force deviation and force deviation change rate; input the force deviation and force deviation change rate into the force control fuzzy neural network model trained in S1 to calculate the force controller parameter adjustment amount; In this embodiment, the force controller can be configured as a PID controller, and the initial parameters Kp, Ki, and Kd of the force controller can be preset. During the control process, the force deviation and the force deviation change rate are input into the force control fuzzy neural network model trained by S1 to calculate the force controller parameter adjustment amounts △Kp, △Ki, and △Kd.

[0019] S4: Obtain the current position value and calculate the deviation and deviation change rate between the current position value and the preset position target value, i.e., the position deviation and position deviation change rate; input the position deviation and position deviation change rate into the position control fuzzy neural network model trained in S1 to calculate the position controller parameter adjustment amount; In this embodiment, the position controller can be configured as a PID controller, and the initial parameters Kp, Ki, and Kd of the position controller can be preset. During the control process, the position deviation and the position deviation change rate are input into the position control fuzzy neural network model trained by S1 to calculate the position controller parameter adjustment amounts △Kp, △Ki, and △Kd.

[0020] S5: Input the force deviation calculated by S3 and the position deviation calculated by S4 into the segmented neural network model trained by S2, calculate the weighted coefficients of the force controller parameter adjustment amount and the position controller parameter adjustment amount, and adjust the force controller parameter adjustment amount and the position controller parameter adjustment amount.

[0021] S6: adjusting the parameters of the force controller and the position controller respectively according to the force controller parameter adjustment amount and the position controller parameter adjustment amount adjusted in S5, and calculating the force control amount and the position control amount respectively.

[0022] In this embodiment, in S5, the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount are calculated based on the force deviation and the position deviation; in S5, the force controller parameter adjustment amounts △Kp, △Ki, △Kd obtained in S3 are multiplied by α, that is, the force controller parameter adjustment amount is adjusted, and then the preset initial parameters Kp, Ki, Kd of the force controller are added in S6 to obtain the final parameters of the force controller, and the force control amount is calculated; in S5, the position controller parameter adjustment amounts △Kp, △Ki, △Kd obtained in S4 are multiplied by β, that is, the position controller parameter adjustment amount is adjusted, and then the preset initial parameters Kp, Ki, Kd of the position controller are added in S6 to obtain the final parameters of the position controller, and the force control amount is calculated.

[0023] A precision force and position control method for a servo electric actuator of this embodiment constructs a force control fuzzy neural network model and a position control fuzzy neural network control model to calculate the parameter adjustment amount of the force controller and the parameter adjustment amount of the position controller; in addition, a segmented neural network model is constructed to comprehensively calculate the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount based on the force deviation and the position deviation; and the final parameters of the force controller and the position controller are determined based on the parameter adjustment amount of the force controller and the parameter adjustment amount of the position controller, as well as the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount, so as to determine the force control amount and the position control amount, thereby improving the end control accuracy.

[0024] In an optional embodiment, the fuzzy neural network control model in S1 has four layers, namely, an input layer, a fuzzification layer, a fuzzy inference layer and an output layer; The input layer consists of two nodes, corresponding to the deviation e and the deviation change rate ec respectively; The fuzzification layer consists of 2*n nodes. The node of the output layer deviation e is connected to the 1~n nodes of the fuzzification layer. The node of the output layer deviation change rate ec is connected to the n+1~2n nodes of the fuzzification layer. n is the number of quantization levels of deviation e and deviation change rate ec. The fuzzy reasoning layer consists of n*n nodes, and the 1~n nodes and n+1~2n nodes of the fuzzification layer are connected to each node of the fuzzy reasoning layer in pairs; The output layer consists of three nodes, corresponding to △Kp, △Ki, and △Kd respectively. Each node in the output layer is connected to each node in the fuzzy inference layer.

[0025] In this optional embodiment, the fuzzy algorithm and the RBF neural network are combined to construct a fuzzy neural network model, which can improve the response speed, stability and accuracy of the control.

[0026] Among them, n is the number of linguistic values ​​when the deviation e and the deviation change rate ec are fuzzified. For example, n=7, the linguistic values ​​are NB, NM, NS, Z, PS, PM, and PB respectively. At this time, the fuzzification layer has a total of 14 nodes and the fuzzy reasoning layer has 49 nodes. The 1st to 7th nodes of the fuzzification layer are connected to the nodes of the input layer e, and the 8th to 14th nodes are connected to the nodes of the input layer ec. The 1st to 7th nodes and the 8th to 14th nodes of the fuzzification layer are connected to the 49 nodes of the fuzzy reasoning layer in pairs.

[0027] In an optional embodiment, the calculation formula for each node in the input layer is as follows: ; Among them, x1 is e, x2 is ec; The calculation formula of each node in the fuzzy layer is as follows: ; Among them, m 1b 、l 1b , b=1,2,...,n are the center and width of the membership function of each language value of e respectively. The center and width of the membership function of each language value can be set as needed in the specific implementation, and will not be elaborated here. m 1b 、l 1b , b=n+1,n+2,...,2n are the center and width of the membership function of each language value of ec respectively; The calculation formula for each node in the reasoning layer is as follows: ; Wherein, b1 and b2 are two nodes in the fuzzification layer connected to the inference layer node c; The calculation formula of each node in the output layer is as follows: ; Among them, w dc is the weight connecting each node of the inference layer with the node d of the output layer, and f4(1), f4(2), and f4(3) are △Kp, △Ki, and △Kd respectively.

[0028] In an optional embodiment, the loss value is calculated by the following formula when training the fuzzy neural network control model: ; Among them, g4(1), g4(2), and g4(3) are the expected outputs of △Kp, △Ki, and △Kd respectively; The fuzzy neural network control model is trained by updating m through the following formula ab 、l ab 、w dc : ; ; ; Where n is the number of iterations, 0<λ<1, 0<Y<1.

[0029] In an optional embodiment, the segmented neural network model in S2 has three layers, namely an input layer, a hidden layer and an output layer; The input layer consists of two nodes, corresponding to stress deviation e1 and position deviation e2 respectively; The hidden layer consists of m nodes, and each node in the input layer is connected to each node in the hidden layer, 3≤m≤12; The output layer consists of two nodes, which are the weighted coefficients α output by the stress control fuzzy neural network model and β output by the position control fuzzy neural network model respectively. Each node in the hidden layer is connected to each node in the output layer; 0<α<1, 0<β<1.

[0030] In this optional embodiment, a segmented neural network model is constructed using a BP neural network to comprehensively calculate the weighted coefficient α of the force controller parameter adjustment amount and the weighted coefficient β of the position controller parameter adjustment amount based on the force deviation and the position deviation.

[0031] In an optional embodiment, when training the segmented neural network model, m is set to 3, 4, ..., 12 for training; an evaluation function is set, and the value of the evaluation function when m is 3, 4, ..., 12 for training is calculated respectively, and m corresponding to the maximum value of the evaluation function is set as the number of nodes in the hidden layer of the segmented neural network model. The form of the evaluation function is as follows: ; Among them, j is the number of evaluation indicators, θ i The sum of is 1, F i is the value of the preset evaluation index i, F i,min is the minimum value of the evaluation index i preset during training for each m value, F i,max It is the maximum value of the evaluation index i preset during training for each m value.

[0032] In this optional embodiment, the range of the m value (i.e., the number of hidden layer nodes) is first determined, and the segmented neural network model is trained using different m values. Then, the optimal m value is determined based on the evaluation function after the training is completed, which is used as the final number of nodes in the hidden layer of the segmented neural network model.

[0033] In an optional embodiment, before training the segmented neural network model, the following steps are performed to obtain θ i : A1: Construct a two-dimensional expert evaluation table, where each row is θ i , each column is an importance level, and each importance level is assigned a value from high to low; A2: Obtain the expert evaluation form and calculate the total score of each indicator H(θ i ), and calculate θ by the following formula i : .

[0034] In this optional embodiment, the importance level of A1 can be very important, generally important, and not important, corresponding to 3 points, 2 points, and 1 point respectively. After constructing the two-dimensional expert evaluation table, 10 experts can be invited to evaluate the importance of each indicator. In A2, the total score of each indicator is calculated based on the expert evaluation and the score corresponding to the importance level, and finally the θ corresponding to each indicator is calculated. i In summary, this optional embodiment can quickly, effectively and accurately obtain the corresponding θ of each indicator. i .

[0035] The above description of the disclosed embodiments will enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein, but is intended to be embodied in the widest possible scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for precise force and position control of a servo electric actuator, characterized in that: The steps include: S1: Construct and train a two-input and three-output force control fuzzy neural network model and a position control fuzzy neural network control model. The input of the force control fuzzy neural network model is the force deviation and the force deviation change rate, and the output is the parameter adjustment amount of the force controller. The input of the position control fuzzy neural network model is the position deviation and the position deviation change rate, and the output is the parameter adjustment amount of the position controller. S2: Construct and train a two-input and two-output segmented neural network model. The input of the segmented neural network model is the force deviation and the position deviation, and the output is the weighted coefficient of the force control fuzzy neural network model and the weighted coefficient of the position control fuzzy neural network model. S3: Obtain the current force value and calculate the deviation and deviation change rate between the current force value and the preset force target value, namely the force deviation and force deviation change rate; input the force deviation and force deviation change rate into the force control fuzzy neural network model trained in S1 to calculate the force controller parameter adjustment amount; S4: Obtain the current position value and calculate the deviation and deviation change rate between the current position value and the preset position target value, i.e., the position deviation and position deviation change rate; input the position deviation and position deviation change rate into the position control fuzzy neural network model trained in S1 to calculate the position controller parameter adjustment amount; S5: Input the force deviation calculated by S3 and the position deviation calculated by S4 into the segmented neural network model trained by S2, calculate the weighted coefficients of the force controller parameter adjustment amount and the position controller parameter adjustment amount, and adjust the force controller parameter adjustment amount and the position controller parameter adjustment amount; S6: adjusting the parameters of the force controller and the position controller respectively according to the force controller parameter adjustment amount and the position controller parameter adjustment amount adjusted in S5, and calculating the force control amount and the position control amount respectively.

2. A servo electric actuator precision force position control method according to claim 1, characterized in that: The fuzzy neural network control model in S1 has four layers, namely input layer, fuzzification layer, fuzzy inference layer and output layer; The input layer consists of two nodes, corresponding to the deviation e and the deviation change rate ec respectively; The fuzzification layer consists of 2*n nodes. The node of the output layer deviation e is connected to the 1~n nodes of the fuzzification layer. The node of the output layer deviation change rate ec is connected to the n+1~2n nodes of the fuzzification layer. n is the number of quantization levels of deviation e and deviation change rate ec. The fuzzy reasoning layer consists of n*n nodes, and the 1~n nodes and n+1~2n nodes of the fuzzification layer are connected to each node of the fuzzy reasoning layer in pairs; The output layer consists of three nodes, corresponding to △Kp, △Ki, and △Kd respectively. Each node in the output layer is connected to each node in the fuzzy inference layer.

3. A servo electric actuator precision force position control method according to claim 2, characterized in that: The calculation formula for each node in the input layer is as follows: ； Among them, x1 is e, x2 is ec; The calculation formula of each node in the fuzzy layer is as follows: ； Among them, m 1b 、l 1b , b=1,2,...,n are the center and width of the membership function of each language value of e, m 1b 、l 1b , b=n+1,n+2,...,2n are the center and width of the membership function of each language value of ec respectively; The calculation formula for each node in the reasoning layer is as follows: ； Among them, b1 and b2 are two nodes in the fuzzification layer connected to the inference layer node c; The calculation formula of each node in the output layer is as follows: ； Among them, w dc is the weight connecting each node of the inference layer with the node d of the output layer, and f4(1), f4(2), and f4(3) are △Kp, △Ki, and △Kd respectively.

4. A servo electric actuator precision force position control method according to claim 3, characterized in that: When training the fuzzy neural network control model, the loss value is calculated using the following formula: ； Among them, g4(1), g4(2), and g4(3) are the expected outputs of △Kp, △Ki, and △Kd respectively; The fuzzy neural network control model is trained by updating m through the following formula ab 、l ab 、w dc : ； ； ； Where n is the number of iterations, 0<λ<1, 0<Y<1.

5. The method for precise force and position control of a servo electric actuator according to claim 1, characterized in that: The segmented neural network model in S2 has three layers, namely input layer, hidden layer and output layer; The input layer consists of two nodes, corresponding to stress deviation e1 and position deviation e2 respectively; The hidden layer consists of m nodes, and each node in the input layer is connected to each node in the hidden layer, 3≤m≤12; The output layer consists of two nodes, which are the weighted coefficients α output by the stress control fuzzy neural network model and β output by the position control fuzzy neural network model respectively. Each node in the hidden layer is connected to each node in the output layer; 0<α<1, 0<β<1.

6. A servo electric actuator precision force position control method according to claim 5, characterized in that: When training the segmented neural network model, m is set to 3, 4, ..., and 12 for training; an evaluation function is set, and the value of the evaluation function is calculated when m is 3, 4, ..., and 12 for training, and m corresponding to the maximum value of the evaluation function is set as the number of nodes in the hidden layer of the segmented neural network model. The form of the evaluation function is as follows: ； Among them, j is the number of evaluation indicators, θ i The sum of is 1, F i is the value of the preset evaluation index i, F i,min is the minimum value of the evaluation index i preset during training for each m value, F i,max It is the maximum value of the evaluation index i preset during training for each m value.

7. A servo electric actuator precision force position control method according to claim 6, characterized in that: Before training the segmented neural network model, the following steps are performed to obtain θ i : A1: Construct a two-dimensional expert evaluation table, where each row is θ i , each column is an importance level, and each importance level is assigned a value from high to low; A2: Obtain the expert evaluation form and calculate the total score of each indicator H(θ i ), and calculate θ by the following formula i : 。

Images