A Behavior Decision Method, Device and Storage Device for Dual Manipulators

Abstract

The present application provides a behavior decision-making method, device and storage device for a dual robotic arm. The method includes: constructing a dual-arm mechanical model according to human arm characteristics, determining an anthropomorphic trajectory mapping process of the robotic arm according to the dual-arm mechanical model, obtaining allocation decision-making data of the human arms, and determining an anthropomorphic decision result of the robotic arm according to the allocation decision-making data and the anthropomorphic trajectory mapping process. By constructing a dual-robotic arm model according to human arm characteristics, and by classifying the operation selection results of the human arms and optimizing the existing greedy algorithm, the anthropomorphic decision result of the robotic arm with stronger anthropomorphic effect is determined, which solves the problem that conventional decision algorithms generally optimize the process of single instruction issuance and it is difficult to effectively optimize the overall process to cope with complex and changeable behavior environments.

Description

Technical Field

[0001] This application relates to the field of industrial robots, and particularly to a behavior decision-making method, device, and storage device for a dual-arm robot. Background Art

[0002] Six degrees of freedom: A six-degree-of-freedom robot can meet general task requirements; by defining a certain spatial task through six degrees of freedom, data information such as the pose of the six-degree-of-freedom robot can be determined.

[0003] Dual-arm: The system consists of two six-degree-of-freedom robotic arms, expanding the operable working space.

[0004] Industrial robots are one of the major high-tech achievements in the 21st century and are widely used in various industries. Most research on robot performance focuses on improving the assembly accuracy of a single robotic arm to enhance the overall performance of the robot. Single robotic arm operation usually adopts a point-to-point operation mode, and a single robotic arm is usually controlled by issuing commands in sequence. In a dual-arm system, there is a common space in the working space of the robotic arms. Using the method of issuing commands in sequence to control the two robotic arms will reduce the working efficiency of the entire dual-arm system.

[0005] For a dual-arm system, behavior decision-making is one of the ways to control the behavior of the dual arms. Behavior decision-making needs to make a predictive judgment on the current behavior environment of the robotic arm based on information such as movement distance, speed requirements, and action time, give a reasonable prediction for each spatial task, and issue a control command measured by physical values by a decision-making algorithm.

[0006] Conventional decision-making algorithms generally optimize the process of issuing a single command and are difficult to effectively optimize the overall process to cope with complex and changing behavior environments. Summary of the Invention

[0007] The purpose of this application is to solve the technical problem that conventional decision-making algorithms generally optimize the process of issuing a single command and are difficult to effectively optimize the overall process to cope with complex and changing behavior environments, and to provide a behavior decision-making method, device, and storage device for a dual-arm robot.

[0008] The above object of this application is achieved through the following technical solutions:

[0009] A behavior decision-making method for a dual-arm robot, the method includes:

[0010] S1: Construct a two-arm mechanical model according to human arm characteristics;

[0011] S2: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the two-arm mechanical model;

[0012] S3: Obtain the allocation decision data of the human arms;

[0013] S4: Determine the anthropomorphic decision result of the robotic arm according to the allocation decision data and the anthropomorphic trajectory mapping process.

[0014] Optionally, the step S1 includes:

[0015] S11: Determine the human arm feature formula according to the human arm features, as follows:

[0016] S H = f{(q E , l1), (q w , l2)}

[0017] where q E is the unit quaternion of the elbow joint relative to the shoulder coordinate system; q w is the unit quaternion of the wrist joint relative to the shoulder coordinate system; l1 is the distance from the shoulder center point of the human arm to the elbow center point of the human arm; l2 is the distance from the elbow center point of the human arm to the wrist center point of the human arm;

[0018] S12: Determine the mapping function according to the length of the human arm and the length of the robotic arm, as follows:

[0019] P i R = KP i H

[0020] where P i R is the center position of the human arm joint; P i H is the center position of the robotic arm joint; K is the proportionality coefficient;

[0021] S13: Construct a human arm triangle according to the shoulder center point, the elbow center point and the wrist center point;

[0022] S14: Determine the elbow joint angle β m of the human arm according to the human arm triangle and the cosine theorem, as follows:

[0023]

[0024] where l3 is the distance from the shoulder center point to the wrist center point;

[0025] S15: Determine the elbow joint angle β m of the robotic arm according to the elbow joint angle β r, as follows:

[0026]

[0027] Denote the 1st joint of the robotic arm as the shoulder joint center of the robotic arm, the 4th joint of the robotic arm as the elbow joint center of the robotic arm, and the 5th joint of the robotic arm as the wrist joint center of the robotic arm;

[0028] Wherein, d1 is the distance from the 1st joint of the robotic arm to the 4th joint of the robotic arm; d2 is the distance from the 4th joint of the robotic arm to the 5th joint of the robotic arm; d3 is the distance from the 1st joint of the robotic arm to the 5th joint of the robotic arm;

[0029] S16: Establish a coordinate system based on the shoulder center, then the parameters of the elbow center point are as follows:

[0030]

[0031] Wherein, (X1, Y1, Z1) are the position coordinate parameters of the elbow center point in the spherical coordinate system; is the azimuth angle, and θ1 is the zenith angle;

[0032] When the arm pose of the human arm is the same as the arm pose of the robotic arm, the position (P x4 , P y4 , P z4 ) of the 4th joint in the Cartesian coordinate system of the 1st joint is as follows:

[0033]

[0034] Denote the position coordinate parameters of the wrist center point as (X2, Y2, Z2), then the position coordinate parameters of the wrist center point relative to the elbow center point are (X2 - X1, Y2 - Y1, Z2 - Z1), that is, the parameters of the wrist center point in the spherical coordinate system of the elbow center point are as follows:

[0035]

[0036] Wherein, is the azimuth angle, and θ2 is the zenith angle;

[0037] When the arm pose of the human arm is the same as the arm pose of the robotic arm, the position (P x5 , P y5 , P z5 ) of the 5th joint in the Cartesian coordinate system of the 4th joint is as follows:

[0038]

[0039] S17: Determine the angles of the first joint and the fourth joint according to the positions of the fourth joint and the fifth joint, and construct the dual-arm mechanical model.

[0040] Optionally, step S2 includes:

[0041] S21: Calculate the shoulder-elbow structure angle of the robotic arm according to the pose matrix;

[0042] S22: Determine the axis rotation angle of the robotic arm according to the human arm angle;

[0043] S23: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the shoulder-elbow structure angle and the axis rotation angle.

[0044] Optionally, step S4 includes:

[0045] S41: Design a Gini index according to the allocation decision data and the anthropomorphic trajectory mapping process. The expression of the Gini index is as follows:

[0046]

[0047] where N is the total number of double operations; P k is the probability of selecting the left robotic arm or the right robotic arm. Denote P1 as the probability of selecting the left robotic arm and P2 as the probability of selecting the right robotic arm, where P1 = P2 = 0.5;

[0048] S42: Obtain the distance data of the human arm and the operation result data of the human arm;

[0049] S43: Determine the anthropomorphic decision result of the robotic arm according to the distance data, the operation result data and the Gini index.

[0050] Optionally, step S43 includes:

[0051] S43a: Denote the distance data as feature A. Denote the distance from the retention position of the human left arm on the workpiece to the zither central axis (the fourth column of the bridge) from left to right as d1. Then, calculated from the seventh column on the leftmost side of the workpiece, d1 = 3, 2, 1, 0, -1, -2. Denote the distance from the retention position of the human right arm on the workpiece to the zither central axis (the fourth column of the bridge) from right to left as d2. Then, calculated from the seventh column on the leftmost side of the workpiece, d2 = -3, -2, -1, 0, 1, 2;

[0052] S43b: Under the condition of feature A, the Gini index of the sample set D is as follows:

[0053]

[0054] Among them, D1 is the case of selecting the left arm of the human when the next operation position of the human arm is to the left of the workpiece, and D2 is the case of selecting the right arm of the human when the next operation position of the human arm is to the right of the workpiece. The expressions for D1 and D2 are as follows:

[0055]

[0056]

[0057] S43c: During the process of selecting the human arm, record the number of times the left arm of the human is selected as k1 and the number of times the right arm of the human is selected as k2. Then, the weight ω of selecting the left arm of the human in the overall operation satisfies the following formula:

[0058]

[0059] N = k1 + k2

[0060] Considering the operation situation where the human arm moves towards the center with the initial pose as a reference, there is:

[0061]

[0062] S43d: Record the operation result data as feature B. If the first operation in the operation result data is the right arm of the human and the current operation is the left arm of the human, then the expression for selecting the left arm of the human for the next operation is as follows:

[0063]

[0064] If the first operation in the operation result data is the right arm of the human and the current operation is the right arm of the human, then the expression for selecting the left arm of the human for the next operation is as follows:

[0065]

[0066] S43e: According to the indexes of feature A and feature B, calculate the influence of the judgment of the two arms being centered and the judgment of the two arms alternating on the two - arm decision - making process, and determine the feature decision result. The expression is as follows:

[0067] CART(D) = S1Gini(D,A) + S2Gini(D,B)

[0068] Among them, S1 is the proportion of the influence of feature A in the two - arm decision - making process, and S2 is the proportion of the influence of feature B in the two - arm decision - making process, satisfying S1 + S2 = 1;

[0069] S43f: Determine the anthropomorphic decision result of the robotic arm according to the characteristic decision result.

[0070] The determination of the anthropomorphic decision result of the robotic arm according to the characteristic decision result specifically includes:

[0071] Obtain the decision result of the greedy algorithm, and the expression of the decision result of the greedy algorithm is as follows:

[0072] Greed(D) = 0.5 + (d l - d r ) / (d l + d r )

[0073] where d l is the distance from the left arm of the robotic arm to the next operation, and d r is the distance from the right arm of the robotic arm to the next operation;

[0074] Determine the anthropomorphic decision result of the robotic arm according to the decision result of the greedy algorithm and the characteristic decision result, as follows:

[0075] P = K1Gini(D) + K2Greed(D)

[0076] where P is the anthropomorphic decision result. When the calculation result of P is greater than 0.5, select the left robotic arm as the operating arm. K1 is the final selection ratio of the output of the Gini index, and K2 is the final selection ratio of the output of the greedy algorithm, satisfying K1 + K2 = 1.

[0077] A storage device, which stores instructions and data for implementing a behavior decision method for a dual robotic arm.

[0078] A behavior decision device for a dual robotic arm, including: a processor and the storage device; the processor loads and executes the instructions and data in the storage device for implementing a behavior decision method for a dual robotic arm.

[0079] The beneficial effects brought by the technical solution provided by this application are:

[0080] By constructing a two-armed mechanical model based on human arm characteristics, the anthropomorphic trajectory mapping process of the robotic arm is determined; the actions of the human arm are matched with those of the robotic arm through joint mapping of the human arm, optimizing the process of the robotic arm learning human actions, reducing the collection time of learning samples, and improving the efficiency of data processing; by classifying the operation selection results of the human arm, the distance data of the human arm and the operation result data of the human arm are obtained, comprehensively considering the influence of dual-arm centering judgment and dual-arm alternating judgment on the dual-arm decision-making process and optimizing the existing greedy algorithm, the anthropomorphic decision result of the robotic arm with stronger anthropomorphic effect is determined, and a high-precision simulation of humanoid operation on the target workpiece can be completed in a short time. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] The present application will be further described below in conjunction with the drawings and embodiments. In the drawings:

[0082] Figure 1 is a flowchart of the steps of a behavior decision-making method for a dual robotic arm in an embodiment of the present application;

[0083] Figure 2 is a human arm feature diagram of a behavior decision-making method for a dual robotic arm in an embodiment of the present application;

[0084] Figure 3 is a schematic diagram of the operation of a hardware device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0085] In order to have a clearer understanding of the technical features, objectives, and effects of the present application, the specific embodiments of the present application will now be described in detail with reference to the drawings.

[0086] Embodiments of the present application provide a behavior decision-making method, device, and storage device for a dual robotic arm.

[0087] Please refer to Figure 1 , Figure 1 which is a flowchart of the steps of a behavior decision-making method for a dual robotic arm in an embodiment of the present application, and specifically includes the following steps:

[0088] S1: Construct a two-armed mechanical model based on human arm characteristics;

[0089] S2: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the two-armed mechanical model;

[0090] S3: Obtain the allocation decision data of the human two arms;

[0091] S4: Determine the anthropomorphic decision result of the robotic arm according to the allocation decision data and the anthropomorphic trajectory mapping process.

[0092] Specifically, the robotic arm is a complex dynamic system with time-varying, strong coupling, and nonlinear characteristics. The system analysis method is used to analyze the dynamic characteristics of the robotic arm. By obtaining the human arm characteristics, a two-arm mechanical model is constructed to determine the anthropomorphic trajectory mapping process of the robotic arm. Then, based on the allocation decision data of the human two arms for workpiece operation and the optimized greedy algorithm, the anthropomorphic decision result of the robotic arm is determined, enabling the robotic arm to continuously optimize the overall operation process according to the operation habits of the human arm, reduce the error rate during the operation process, and solve the technical problem that the conventional decision algorithm generally optimizes the process of a single instruction issuance and is difficult to effectively optimize the overall process to cope with the complex and changeable behavior environment.

[0093] Please refer to Figure 2 , Figure 2 which is the human arm feature diagram of a behavior decision method for a dual robotic arm in an embodiment of the present application. Step S1 includes:

[0094] S11: According to the human arm characteristics, determine the human arm characteristic formula as follows:

[0095] S H = f{(q E , l1), (q w , l2)}

[0096] where q E is the unit quaternion of the elbow joint relative to the shoulder coordinate system; q w is the unit quaternion of the wrist joint relative to the shoulder coordinate system; l1 is the distance from the shoulder center point of the human arm to the elbow center point of the human arm; l2 is the distance from the elbow center point of the human arm to the wrist center point of the human arm;

[0097] S12: According to the lengths of the human arm and the robotic arm, determine the mapping function as follows:

[0098] P i R = KP i H

[0099] where P i R is the central position of the human arm joint; P i H is the central position of the robotic arm joint; K is the proportionality coefficient;

[0100] S13: Construct a human arm triangle according to the shoulder center point, elbow center point, and wrist center point;

[0101] S14: Determine the elbow joint angle β m of the human arm according to the human arm triangle and the cosine theorem as follows:

[0102]

[0103] Among them, l3 is the distance from the center point of the shoulder to the center point of the wrist;

[0104] S15: Determine the elbow joint angle β of the robotic arm according to the elbow joint angle β of the human arm m , and the elbow joint angle β of the robotic arm is determined as follows: r as follows:

[0105]

[0106] Denote the 1st joint of the robotic arm as the shoulder joint center of the robotic arm, the 4th joint of the robotic arm as the elbow joint center of the robotic arm, and the 5th joint of the robotic arm as the wrist joint center of the robotic arm;

[0107] Among them, d1 is the distance from the 1st joint of the robotic arm to the 4th joint of the robotic arm; d2 is the distance from the 4th joint of the robotic arm to the 5th joint of the robotic arm; d3 is the distance from the 1st joint of the robotic arm to the 5th joint of the robotic arm;

[0108] S16: Establish a coordinate system based on the shoulder center, and the parameters of the elbow center point are as follows:

[0109]

[0110] Among them, (X1, Y1, Z1) are the position coordinate parameters of the elbow center point in the spherical coordinate system; is the azimuth angle, and θ1 is the zenith angle;

[0111] When the arm posture of the human arm is the same as that of the robotic arm, the position (P x4 , P y4 , P z4 ) of the 4th joint in the Cartesian coordinate system of the 1st joint is as follows:

[0112]

[0113] Denote the position coordinate parameters of the wrist center point as (X2, Y2, Z2), then the position coordinate parameters of the wrist center point relative to the elbow center point are (X2 - X1, Y2 - Y1, Z2 - Z1), that is, the parameters of the wrist center point in the spherical coordinate system of the elbow center point are as follows:

[0114]

[0115] Among them, is the azimuth angle, and θ2 is the zenith angle;

[0116] When the arm posture of the human arm is the same as that of the robotic arm, the position of the 5th joint in the Cartesian coordinate system of the 4th joint (P x5 , P y5 , P z5 ) is as follows:

[0117]

[0118] S17: Determine the angle of the 1st joint and the angle of the 4th joint according to the position of the 4th joint and the position of the 5th joint, and construct a dual-arm mechanical model.

[0119] Specifically, the human arm can be a seven-degree-of-freedom robotic arm with an S-R-S configuration. When the end movement is simple, a six-degree-of-freedom robotic arm can also complete simple grasping, moving and other functions. The joints of the human arm are driven by bones and muscles. Taking the shoulder joint as an example, this joint contains three degrees of freedom, and its movement on the human body is mainly manifested as two rotational movements around axes to change the position and one rotational movement around an axis to adjust the posture. Since the axis for changing the position can vary with the movement, the flexibility of the human arm is very strong. While the axis position of the robotic arm is fixed, and there are offsets between the three axes. In this application, from the perspective of physiological structure, the robotic arm is also defined as having three joints. Denote the 1st joint of the robotic arm as the center of the shoulder joint of the robotic arm, the 4th joint of the robotic arm as the center of the elbow joint of the robotic arm, and the 5th joint of the robotic arm as the center of the wrist joint of the robotic arm, and determine the arm length of the robotic arm and the corresponding included angles, so as to construct a dual-arm mechanical model.

[0120] Step S2 includes:

[0121] S21: Calculate the shoulder-elbow structure angle of the robotic arm according to the pose matrix;

[0122] S22: Determine the axis rotation angle of the robotic arm according to the human arm angle;

[0123] S23: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the shoulder-elbow structure angle and the axis rotation angle.

[0124] Step S4 includes:

[0125] S41: Design a Gini index according to the allocation decision data and the anthropomorphic trajectory mapping process. The expression of the Gini index is as follows:

[0126]

[0127] where N is the total number of double operations; P k is the probability of selecting the left robotic arm or the right robotic arm. Denote P1 as the probability of selecting the left robotic arm and P2 as the probability of selecting the right robotic arm, where P1 = P2 = 0.5;

[0128] S42: Obtain the distance data of the human arm and the operation result data of the human arm;

[0129] S43: Determine the anthropomorphic decision result of the robotic arm according to the distance data, the operation result data, and the Gini index.

[0130] Step S43 includes:

[0131] S43a: Denote the distance data as feature A. Denote the distance from the retention position of the human left arm on the workpiece in the distance data to the central axis of the dulcimer (the fourth column of the bridge) from left to right as d1. Then, starting from the seventh column on the leftmost side of the workpiece, we have d1 = 3, 2, 1, 0, -1, -2. Denote the distance from the retention position of the human right arm on the workpiece in the distance data to the central axis of the dulcimer (the fourth column of the bridge) from right to left as d2. Then, starting from the seventh column on the leftmost side of the workpiece, we have d2 = -3, -2, -1, 0, 1, 2;

[0132] S43b: Under the condition of feature A, the Gini index of the sample set D is as follows:

[0133]

[0134] Among them, D1 is the situation of selecting the human left arm when the next operation position of the human arm is biased to the left side of the workpiece, and D2 is the situation of selecting the human right arm when the next operation position of the human arm is biased to the right side of the workpiece. The expressions of D1 and D2 are as follows:

[0135]

[0136]

[0137] S43c: During the process of selecting the human arm, denote the number of times the human left arm is selected as k1, and the number of times the human right arm is selected as k2. Then, the weight ω of selecting the human left arm in the overall operation satisfies the following formula:

[0138]

[0139] N = k1 + k2

[0140] In the case of considering the operation when the human arms move towards the center with the initial pose as the reference, we have:

[0141]

[0142] S43d: Denote the operation result data as feature B. If the first operation in the operation result data is by the human right arm and the current operation is by the human left arm, then the expression for selecting the human left arm for the next operation is as follows:

[0143]

[0144] If the first operation in the operation result data is the right arm of a human and the current operation is the right arm of a human, then the following expression is used to select the left arm of a human for the next operation:

[0145]

[0146] S43e: Calculate the influence of the two-arm centering judgment and the two-arm alternating judgment on the two-arm decision-making process according to the indexes of feature A and feature B, and determine the feature decision result. The expression is as follows:

[0147] CART(D) = S1Gini(D,A) + S2Gini(D,B)

[0148] Where S1 is the influence proportion of feature A in the two-arm decision-making process, and S2 is the influence proportion of feature B in the two-arm decision-making process, and S1 + S2 = 1;

[0149] S43f: Determine the anthropomorphic decision result of the robotic arm according to the feature decision result.

[0150] Determine the anthropomorphic decision result of the robotic arm according to the feature decision result, which specifically includes:

[0151] Obtain the decision result of the greedy algorithm. The expression of the decision result of the greedy algorithm is as follows:

[0152] Greed(D) = 0.5 + (d l -d r ) / (d l +d r )

[0153] Where d l is the distance from the left arm of the robotic arm to the next operation, and d r is the distance from the right arm of the robotic arm to the next operation;

[0154] Determine the anthropomorphic decision result of the robotic arm according to the decision result of the greedy algorithm and the feature decision result, as follows:

[0155] P = K1Gini(D) + K2Greed(D)

[0156] Where P is the anthropomorphic decision result. When the calculation result of P is greater than 0.5, select the left arm of the robot as the operating arm. K1 is the final selection proportion of the output of the Gini index, and K2 is the final selection proportion of the output of the greedy algorithm, and K1 + K2 = 1.

[0157] Specifically, the way to obtain the allocation decision data of the human arms can be to obtain the data through a motion capture device; a motion capture device is an instrument that records the activities of marker points in a three-dimensional space by capturing reflected light signals. This instrument can record the trajectories of the marked points in space, including information such as the positions, speeds, and accelerations of the marker points. The motion capture process is recorded at 60 frames per second. The recorded demo can be used to view the visualized trajectory through the corresponding software, or tabular data arranged in the order of the recorded frames can be obtained by parsing.

[0158] The algorithm of this application uses the NOKOV optical three-dimensional motion capture system to record human actions. This system consists of 8 high-speed cameras. The cameras are hung clockwise around the shooting area according to the built-in number sequence. The static repeat accuracy of the cameras reaches 0.037 mm, the absolute accuracy can reach 0.087 mm, the linear dynamic trajectory error can reach 0.2 mm, and the circular arc trajectory error is 0.22 mm. The high-speed cameras receive the reflected light of the reflective marker points to determine the positions of the marker points. By adjusting the aperture and focal length of the cameras, other invalid reflective points can be excluded from the recording area. The algorithm of this application has higher accuracy compared with the conventional algorithms, improves the effectiveness of subsequent data processing, and improves the accuracy of the anthropomorphic decision results of the robotic arms.

[0159] Specifically, the conventional robotic arm double-arm allocation process only adopts the idea of the greedy algorithm, that is, after one operation is completed, the arm selected for the next operation only makes a decision based on the principle of the nearest distance. This method is simple and effective, can quickly generate a complete double-arm allocation table, and can complete the robotic arm operation process when the number of task steps is relatively small and the double-arm work process is relatively independent. However, errors often occur during the operation. For example, when the operation task points are relatively concentrated in some areas, there will be a situation where one robotic arm operates continuously while the other robotic arm remains stationary. Another example is that when the cumulative results of multiple decisions cause the entire robotic arms to deviate to the left or right side of the workpiece, once the next position is relatively far away, in order to fit the operation rhythm, the robotic arm joints will exceed the speed limit and get stuck.

[0160] This application intends to use the Gini index to measure whether the double-arm state is close to the center of the workpiece, and whether two adjacent actions of the robotic arm system are completed by different robotic arms in the same system. The overall robotic arm double-arm allocation process is optimized by adjusting the weight of the Gini index in the overall decision result.

[0161] Specifically, by studying the sample data of the human arms operating on workpieces, the sample data is inductively analyzed; the torso movements and joint movements of the human arms are extracted and mapped to the movements of each axis of the robotic arm to determine the allocation data of the robotic arm. According to the pre-organized data, the Gini index is designed to determine the Gini index; and the Gini evaluation index is combined with the improved greedy algorithm to obtain a human-like decision-making result of the robotic arm that is closer to human operation.

[0162] Please refer to Figure 3 , Figure 3 which is a schematic diagram of the operation of the hardware device according to an embodiment of the present application. The hardware device specifically includes: a tree-shaped decision-making device 401 for the work allocation of a six-degree-of-freedom dual robotic arm, a processor 402, and a storage device 403.

[0163] A tree-shaped decision-making device 401 for the work allocation of a six-degree-of-freedom dual robotic arm: The tree-shaped decision-making device 401 for the work allocation of a six-degree-of-freedom dual robotic arm implements the behavior decision-making method of the dual robotic arm.

[0164] Processor 402: The processor 402 loads and executes the instructions and data in the storage device 403 to implement the behavior decision-making method of the dual robotic arm.

[0165] Storage device 403: The storage device 403 stores instructions and data; the storage device 403 is used to implement the behavior decision-making method of the dual robotic arm.

[0166] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A behavior decision-making method for a dual robotic arm, characterized in that, The method includes the following steps: S1: Construct a two-arm mechanical model according to the characteristics of the human arm; The step S1 includes: S11: Determine the human arm characteristic formula according to the characteristics of the human arm as follows: S H = f{(q E , l1), (q w , l2)} where q E is the unit quaternion of the elbow joint relative to the shoulder coordinate system; q w is the unit quaternion of the wrist joint relative to the shoulder coordinate system; l1 is the distance from the shoulder center point to the elbow center point of the human arm; l2 is the distance from the elbow center point to the wrist center point of the human arm; S12: Determine the mapping function according to the length of the human arm and the length of the robotic arm as follows: Among them, is the central position of the human arm joint; is the central position of the robotic arm joint; K is the proportionality coefficient; S13: Construct a human arm triangle according to the shoulder center point, the elbow center point, and the wrist center point; S14: Determine the elbow joint angle β of the human arm according to the human arm triangle and the cosine theorem, as follows: m , as follows: Wherein, l3 is the distance from the shoulder center point to the wrist center point; S15: Determine the elbow joint angle β of the robotic arm according to the elbow joint angle β of the human arm m ; The elbow joint angle β r of the robotic arm is determined as follows: r The calculation formula for the elbow joint angle β is as follows: Denote the 1st joint of the robotic arm as the shoulder joint center of the robotic arm, the 4th joint of the robotic arm as the elbow joint center of the robotic arm, and the 5th joint of the robotic arm as the wrist joint center of the robotic arm; Wherein, d1 is the distance from the 1st joint of the robotic arm to the 4th joint of the robotic arm; d2 is the distance from the 4th joint of the robotic arm to the 5th joint of the robotic arm; d3 is the distance from the 1st joint of the robotic arm to the 5th joint of the robotic arm; S16: Establish a coordinate system based on the shoulder center, and the parameters of the elbow center point are as follows: Among them, (X1, Y1, Z1) are the position coordinate parameters of the center point of the elbow in the spherical coordinate system; is the azimuth angle, and θ1 is the zenith angle; When the arm posture of the human arm is the same as that of the robotic arm, the position of the No. 4 joint in the Cartesian coordinate system of the No. 1 joint (P x4 , P y4 , P z4 ), is as follows: Denote the position coordinate parameters of the wrist center point as (X2, Y2, Z2), then the position coordinate parameters of the wrist center point relative to the elbow center point are (X2 - X1, Y2 - Y1, Z2 - Z1), that is, the parameters of the wrist center point in the spherical coordinate system of the elbow center point are as follows: wherein, is the azimuth angle, and θ2 is the zenith angle; When the arm posture of the human arm is the same as that of the robotic arm, the position of the No. 5 joint in the Cartesian coordinate system of the No. 4 joint (P x5 , P y5 , P z5 ) is as follows: S17: Determine the angle of the 1st joint and the angle of the 4th joint according to the positions of the 4th joint and the 5th joint, and construct the two-arm mechanical model; S2: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the two-arm mechanical model; S3: Obtain the allocation decision data of the human arms; S4: Determine the anthropomorphic decision result of the robotic arm according to the allocation decision data and the anthropomorphic trajectory mapping process.

2. The behavior decision-making method of a dual robotic arm according to claim 1, wherein: The step S2 includes: S21: Calculate the shoulder-elbow structure angle of the robotic arm according to the pose matrix; S22: Determine the axis rotation angle of the robotic arm according to the human arm angle; S23: Determine the anthropomorphic trajectory mapping process of the robotic arm according to the shoulder-elbow structure angle and the axis rotation angle.

3. The behavior decision-making method of a dual robotic arm according to claim 1, characterized in that: The step S4 includes: S41: Design a Gini index according to the allocation decision data and the anthropomorphic trajectory mapping process, and the expression of the Gini index is as follows: where N is the total number of double operations; P k is the probability of selecting the left robotic arm or the right robotic arm. Denote P1 as the probability of selecting the left robotic arm and P2 as the probability of selecting the right robotic arm, where P1 = P2 = 0.5; S42: Obtain the distance data of the human arm and the operation result data of the human arm; S43: Determine the anthropomorphic decision result of the robotic arm according to the distance data, the operation result data, and the Gini index.

4. A storage device, characterized in that: The storage device stores instructions and data for implementing a behavior decision method of a dual robotic arm according to any one of claims 1 to 3.

5. A tree-shaped decision-making device for the work allocation of a six-degree-of-freedom dual robotic arm, characterized in that: Including: A processor and a storage device; the processor loads and executes the instructions and data in the storage device for implementing a behavior decision method of a dual robotic arm according to any one of claims 1 to 3.

Images