Series mechanical arm configuration optimization method

Abstract

The embodiment of the invention provides a series mechanical arm configuration optimization method, which realizes configuration optimization of a series mechanical arm, and comprises the following steps: deducing a configuration coding model according to a configuration expression mode of the series mechanical arm, constructing a structure tree without repeated configuration, and constructing a structure tree with repeated configuration according to a typical performance index mathematical expression of the mechanical arm. Comprising a flexible index, operability, a condition number, a minimum singular value, a working space index and terminal precision, forward normalization processing is carried out according to monotonicity, an expert scoring system comprehensive index evaluation system is established, a hierarchical pruning method is fused to carry out configuration screening in a configuration structure tree growth process, and an optimal solution set is obtained. According to the technical scheme provided by the embodiment of the invention, multi-performance index evaluation and rapid configuration optimization of the uncertain-degree-of-freedom series mechanical arm can be realized.

Description

[Technical field]

[0001] The invention relates to a method for optimizing the configuration of a serial robot arm, and belongs to the field of robot arm configuration optimization. [Background technology]

[0002] As human beings continue to explore space, the complexity of space missions increases, and the difficulty also increases. In terms of safety, efficiency, and cost, it is not enough to rely solely on astronauts to complete space exploration missions. Space robots, with their diverse application scenarios, have gradually become the main force of future space exploration. Therefore, it has become an inevitable trend to use space robots to replace astronauts to complete on-orbit missions. Space robots are mainly based on tandem robotic arms, have multiple degrees of freedom and certain flexibility, and can perform precise operations in small and complex environments. Therefore, the design of space robots with tandem robotic arms as the main body has become a major trend.

[0003] For the configuration design of serial manipulators, the existing methods are empirical method and traversal method.

[0004] The empirical method is based on the existing experience in manipulator configuration design to form a configuration database of the manipulator, and searches for configurations close to the design goals during design. However, the empirical method only considers typical manipulator configurations and has certain limitations. The traversal rule first calculates all possible configurations and then screens them according to performance indicators, but often its degrees of freedom are given, and the calculation is large and the efficiency is low. Therefore, establishing a comprehensive and efficient configuration optimization method for serial manipulators is of great significance to the problem of manipulator selection. [Summary of the invention]

[0005] In view of this, the present invention provides a method for optimizing the configuration of a serial robot arm, which enables a serial robot arm with indefinite degrees of freedom to be quickly classified when facing multiple performance index design requirements, and then obtains an optimal configuration solution set.

[0006] The present invention provides a method for optimizing the configuration of a tandem robotic arm, comprising:

[0007] According to the joint motion pair types and connection relationships, a serial robot arm coding configuration model is constructed to create a non-repetitive configuration structure tree;

[0008] The expressions of typical robot arm performance evaluation indicators are sorted out, including flexibility index, maneuverability, condition number, minimum singular value, workspace index, and end precision, and positive normalization is performed according to their monotonicity;

[0009] Establish an expert scoring system, and design the classification standards of each indicator in the configuration optimization process according to the experts' scores on the importance of each indicator;

[0010] A serial robot configuration optimization method based on hierarchical pruning method is designed. The robot configurations are classified according to different design requirements to obtain the optimal configuration solution set.

[0011] In the above method, the method of constructing a serial robot arm coding configuration model based on the joint motion pair type and connection relationship, and creating a non-repetitive configuration structure tree includes:

[0012] (1) Joint motion type

[0013] The joint is the basic unit of the robot arm configuration. In the actual design, the axes of adjacent joints are generally parallel or orthogonal. With the base coordinate system as the reference, the rotation around the X-axis is the yaw, the rotation around the Y-axis is the pitch, and the rotation around the Z-axis is the roll.

[0014] (2) Coding Configuration Model

[0015] In order to describe the relative position relationship between adjacent joints, the encoding configuration model e is defined. i :

[0016] e i =(Rx i R i R i ux i uy i uz i )|(i=1,…,n)

[0017] Where:

[0018] Rx i , Ry i , Rz i ——Rotation direction of the i-th joint, if Rx i =1, Ry i =0, Rz i = 0, it means that the i-th joint rotates around the X-axis. If Rx i =0, Ry i =1, Rz i = 0, it means that the i-th joint rotates around the Y axis, Rx i =1, Ry i =0, Rz i =1, it means that the i-th joint rotates around the Z axis;

[0019] ux i 、uy i ,uz i——Hyperparameters, which respectively indicate whether there is a distance between the positions of the i-1 joint and the i joint in the X, Y, and Z directions of the base coordinate system. When the value is 0, it means that there is no distance between the i-1 joint and the i joint in this direction; when the value is 1, it means that there is a distance between the i-1 joint and the i joint in this direction;

[0020] (3) Configuration structure tree

[0021] A tree diagram is a graphical tool that represents data or information in a hierarchical structure. It consists of nodes and edges. Nodes represent data or information, and edges represent the relationship between nodes.

[0022] Use Config n Represents the configuration information of n-DOF series branched chain joints:

[0023] Config n =(e1; e2; ...; e n )|i=(1,…,n)

[0024] In the formula, e i represents the joint i configuration;

[0025] In n-DOF serial robot Config n Add joint e at the end joint n+1 , get the n+1 degree of freedom serial robot Config n+1 , the configuration generation expression is as follows:

[0026]

[0027] Define the connection matrix T i i+1 Represents the configuration Config n With joints n+1 Connection relationship:

[0028]

[0029] Where t1, t2, and t3 represent the rotation of the joint around the X, Y, and Z axes, respectively. Indicates t j Can the type joint be connected with t k Type of joint connection, when When Config n The end joint is t j , then t k Cannot be used as a subsequent joint; if When Config n The end joint is t j , then t k Can be used as a follow-up joint;

[0030] Configuration repetition and failure conditions are:

[0031] (1) When the axes of adjacent joints are perpendicular, all the revolute pairs orthogonal to them are equivalent.

[0032] (2) When the axes of two adjacent revolute pairs are collinear, the two adjacent degrees of freedom degenerate into one degree of freedom. Considering the above situation as a joint connection constraint condition, the connection matrix under the constraint condition is established:

[0033]

[0034] in:

[0035]

[0036] ∨ is or, is XOR.

[0037] In the above method, the typical robot arm performance evaluation index expression includes flexibility index, maneuverability, condition number, minimum singular value, workspace index, and end precision, and is forward normalized according to its monotonicity, including:

[0038] (1) Flexibility index

[0039] The flexibility matrix is ​​used to describe the displacement response characteristics of the end of the robot under force. It is an important tool for analyzing the deformation of the robot under external force, causing the end to generate ΔX e The generalized force required for a small change in is:

[0040] F=K x ΔX e

[0041] Among them, K x is the Cartesian stiffness matrix

[0042] Cartesian stiffness matrix K x and the joint stiffness matrix K θ The following transformations exist:

[0043] K x =J -T K θ J -1

[0044] Where J represents the Jacobian matrix of the robot arm

[0045] therefore,

[0046] F=K x ΔX e =J -T K θ J-1 ΔX e

[0047] K x The inverse gives the flexibility matrix E flex :

[0048]

[0049] When the manipulator is at a boundary singularity, its stiffness is often relatively poor, and the maximum eigenvalue f of the flexibility matrix is ​​selected as the performance evaluation index;

[0050] (2) Operability

[0051] Maneuverability is used to indicate the flexibility of the robot's movement:

[0052]

[0053] Among them, det represents the determinant of the calculation matrix

[0054] (3) Condition number

[0055] Perform singular value decomposition on the Jacobian matrix:

[0056] J=U∑V

[0057] Among them, U and V are orthogonal matrices, and ∑ has the following form:

[0058]

[0059] Among them, σ1, σ2, …, σ m is a singular value of J, and σ1>σ2>…>σ m , σ1 is the largest singular value, σ m is the minimum singular value;

[0060] The condition number is the ratio of the largest singular value to the smallest singular value:

[0061]

[0062] (4) Minimum singular value

[0063] The minimum singular value indicates the movement ability of the robot in the worst direction. At the singular point of the configuration, the minimum singular value of the robot is zero, and the expression is:

[0064] s=σ m

[0065] (5) Workspace indicators

[0066] The workspace index Q is defined as the ratio of the sum of the lengths of the robot arm links to the cube root of the volume of the workspace that can be reached at the end. The calculation formula is:

[0067]

[0068] Where:

[0069] L – the sum of the connecting rod lengths;

[0070] V – volume of accessible working space;

[0071] a——connecting rod length;

[0072] d——connecting rod offset;

[0073] (6) Terminal error index

[0074] In order to calculate the position accuracy of the robot arm and simulate the joint repeatability, the joint angle error is set to +0.001°. In addition, errors in the manufacturing and assembly processes will cause errors in the robot's connecting rod structural parameters. The dimensional accuracy of the robot arm is set to ±0.01% of the actual value. Therefore, the error of the connecting rod parameters (a, d) is set to ±0.01%.

[0075] According to the forward kinematics equation, the position of the end relative to the reference coordinate system can be obtained, and the actual position is P N =(x N ,y N , z N ), the theoretical position is P A =(x A ,y A , z A ), then the end position deviation index is:

[0076]

[0077] Among them, aver means calculating the average

[0078] (7) Forward normalization

[0079] The six performance indicators of the serial robot arm are: flexibility index f, operability ω, condition number c, minimum singular value s, workspace index Q, and end accuracy E. P ; All performance indicators can be divided into two categories: positive indicators and reverse indicators. The larger the value of the positive indicator, the better the performance is considered; the larger the reverse indicator, the worse the performance is considered. When evaluating the configuration, it must first be trended. Generally, the reverse indicator is converted into a positive indicator. The formula is:

[0080] h'=1 / h

[0081] After all evaluation indicators are positively converted, in order to eliminate the incommensurability caused by the different dimensions of each indicator, each indicator should be dimensionless through normalization. The formula is:

[0082]

[0083] The six performance indicators (f, ω, c, s, Q, E P ) are normalized and correspond to (h1, h2, h3, h4, h5, h6) respectively.

[0084] In the above method, the expert scoring system is established, and the classification standards of each indicator in the configuration optimization process are designed according to the expert's score on the importance of each indicator, including:

[0085] Assuming that task requirement a is faced, n experts are asked to use a scale of 0 to 10 to score the importance of each performance indicator according to their own experience, and then the scale sequence of each evaluation indicator can be obtained as follows: The average of the expert scores is taken as the final score of each indicator. i The final score is:

[0086]

[0087] Determine the classification threshold of each performance indicator based on the value range of the scoring results Establish classification criteria for configuration screening:

[0088]

[0089] When the final score of indicator i is When the index evaluation level is "poor", the index classification threshold Q1 is the lower quartile; when the final score of index i is When the index evaluation level is "good", the index classification threshold Q2 is the divisor; when the final score of index i is When , the indicator evaluation level is "excellent" and the indicator classification threshold Q3 is the upper quartile.

[0090] In the above method, the design is based on a serial manipulator configuration optimization method of a hierarchical pruning method, and the manipulator configuration is classified according to different design requirements to obtain a configuration optimization solution set, including:

[0091] For the optimization problem of comprehensive configuration of serial manipulator with multiple indicators, the performance indicators should be classified first: for the optimization indicators with monotonicity, the pruning method is used for screening during the tree growth process to effectively exclude the configurations that do not meet the requirements; for the optimization indicators without monotonicity, the structure tree is traversed and layered to ensure that all possible configurations are considered, and finally a configuration set that meets all performance requirements is obtained. The specific optimization process is as follows:

[0092] ① The number of degrees of freedom n, the set of performance indicators ψ, and the set of performance indicator thresholds χ;

[0093] ② Divide the six performance indicators into two categories, with a monotonic change set ψ1 = {κ1, κ2, …, κ i}, does not have a monotonic variation set ψ2 = {λ1, λ2, ..., λ i};

[0094] ③ Input the minimum degree of freedom n min , establish the configuration structure tree Tree of the serial robot arm min ;

[0095] ④ Get all leaf nodes e under the structure tree i , and calculate the monotonic performance index value of each branch

[0096] ⑤Compare the monotonic performance index value and the corresponding threshold like hour, like hour,

[0097] ⑥Calculate node growth factor

[0098] ⑦ Design a new connection matrix: When the node growth factor E1 = 0, it indicates that the current node branch is cut and no new child nodes are generated;

[0099] ⑧Increase the depth of the tree layer depth = depth + 1;

[0100] ⑨ Determine whether the depth reaches n. If so, stop the calculation and get the structure tree Tree_Mid at this time n , if not, return to step ④;

[0101] ⑩Get the current tree Tree_Mid through recursive traversal n All configurations under Config_Mid n , and calculate the performance index value that is not monotonic

[0102] Calculate categorical factors For Config_Mid n It is layered, retaining all configurations that satisfy E2 = 1, eliminating all configurations that satisfy E2 = 0, and finally obtaining the configuration set Config_Dis that satisfies all optimization objectives n .

Brief Description of the Drawings

[0103] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creativity and labor.

[0104] Figure 1 is a schematic flow chart of a method for optimizing the configuration of a serial robot arm provided in an embodiment of the present invention;

[0105] Figure 2 This is a schematic diagram comparing the computational efficiency of the traversal method and the hybrid hierarchical pruning method under the expert scoring situation in the example of the present invention;

[0106] Figure 3 1 is a schematic diagram comparing the computational efficiency of the traversal method and the hybrid hierarchical pruning method under the second expert scoring situation in the example of the present invention;

[0107] Figure 4 1. It is a schematic diagram comparing the computational efficiency of the three-pass method and the hybrid hierarchical pruning method for expert scoring in an example of the present invention;

[0108] Figure 5 1. It is a schematic diagram comparing the computational efficiency of the four-pass method and the hybrid hierarchical pruning method for expert scoring in an example of the present invention;

[0109] Figure 6 It is a schematic diagram of average calculation efficiency comparison under the condition of 100 groups of expert scoring in the example of the present invention; [Specific embodiment]

[0110] Invite 5 experts to score the performance design of serial manipulators with 4 to 7 degrees of freedom, and calculate their comprehensive scores to design the thresholds of each performance indicator. Since each scoring result is divided into 3 levels, 3 levels will be generated for the 6 performance indicators. 6 Let H = [q1, q2, q3, q4, q5, q6] represent the classification level of each indicator, q i It can be divided into three levels: {'excellent', 'good', 'poor'}. 100 groups of scoring results are randomly generated for configuration optimization. Figures 2 to 5The number of configurations that need to be calculated by the hierarchical pruning method under four scoring conditions is given. The horizontal axis represents the corresponding degree of freedom, the vertical axis represents the number of configurations that need to calculate the performance, the left column in the figure represents the number of calculations by the traversal method, and the right column represents the number of calculations by the hierarchical pruning method. The curve in the figure represents the ratio of the right column to the left column. When the ratio is greater than 1, it means that the number of calculations by the traversal method is less than that of the hierarchical pruning method, and the efficiency is higher. When the ratio is less than 1, it means that the number of calculations by the hierarchical pruning method is less than that of the traversal method, and the efficiency is higher.

[0111] Figure 2 The corresponding scoring result H = [excellent, good, bad, good, bad], Figure 3 The corresponding scoring result is H = [excellent, good, poor, good, excellent], Figure 4 The corresponding scoring result H = [good, bad, good, good, bad], Figure 5 The corresponding scoring result is H = [bad bad good bad good, good]. The flexibility index h1 and the operability index h2 are monotonic, while the condition number h3, the minimum singular value h4, the terminal error h5, and the workspace index h6 are not monotonic. Figure 2 It can be seen that as the degree of freedom changes, the computational efficiency of the hierarchical pruning method is always higher than that of the traditional traversal method. Figure 5 It can be seen that when the classification level of the flexibility index h1 and the maneuverability index h2 is low, the calculation amount of the hierarchical pruning method will be greater than that of the traversal method. This is because there are more configurations of the upper degree of freedom used to calculate h1 and h2. In addition, this situation may only occur in low-degree-of-freedom configurations, and when there are 6 or 7 degrees of freedom, the efficiency of the hierarchical pruning method is significantly higher than that of the traditional traversal method.

[0112] In addition, 100 groups of scoring results H are randomly generated for configuration optimization. Figure 6 A comparison of the average number of calculations between the traversal method and the layered pruning method is given.

[0113] according to Figure 6 As a result, regardless of the H evaluation, the computational complexity of both methods increases with the increase of degrees of freedom, but the computational efficiency of the hierarchical pruning method is significantly higher than that of the traditional traversal method, and the advantage becomes more obvious with the increase of degrees of freedom. The computational efficiency is improved by 36.8% at 7R.

[0114] The technical solution of the embodiment of the present invention has the following beneficial effects:

[0115] Based on the types of joint motion pairs and their connection relationships, a coded configuration model of a serial robot arm was constructed, and a configuration structure tree with no failure and repeated configuration was designed. A variety of robot arm operation performance evaluation indicators were sorted out and normalized. The multi-expert scoring system was combined with the hierarchical pruning method to design a serial robot arm configuration optimization for multiple indicators. This optimization scheme can screen the configuration of serial robot arms according to different needs, and is efficient and feasible.

[0116] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0117] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A method for optimizing the configuration of a serial robot arm, characterized in that: The method comprises: According to the joint motion pair types and connection relationships, a serial robot arm coding configuration model is constructed to create a non-repetitive configuration structure tree; The expressions of typical robot arm performance evaluation indicators are sorted out, including flexibility index, maneuverability, condition number, minimum singular value, workspace index, and end precision, and positive normalization is performed according to their monotonicity; an expert scoring system is established, and the classification standards of each indicator in the configuration optimization process are designed according to the expert's score on the importance of each indicator; A serial robot configuration optimization method based on hierarchical pruning method is designed. The robot configurations are classified according to different design requirements to obtain the optimal configuration solution set.

2. The method according to claim 1, characterized in that The method of constructing a serial robot arm coding configuration model based on the joint motion pair type and connection relationship and creating a non-repetitive configuration structure tree includes: (1) Joint motion type The joint is the basic unit of the robot arm configuration. In the actual design, the axes of adjacent joints are generally parallel or orthogonal. With the base coordinate system as the reference, the rotation around the X-axis is the yaw, the rotation around the Y-axis is the pitch, and the rotation around the Z-axis is the roll. (2) Coding Configuration Model In order to describe the relative position relationship between adjacent joints, the encoding configuration model e is defined. i : have been i =(Rx i Ry i Rz i ux i oops i uz i )|(i=1,…,n) Where: Rx i , Ry i , Rz i ——Rotation direction of the i-th joint, if Rx i =1, Ry i =0, Rz i = 0, it means that the i-th joint rotates around the X-axis. If Rx i =0, Ry i =1, Rz i = 0, it means that the i-th joint rotates around the Y axis, Rx i =1, Ry i =0, R z i=1, it means the i-th joint rotates around the Z axis; ux i 、uy i ,uz i ——Hyperparameters, which respectively indicate whether there is a distance between the positions of the i-1 joint and the i joint in the X, Y, and Z directions of the base coordinate system. When the value is 0, it means that there is no distance between the i-1 joint and the i joint in this direction; when the value is 1, it means that there is a distance between the i-1 joint and the i joint in this direction; (3) Configuration structure tree A tree diagram is a graphical tool that represents data or information in a hierarchical structure. It consists of nodes and edges. Nodes represent data or information, and edges represent the relationship between nodes. Use Config n Represents the configuration information of n-DOF series branched chain joints: Config n =(e1;e2;…;e n )|i=(1,…,n) In n-DOF serial robot Config n Add joint e at the end joint n+1 , get the n+1 degree of freedom serial robot Config n+1 , the configuration generation expression is as follows: Define the connection matrix T i i+1 Represents the configuration Config n With joints n+1 Connection relationship: Where t1, t2, and t3 represent the rotation of the joint around the X, Y, and Z axes, respectively. Indicates t j Can the type joint be connected with t k Type of joint connection, when When Config n The end joint is t j , then t k Cannot be used as a subsequent joint; if When Config n The end joint is t j , then t k Can be used as a follow-up joint; Configuration repetition and failure conditions are: (1) When the axes of adjacent joints are perpendicular, all the revolute pairs orthogonal to them are equivalent. (2) When the axes of two adjacent revolute pairs are collinear, the two adjacent degrees of freedom degenerate into one degree of freedom. Considering the above situation as a joint connection constraint condition, the connection matrix under the constraint condition is established: in: ∨ is or, is XOR.

3. The method according to claim 1, characterized in that The typical robot arm performance evaluation index expression includes flexibility index, maneuverability, condition number, minimum singular value, workspace index, and end precision, and is forward normalized according to its monotonicity, including: (1) Flexibility index The flexibility matrix is ​​used to describe the displacement response characteristics of the end of the robot under force. It is an important tool for analyzing the deformation of the robot under external force, causing the end to generate ΔX e The generalized force required for a small change in is: F=K x ΔX e Among them, K x is the Cartesian stiffness matrix Cartesian stiffness matrix K x and the joint stiffness matrix K θ The following transformations exist: K x =J -T K θ J -1 Where J represents the Jacobian matrix of the robot therefore, F=K x ΔX e =J -T K θ J -1 ΔX e K x The inverse gives the flexibility matrix E flex : When the manipulator is at a boundary singularity, its stiffness is often relatively poor, and the maximum eigenvalue f of the flexibility matrix is ​​selected as the performance evaluation index; (2) Operability Maneuverability is used to indicate the flexibility of the robot's movement: Among them, det represents the determinant of the calculation matrix (3) Condition number Perform singular value decomposition on the Jacobian matrix: J=U∑V Among them, U and V are orthogonal matrices, and ∑ has the following form: Among them, σ1, σ2, …, σ m is a singular value of J, and σ1>σ2>…>σ m , σ1 is the largest singular value, σ m is the minimum singular value; The condition number is the ratio of the largest singular value to the smallest singular value: (4) Minimum singular value The minimum singular value indicates the movement ability of the robot in the worst direction. At the singular point of the configuration, the minimum singular value of the robot is zero, and the expression is: s=σ m (5) Workspace indicators The workspace index Q is defined as the ratio of the sum of the lengths of the robot arm links to the cube root of the volume of the workspace that can be reached at the end. The calculation formula is: Where: L – the sum of the connecting rod lengths; V – volume of accessible working space; a——connecting rod length; d——connecting rod offset; (6) Terminal error index In order to calculate the position accuracy of the robot arm and simulate the joint repeatability, the joint angle error is set to +0.001°. In addition, errors in the manufacturing and assembly processes will cause errors in the robot's connecting rod structural parameters. The dimensional accuracy of the robot arm is set to ±0.01% of the actual value. Therefore, the error of the connecting rod parameters (a, d) is set to ±0.01%. According to the forward kinematics equation, the position of the end relative to the reference coordinate system can be obtained, and the actual position is P N =(x N ,y N , z N ), the theoretical position is P A =(x A ,y A , z A ), then the end position deviation index is: Among them, aver means calculating the average (7) Forward normalization The six performance indicators of the serial robot arm are: flexibility index f, operability ω, condition number c, minimum singular value s, workspace index Q, and end accuracy E. P ; All performance indicators can be divided into two categories: positive indicators and reverse indicators. The larger the value of the positive indicator, the better the performance is considered; the larger the reverse indicator, the worse the performance is considered. When evaluating the configuration, it must first be trended. Generally, the reverse indicator is converted into a positive indicator. The formula is: h'=1 / h After all evaluation indicators are positively converted, in order to eliminate the incommensurability caused by the different dimensions of each indicator, each indicator should be dimensionless through normalization. The formula is: The six performance indicators (f, ω, c, s, Q, E P ) are normalized and correspond to (h1, h2, h3, h4, h5, h6) respectively.

4. The method according to claim 1, characterized in that: The expert scoring system is established, and the classification standards of each indicator in the design configuration optimization process are designed according to the expert's score on the importance of each indicator, including: Assuming that task requirement a is faced, n experts are asked to use a scale of 0 to 10 to score the importance of each performance indicator according to their own experience, and then the scale sequence of each evaluation indicator can be obtained as follows: The average of the expert scores is taken as the final score of each indicator. i The final score is: Determine the classification threshold of each performance indicator based on the value range of the scoring results Establish classification criteria for configuration screening: When the final score of indicator i is When the index evaluation level is "poor", the index classification threshold Q1 is the lower quartile; when the final score of index i is When the index evaluation level is "good", the index classification threshold Q2 is the divisor; when the final score of index i is When , the indicator evaluation level is "excellent" and the indicator classification threshold Q3 is the upper quartile.

5. The method according to claim 1, characterized in that The design is based on a serial manipulator configuration optimization method of a hierarchical pruning method, which classifies the manipulator configuration according to different design requirements and obtains a configuration optimization solution set, including: For the optimization problem of comprehensive configuration of serial manipulator with multiple indicators, the performance indicators should be classified first: for the optimization indicators with monotonicity, the pruning method is used for screening during the tree growth process to effectively exclude the configurations that do not meet the requirements; for the optimization indicators without monotonicity, the structure tree is traversed and layered to ensure that all possible configurations are considered, and finally a configuration set that meets all performance requirements is obtained. The specific optimization process is as follows: ① The number of degrees of freedom n, the set of performance indicators ψ, and the set of performance indicator thresholds χ; ② Divide the six performance indicators into two categories, with a monotonic change set ψ1 = {κ1, κ2, …, κ i }, does not have a monotonic variation set ψ2 = {λ1, λ2, ..., λ i }; ③ Input the minimum degree of freedom n min , establish the configuration structure tree Tree of the serial robot arm min ; ④ Get all leaf nodes e under the structure tree i , and calculate the monotonic performance index value of each branch ⑤Compare the monotonic performance index value and the corresponding threshold like hour, like hour, ⑥Calculate node growth factor ⑦ Design a new connection matrix: When the node growth factor E1 = 0, it indicates that the current node branch is cut and no new child nodes are generated; ⑧Increase the depth of the tree layer depth = depth + 1; ⑨ Determine whether the depth reaches n. If so, stop the calculation and get the structure tree Tree_Mid at this time n , if not, return to step ④; ⑩Get the current tree Tree_Mid through recursive traversal n All configurations under Config_Mid n , and calculate the performance index value that is not monotonic Calculate categorical factors For Config_Mid n It is layered, retaining all configurations that satisfy E2 = 1, eliminating all configurations that satisfy E2 = 0, and finally obtaining the configuration set Config_Dis that satisfies all optimization objectives n .

Images