Transient response characteristic-based mechanical arm intermittent control parameter optimization method
By establishing the impulse response model of the robot arm and analyzing the transient response characteristics, and optimizing the intermittent control parameters, the oscillation problem caused by transient response in the robot arm control is solved, the control accuracy and system stability are improved, and the model representation is simplified.
Patent Information
- Application Number
- CN202411022188.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-12
AI Technical Summary
Ignoring transient response processes in robotic arm control may lead to reduced oscillation and control accuracy, and even disrupt system stability.
By establishing the impulse response model of the robot arm, the transient response characteristics are analyzed, and the intermittent control parameters are optimized using fractional power function and cuckoo optimization algorithm to accurately characterize the transient response mechanism and optimize control performance of the robot arm.
Improves the robotic arm control accuracy, avoids oscillation, enhances system stability, and simplifies model representation, reduces computational complexity, and achieves better engineering implementation.
Smart Images

Figure CN118938678A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of robot control technology, and in particular to a method for optimizing intermittent control parameters of a robot arm based on transient response characteristics. Background Art
[0002] The robot arm is the actuator of the robot, and its control accuracy directly affects the robot's operation accuracy. Therefore, how to improve the control accuracy of the robot arm has become an important research topic. Usually, sensors are used to detect the output of the robot arm, and feedback controllers are designed according to classical control theory (such as Figure 1 ) Study the system output stability. Design a feedback controller based on the system stability constraints (such as Figure 1 ), and discretize the designed controller for computer implementation.
[0003] Since the response speed of the robot arm is relatively slow, in engineering, the output is often not continuously fed back to the input end. Figure 2 The intermittent control method shown is one of the more commonly used methods, and is studied using the following state transfer equation.
[0004]
[0005] according to This property (δ(t) is an ideal impulse response) shows that the impulse response model of the system is an idealized model, which is an idealized treatment of the actual system. Changes in the input immediately affect the output, and changes in the output are only related to the current input and have nothing to do with the historical process.
[0006] However, any physical process cannot be completed instantly, and there is a response process. Different systems have different response characteristics, and the duration of this process varies. The response of the robot arm is often slow, and there is often a transient response process when the input switches. Ignoring this transient response process and studying the robot arm control may cause the robot arm control oscillation, reduce the robot arm control accuracy, and even destroy the stability of the robot arm system, causing the system to crash. Summary of the invention
[0007] In order to solve the problems of mechanical arm oscillation and reduced control accuracy caused by the transient process of intermittent control, the purpose of the present invention is to provide a method for optimizing mechanical arm intermittent control parameters based on transient response characteristics. In view of the problem of intermittent control mechanical arm system oscillation caused by transient response, the present invention establishes an impulse response model of the mechanical arm control system, memorizes the system response characteristics, and analyzes the transient response characteristics of the mechanical arm under intermittent control on this basis, establishes the relationship between transient response and control parameters and control process, optimizes the mechanical arm intermittent control parameters, and optimizes the mechanical arm control performance.
[0008] The technical solution of the present invention to solve the above technical problems is as follows:
[0009] The intermittent control parameter optimization method of the robot arm based on transient response characteristics includes the following steps:
[0010] Step 1: Use fractional power function to establish the input and output response model of the robot arm;
[0011] Step 2: Based on the system response model, combined with the relationship between the system transient process under intermittent control and the current input and historical process, the transient response characteristics of the intermittent control of the manipulator are obtained;
[0012] Step 3: Based on the transient response characteristics of the intermittent control of the robot arm, the intermittent control parameters of the robot arm are optimized to obtain the optimal control parameters.
[0013] Furthermore, the fractional power function is used to establish the input-output response model of the robot arm. The process is as follows:
[0014] Step 1.1: Establish the input and output response model of the robot arm:
[0015] y(t)=h(t)*x(t), where h(t) is the system impulse response, x(t) is the robot control input, y(t) is the robot output, and t is time;
[0016] Step 1.2: Express h(t) as a fractional power function: h(t) = βt -α , where α and β are constants to be determined;
[0017] Step 1.3: Discretize y(t), x(t) and h(t) according to the sampling interval T to obtain N-point sequences Y N , X N and H N , the discrete convolution model is: Y N =H N *X N ;
[0018] Step 1.4: Use Legendre transform to transform the discrete convolution model in step 1.3 into an algebraic model and get:
[0019] (H N *X N ) T =βΨ N,N X N T (3)
[0020] Among them Ψ N,N is the Legendre transformation matrix,
[0021]
[0022] Where Γ is the gamma function, η i =(i+1) α+1 -2i α+1 +(i-1) α+1 , i=1,2,…,N;
[0023] Step 1.5: Experimental data collection to obtain Y N and X N ;
[0024] Step 1.6: Construct the objective function, use the cuckoo optimization search algorithm to optimize the objective function, identify the unknown parameters α and β, and establish the system impulse response model.
[0025] Furthermore, in step 1.6, the objective function is constructed and optimized using the cuckoo optimization search algorithm to identify the unknown parameters α and β. The process is as follows:
[0026] Using the input and output data collected in step 1.5 and the mean square error of the output estimate obtained by solving the model according to formula (3) as the objective function, M bird nest positions are generated for the parameters α and β to be identified, and the objective function of the current bird nest is calculated;
[0027] Set the step vector, randomly walk around the current bird's nest position with the step vector as the unit, update the bird's nest position, calculate the objective function of the bird's nest position, and realize global update;
[0028] According to the current bird's nest position, two bird's nest positions are randomly selected, the bird's nest position is locally updated, the objective function is calculated, and the local update is realized;
[0029] The above steps are repeated until the objective function is less than the threshold value, and the iteration ends to obtain the optimal estimate of the unknown parameters α and β.
[0030] Furthermore, the transient response characteristics of the intermittent control of the robot arm are analyzed as follows:
[0031] Step 2.1: Based on the upper and lower limits of the input and the allowable error range, h(t) with a length of L is intercepted as the approximate impulse response.
[0032] Step 2.2: Express the output as a unit step sequence: in is the system output estimate obtained from the model solution, yes The discretization result has a length of L', which represents the number of memory sequences and satisfies L' = L / T, where T is the sampling time interval;
[0033] Introducing the unit step sequence, we get: u(n) is the discrete sampling of the unit step function u(t);
[0034] According to discrete signal convolution, The length of is L'+N-1, The first section with a length of N is a steady-state term, and the last section with a length of L'-1 is a transient term;
[0035] Step 2.3: Separate the steady-state term and the transient term;
[0036]
[0037] Among them, the output before the intermittent control signal stops working is the steady-state term; the transient term u(nN) means u(n) shifted right by N bits;
[0038] Step 2.4: Introduce the unit step sequence, extract the historical information of length L' to analyze the system transient, and simplify the transient term;
[0039] Step 2.5: Under the input dynamics, according to the simplified transient term in step 2.4, calculate the difference between the maximum and minimum values of the transient term to obtain the amplitude of the transient oscillation of the robot arm.
[0040] Furthermore, a unit step sequence is introduced in step 2.4, and the transient term is expressed as:
[0041]
[0042] in Indicates the impact of the input before L' number of the current sequence on the transient state, according to The length of is L', then the influence of the previous input on the transient state is zero;
[0043] Get the transient item of the robot arm
[0044] Furthermore, on the basis of obtaining the transient response characteristics of the intermittent control of the robot arm, the control parameters are optimized, and the steps are as follows:
[0045] Constructing cost function in is the transient item of the robot arm;
[0046] Utilize the cuckoo search algorithm to optimize control parameters, including:
[0047] Generate M bird nest locations for the control parameters and calculate the cost function of the current bird nest
[0048] Set the step vector, randomly walk around the current bird's nest position with the step vector as the unit, update the bird's nest position, calculate the cost function of the bird's nest position, and achieve global update;
[0049] According to the current bird nest position, two bird nest positions are randomly selected, the bird nest positions are locally updated, and the cost function is calculated to achieve local update;
[0050] The above steps are repeated until the cost function is less than the threshold, the iteration ends, and the optimal control parameters are obtained.
[0051] Beneficial effects: Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects:
[0052] 1. The present invention establishes a transient model of intermittent control of a robotic arm based on the system impulse response, which is more in line with the response characteristics of the robotic arm, accurately describes the transient response mechanism of the robotic arm, and has a clearer physical meaning.
[0053] 2. The present invention uses fractional power functions to establish an impact response model, and simply represents the complex response model of the robotic arm with very few parameters, making the model simpler.
[0054] 3. The present invention intercepts finite-length historical information according to the attenuation characteristics to analyze the intermittent control and optimization of the robot arm, taking into account the historical dynamic information and calculation complexity of the robot arm, and has both good performance and easy engineering implementation. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] Figure 1 This is the structural diagram of the continuous feedback system.
[0056] Figure 2 This is the structural diagram of the discontinuous feedback system.
[0057] Figure 3 This is a framework diagram of the method of the present invention.
[0058] Figure 4 Create a flow chart for the robot arm response model.
[0059] Figure 5 This is the transient response characteristic analysis diagram of the intermittent control of the robotic arm.
[0060] Figure 6 Flowchart for optimizing intermittent control parameters of the robotic arm.
[0061] Figure 7 This is the transient response diagram when the system response length is 2 and the input signal length is 3. DETAILED DESCRIPTION
[0062] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0063] The method for optimizing intermittent control parameters of a robot arm based on transient response characteristics of the present invention comprises the following steps: Figure 3 The three steps shown are implemented as follows:
[0064] Step 1: Build a robot response model.
[0065] The structure of the robot arm is complex, and it is difficult to establish a system impulse response model based on its mechanism. The present invention uses a fractional power function to establish an input-output response model of the robot arm. Figure 4 As shown, the steps are as follows;
[0066] Step 1.1: Establish the input-output response model of the robot arm.
[0067] Any system input and output can be studied using input-output response characteristics. The impulse response of any physically realizable system is not an ideal impulse response model. Generally, the impulse response function can be represented by a time function h(t). Based on this, the input-output response model of the robot arm is established as follows:
[0068] y(t)=h(t)*x(t) (1)
[0069] Where h(t) is the system impulse response, x(t) represents the robot control input, and y(t) represents the robot output.
[0070] Step 1.2: Express h(t) as a fractional power function.
[0071] The impulse response of any physically realizable system must be absolutely integrable. At the same time, under the action of the pulse signal, its response decays with time. According to the intermediate equivalent process theory, it can always be equivalently represented by a fractional power function. Based on the above two points, the present invention assumes that the impact response of the robot arm is: h(t) = βt -α , where α and β are unknown constants. Once the unknown constants are identified, the robot response model can be established.
[0072] Step 1.3: Discretize y(t), x(t) and h(t) according to the sampling interval T to obtain N-point sequences Y N , X N and H N , respectively satisfying the following relations:
[0073] Y N =[y(T),y(2T),…,y(NT)],
[0074] X N =[x(T),x(2T),…,x(NT)],
[0075] H N =[h(T),h(2T),…,h(NT)].
[0076] Y N =H N *X N (2).
[0077] Step 1.4: Use Legendre transform to transform the discrete convolution model of formula (2) into an algebraic model.
[0078] According to Legendre transformation:
[0079] (H N *X N ) T =βΨ N,N X N T (3)
[0080] Among them Ψ N,N is the Legendre transformation matrix,
[0081]
[0082] Where Γ is the gamma function, η i =(i+1) α+1 -2i α+1 +(i-1) α+1 , i=1,2,…,N.
[0083] Step 1.5: Experimental data collection to obtain Y N and X N .
[0084] Under working conditions, the robot arm is controlled and the input and output data are sampled at time intervals T to obtain Y N and X N .
[0085] Step 1.6: Construct the objective function, use the cuckoo optimization search algorithm to optimize the objective function, identify the unknown parameters α and β, and establish the system impulse response model.
[0086] Using the mean square error of the experimental output data and the output estimate obtained by solving the model as the objective function, M bird nest positions are generated for the parameters to be identified α and β, and the objective function of the current bird nest is calculated. Set the step vector, randomly walk around the current bird nest position with the step vector as the unit, update the bird nest position, calculate the objective function of the bird nest position, and achieve global update; according to the contemporary bird nest position, randomly select two bird nest positions, locally update the bird nest position, calculate the objective function, and achieve local update. Repeat the above steps until the objective function is less than the threshold, end the iteration, and obtain the optimal estimate of the unknown parameters α and β.
[0087] Step 2: Analysis of transient response characteristics of intermittent control of the robot arm, such as Figure 5 As shown. The attenuation characteristics of the analytical response model h(t) reflect the impact of the current input on the future output, and also reflect the impact of the input before the current moment on the current output.
[0088] Step 2.1: Analyze the half-cycle transient response characteristics under intermittent control.
[0089] One cycle of intermittent control includes a working period and an idle period, two half-cycles. Since the historical process of each half cycle is different, its transient response characteristics are different and are studied separately.
[0090] Due to the system response characteristics, the output does not cut off when the input is cut off (such as Figure 7 , the transient response diagram of the system response length is 2 and the input signal length is 3), and the transient process of the system under the current control width is analyzed.
[0091] The transient process of the system is not only related to the current input, but also to the historical process. If the length of the system impulse response is L, the input within the time period of length L before the current moment will affect the system output.
[0092] For a given system, the transient response is related to the input. According to the upper and lower limits of the input and the allowable error range, L is reasonably selected, and h(t) with a length of L is taken as the approximate impulse response function.
[0093] Step 2.2: Express the output as a sequence of unit steps.
[0094]
[0095] in is the system output estimate obtained from the model solution, yes The discretization result has a length of L', which represents the number of memory sequences and satisfies L'=L / T, where T is the sampling time interval.
[0096] Introduce the unit step sequence u(n):
[0097]
[0098] u(n) is a discrete sample of the unit step function u(t).
[0099] According to the discrete signal convolution, The length is L'+N-1. The first section with a length of N is a steady-state term, and the second section with a length of L'-1 is a transient term.
[0100] Step 2.3: Separate the steady-state term and the transient term.
[0101]
[0102] Where u(nN) means u(n) shifted right by N bits; The segment is the output before the intermittent control signal stops acting, which is a steady-state item; It is a transient term and gradually decays over time.
[0103] Step 2.4: Simplify the transient items and extract the historical information of length L' to analyze the system transient.
[0104] Further introducing the unit step sequence, the transient term can be expressed as:
[0105]
[0106] in Indicates the impact of the input before L' number of the current sequence on the transient state. The approximate length of is taken as L', and the influence of the previous input on the transient is 0 or negligible. Therefore, the transient term of the robot arm can be obtained
[0107] Step 2.5: Under dynamic input (according to the memory decay characteristics, only the data with a length of L' before the current moment is selected), according to By calculating the difference between the maximum and minimum values of the transient term, we can get the amplitude of the transient oscillation of the manipulator. The transient oscillation of the manipulator will inevitably affect the control uncertainty, and the oscillation amplitude determines the range of control uncertainty. The smaller the transient oscillation amplitude, the higher the control accuracy.
[0108] Step 3: Optimization of intermittent control parameters of the robotic arm.
[0109] On the basis of obtaining the transient response characteristics of the intermittent control of the robot arm, the control parameters are further optimized to optimize the transient performance of the robot arm. Figure 6 As shown, the specific steps are as follows:
[0110] Step 3.1: Construct the cost function.
[0111] Transient performance is optimal, that is, the cost function Minimum.
[0112] Step 3.2: Optimize control parameters using the cuckoo search algorithm.
[0113] For a given robot arm, its impulse response is determined, and the transient performance can only be adjusted by changing the control input. For intermittent control, the control parameters include control period, control width and control gain. Search with the cuckoo search algorithm:
[0114] Generate M bird nest locations for the control parameters and calculate the cost function of the current bird nest Set the step vector, randomly walk around the current nest position with the step vector as the unit, update the nest position, calculate the objective function of the nest position, and achieve global update; according to the current nest position, randomly select two nest positions, locally update the nest position, calculate the objective function, and achieve local update. Repeat the above steps until the objective function is less than the threshold, end the iteration, and get the optimal control parameters.
Claims
1. A method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics, characterized in that: The following steps are involved: Step 1: Use fractional power function to establish the input and output response model of the robot arm; Step 2: Based on the system response model, combined with the relationship between the system transient process under intermittent control and the current input and historical process, the transient response characteristics of the intermittent control of the manipulator are obtained; Step 3: Based on the transient response characteristics of the intermittent control of the robot arm, the intermittent control parameters of the robot arm are optimized to obtain the optimal control parameters.
2. The method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics according to claim 1, characterized in that: The input-output response model of the robot arm is established using the fractional power function. The process is as follows: Step 1.1: Establish the input and output response model of the robot arm: y(t)=h(t)*x(t), Where h(t) is the system impulse response, x(t) is the robot control input, y(t) is the robot output, and t is time; Step 1.2: Express h(t) as a fractional power function: h(t) = βt -α , where α and β are constants to be determined; Step 1.3: Discretize y(t), x(t) and h(t) according to the sampling interval T to obtain N-point sequences Y N , X N and H N , the discrete convolution model is: Y N =H N *X N ; Step 1.4: Use Legendre transform to transform the discrete convolution model in step 1.3 into an algebraic model and get: (H N *X N ) T =βΨ N,N X N T (3), Among them Ψ N,N is the Legendre transformation matrix, Where Γ is the gamma function, η i =(i+1) α+1 -2i α+1 +(i-1) α+1 , i=1,2,…,N; Step 1.5: Experimental data collection to obtain Y N and X N ; Step 1.6: Construct the objective function, use the cuckoo optimization search algorithm to optimize the objective function, identify the unknown parameters α and β, and establish the system impulse response model.
3. The method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics according to claim 2, characterized in that: In step 1.6, construct the objective function and use the cuckoo optimization search algorithm to optimize the objective function and identify the unknown parameters α and β. The process is as follows: Using the input and output data collected in step 1.5 and the mean square error of the output estimate obtained by solving the model according to formula (3) as the objective function, M bird nest positions are generated for the parameters α and β to be identified, and the objective function of the current bird nest is calculated; Set the step vector, randomly walk around the current bird's nest position with the step vector as the unit, update the bird's nest position, calculate the objective function of the bird's nest position, and realize global update; According to the current bird nest position, two bird nest positions are randomly selected, the bird nest position is locally updated, the objective function is calculated, and the local update is realized; The above steps are repeated until the objective function is less than the threshold value, and the iteration ends to obtain the optimal estimate of the unknown parameters α and β.
4. The method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics according to claim 2, characterized in that: The process of transient response characteristic analysis of intermittent control of the robotic arm is as follows: Step 2.1: Based on the upper and lower limits of the input and the allowable error range, h(t) with a length of L is intercepted as the approximate impulse response. Step 2.2: Express the output as a unit step sequence: in is the system output estimate obtained from the model solution, yes The discretization result has a length of L', which represents the number of memory sequences and satisfies L' = L / T, where T is the sampling time interval; Introducing the unit step sequence, we get: u(n) is the discrete sampling of the unit step function u(t); According to discrete signal convolution, The length of is L'+N-1, The first section with a length of N is the steady-state term, and the last section with a length of L'-1 is the transient term; Step 2.3: Separate the steady-state term and the transient term; Among them, the output before the intermittent control signal stops working is the steady-state term; the transient term u(nN) means u(n) shifted right by N bits; Step 2.4: Introduce the unit step sequence, extract the historical information of length L' to analyze the system transient, and simplify the transient term; Step 2.5: Under the input dynamics, according to the simplified transient term in step 2.4, calculate the difference between the maximum and minimum values of the transient term to obtain the amplitude of the transient oscillation of the robot arm.
5. The method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics according to claim 4, characterized in that: In step 2.4, a unit step sequence is introduced and the transient term is expressed as: in Indicates the impact of the input before L' number of the current sequence on the transient state, according to The length of is L', then the influence of the previous input on the transient state is zero; Get the transient item of the robot arm 6. The method for optimizing intermittent control parameters of a robotic arm based on transient response characteristics according to claim 1 or 4, characterized in that: On the basis of obtaining the transient response characteristics of the intermittent control of the robot arm, the control parameters are optimized. The steps are as follows: Constructing cost function in is the transient item of the robot arm; Utilize the cuckoo search algorithm to optimize control parameters, including: Generate M bird nest locations for the control parameters and calculate the cost function of the current bird nest Set the step vector, randomly walk around the current nest position with the step vector as the unit, update the nest position, calculate the cost function of the nest position, and achieve global update; According to the current bird nest position, two bird nest positions are randomly selected, the bird nest positions are locally updated, and the cost function is calculated to achieve local update; The above steps are repeated until the cost function is less than the threshold, the iteration ends, and the optimal control parameters are obtained.
