A design optimization method of a linear joint permanent magnet motor of a humanoid robot

Abstract

The application discloses a kind of humanoid robot linear joint permanent magnet motor design optimization method, this method includes: based on the mechanical characteristics of human gait cycle to establish knee joint connecting rod mechanism dynamics model, extract characteristic working condition point and establish screw-motor electromechanical matching criterion, using Latin hypercube sampling to obtain design sample point, establish permanent magnet motor multi-objective optimization model with limit torque, torque ripple, efficiency as target, then introduce evolutionary algorithm for global optimization, obtain Pareto non-inferior solution set;Finally, based on CRITIC-TOPSIS evaluation method, fusion objective data characteristics and subjective working condition demand, determine the optimal design scheme from Pareto non-inferior solution set.The design optimization method of the application can define the design requirements of permanent magnet motor, and can also scientifically distribute the target weight, ultimately can obtain the permanent magnet motor scheme highly matched with specific joint working condition, effectively improves the power performance and operating stability of humanoid robot linear joint.

Description

Technical Field

[0001] This invention relates to the field of motor design and optimization technology, and in particular to a design and optimization method for a linear joint permanent magnet motor for a humanoid robot. Background Technology

[0002] With the rise of the concept of "embodied intelligence," humanoid robots are seen as the next disruptive product after computers, smartphones, and new energy vehicles. As the core power unit of humanoid robots, joint modules provide flexible movement capabilities through deep integration of structure and function. Among them, the shape of linear joints perfectly matches the narrow installation space requirements of humanoid robots' thighs, arms, etc. However, there is a significant technical contradiction between the extremely limited installation space of linear joints and the demand for high-performance drives. Under stringent space constraints, simply pursuing high torque density often leads to deep magnetic circuit saturation, uncontrolled temperature rise suppression, and process interference caused by excessively high stator slot fill factor.

[0003] In current technologies, the performance indicators of linear joints in humanoid robots are deeply coupled and mutually restrictive under strict spatial constraints. Especially under complex working conditions, if the performance margin is redundant or insufficient, it will affect the reliability of the system operation. Traditional design methods based on a single indicator are no longer suitable for their complex and ever-changing dynamic operation requirements, which can easily lead to problems such as insufficient explosive force, uncontrollable vibration and temperature rise in joint modules in practical applications.

[0004] Therefore, there is an urgent need to design an optimization design method that can define the design requirements of permanent magnet motors and achieve a reasonable allocation of multi-objective weights, so that permanent magnet motors are better matched with the application conditions of linear joints. Summary of the Invention

[0005] To address the shortcomings of the prior art, this invention provides a design optimization method for a linear joint permanent magnet motor for humanoid robots, aiming to solve the problem mentioned in the background art where current motor designs struggle to balance stringent spatial constraints with high-performance dynamic requirements.

[0006] To achieve the above objectives, the present invention provides the following solution: A design optimization method for a linear joint permanent magnet motor for a humanoid robot includes the following steps: Step 1: Establish a dynamic model of the knee joint linkage mechanism based on the mechanical characteristics of the human gait cycle, and obtain the joint characteristic working conditions; Step 2: Based on the joint characteristic working points and inertial torque, establish the electromechanical matching criteria between the lead screw and the motor; Step 3: Use Latin hypercube sampling to obtain representative sample points within the design space; Step 4: Using the limiting torque, torque ripple, and efficiency of the permanent magnet motor as optimization objectives, and the magnetic flux density values ​​of each part of the motor as constraints, establish an optimization model for the permanent magnet motor. Step 5: Perform global optimization on the permanent magnet motor optimization model based on the evolutionary algorithm to obtain the Pareto non-dominated solution set; Step 6: Based on the CRITIC-TOPSIS evaluation method, integrate objective data characteristics with subjective working condition requirements, and select the optimal combination of motor parameters from the Pareto non-dominated solution set; Step 7: Based on the electromechanical matching criteria described in Step 2, verify the optimization effect of the selected optimal motor parameter combination.

[0007] Furthermore, in step 1, a dynamic model of the knee joint linkage mechanism is established based on the mechanical characteristics of the human gait cycle to obtain the joint characteristic working conditions, specifically including: Based on the mechanical characteristics of the human body under the actual gait cycle, a dynamic model of the knee joint linkage mechanism is established. Then, the established humanoid robot knee joint structure is simplified into a linkage mechanism. The mechanical characteristic curve of the knee joint is converted using the Lagrange method to obtain the motion curve of the joint bearing at the output end of the linear joint. Four working points are extracted from the motion curve as characteristic working points, thereby realizing the determination of the working conditions of the linear joint.

[0008] Furthermore, the four operating points include: Working condition 1: Equivalent continuous working condition. Considering the significant load fluctuations during gait, the root mean square value is used to characterize the equivalent load force in the complete walking process. Operating point 2: High acceleration condition, corresponding to the initial stage of the oscillating phase; Operating point 3: High-speed operating condition, corresponding to the mid-phase of the oscillation; Operating point 4: High load condition, corresponding to the foot contact stage at the end of the swing.

[0009] Furthermore, in step 2, the specific method for establishing the electromechanical matching criteria between the lead screw and the motor is as follows: Based on the kinematic transmission principle of the lead screw pair, the rotational performance index at the motor end and the linear performance index at the load end are converted using the following formula: ; Introducing the inertia compensation torque generated by the rotational inertia of the transmission components, the total torque demand equation at the motor end considering the influence of inertia is calculated. The specific formula is as follows: T total =T load + J α; in, This is the motor torque. FFor the lead screw thrust, L For the lead screw, Let n be the lead screw efficiency and n be the motor speed. v T is the linear velocity at the load end. total T represents the total torque at the motor terminals. load For load torque, J Let α be the moment of inertia on the motor side, and α be the angular acceleration. Using the extreme boundary of the motor's TN characteristic curve as the core criterion, we verify whether the motor's output envelope curve can completely cover all dynamic operating points in the gait cycle.

[0010] Furthermore, in step 3, representative sample points within the design space are obtained using Latin hypercube sampling, specifically including: Step 301: Divide the [0,1] value range of each design variable into... n A number of non-overlapping intervals with equal probability; Step 302: For the first j The first dimension of the variable i Each hierarchical interval is mapped to its corresponding upper and lower physical constraint limits using linear interpolation. The specific formula is as follows: ; in, Represents the lower limit of the physical constraints of the interval. A represents the upper limit of the physical constraints of the interval. j With B j Representing the first j The minimum and maximum values ​​of each design variable within the allowable range of the project; and Let be an interval partitioning function, satisfying =( i -1) / n and = i / n , ( i =1,2,... n ); Step 303: Use the Latin hypercube sampling algorithm to randomly select an independent sample point in each sub-interval, and associate the sample points of each dimension through random permutation and combination to form a sample matrix covering the whole. The operator for generating sample points is represented as: ; in, For the first j The variable in the first... i The specific values ​​in each sample; Let be an independent random variable that follows a uniform distribution (0,1). Step 304: Remove invalid samples caused by geometric interference or inability to divide the mesh, and use the selected valid individuals to form the first generation population.

[0011] Furthermore, in step 4, the optimization objectives are specifically: to maximize the limiting torque, minimize the cogging torque, and maximize the operating efficiency; The constraint is as follows: the magnetic flux density saturation threshold of the permanent magnet motor stator, rotor and air gap is introduced as a penalty function constraint.

[0012] Furthermore, in step 5, the permanent magnet motor optimization model is globally optimized based on an evolutionary algorithm, specifically including: Step 501: Use the valid individuals obtained from Latin hypercube sampling as the parent population, and set the crossover probability P. c =0.98; Step 502: Introduce a low-probability mutation mechanism and set the mutation rate P. m =0.01; Step 503: Perform non-dominated sorting and elite retention, and re-perform finite element simulation calculations for each newly generated population to obtain the limit torque, cogging torque and efficiency of the permanent magnet motor. Step 504: Stratify and classify the population according to the Pareto dominance relationship between individuals, and individuals on the Pareto front will be retained to the next generation of the population; Step 505: Set the maximum number of generations. When the number of iterations reaches the maximum number of generations, the algorithm terminates and outputs the Pareto non-dominated solution set.

[0013] Furthermore, step 6 specifically includes: Step 601: Construct the initial decision matrix based on the three evaluation indicators: limiting torque, cogging torque, and efficiency; Step 602: Normalize the original data to obtain a dimensionless matrix; Step 603: Calculate the weights of each evaluation indicator using information entropy; Step 604: Based on the different joint types of the humanoid robot, introduce the corresponding preference weight vector, and then use the multiplication synthesis normalization principle to calculate the final comprehensive weight; Step 605: Construct a weighted decision matrix using comprehensive weights, calculate the Euclidean distance from each alternative to the positive ideal solution and the negative ideal solution, then calculate the relative proximity of each alternative based on the obtained Euclidean distance, and finally sort the Pareto solution set according to the magnitude of the relative proximity to select the optimal combination of motor structure parameters.

[0014] Furthermore, step 7 specifically includes: The TN characteristic curve of the optimized permanent magnet motor is established, and envelope matching analysis is performed with the linear joint operating points under different lead screw side effects to generate the efficiency map of the optimized motor. The loss characteristics within the operating range are evaluated based on the efficiency map.

[0015] The present invention also provides a humanoid robot linear joint permanent magnet motor, which is controlled based on the above-mentioned design optimization method.

[0016] Compared with the prior art, the beneficial technical effects of the present invention are as follows: This invention establishes a dynamic model based on real human gait characteristics, extracting more engineering-representative characteristic working points, making the design goals more aligned with practical application needs. It also establishes a screw-motor electromechanical matching criterion considering inertial torque, enabling more accurate assessment of the motor's torque requirements during actual dynamic processes, achieving deep adaptation of motor performance to the complex working conditions of humanoid robot linear joints. Based on the three core objectives of permanent magnet motors—limit torque, torque ripple, and efficiency—this invention establishes a multi-objective optimization model, while introducing magnetic flux density saturation thresholds for each component as penalty function constraints, effectively avoiding local magnetic circuit oversaturation that may be caused by adjustments to design variables.

[0017] This invention introduces an improved CRITIC-TOPSIS evaluation method. On one hand, it objectively measures the data characteristics of each indicator through information entropy; on the other hand, it introduces adjustable subjective preference weight vectors based on different joint types, and uses the multiplicative synthesis normalization principle to determine the final comprehensive weight. This approach avoids the arbitrariness of purely subjective weighting and flexibly adapts to the performance emphasis of different application scenarios, thus more reliably selecting the optimal combination of motor structure parameters that highly fits the specific operating conditions from the Pareto non-dominated solution set. After optimization, this invention reuses the electromechanical matching criteria established in step 2, and uses TN envelope matching and efficiency map analysis to verify the effectiveness of the optimization scheme from multiple dimensions, further improving the credibility of the design results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the design and optimization method for a linear joint permanent magnet motor of a humanoid robot according to an embodiment of the present invention. Figure 2 This is a full-cycle time-domain response curve of the motor side according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the TN characteristics and operating envelope of a permanent magnet motor according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the optimized variables in an embodiment of the present invention; Figure 5 The flowchart is optimized for embodiments of the present invention; Figure 6 This is a diagram illustrating the determination of the optimal Pareto solution set in an embodiment of the present invention. Figure 7 This is a diagram showing the envelope matching results of the lead screw motor operating conditions in an embodiment of the present invention; Figure 8 This is a MAP result diagram of motor efficiency in an embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] like Figure 1 As shown, this invention discloses a design optimization method for a linear joint permanent magnet motor for a humanoid robot, comprising the following steps: Step 1: Establish a dynamic model of the knee joint linkage mechanism based on the mechanical characteristics of the human gait cycle, and obtain the joint characteristic working conditions; Specifically, this invention proposes an efficient and reliable method for determining joint working conditions: combining the mechanical characteristics of the human body under the actual gait cycle, a dynamic model of the knee joint linkage mechanism is established. Then, based on the established humanoid robot knee joint structure, it is simplified into a linkage mechanism. The mechanical characteristic curve of the knee joint is converted using the Lagrange method to obtain the motion curve of the joint bearing at the output end of the linear joint. Four working conditions are extracted from the motion curve as characteristic working conditions, thereby realizing the determination of the working conditions of the linear joint.

[0023] The four operating points are as follows: Operating point 1: Equivalent continuous operating condition, which uses the root mean square value to address the significant load fluctuations during gait cycles. F rmsThis represents the equivalent load force throughout the entire walking process. Working point 2: High acceleration working point, corresponding to the initial stage of the swing phase. At this time, the limb needs to rapidly accelerate from rest to peak swing speed, placing extremely high demands on the instantaneous acceleration and explosive force of the linear joints. Working point 3: High speed working point, corresponding to the mid-stage of the swing phase. To achieve rapid folding and extension of the joints to complete posture adjustments, the linear joints are required to have high operating speeds. Working point 4: High load working point, corresponding to the foot-to-ground contact stage at the end of the swing. At this time, it needs to withstand the reverse impact generated at the moment of ground contact, placing stringent demands on the peak thrust and mechanical stiffness of the linear joints.

[0024] Step 2: Based on the joint characteristic operating points and inertial torque, establish the electromechanical matching criteria between the lead screw and the motor, specifically as follows: Based on the kinematic transmission principle of the lead screw pair, the rotational performance index at the motor end and the linear performance index at the load end are converted using the following formula: (1) Considering the inertial compensation torque introduced by the rotational inertia of the transmission components when the humanoid robot performs high-frequency commutation or violent acceleration, the total torque demand equation at the motor end considering the influence of inertia is calculated, and the specific formula is as follows: T total =T load + J α (2) in, This is the motor torque. F For the lead screw thrust, L For the lead screw, Let n be the lead screw efficiency and n be the motor speed. v T is the linear velocity at the load end. total T represents the total torque at the motor terminals. load For load torque, J Let α be the moment of inertia on the motor side and α be the angular acceleration; the resulting full-cycle time-domain response curve on the motor side is shown in Figure 1. Figure 2 As shown.

[0025] Using the extreme boundaries of the motor's TN characteristic curve as the core criterion, we verify whether the motor's output envelope curve can completely cover all dynamic operating points in the gait cycle, referencing... Figure 3 .

[0026] Step 3: Obtain representative sample points within the design space using Latin hypercube sampling, specifically: Step 301: Divide the [0,1] value range of each design variable into... n A number of non-overlapping intervals with equal probability; Step 302: For the first j The first dimension of the variablei Each hierarchical interval is mapped to its corresponding upper and lower physical constraint limits using linear interpolation. The specific formula is as follows: (3) in, Represents the lower limit of the physical constraints of the interval. A represents the upper limit of the physical constraints of the interval. j With B j Representing the first j The minimum and maximum values ​​of each design variable within the allowable range of the project; and Let be an interval partitioning function, satisfying =( i -1) / n and = i / n , ( i =1,2,... n ); Step 303: Use the LHS algorithm (Latin Hypercube Sampling Algorithm) to randomly select an independent sample point in each sub-interval, and associate the sample points of each dimension through random permutation and combination to finally form a sample matrix covering the whole. The operator for generating sample points is represented as: (4) in, For the first j The variable in the first... i The specific values ​​in each sample; Let be an independent random variable that follows a uniform distribution (0,1). Step 304: Set the initial sample size to... n =600, invalid samples due to geometric interference or inability to mesh are removed, and the selected valid individuals constitute the first generation population. N 1.

[0027] Step 4: Using the limiting torque, torque ripple, and efficiency of the permanent magnet motor as optimization objectives, and the magnetic flux density values ​​of various parts of the motor as constraints, establish an optimization model for the permanent magnet motor, specifically as follows: First, identify the key structural parameters, such as Figure 4 As shown, since the stator slot structure directly modulates the air gap permeability and determines the magnetic saturation degree, the slot opening width, slot bottom width, and tooth height are selected as optimization variables on the stator side; on the rotor side, the focus is on magnetic field morphology control, and the magnetic pole thickness, pole arc coefficient, and eccentricity are selected as optimization objects to achieve the improvement of torque density and the suppression of pulsation. Considering that random combinations of design variables may lead to topological failures in the geometric model during multi-objective optimization, such as insufficient winding space or edge interference between magnetic poles and rotors, resulting in mesh generation failure, this embodiment introduces a set of normalized scaling coefficients to correlate structural dimensions. By converting absolute dimensions into dimensionless scaling coefficients, it ensures that all generated geometric variants satisfy physical constraints and assembly logic within the parameter fluctuation range, significantly improving the automation robustness of the simulation process. See Table 1.

[0028] Table 1. Mapping Table of Design Variables and Corresponding Coefficients

[0029] Furthermore, based on the high burst speed and precise transmission characteristics of humanoid robot joints, the optimization model establishes three core objectives: maximizing the limit torque. T max To meet the robot's instantaneous high overload motion requirements; to minimize cogging torque. T cog To reduce torque ripple in the transmission system, improve operational stability, and maximize operating efficiency. To reduce heat loss and extend the robot's runtime, the objective function is defined as follows: .

[0030] In addition, the saturation threshold of magnetic flux density of each part of the permanent magnet motor is introduced as a penalty function constraint. By strictly limiting the maximum magnetic flux density of the stator (tooth part, yoke part), rotor and air gap, local magnetic circuit oversaturation caused by the adjustment of design variables is prevented.

[0031] Step 5: Perform global optimization on the permanent magnet motor optimization model based on the evolutionary algorithm to obtain the Pareto non-dominated solution set. The detailed process is as follows: Figure 5 As shown, specifically: Step 501: Use the valid individuals obtained from LHS screening as the parent population, and set the crossover probability P. c =0.98; Step 502: To prevent the algorithm from prematurely converging to local extrema during evolution, a low-probability mutation mechanism is introduced, with a mutation rate P set. m =0.01; Step 503: Perform non-dominated sorting and elite retention, and sort the newly generated population in each generation. N 2. Finite element simulation calculations were performed again to obtain the target function values ​​such as the limiting torque, cogging torque and efficiency of the permanent magnet motor. Step 504: Assess the population based on Pareto dominance relationships among individuals. N 3. A stratified and graded system will be implemented, with elite individuals at the Pareto front being preserved for the next generation of the population. N 4; Step 505: Set the maximum number of generations. In this embodiment, the maximum number of generations is set. G max =60. When the number of iterations reaches the threshold, the algorithm terminates and outputs a Pareto non-dominated solution set containing a series of trade-off solutions.

[0032] Step 6: Based on the CRITIC-TOPSIS evaluation method, integrate objective data characteristics with subjective operating condition requirements, and select the optimal motor parameter combination from the Pareto non-dominated solution set, specifically as follows: Step 601: Based on the limiting torque T max Cogging torque T cog and efficiency These three evaluation metrics are used to construct the initial decision matrix X; Step 602: Since the dimensions of each physical quantity are different (for example, efficiency is a dimensionless percentage, and torque is measured in N·m), and there are conflicts in the index attributes, the original data is normalized and transformed into a standardized dimensionless matrix. ; Step 603: Determine the weights by measuring the information increment carried by each evaluation indicator. j Information entropy of each evaluation indicator e j The calculation formula is: (5) in, P ij For the first i The first scheme is in the j The proportion of features under each indicator k The objective weights are determined by setting them as constants. w j The calculation formula is: (6) Step 604: Based on the different joint types of the humanoid robot, introduce corresponding preference weight vectors. ,in, The value is adjusted based on the actual working conditions of different joints. For example, for hip and knee joints, which are power-driven joints, they need to directly bear the weight of the machine body and cope with the impact during walking. They are often in heavy-load start-up or high-burst working conditions. Therefore, the weighting coefficient of the limit torque needs to be increased during optimization. To ensure the robot's resistance to disturbances and high dynamic response performance; for control-driven joints such as the shoulder and wrist, the focus is on the trajectory tracking accuracy and motion smoothness of the end effector, and the weighting factor of torque stability needs to be increased during optimization. .

[0033] The final composite weight is calculated using the multiplicative composition normalization principle. The calculation formula is: (7) This embodiment provides a knee joint weight allocation table for a humanoid robot, as shown in Table 2: Table 2 Target Scope and Weight Allocation

[0034] Step 605: Utilize the comprehensive weights Construct a weighted decision matrix and define the positive ideal solution Z in the multidimensional space. + With negative ideal solution Z - : (8) Calculate the Euclidean distance from each alternative solution to the positive ideal solution. D i + Euclidean distance to the negative ideal solution D i - Then, based on the obtained Euclidean distance, the relative closeness C of each scheme is calculated. i : (9) C i The closer the value is to 1, the closer the solution is to the theoretical optimal solution, and the farther it is from the theoretical worst-case solution. Finally, based on the relative closeness C... i The Pareto solution set is sorted by its size, as shown in Table 3. The solution ranked first is the optimal combination of motor structure parameters under the weight configuration of this specific operating condition. Figure 6 As shown.

[0035] Table 3 Ranking of Pareto solution set fit

[0036] Step 7: Based on the electromechanical matching criteria described in Step 2, verify the optimization effect of the selected optimal motor parameter combination, specifically as follows: After comparing the optimization effects of each optimization objective, the TN characteristic curve of the optimized permanent magnet motor is established, and envelope matching analysis is performed with the linear joint operating points under different lead screw side effects, such as... Figure 7 As shown, to further evaluate the impact of different leads, an efficiency map of the optimized motor is generated to assess the loss characteristics within the operating range, such as... Figure 8 As shown.

[0037] This invention provides a design optimization method for a humanoid robot linear joint permanent magnet motor. Based on the real mechanical characteristics of the human gait cycle and considering the spatial arrangement of the linear joint at the knee joint, the method simplifies the linkage mechanism to calculate the characteristic working conditions of the linear joint and uses these as design objectives to establish an electromechanical matching criterion between the lead screw and the motor. Latin hypercube sampling is used to extract design parameter variations within the design variable range, randomly combining design parameters such as slot shoulder height, slot depth, upper slot width, lower slot width, magnetic pole thickness, and eccentric reduction. A multi-objective optimization model for the permanent magnet motor is established, with the maximum objective of limiting torque and efficiency, the minimum objective of torque ripple, and the magnetic flux density of each part as constraints. The optimization model is solved using an evolutionary algorithm and an improved CRITIC-TOPSIS evaluation method, comprehensively considering objective data characteristics and subjective working condition requirements to obtain an optimized permanent magnet motor scheme that matches the working conditions of the humanoid robot's linear joint. The optimization effect is verified through the electromechanical matching criterion, demonstrating significant research and engineering application value.

[0038] Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. Furthermore, those skilled in the art will recognize that, based on the ideas of this invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A design optimization method for a linear joint permanent magnet motor for a humanoid robot, characterized in that, Includes the following steps: Step 1: Establish a dynamic model of the knee joint linkage mechanism based on the mechanical characteristics of the human gait cycle, and obtain the joint characteristic working conditions; Step 2: Based on the joint characteristic working points and combined with the inertial torque, establish the electromechanical matching criteria between the lead screw and the motor; Step 3: Use Latin hypercube sampling to obtain representative sample points within the design space; Step 4: Using the limiting torque, torque ripple, and efficiency of the permanent magnet motor as optimization objectives, and the magnetic flux density values ​​of each part of the motor as constraints, establish an optimization model for the permanent magnet motor. Step 5: Perform global optimization on the permanent magnet motor optimization model based on the evolutionary algorithm to obtain the Pareto non-dominated solution set; Step 6: Based on the CRITIC-TOPSIS evaluation method, integrate objective data characteristics with subjective working condition requirements, and select the optimal combination of motor parameters from the Pareto non-dominated solution set; Step 7: Based on the electromechanical matching criteria described in Step 2, verify the optimization effect of the selected optimal motor parameter combination.

2. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, In step 1, a dynamic model of the knee joint linkage mechanism is established based on the mechanical characteristics of the human gait cycle to obtain the joint characteristic working conditions, specifically including: Based on the mechanical characteristics of the human body under the actual gait cycle, a dynamic model of the knee joint linkage mechanism is established. Then, the established humanoid robot knee joint structure is simplified into a linkage mechanism. The mechanical characteristic curve of the knee joint is converted using the Lagrange method to obtain the motion curve of the joint bearing at the output end of the linear joint. Four working points are extracted from the motion curve as characteristic working points, thereby realizing the determination of the working conditions of the linear joint.

3. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 2, characterized in that, The four operating points include: Working condition 1: Equivalent continuous working condition. Considering the significant load fluctuations during gait, the root mean square value is used to characterize the equivalent load force in the complete walking process. Operating point 2: High acceleration condition, corresponding to the initial stage of the oscillating phase; Operating point 3: High-speed operating condition, corresponding to the mid-phase of the oscillation; Operating point 4: High load condition, corresponding to the foot contact stage at the end of the swing.

4. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, In step 2, the specific method for establishing the electromechanical matching criteria between the lead screw and the motor is as follows: Based on the kinematic transmission principle of the lead screw pair, the rotational performance index at the motor end and the linear performance index at the load end are converted using the following formula: ； Introducing the inertia compensation torque generated by the rotational inertia of the transmission components, the total torque demand equation at the motor end considering the influence of inertia is calculated. The specific formula is as follows: T total =T load + J α; in, This is the motor torque. F For the lead screw thrust, L For the lead screw, Let n be the lead screw efficiency and n be the motor speed. v T is the linear velocity at the load end. total T represents the total torque at the motor terminals. load For load torque, J Let α be the moment of inertia on the motor side, and α be the angular acceleration. Using the extreme boundary of the motor's TN characteristic curve as the core criterion, we verify whether the motor's output envelope curve can completely cover all dynamic operating points in the gait cycle.

5. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, In step 3, representative sample points within the design space are obtained using Latin hypercube sampling, specifically including: Step 301: Divide the [0,1] value range of each design variable into... n A number of non-overlapping intervals with equal probability; Step 302: For the first j The first dimension of the variable i Each hierarchical interval is mapped to its corresponding upper and lower physical constraint limits using linear interpolation. The specific formula is as follows: ； in, Represents the lower limit of the physical constraints of the interval. A represents the upper limit of the physical constraints of the interval. j With B j Representing the first j The minimum and maximum values ​​of each design variable within the allowable range of the project; and Let be an interval partitioning function, satisfying =( i -1) / n and = i / n , ( i =1,2,... n ); Step 303: Use the Latin hypercube sampling algorithm to randomly select an independent sample point in each sub-interval, and associate the sample points of each dimension through random permutation and combination to form a sample matrix covering the whole. The operator for generating sample points is represented as: ; in, For the first j The variable in the first... i The specific values ​​in each sample; Let be an independent random variable that follows a uniform distribution (0,1). Step 304: Remove invalid samples caused by geometric interference or inability to divide the mesh, and use the selected valid individuals to form the first generation population.

6. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, In step 4, the optimization objectives are specifically: to maximize the limit torque, minimize the cogging torque, and maximize the operating efficiency; The constraint is as follows: the magnetic flux density saturation threshold of the permanent magnet motor stator, rotor and air gap is introduced as a penalty function constraint.

7. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 5, characterized in that, In step 5, the permanent magnet motor optimization model is globally optimized based on an evolutionary algorithm, specifically including: Step 501: Use the valid individuals obtained from Latin hypercube sampling as the parent population, and set the crossover probability P. c =0.98; Step 502: Introduce a low-probability mutation mechanism and set the mutation rate P. m =0.01; Step 503: Perform non-dominated sorting and elite retention, and re-perform finite element simulation calculations for each newly generated population to obtain the limit torque, cogging torque and efficiency of the permanent magnet motor. Step 504: Stratify and classify the population according to the Pareto dominance relationship between individuals, and individuals on the Pareto front will be retained to the next generation of the population; Step 505: Set the maximum number of generations. When the number of iterations reaches the maximum number of generations, the algorithm terminates and outputs the Pareto non-dominated solution set.

8. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, Step 6 specifically includes: Step 601: Construct the initial decision matrix based on the three evaluation indicators: limiting torque, cogging torque, and efficiency; Step 602: Normalize the original data to obtain a dimensionless matrix; Step 603: Calculate the weights of each evaluation indicator using information entropy; Step 604: Based on the different joint types of the humanoid robot, introduce the corresponding preference weight vector, and then use the multiplication synthesis normalization principle to calculate the final comprehensive weight; Step 605: Construct a weighted decision matrix using comprehensive weights, calculate the Euclidean distance from each alternative to the positive ideal solution and the negative ideal solution, then calculate the relative proximity of each alternative based on the obtained Euclidean distance, and finally sort the Pareto solution set according to the magnitude of the relative proximity to select the optimal combination of motor structure parameters.

9. The design and optimization method for a linear joint permanent magnet motor for a humanoid robot according to claim 1, characterized in that, Step 7 specifically includes: The TN characteristic curve of the optimized permanent magnet motor is established, and envelope matching analysis is performed with the linear joint operating points under different lead screw side effects to generate the efficiency map of the optimized motor. The loss characteristics within the operating range are evaluated based on the efficiency map.

10. A linear joint permanent magnet motor for a humanoid robot, characterized in that, The humanoid robot's linear joint permanent magnet motor is controlled based on the design optimization method described in any one of claims 1-9.

Images