Quadruped robot joint motor parameter adaptive adjustment method based on fuzzy control

Abstract

The invention relates to the field of quadruped robots, in particular to a fuzzy control-based quadruped robot joint motor parameter adaptive adjustment method. Comprising the following steps: S1, calculating a current error and an error change rate according to a current reference instruction and a feedback current; s2, dividing the current error and the error change rate into a minimum value, a small value, a medium value and a large value, performing fuzzification through a triangular membership function, forming a fuzzy set, and calculating the membership of the fuzzy set; s3, according to a fuzzy rule, obtaining fuzzy subsets and membership degrees of the proportionality coefficient and the integral coefficient; and S4, clear values of the proportionality coefficient and the integral coefficient are obtained through an inverse triangle membership function, and the clear values are transmitted to a current loop proportional-integral parameter. The control precision can be improved, the adaptability is enhanced, the electromagnetic noise influence is reduced, and the parameter adjusting process is simplified.

Description

Technical Field

[0001] The invention relates to the field of quadruped robots, in particular to a method for adaptively adjusting joint motor parameters of quadruped robots based on fuzzy control. Background Art

[0002] In the control of quadruped robots, precise control of joint motors is crucial. They are responsible for driving the joints of the robot's limbs to achieve complex gait and motion control. Since quadruped robots will encounter various complex terrains and external interference during walking, the motors need to have strong responsiveness to adapt to these changes and ensure that the robot walks stably. Especially when dealing with high dynamic response requirements, the performance of the current loop controller will directly affect the stability and motion performance of the entire system. In the prior art, on the one hand, the traditional fixed-parameter PI controller cannot automatically adjust parameters according to changes in load, resulting in a decrease in control performance in some cases, or even instability; on the other hand, the traditional PI control method requires a long adjustment time to restore the system to stability when dealing with these interferences, affecting the robot's motion performance and response speed. Summary of the invention

[0003] To solve the above problems, the present invention provides a method for adaptively adjusting the joint motor parameters of a quadruped robot based on fuzzy control, which can improve control accuracy, enhance adaptability, reduce the influence of electromagnetic noise and simplify the parameter adjustment process.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] A method for adaptively adjusting the parameters of a quadruped robot joint motor based on fuzzy control comprises the following steps:

[0006] S1: Calculate the current error and error change rate based on the current reference command and feedback current;

[0007] S2: Divide the current error and error change rate into extremely small, small, medium and large, fuzzify them through the triangular membership function to form a fuzzy set, and calculate the membership of the fuzzy set;

[0008] S3: According to the fuzzy rules, the fuzzy subsets and membership degrees of the proportional coefficient and the integral coefficient are obtained;

[0009] S4: Obtain clear values ​​of the proportional coefficient and the integral coefficient through the inverse trigonometric membership function, and transfer the clear values ​​to the current loop proportional integral parameters.

[0010] Furthermore, the current error and error change rate include: the error change rate is filtered out with a first-order low-pass filter to remove high-frequency sampling errors, and the error and error change rate are normalized within the range of 0 to 1.

[0011] Furthermore, the calculation of the membership of the fuzzy set includes the following steps:

[0012] S21: Divide the current error and error change rate into four fuzzy sets of extremely small, small, medium and large, represented as {Z0, PS, PM, PB}, and determine the domain in the range of 0 to 1 according to the normalization of the error and error change rate;

[0013] S22: For each fuzzy set, determine the membership change of the fuzzy set in different intervals through the coordinates of the three vertices (a, b, c) of the triangle in the triangle membership function;

[0014] S23: Given the value x of the current error or error change rate, determine the interval to which the value x belongs in {Z0, PS, PM, PB}, and calculate the membership according to the calculation formula of the triangle membership function;

[0015] In step S23, the calculation formula of the triangle membership function is:

[0016] When a≤x≤b, When b≤x≤c, Among them, U represents the degree of membership.

[0017] Furthermore, the fuzzy rules include: a current loop open-loop system transfer function, the formula is: where k p is the proportionality coefficient, k i is the integral coefficient, s is the Laplace operator, R is the motor resistance, L q is the motor q-axis inductance;

[0018] k p and k i The setting formula is: where ω c is the system bandwidth;

[0019] The formula for adjusting the current loop closed-loop transfer function to a first-order inertial system is:

[0020] Furthermore, the system bandwidth ω c ,include:

[0021] When c When it is large, the system responds quickly and the motor generates large electromagnetic noise after being enabled;

[0022] When c hours, the system responds slowly, and the motor generates small electromagnetic noise after being enabled;

[0023] According to ω cThe size of the joint motor has four states: extremely small, small, medium, and large. The bandwidth values ​​corresponding to the four states are 0.5e3, 1e3, 1.5e3, and 2e3. The fuzzy sets of the four states are {Z0, PS, PM, and PB}.

[0024] K p The complete membership element of is:

[0025] {kp.ZO=0.5e3*R, kp.PS=1e3*R, kp.PM=1.5e3*R, kp.PB=2e3*R};

[0026] k i The complete membership element of is:

[0027] {ki.ZO=0.5e3*R, ki.PS=1e3*R, ki.PM=1.5e3*R, ki.PB=2e3*R}.

[0028] Furthermore, the fuzzy rules include:

[0029] When the quadruped robot is in motion, the joint motors are switched to large system bandwidth;

[0030] When the quadruped robot is in the standing state, the joint motors are switched to a small system bandwidth;

[0031] When the quadruped robot switches from a standing state to a moving state, the joint motors immediately switch from a small system bandwidth to a large system bandwidth.

[0032] Furthermore, the fuzzy rules include:

[0033] When there is a large current error and a small error change rate, the joint motor will increase the integral coefficient individually;

[0034] When a small current error or a large error change rate occurs, the joint motor will reduce the proportional coefficient and the integral coefficient.

[0035] Furthermore, the obtaining of the fuzzy subsets and the membership degree of the proportional coefficient and the integral coefficient includes:

[0036] Traversing step S2, the current error and the error change rate obtain the membership degrees respectively, select the maximum membership degree, and judge the fuzzy subsets of extremely small, small, medium and large in which the current error and the error change rate are located according to the maximum membership degree. By querying the fuzzy subsets of the proportional coefficient and the integral coefficient belonging to the fuzzy subsets of the current error and the error change rate, the control parameter fuzzy subset is determined, and the maximum value of the error and the error change rate membership degree is taken as the membership degree of the proportional coefficient and the integral coefficient.

[0037] Furthermore, obtaining clear values ​​of the proportional coefficient and the integral coefficient through the inverse triangular membership function includes:

[0038] According to the obtained fuzzy subsets and membership of the proportional coefficient and the integral coefficient, the clear values ​​of the proportional coefficient and the integral coefficient are obtained through the inverse triangular membership function;

[0039] The inverse trigonometric membership function, in the trigonometric function, when one membership corresponds to two proportional coefficients or integral coefficients, selects the smaller coefficient of the two coefficients as the clear value of the proportional coefficient, and selects the larger coefficient of the two coefficients as the clear value of the integral coefficient.

[0040] Further, the transmitting of the clear value to the current loop proportional integral parameter includes:

[0041] Determine that the clarity value is within a reasonable range, update the proportional part parameters and the integral part parameters of the PI controller, and after updating the PI controller parameters, start the joint motors of the quadruped robot and observe the operating status, which includes current response, speed change and position accuracy.

[0042] The beneficial effects of the present invention are:

[0043] 1. Through the application of fuzzy processing and fuzzy rules, various system states are considered more carefully, avoiding the inadaptability of traditional fixed parameter control methods in complex situations, making the control more accurate. The proportional coefficient and integral coefficient can be automatically adjusted according to the real-time current error and error change rate, so that the joint motor can track the expected current value more accurately under different working conditions, thereby improving the control accuracy of the quadruped robot. The fuzzy control-based method does not require manual adjustment of parameters. Adaptive adjustment of parameters can be achieved through fuzzy rules and automatic reasoning processes, which greatly simplifies the parameter adjustment process, not only saving debugging time and labor costs, but also improving the accuracy and efficiency of parameter adjustment, so that the quadruped robot can be put into practical application more quickly.

[0044] 2. The characteristics of the joint motors of the quadruped robot will change under different terrains, loads and motion states. The fuzzy control-based method automatically adapts to these changes without manual intervention to adjust parameters, greatly enhancing the adaptability of the quadruped robot to different working environments. Whether on flat ground or rugged terrain, whether light or heavy load, this method can adjust parameters to achieve the best control effect, improving the reliability and practicality of the robot.

[0045] 3. Find a balance between system response speed and electromagnetic noise by reasonably adjusting the system bandwidth. When the system needs to respond quickly, the bandwidth can be appropriately increased, but it will also generate greater electromagnetic noise; when the noise requirement is high, the bandwidth can be reduced. Although the system response speed will be slower, the electromagnetic noise will also be reduced. This adaptive adjustment capability is optimized according to actual needs, reducing the interference of electromagnetic noise on other electronic components of the quadruped robot and improving the stability and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a control block diagram of the adaptive adjustment method of the joint motor parameters of a quadruped robot.

[0047] Figure 2 It is a schematic diagram of the fuzzification of error and error change rate.

[0048] Figure 3 It is the current loop control block diagram of the joint motor.

[0049] Figure 4 is a fuzzy rule table.

[0050] Figure 5 It is a schematic diagram of defuzzification of error and error change rate. DETAILED DESCRIPTION

[0051] See also Figure 1-5 As shown, the present invention relates to a method for adaptively adjusting the parameters of a quadruped robot joint motor based on fuzzy control, comprising the following steps:

[0052] S1: Calculate the current error and error change rate based on the current reference command and feedback current;

[0053] S2: Divide the current error and error change rate into extremely small, small, medium and large, fuzzify them through the triangular membership function to form a fuzzy set, and calculate the membership of the fuzzy set;

[0054] S3: According to the fuzzy rules, the fuzzy subsets and membership degrees of the proportional coefficient and the integral coefficient are obtained;

[0055] S4: Obtain clear values ​​of the proportional coefficient and the integral coefficient through the inverse trigonometric membership function, and transfer the clear values ​​to the current loop proportional integral parameters.

[0056] Specifically, this method is as follows Figure 1 As shown in the figure, i q represents the q-axis current of the motor, represents the error change rate, k p represents the proportionality coefficient, k iRepresents the integral coefficient. First, the current error is calculated. According to the current reference instruction and the feedback current, the current error and the error change rate are calculated. Then, fuzzification is performed to fuzzify the current error and the error change rate to obtain the membership of the corresponding fuzzy set. For the application of fuzzy rules, the fuzzy subset with the largest membership in the fuzzification process is selected, and the fuzzy subsets of the proportional coefficient and the integral coefficient and their corresponding membership are determined according to the designed fuzzy rules. Finally, defuzzification is performed. Through the defuzzification process, the specific proportional coefficient and integral coefficient values ​​are obtained from the fuzzy set and its membership, and the current loop PI parameters are changed. The proportional coefficient and integral coefficient parameters calculated by the fuzzy controller are applied to the PI controller of the current loop to perform current loop control.

[0057] In the control of the joint motor of the quadruped robot, the PI controller automatically adjusts the values ​​of the proportional coefficient and the integral coefficient according to the real-time motor operation status and the desired target to achieve precise control of the motor. Through the fuzzy control method, the PI parameters are dynamically adjusted according to different working conditions and load changes to improve the adaptability and stability of the robot. For example, when the quadruped robot walks on different terrains, the load will change. At this time, the PI controller adjusts the parameters to make the motor output the appropriate torque to ensure the smooth walking of the robot. When climbing a slope, a larger torque is required. The PI controller increases the values ​​of the proportional coefficient and the integral coefficient to increase the output power of the motor; when walking on flat ground, the parameter value is appropriately reduced to reduce the energy consumption of the motor.

[0058] Furthermore, the current error and error change rate include: the error change rate is filtered out with a first-order low-pass filter to remove high-frequency sampling errors, and the error and error change rate are normalized within the range of 0 to 1.

[0059] Specifically, for obtaining the current reference command and feedback current, the current reference command is usually the expected current value determined according to the motion control requirements of the quadruped robot, while the feedback current is the actual current value of the joint motor measured in real time by devices such as current sensors. When calculating the current error, the current error is defined as the difference between the current reference command and the feedback current, that is: current error (e) = current reference command - feedback current. For calculating the error change rate, it is necessary to obtain the current error value at different time points. The error change rate is defined as the difference between the current errors at two adjacent time points divided by the time interval, that is: error change rate (ec) = (current error at the current time point - current error at the previous time point) / time interval.

[0060] In the actual sampling process, high-frequency noise and interference may be introduced due to various factors, and these high-frequency sampling errors will affect the accuracy of the error change rate. Through a first-order low-pass filter, these high-frequency errors can be effectively filtered out, making the error change rate data more stable. High-frequency sampling errors can cause instability in the control system, especially in the parameter adaptive adjustment based on fuzzy control. Inaccurate error change rate will cause errors in the judgment of fuzzy rules, thus affecting the control effect. Filtering high-frequency errors can enhance the robustness of the system and improve the system's resistance to various interferences. An accurate error change rate is very important for adjusting the proportional coefficient and the integral coefficient. The error change rate after low-pass filtering provides more reliable information, enabling the fuzzy control algorithm to more accurately determine the appropriate parameter adjustment direction and amplitude, thereby improving the control performance of the system.

[0061] After normalization, errors and error change rates of different magnitudes are unified into the same scale range, making subsequent fuzzification processing more convenient and accurate, and simplifying the design of fuzzy rules. Regardless of the original error and error change rate values, they can be fuzzily divided within the range of 0 to 1 to determine the degree to which they belong to each fuzzy set. If normalization is not performed, the error and error change rate may have very large or very small values. In fuzzy control, these extreme values ​​may lead to unstable or unreasonable results in the fuzzy reasoning process. Through normalization, the values ​​can be limited to a reasonable range, reducing the occurrence of such instability and facilitating comparison and analysis.

[0062] Furthermore, the calculation of the membership of the fuzzy set includes the following steps:

[0063] S21: Divide the current error and error change rate into four fuzzy sets of extremely small, small, medium and large, represented as {Z0, PS, PM, PB}, and determine the domain in the range of 0 to 1 according to the normalization of the error and error change rate;

[0064] S22: For each fuzzy set, determine the membership change of the fuzzy set in different intervals through the coordinates of the three vertices (a, b, c) of the triangle in the triangle membership function;

[0065] S23: Given the value x of the current error or error change rate, determine the interval to which the value x belongs in {Z0, PS, PM, PB}, and calculate the membership according to the calculation formula of the triangle membership function;

[0066] In step S23, the calculation formula of the triangle membership function is:

[0067] When a≤x≤b, When b≤x≤c, Among them, U represents the degree of membership.

[0068] Specifically, the current error and error change rate are divided into four fuzzy sets: extremely small (Z0), small (PS), medium (PM), and large (PB). The definitions of these fuzzy sets are relative and are used to describe the magnitude of the error and the error change rate. The error and error change rate are normalized in the range of 0 to 1, which means that all error and error change rate values ​​are mapped to this interval for fuzzy processing, such as Figure 2 Schematic diagram of the fuzzification of error and error change rate shown.

[0069] For each fuzzy set, through the triangle membership function, the shape of the triangle membership function is determined by three vertices, corresponding to different degrees of membership of the fuzzy set. Taking the error as an example, assume that the coordinates of the three vertices of the triangle are (a1, b1, c1), (a2, b2, c2), (a3, b3, c3), etc. The values ​​of these vertices are determined according to the specific fuzzy set and the domain range. For example: for the Z0 (extremely small) fuzzy set, the coordinates of the triangle vertices can be (0, 0.1, 0.2), which means that when the error is between 0 and 0.1, the membership gradually increases from 0 to 1; when it is between 0.1 and 0.2, the membership gradually decreases from 1 to 0, and beyond this range, the membership is 0. For the PS (small) fuzzy set, the vertex coordinates can be (0.1, 0.3, 0.5), when the error is between 0.1 and 0.3, the membership gradually increases from 0 to 1; when it is between 0.3 and 0.5, the membership gradually decreases from 1 to 0. For PM (medium) fuzzy sets, the vertex coordinates can be (0.4, 0.6, 0.8). When the error is between 0.4 and 0.6, the membership gradually increases from 0 to 1; when it is between 0.6 and 0.8, the membership gradually decreases from 1 to 0. For PB (large) fuzzy sets, the vertex coordinates can be (0.7, 0.9, 1). When the error is between 0.7 and 0.9, the membership gradually increases from 0 to 1; when it is between 0.9 and 1, the membership gradually decreases from 1 to 0.

[0070] In calculating the membership, for example, taking the error as an example, let's assume that an error value e = 0.15 is given, and determine the interval in which the error value e is located. 0.15 is in the overlapping interval of the PS (small) fuzzy set and the PM (medium) fuzzy set. Calculate the membership in the PS fuzzy set: The coordinates of the triangle vertices of the PS fuzzy set are: (0.1, 0.3, 0.5). Since 0.1≤0.15≤0.3, according to the formula Where a=0.1, b=0.3, x=0.15. The membership of the error value in PS (small) is 0.25. Similarly, the membership in the PM fuzzy set is calculated. The coordinates of the triangle vertices of the PM fuzzy set are (0.4, 0.6, 0.8). 0.15 is not in the main interval of the PM fuzzy set, but the membership can be determined by judging its relationship with the adjacent interval. Since 0.15 is less than the minimum vertex value of 0.4 of the PM fuzzy set, its membership in the PM fuzzy set is 0. The same method is used to calculate the error change rate ec. Suppose an error change rate value ec=0.45 is given. Determine the interval where ec is located, 0.45 is in the PM (medium) fuzzy set. Calculate the membership in the PM fuzzy set: The coordinates of the triangle vertices of the PM fuzzy set are (0.4, 0.6, 0.8). Since 0.4≤0.45≤0.6, according to the formula Substituting this into the equation, we get 0.25.

[0071] Furthermore, the fuzzy rules include: a current loop open-loop system transfer function, the formula is: where k p is the proportionality coefficient, k i is the integral coefficient, s is the Laplace operator, R is the motor resistance, L q is the motor q-axis inductance;

[0072] k p and k i The setting formula is: where ω c is the system bandwidth;

[0073] The formula for adjusting the current loop closed-loop transfer function to a first-order inertial system is:

[0074] Specifically, Figure 3 The current loop control block diagram of the joint motor shown in the figure shows the current loop open-loop system transfer function, which is used to describe the response characteristics of the joint motor current control system at different frequencies. By analyzing the frequency response of the transfer function, the important characteristics of the system such as stability, response speed and bandwidth can be understood. In practical applications, the current loop open-loop system transfer function is used to predict the performance of the joint motor under different working conditions. For the current loop closed-loop transfer function, it is adjusted to a first-order inertial system, which has a relatively simple mathematical form and is convenient for system analysis and design. By adjusting the current loop closed-loop transfer function to a first-order inertial system, some mature control theory methods are used to analyze the stability, response speed and performance indicators of the system. In the fuzzy control-based method, after the current loop closed-loop transfer function is adjusted to a first-order inertial system, it is more convenient to design fuzzy rules and determine the parameters of the fuzzy controller to achieve adaptive control of the joint motor, and it has the characteristics of good stability and reliability.

[0075] Furthermore, the system bandwidth ω c ,include:

[0076] When c When it is large, the system responds quickly and the motor generates large electromagnetic noise after being enabled;

[0077] When c hours, the system responds slowly, and the motor generates small electromagnetic noise after being enabled;

[0078] According to ω c The size of the joint motor has four states: extremely small, small, medium, and large. The bandwidth values ​​corresponding to the four states are 0.5e3, 1e3, 1.5e3, and 2e3. The fuzzy sets of the four states are {ZO, PS, PM, PB};

[0079] k p The complete membership element of is:

[0080] {kp.ZO=0.5e3*R, kp.PS=1e3*R, kp.PM=1.5e3*R, kp.PB=2e3*R};

[0081] k i The complete membership element of is:

[0082] {ki.ZO=0.5e3*R, ki.PS=1e3*R, ki.PM=1.5e3*R, ki.PB=2e3*R}.

[0083] Specifically, in the current loop control of the joint motor, the PI controller function and the motor function are multiplied to obtain the current loop open-loop system transfer function. The parameters of the proportional coefficient and the integral coefficient have specific settings. The system bandwidth has an important influence on the performance of the joint motor. When the system bandwidth is large, the system response speed is accelerated, but the motor will generate large electromagnetic noise after enabling; conversely, when the system bandwidth is small, the system response speed is slowed down, and the electromagnetic noise generated after the motor is enabled is small. More specifically, according to the size of the system bandwidth, the response performance of the joint motor is divided into four states: extremely small, small, medium, and large. The bandwidth values ​​corresponding to these four states are 0.5e3, 1e3, 1.5e3, and 2e3, respectively, and their fuzzy sets are expressed as {ZO (extremely small), PS (small), PM (medium), PB (large)}. By adjusting the system bandwidth, a trade-off is made between response speed and electromagnetic noise to meet the needs of different application scenarios. In the process, the proportional coefficient and the integral coefficient play a key role. The current error and error change rate are fuzzified, and the fuzzy subsets of the proportional coefficient and the integral coefficient and their membership are determined by fuzzy rules. Among them, the full membership element is of great significance in the process of determining the proportional coefficient and the integral coefficient: the full membership element represents the degree of complete belonging of the proportional coefficient and the integral coefficient to a certain fuzzy subset under specific conditions. When the error and the error change rate are in a specific state, the fuzzy subset to which the proportional coefficient and the integral coefficient belong is determined by querying the fuzzy rule table. At this time, the full membership element can help determine the value of the proportional coefficient and the integral coefficient more accurately, thereby achieving more precise adjustment of the joint motor parameters. When the fuzzy subset of the error and the error change rate is determined to be a certain state, if there is a full membership element in the corresponding fuzzy subset of the proportional coefficient and the integral coefficient, then the proportional coefficient and the integral coefficient value corresponding to the element can be more confidently applied to the adjustment of the current loop PI parameters to improve the control performance and stability of the joint motor.

[0084] Furthermore, the fuzzy rules include:

[0085] When the quadruped robot is in motion, the electronic control system needs to respond quickly, and the mechanical noise during the motion process is much greater than the electromagnetic noise, so the joint motor switches to a large system bandwidth;

[0086] When the quadruped robot is in a standing state, the electric control system is in a steady state, and the quadruped robot has only electromagnetic noise, then the joint motor switches to a small system bandwidth;

[0087] When the quadruped robot switches from a standing state to a moving state, the electronic control system needs to respond quickly. The mechanical noise during the movement process is much greater than the electromagnetic noise, so the joint motor immediately switches from a small system bandwidth to a large system bandwidth.

[0088] Specifically, regarding the impact of noise on quadruped robots, quadruped robots are usually equipped with various sensors to monitor parameters such as joint position, speed, and current. Noise can interfere with the signals of these sensors, resulting in inaccurate measurements, and noise can affect the stability of the control algorithm. Regarding the impact of electromagnetic noise on quadruped robots, it can interfere with the control signal of the motor and affect the normal operation of the motor. It directly affects the performance of the joint motor, and thus affects the motion control of the quadruped robot. Inaccurate motor parameter measurements can affect the accuracy of the parameter adaptive adjustment algorithm.

[0089] When the quadruped robot is in motion, the joint motors are switched to a large system bandwidth. Under a large system bandwidth, the system response speed is accelerated, which can meet the requirements for fast response of the electronic control system in motion. Although a large electromagnetic noise will be generated at this time, the increase in electromagnetic noise caused by the large system bandwidth is relatively insignificant because the mechanical noise during motion is much greater than the electromagnetic noise.

[0090] When the quadruped robot is in the standing state, the joint motor is switched to a small system bandwidth. Under the small system bandwidth, the system response speed is relatively slow, but the electromagnetic noise generated is small. Since the electronic control system is in a steady state in the standing state, the response speed requirement is not high, and electromagnetic noise becomes the main noise source of concern at this time. The small system bandwidth can effectively reduce electromagnetic noise.

[0091] When the quadruped robot switches from a standing state to a moving state, the response of the electronic control system needs to change quickly. To meet this requirement, the joint motor should immediately switch from a small system bandwidth to a large system bandwidth. This ensures that when the quadruped robot switches states, the electronic control system can quickly adjust to the needs of the moving state. At the same time, since the mechanical noise during the movement is much greater than the electromagnetic noise, the increase in electromagnetic noise caused by the large system bandwidth will not have a significant impact on the system performance.

[0092] Furthermore, the fuzzy rules include:

[0093] When there is a large current error and a small error change rate, the motor is in a large steady-state error state, and the joint motor alone increases the integral coefficient;

[0094] When a small current error and a large error change rate occur, the motor is in an unstable state, and the joint motor reduces the proportional coefficient and the integral coefficient.

[0095] Specifically, when a large current error and a small error change rate occur, the motor is in a large steady-state error state. At this time, in order to reduce the steady-state error, the joint motor individually increases the integral coefficient. The integral action is mainly used to eliminate the steady-state error. By increasing the integral coefficient, the system's accumulation effect on the error can be enhanced, so that the system gradually reduces the steady-state error and the motor tends to the desired operating state more quickly.

[0096] When a small current error and a large error change rate occur, the motor is in an unstable state. In order to stabilize the system again, the joint motor adjusts the proportional coefficient and the integral coefficient to a smaller value. The proportional coefficient mainly affects the system's response speed and the immediate response to the error. The integral coefficient is used to eliminate the steady-state error as mentioned above. Under a large error change rate, the system may oscillate or become unstable. Reducing the proportional coefficient can reduce the system's excessive response to the error and reduce the oscillation amplitude; at the same time, reducing the integral coefficient can avoid excessive accumulation of errors due to the integral action.

[0097] Through the above-mentioned joint motor parameter adjustment method for different current error and error change rate combinations, it is possible to effectively perform adaptive adjustment according to the actual operating state of the motor, improve the stability and reliability of the system, and enhance the performance of the quadruped robot under different working conditions. At the same time, the method of the present invention is adaptive and flexible, and can be dynamically adjusted according to the real-time operating state without manual intervention, providing strong technical support for the intelligent control of the quadruped robot.

[0098] Furthermore, the obtaining of the fuzzy subsets and the membership degree of the proportional coefficient and the integral coefficient includes:

[0099] Traversing step S2, the current error and the error change rate obtain the membership degrees respectively, select the maximum membership degree, and judge the fuzzy subsets of extremely small, small, medium and large in which the current error and the error change rate are located according to the maximum membership degree. By querying the fuzzy subsets of the proportional coefficient and the integral coefficient belonging to the fuzzy subsets of the current error and the error change rate, the control parameter fuzzy subset is determined, and the maximum value of the error and the error change rate membership degree is taken as the membership degree of the proportional coefficient and the integral coefficient.

[0100] Specifically, the current error and the error change rate are first fuzzified. Through a specific fuzzification method, the current error and the error change rate are respectively assigned memberships of different fuzzy sets (extremely small, small, medium, and large). Then, the memberships of the current error and the error change rate obtained after the fuzzification process are traversed, and the maximum membership is selected. According to this maximum membership, the specific fuzzy subset (one of the extremely small, small, medium, and large) in which the current error and the error change rate are located is determined. Then, by querying the pre-established fuzzy rule table, the fuzzy subset of the proportional coefficient and the integral coefficient to which the fuzzy subset of the current error and the error change rate belongs is determined. In this way, the fuzzy subset of the control parameter can be determined, providing a basis for further determining the specific control parameter value. Finally, the maximum value of the error and error change rate membership is used as the membership of the proportional coefficient and the integral coefficient. This approach enables the membership of the proportional coefficient and the integral coefficient to fully reflect the main characteristics of the current error and the error change rate, and improves the accuracy and reliability of the control parameter determination. The fuzzy rule table is as follows: Figure 4 shown.

[0101] In this way, the fuzzy subsets and membership of the proportional coefficient and integral coefficient can be adaptively determined according to the real-time status of the current error and the error change rate, thereby achieving accurate adjustment of the parameters of the quadruped robot joint motor. This method improves the adaptability and robustness of the system and can maintain good control performance under different working conditions. At the same time, through the application of fuzzy processing and fuzzy rules, the dependence on precise mathematical models is reduced, and the flexibility and versatility of the system are improved.

[0102] Furthermore, obtaining clear values ​​of the proportional coefficient and the integral coefficient through the inverse triangular membership function includes:

[0103] According to the obtained fuzzy subsets and membership of the proportional coefficient and the integral coefficient, the clear values ​​of the proportional coefficient and the integral coefficient are obtained through the inverse triangular membership function;

[0104] The inverse trigonometric membership function, in the trigonometric function, when one membership corresponds to two proportional coefficients or integral coefficients, selects the smaller coefficient of the two coefficients as the clear value of the proportional coefficient, and selects the larger coefficient of the two coefficients as the clear value of the integral coefficient.

[0105] Specifically, in the application of the inverse triangular membership function, the vertex coordinates of the triangular membership function of the fuzzy subset corresponding to the proportional coefficient are assumed to be (a1, b1, c1). According to the properties of the inverse triangular membership function, when the membership is When: When a1≤x≤b1, according to the formula Can get When b1≤x≤c1, according to the formula Can get This gives two possible solutions. For the integral coefficient, similar to the above, assume that the vertex coordinates of the triangle membership function of its fuzzy subset are (a2, b2, c2). When the membership is U ki When , two possible solutions can be obtained through similar derivation, such as Figure 5 shown.

[0106] For the proportionality coefficient, if the two solutions obtained by the previous derivation are k p 1 and k p 2, and k p 1<k p 2, then select k p 1 as k p For the integration coefficient, if the two solutions obtained are k i 1 and k i 2, and k i 1<k i 2, then select k i2 as k i clarity value.

[0107] Through the above steps, we can p , k i The fuzzy subsets and their memberships are used to obtain clear values ​​through the inverse triangular membership function, and a reasonable choice is made when there are two solutions, providing specific control parameter values ​​for the adaptive adjustment of the joint motors of the quadruped robot.

[0108] Further, the transmitting of the clear value to the current loop proportional integral parameter includes:

[0109] Determine that the clarity value is within a reasonable range, update the proportional part parameters and the integral part parameters of the PI controller, and after updating the PI controller parameters, start the joint motors of the quadruped robot and observe the operating status, which includes current response, speed change and position accuracy.

[0110] Specifically, after obtaining clear proportional coefficient and integral coefficient values, we must first check whether these values ​​are within a reasonable range. If the values ​​are too large or too small, it may cause system instability or poor control effect. Through the PI controller part of the current loop in the quadruped robot joint motor control system, the clear k p The value is assigned to the proportional coefficient, that is, the proportional part parameter in the PI controller is updated. i The value is assigned to the integral coefficient and the integral part parameters in the PI controller are updated.

[0111] After updating the PI parameters, perform a system test, start the quadruped robot joint motor, and observe its operating status, including current response, speed change, position accuracy, etc. If the system is found to be unstable, oscillating, or the response speed is not ideal, readjust the fuzzy control rules or the selection strategy of the inverse triangular membership function to obtain a more appropriate k p and k i value.

[0112] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for adaptively adjusting the joint motor parameters of a quadruped robot based on fuzzy control, characterized in that: The following steps are involved: S1: Calculate the current error and error change rate based on the current reference command and feedback current; S2: Divide the current error and error change rate into extremely small, small, medium and large, fuzzify them through the triangular membership function to form a fuzzy set, and calculate the membership of the fuzzy set; S3: According to the fuzzy rules, the fuzzy subsets and membership degrees of the proportional coefficient and the integral coefficient are obtained; S4: Obtain clear values ​​of the proportional coefficient and the integral coefficient through the inverse trigonometric membership function, and transfer the clear values ​​to the current loop proportional integral parameters.

2. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 1 is characterized in that: The current error and error change rate include: the error change rate is filtered out by a first-order low-pass filter to remove high-frequency sampling errors, and the error and error change rate are normalized within the range of 0 to 1.

3. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 2 is characterized in that: The calculation of the membership degree of the fuzzy set comprises the following steps: S21: Divide the current error and error change rate into four fuzzy sets of extremely small, small, medium and large, represented as {ZO, PS, PM, PB}, and determine the domain according to the normalization of the error and error change rate in the range of 0 to 1; S22: For each fuzzy set, determine the membership change of the fuzzy set in different intervals through the coordinates of the three vertices (a, b, c) of the triangle in the triangle membership function; S23: Given the value x of the current error or error change rate, determine the interval to which the value x belongs in {ZO, PS, PM, PB}, and calculate the membership according to the calculation formula of the triangle membership function; In step S23, the calculation formula of the triangle membership function is: When a≤x≤b, When b≤x≤c, Among them, U represents the degree of membership.

4. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 1 is characterized in that: The fuzzy rules include: the current loop open loop system transfer function, the formula is: where k p is the proportionality coefficient, k i is the integral coefficient, s is the Laplace operator, R is the motor resistance, L q is the motor q-axis inductance; k p and k i The setting formula is: where ω c is the system bandwidth; The formula for adjusting the current loop closed-loop transfer function to a first-order inertial system is:

5. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 4 is characterized in that: The system bandwidth ω c ,include: When c When it is large, the system responds quickly and the motor generates large electromagnetic noise after being enabled; When c hours, the system responds slowly, and the motor generates small electromagnetic noise after being enabled; According to ω c The size of the joint motor has four states: extremely small, small, medium, and large. The bandwidth values ​​corresponding to the four states are 0.5e3, 1e3, 1.5e3, and 2e3. The fuzzy sets of the four states are {ZO, PS, PM, PB}; k p The complete membership element of is: {kp.ZO=0.5e3*R, kp.PS=1e3*R, kp.PM=1.5e3*R, kp.PB=2e3*R}; k i The complete membership element of is: {ki.ZO=0.5e3*R, ki.PS=1e3*R, ki.PM=1.5e3*R, ki.PB=2e3*R}.

6. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 1 is characterized in that: The fuzzy rules include: When the quadruped robot is in motion, the joint motors are switched to large system bandwidth; When the quadruped robot is in the standing state, the joint motors are switched to a small system bandwidth; When the quadruped robot switches from a standing state to a moving state, the joint motors immediately switch from a small system bandwidth to a large system bandwidth.

7. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 1 is characterized in that: The fuzzy rules include: When there is a large current error and a small error change rate, the joint motor will increase the integral coefficient individually; When a small current error or a large error change rate occurs, the joint motor will reduce the proportional coefficient and the integral coefficient.

8. The method for adaptively adjusting the joint motor parameters of a quadruped robot based on fuzzy control according to claim 1, characterized in that: The obtaining of the fuzzy subsets and the membership degree of the proportional coefficient and the integral coefficient comprises: Traversing step S2, the current error and the error change rate obtain the membership degrees respectively, select the maximum membership degree, and judge the fuzzy subsets of extremely small, small, medium and large in which the current error and the error change rate are located according to the maximum membership degree. By querying the fuzzy subsets of the proportional coefficient and the integral coefficient belonging to the fuzzy subsets of the current error and the error change rate, the control parameter fuzzy subset is determined, and the maximum value of the error and the error change rate membership degree is taken as the membership degree of the proportional coefficient and the integral coefficient.

9. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 8, characterized in that: The method of obtaining clear values ​​of the proportional coefficient and the integral coefficient by using the inverse triangular membership function includes: According to the obtained fuzzy subsets and membership of the proportional coefficient and the integral coefficient, the clear values ​​of the proportional coefficient and the integral coefficient are obtained through the inverse triangular membership function; The inverse trigonometric membership function, in the trigonometric function, when one membership corresponds to two proportional coefficients or integral coefficients, selects the smaller coefficient of the two coefficients as the clear value of the proportional coefficient, and selects the larger coefficient of the two coefficients as the clear value of the integral coefficient.

10. The method for adaptively adjusting the parameters of the joint motors of a quadruped robot based on fuzzy control according to claim 1, characterized in that: The clear value is transmitted to the current loop proportional integral parameter, including: Determine that the clarity value is within a reasonable range, update the proportional part parameters and the integral part parameters of the PI controller, and after updating the PI controller parameters, start the joint motors of the quadruped robot and observe the operating status, which includes current response, speed change and position accuracy.

Images