A method and system for automatic spraying of jeans based on manual teaching

Abstract

The present invention discloses a method and system for automatic spraying of jeans based on manual teaching, the method comprising: initializing a teaching robot and a spray gun, obtaining a teaching posture data set of the robot through a motion capture system; obtaining a DMP model; reproducing a motion trajectory through the DMP model; modeling according to the connecting rod parameters of the robot, obtaining a Jacobian matrix, passing the motion trajectory into a zeroing neural network as an expectation, calculating the joint angle information of the robot according to the expected position and the actual position at each moment; controlling the robot to perform simulated motion according to the joint angle information, obtaining the actual position and feeding it back to the zeroing neural network, completing the optimization of the zeroing neural network; controlling the robot to automatically spray jeans according to the real-time joint angle information. The present invention utilizes the DMP model and the zeroing neural network, improves the accuracy of trajectory generation, thereby improving the accuracy of automatic spraying, has strong generalization, and can be widely used in the field of computer technology.

Description

Technical Field

[0001] The invention relates to the field of computer technology, in particular to an automatic spraying method and system for jeans based on manual teaching. Background Art

[0002] Robotic arms have broad potential application prospects in various fields, especially in industrial production, where they can play a huge role, such as reducing manpower burden, improving production efficiency, and reducing production costs. Especially in the jeans spraying industry, robots can replace workers to perform repetitive work.

[0003] In the jeans spraying industry, workers usually hold spray guns to spray jeans and render patterns; while robots grab spray guns to spray, which requires consideration of trajectory planning, posture control and other issues. The first is trajectory planning. Different brands or models of jeans have different patterns, so it is necessary to plan the running trajectory of the robot end in three-dimensional space in advance.

[0004] At present, there is no relevant technology to solve the technical problems corresponding to the field of automatic spraying of jeans.

[0005] In addition, related technologies mention the use of GMM / GMR (Gaussian mixture model / Gaussian mixture regression) as an imitation learning algorithm. First, workers are required to grasp the robot for multiple teachings, use sensors to collect multi-modal information such as the robot's position, torque, stiffness, etc., obtain multiple teaching trajectories, use HSMM (hidden semi-Markov model) to encode the teaching data, learn and reproduce the trajectory through GMM / GMR, and finally use impedance control to control the robot.

[0006] However, the above method has the following shortcomings:

[0007] 1. The GMM / GMR algorithm requires multiple teaching trajectories as input, requiring workers to repeat the teaching, and the trajectory reproduced by the model is the weighted mean of the teaching trajectory set. It is difficult to obtain the required reproduction trajectory, with poor accuracy and lack of generalization;

[0008] 2. The above control scheme does not take the robot's posture into consideration, and only plans the robot's trajectory based on the position of the end. In the spraying industry, if the spray gun's posture is incorrect, it will result in different jeans spray patterns, leading to production problems. Summary of the invention

[0009] In view of this, the embodiments of the present invention provide a method and system for automatic spraying of jeans based on manual teaching with high precision and strong generalization, so as to achieve accurate spraying of jeans.

[0010] One aspect of an embodiment of the present invention provides a method for automatically spraying jeans based on manual teaching, comprising:

[0011] Initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system;

[0012] According to the teaching posture data set, a 6-dimensional DMP learning teaching data is established through a DMP model to obtain a DMP model with extracted motion features;

[0013] Reproducing the motion trajectory through the DMP model, and comparing the reproduced motion trajectory with the initial teaching trajectory, and determining that the DMP model is constructed according to the comparison result;

[0014] The robot's Jacobian matrix is ​​obtained by modeling the robot's connecting rod parameters, and the motion trajectory reproduced by the DMP model is passed into the annihilation neural network as the expectation, and the annihilation neural network calculates the robot's joint angle information according to the expected position and actual position at each moment;

[0015] The VPER simulation software controls the robot to perform simulation motion according to the joint angle information, and feeds back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network;

[0016] The real-time joint angle information of the robot is calculated according to the optimized annihilation neural network, and the robot is controlled to automatically spray the jeans according to the real-time joint angle information.

[0017] Optionally, the kinematic model of the robot's multiple joints of the robot's mechanical arm is expressed as:

[0018] f(θ(t))=q(t)

[0019] Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle values ​​of the m joint angles of the robot arm at time t; q(t) represents the vector consisting of the position and posture of the end of the robot arm in the n-dimensional task space.

[0020] Optionally, the calculation formula of the joint angle is:

[0021] θ(t)=g(q d (t))

[0022] Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the zeroing neural network formula.

[0023] Another aspect of the embodiment of the present invention further provides a jeans automatic spraying system based on manual teaching, comprising:

[0024] The first module is used to initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system;

[0025] The second module is used to establish 6-dimensional DMP learning teaching data through the DMP model according to the teaching posture data set, and obtain the DMP model with extracted motion features;

[0026] The third module is used to reproduce the motion trajectory through the DMP model, and compare the reproduced motion trajectory with the initial teaching trajectory, and determine that the DMP model is built according to the comparison result;

[0027] The fourth module is used to model the robot according to the connecting rod parameters, obtain the Jacobian matrix of the robot, and pass the motion trajectory reproduced by the DMP model into the annihilation neural network as the expectation, and the annihilation neural network calculates the joint angle information of the robot according to the expected position and actual position at each moment;

[0028] The fifth module is used to control the robot to perform simulated motion according to the joint angle information by the VPER simulation software, and feed back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network;

[0029] The sixth module is used to calculate the real-time joint angle information of the robot according to the optimized zeroing neural network, and control the robot to automatically spray jeans according to the real-time joint angle information.

[0030] Optionally, the kinematic model of the robot's multiple joints of the robot's mechanical arm is expressed as:

[0031] f(θ(t))=q(t)

[0032] Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle values ​​of the m joint angles of the robot arm at time t; q(t) represents the vector consisting of the position and posture of the end of the robot arm in the n-dimensional task space.

[0033] Optionally, the calculation formula of the joint angle is:

[0034] θ(t)=g(q d (t))

[0035] Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the zeroing neural network formula.

[0036] Another aspect of an embodiment of the present invention further provides an electronic device, including a processor and a memory;

[0037] The memory is used to store programs;

[0038] The processor executes the program to implement the method described above.

[0039] Another aspect of the embodiment of the present invention further provides a computer-readable storage medium, wherein the storage medium stores a program,

[0040] The program is executed by a processor to implement the above-mentioned method.

[0041] The embodiment of the present invention also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above method.

[0042] The embodiment of the present invention initializes a teaching robot and a spray gun, and obtains a teaching posture data set of the robot through a motion capture system; according to the teaching posture data set, a 6-dimensional DMP learning teaching data is established through a DMP model to obtain a DMP model with extracted motion features; the motion trajectory is reproduced through the DMP model, and the reproduced motion trajectory is compared with the initial teaching trajectory, and the construction of the DMP model is determined to be completed according to the comparison result; the robot's connecting rod parameter model is modeled to obtain the Jacobian matrix of the robot, and the motion trajectory reproduced by the DMP model is passed into a zeroing neural network as an expectation, and the zeroing neural network calculates the joint angle information of the robot according to the expected position and the actual position at each moment; the VPER simulation software controls the robot to perform simulation motion according to the joint angle information, and the position of the robot during the simulation is fed back to the zeroing neural network as the actual position, so as to complete the optimization of the zeroing neural network; the real-time joint angle information of the robot is calculated according to the optimized zeroing neural network, and the robot is controlled to automatically spray jeans according to the real-time joint angle information. The present invention utilizes the DMP model and the zeroing neural network to improve the accuracy of trajectory generation, thereby improving the accuracy of automatic spraying and having strong generalization. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 Flow chart of automatic spraying system for jeans;

[0045] Figure 2 It is a schematic diagram of the automatic spraying system for jeans;

[0046] Figure 3 Schematic diagram for position visualization trajectory;

[0047] Figure 4 Schematic diagram of the trajectory for posture visualization;

[0048] Figure 5 Schematic diagram of the expected trajectory and simulated trajectory in the X direction;

[0049] Figure 6 Schematic diagram of the expected trajectory and simulated trajectory in the Y direction;

[0050] Figure 7 Schematic diagram of the expected trajectory and the simulated trajectory in the Z direction;

[0051] Figure 8 Schematic diagram of position error between expected trajectory and simulated trajectory;

[0052] Fig. 9 Schematic diagram of the angle error between the expected trajectory and the simulated trajectory;

[0053] Fig.10 Schematic diagram of expected trajectory and simulated trajectory. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] In view of the problems existing in the prior art, an embodiment of the present invention provides an automatic spraying method for jeans based on manual teaching, comprising:

[0056] Initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system;

[0057] According to the teaching posture data set, a 6-dimensional DMP learning teaching data is established through a DMP model to obtain a DMP model that extracts motion features; wherein the DMP model refers to Dynamic movement primitives, a dynamic motion primitive model.

[0058] Reproducing the motion trajectory through the DMP model, and comparing the reproduced motion trajectory with the initial teaching trajectory, and determining that the DMP model is constructed according to the comparison result;

[0059] The robot's Jacobian matrix is ​​obtained by modeling the robot's connecting rod parameters, and the motion trajectory reproduced by the DMP model is passed into the annihilation neural network as the expectation, and the annihilation neural network calculates the robot's joint angle information according to the expected position and actual position at each moment;

[0060] The VPER simulation software controls the robot to perform simulation motion according to the joint angle information, and feeds back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network;

[0061] The real-time joint angle information of the robot is calculated according to the optimized annihilation neural network, and the robot is controlled to automatically spray the jeans according to the real-time joint angle information.

[0062] Optionally, the kinematic model of the robot's multiple joints of the robot's mechanical arm is expressed as:

[0063] f(θ(t))=q(t)

[0064] Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle values ​​of the m joint angles of the robot arm at time t; q(t) represents the vector consisting of the position and posture of the end of the robot arm in the n-dimensional task space.

[0065] Optionally, the calculation formula of the joint angle is:

[0066] θ(t)=g(q d (t))

[0067] Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the ZNN (Zeroing Neural Network) formula.

[0068] Another aspect of the embodiment of the present invention further provides a jeans automatic spraying system based on manual teaching, comprising:

[0069] The first module is used to initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system;

[0070] The second module is used to establish 6-dimensional DMP learning teaching data through the DMP model according to the teaching posture data set, and obtain the DMP model with extracted motion features;

[0071] The third module is used to reproduce the motion trajectory through the DMP model, and compare the reproduced motion trajectory with the initial teaching trajectory, and determine that the DMP model is built according to the comparison result;

[0072] The fourth module is used to model the robot according to the connecting rod parameters, obtain the Jacobian matrix of the robot, and pass the motion trajectory reproduced by the DMP model into the annihilation neural network as the expectation, and the annihilation neural network calculates the joint angle information of the robot according to the expected position and actual position at each moment;

[0073] The fifth module is used to control the robot to perform simulated motion according to the joint angle information by the VPER simulation software, and feed back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network;

[0074] The sixth module is used to calculate the real-time joint angle information of the robot according to the optimized zeroing neural network, and control the robot to automatically spray jeans according to the real-time joint angle information.

[0075] Optionally, the kinematic model of the robot's multiple joints of the robot's mechanical arm is expressed as:

[0076] f(θ(t))=q(t)

[0077] Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle values ​​of the m joint angles of the robot arm at time t; q(t) represents the vector consisting of the position and posture of the end of the robot arm in the n-dimensional task space.

[0078] Optionally, the calculation formula of the joint angle is:

[0079] θ(t)=g(q d (t))

[0080] Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the ZNN formula.

[0081] Another aspect of an embodiment of the present invention further provides an electronic device, including a processor and a memory;

[0082] The memory is used to store programs;

[0083] The processor executes the program to implement the method described above.

[0084] Another aspect of the embodiment of the present invention further provides a computer-readable storage medium, wherein the storage medium stores a program,

[0085] The program is executed by a processor to implement the above-mentioned method.

[0086] The embodiment of the present invention also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the above method.

[0087] The specific implementation process of the present invention is described in detail below in conjunction with the accompanying drawings:

[0088] In view of the shortcomings of the prior art, the present invention has the following two purposes:

[0089] First, the present invention is to add a self-learning function to the automatic spraying system for jeans. To this end, the present invention adopts the DMP algorithm (Dynamic movement primimtives) to obtain the posture information of the end through a single teaching, and creates a 6-dimensional DMP model to learn the posture information of the robot end holding the spray gun. After training the model, DMP can successfully reproduce the posture trajectory of the spraying process and can also generalize to trajectories of different movement speeds.

[0090] Second, we need to control the robot’s posture during the spraying process. The change in the robot’s posture during movement is the key to spraying patterns. We use a model-based ZNN (Zeroing Neural Network) control method to constrain the robot’s posture at the end of the operation process and achieve accurate spraying.

[0091] (1) Dynamic motion primitives

[0092] First, imitation learning refers to a type of algorithm in which robots acquire motion skills by learning teaching samples. The learning process is generally to extract motion features from a single or a small number of teaching trajectories. For the teaching method, this embodiment uses a worker holding a spray gun to perform a spraying demonstration, and collects teaching data through the posture sensor on the end of the robot. This embodiment uses the common DMP as the learning algorithm, which is based on an improved second-order spring damping system:

[0093]

[0094]

[0095]

[0096] In formula (1), y is the learning teaching trajectory, is the first-order derivative of y, g is the target point, τ is the scaling term, and adjusting this value can obtain trajectories with different convergence speeds. y ,β y is the control constant, f is a nonlinear function; in formula (2), x is used to construct the nonlinear function f(x) of formula (1), τ also plays the role of speed modulation, α x is a constant; in formula (3), ψ i (x) is the radial basis function, which is defined as: σ i and c i They are the basis functions ψ i (x) width and center position, w i is the basis function weight, and the above parameters are obtained by learning the teaching trajectory using local least squares regression in the DMP learning stage.

[0097] (2) Zeroing Neural Network

[0098] The kinematic model of a robot arm with m joints can be expressed by the following formula:

[0099] f(θ(t))=q(t)#(4)

[0100] Among them, the input column vector θ(t) = [θ 1 ,θ 2 ,…,θ m ] T ∈R m represents the angle value of the m joint angles of the robot at time t, (·) T represents the transposition operator, column vector q(t)∈R n The vector representing the position and posture of the end of the robot in the n-dimensional task space (usually 6, including 3D position coordinates and Euler angles), f(·): R m →R n is a nonlinear mapping function that describes the kinematics of the robot. The kinematic model of the robot can usually also be described at the velocity level. By taking the time derivatives of both sides of formula (4), we can obtain

[0101]

[0102] in, represents the partial derivative of the function f(·) with respect to the joint angle θ(t), i.e., the Jacobian matrix of the manipulator, and the column vector Represents the rotation speed of each joint of the robot arm, column vector The tracking control problem of the robot arm (i.e., the inverse kinematics problem of the robot arm) refers to the control of the robot arm according to the given target path and posture vector q. d (t), and obtain the required joint angle θ(t), that is:

[0103] θ(t)=g(q d (t))##(6)

[0104] The obtained θ(t) is then used to control the robotic arm so that its end follows the target position q d (t) Movement, that is, the actual position q of the end of the robot arm a(t) Approaching the target pose q d (t), we need to deduce g(·):R n →R m Therefore, this embodiment defines the error e(t)=q d(t) -q a(t) , using the ZNN formula It can be introduced:

[0105]

[0106] Wherein, F(·) is an n-order nonlinear activation function. g(·) is a ZNN formula, which is given by formulas (8) and (9) in the handbook. In summary, the formula for solving the joint angle can be obtained in this embodiment:

[0107]

[0108]

[0109] J * represents the Jacobian matrix pseudoinverse.

[0110] like Figure 1 and 2 As shown, the invented automatic spraying system comprises the following steps:

[0111] It should be noted that: Figure 1 and 2 The trajectory mentioned in the above contains position and attitude information.

[0112] A. Take LBR iiwa 7 R800 (number of joint angles m = 7) as the teaching robot arm, and the end of the robot arm grasps the spray gun. First, initialize the two robots t 0 = Joint angle θ at time 0 L (t0 ) = [0,25,0,10,6.19,-1.8,0] T *π / 180, using this as the initial state to control the movement of the teaching robot, with a time interval of 50ms, using a motion capture camera to record the position and posture information of the end, and a total of 123 trajectory points were collected.

[0113] B. After teaching, use the 6-dimensional DMP to learn the teaching data, and the DMP hyperparameter value is α y =60,β y =4,τ=1,α x =1, the number of basis functions is 100, and the following results are obtained. Figure 3 and Figure 4 As shown, Figure 3 Schematic diagram of the position visualization trajectory, Figure 4 It is a schematic diagram of the posture visualization trajectory; the red curve represents the teaching trajectory collected by the motion capture system, and the blue curve represents the reproduction trajectory. The error of the result graph is small enough, which verifies the feasibility of the DMP algorithm.

[0114] C. The reproducing trajectory output by DMP is used as the input of ZNN. First, the LBR iiwa 7 R800 is modeled to obtain the initial transformation matrix and Jacobian matrix information of the robot arm. The control process of this embodiment starts at time t (t=0), and the position at time t in the reproducing trajectory is taken as the expected position q d (t), the position corresponding to the joint angle of the robot at time t is taken as the actual position q a (t), and use formulas (8) and (9) to calculate the robot's joint angle θ at time t+1 t+1 and angular velocity The joint angle and joint angular velocity information at time t+1 are used as input and passed to the VREP simulation software to control the movement of the LBR iiwa 7R800 robot arm. At this time, the robot moves to the actual position q a (t+1), which is sent back to ZNN as a feedback signal, and the position at time t+1 in the reproduced trajectory is taken as the expected position q d (t+1), used to calculate the joint angle θ at the next moment t+2 t+2 and joint angular velocity The above process repeats itself until the last moment.

[0115] Related experimental results such as Figure 5-10 As shown, Figure 5 Schematic diagram of the expected trajectory and simulated trajectory in the X direction; Figure 6 Schematic diagram of the expected trajectory and simulated trajectory in the Y direction; Figure 7 Schematic diagram of the expected trajectory and the simulated trajectory in the Z direction; Figure 8Schematic diagram of position error between expected trajectory and simulated trajectory; Fig. 9 Schematic diagram of the angle error between the expected trajectory and the simulated trajectory; Fig.10 The diagram is a schematic diagram of the expected trajectory and the simulated trajectory. The experimental results show that the position error between the reproduced trajectory and the taught trajectory of the jeans robot automatic spraying system is within 1.4 cm, and the posture error is within 0.5 rad. The simulation verifies the feasibility and meets the actual production requirements.

[0116] This experimental example specifically introduces the process of DMP learning the teaching trajectory and the principle and process of ZNN controlling the robot's motion. First, this experimental example can effectively extract the motion characteristics of a teaching trajectory, and the obtained reproduced trajectory fits the teaching trajectory, solving the inconvenience and inaccuracy caused by multiple teachings. In addition, DMP can reproduce the motion trajectory of different speeds and different target points, and has a certain generalization, which expands the application scenarios of the spraying system. Secondly, compared with the method of using inverse kinematics control, the model-based ZNN will receive feedback on the actual position and posture of the robot during operation, effectively controlling the stability of the spray gun during the motion process, keeping the spray gun level and reducing interference factors caused by the environment.

[0117] In summary, the present invention provides an automatic jeans spraying system with learning ability, which only needs one teaching data from the worker to reproduce the worker's movement process, replace the worker's labor, and liberate labor. The multimodal imitation learning of the present invention collects teaching data from the position and posture dimensions of the robot end as the input of the six-dimensional DMP model, uses local least squares regression learning to obtain the radial basis function and weight parameters of the DMP, and the trained DMP model reproduces the posture trajectory as the input of the ZNN. The present invention uses a ZNN-based method control to ensure that the spray gun posture during reproduction is consistent with the posture of the worker during teaching, preventing the spray gun posture from being incorrectly sprayed to places outside the jeans.

[0118] Compared with the prior art, the present invention has the following advantages:

[0119] First, the main purpose of the present invention is to replace workers in spraying patterns on jeans. In order to maximize the benefits, the DMP model based on a one-time teaching process can ensure accuracy and generalization. It has the advantages of one-time teaching and multiple uses, and can generalize trajectories of different speeds and trajectories of different target points, which can be applied to the production of different jeans models.

[0120] Secondly, the control method of the present invention adopts a model-based ZNN, which can control the robot operation from two aspects of position and posture. If the robot joint shakes during production, ZNN receives feedback of the actual position and makes adjustments, thereby ensuring overall stability during operation.

[0121] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.

[0122] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0123] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0124] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0125] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0126] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0127] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0128] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0129] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A method for automatic spraying of jeans based on manual teaching, characterized in that: include: Initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system; According to the teaching posture data set, a 6-dimensional DMP learning teaching data is established through a DMP model to obtain a DMP model with extracted motion features; Reproducing the motion trajectory through the DMP model, and comparing the reproduced motion trajectory with the initial teaching trajectory, and determining that the DMP model is constructed according to the comparison result; The robot's Jacobian matrix is ​​obtained by modeling the robot's connecting rod parameters, and the motion trajectory reproduced by the DMP model is passed into the annihilation neural network as the expectation, and the annihilation neural network calculates the robot's joint angle information according to the expected position and actual position at each moment; The VREP simulation software controls the robot to perform simulation motion according to the joint angle information, and feeds back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network; Calculating the real-time joint angle information of the robot according to the optimized zeroing neural network, and controlling the robot to automatically spray the jeans according to the real-time joint angle information; The expression of the kinematic model of the robot's multiple joints of the mechanical arm is: f(θ(t))=q(t) Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle value of the m joint angles of the robot arm at time t; q(t) represents the vector composed of the position and posture of the end of the robot arm in the n-dimensional task space; the calculation formula of the joint angle is: θ(t)=g(q d (t)) Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the zeroing neural network formula.

2. An automatic spraying system for jeans based on manual teaching, characterized in that: include: The first module is used to initialize the teaching robot and spray gun, and obtain the robot's teaching posture data set through the motion capture system; The second module is used to establish 6-dimensional DMP learning teaching data through the DMP model according to the teaching posture data set, and obtain the DMP model with extracted motion features; The third module is used to reproduce the motion trajectory through the DMP model, and compare the reproduced motion trajectory with the initial teaching trajectory, and determine that the DMP model is built according to the comparison result; The fourth module is used to model the robot according to the connecting rod parameters, obtain the Jacobian matrix of the robot, and pass the motion trajectory reproduced by the DMP model into the annihilation neural network as the expectation, and the annihilation neural network calculates the joint angle information of the robot according to the expected position and actual position at each moment; The fifth module is used to control the robot to perform simulated motion according to the joint angle information by VREP simulation software, and feed back the position of the robot during the simulation as the actual position to the zeroing neural network to complete the optimization of the zeroing neural network; The sixth module is used to calculate the real-time joint angle information of the robot according to the optimized zeroing neural network, and control the robot to automatically spray the jeans according to the real-time joint angle information; The expression of the kinematic model of the robot's multiple joints of the mechanical arm is: f(θ(t))=q(t) Among them, f(·) is a nonlinear mapping function used to describe the kinematics of the robot arm; θ(t) represents the angle value of the m joint angles of the robot arm at time t; q(t) represents the vector composed of the position and posture of the end of the robot arm in the n-dimensional task space; the calculation formula of the joint angle is: θ(t)=g(q d (t)) Among them, θ(t) represents the angle value of the m joint angles of the robot arm at time t; q d (t) represents the target path and posture vector; g(·) represents the zeroing neural network formula.

3. An electronic device, characterized in that: including a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method as claimed in claim 1.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and the program is executed by a processor to implement the method as claimed in claim 1.

5. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method as claimed in claim 1 is implemented.

Images