Data processing method and device of mechanical arm, electronic equipment, computer readable storage medium and computer program product

Abstract

The invention provides a data processing method and device of a mechanical arm, electronic equipment, a computer readable storage medium and a computer program product. The method comprises the steps that when the mechanical arm moves along a preset excitation track at multiple moments, motion data vectors of n joints and moment vectors of h joints are collected; based on the motion data vectors of the n joints at the multiple moments, multiple first regression matrixes corresponding to the n joints at the multiple moments are generated; modifying the plurality of first regression matrixes by using a preset selection matrix to obtain a plurality of second regression matrixes corresponding to the h joints at a plurality of moments, and establishing a first association relationship among the second regression matrixes, the kinetic parameters and the moment vectors of the h joints; and on the basis of the multiple second regression matrixes, the multiple torque vectors corresponding to the h joints and the first incidence relation, the kinetic parameter vector of the mechanical arm is determined. By means of the method and device, the complete kinetic parameters can be accurately obtained only through the moment vectors of part of the joints.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical arms, in particular to a data processing method and device of a mechanical arm, an electronic device, a computer readable storage medium and a computer program product. BACKGROUND

[0002] The dynamic parameters of a mechanical arm are the data basis for the mechanical arm to realize high-performance control (such as admittance control, impedance control and other compliant control). The traditional dynamic parameter identification method usually needs to install a torque sensor at each joint of the mechanical arm, so as to obtain complete moment information of all joints during the movement of the mechanical arm. However, due to the high purchase cost of multi-joint torque sensors, and the installation and debugging of torque sensors, long-term maintenance not only requires a large amount of time and manpower, but also interferes with the lightweight design and motion flexibility of the mechanical arm, resulting in high overall cost of traditional dynamic parameter identification. SUMMARY

[0003] The embodiments of the present application provide a data processing method, device, electronic device, computer readable storage medium and computer program product of a mechanical arm, which can obtain the dynamic parameters of the mechanical arm by using only the moment vectors of part of the joints.

[0004] The technical solutions of the embodiments of the present application are as follows: The embodiments of the present application provide a data processing method of a mechanical arm, which comprises: controlling the mechanical arm to move along a preset excitation trajectory, and collecting motion data vectors of n joints of the mechanical arm at multiple time points and moment vectors of h joints of the mechanical arm at multiple time points during the movement, wherein n is the total number of joints of the mechanical arm, and h is a positive integer less than n; generating multiple first regression matrices respectively corresponding to the n joints at the multiple time points based on the motion data vectors of the n joints at the multiple time points; modifying the multiple first regression matrices by using a preset selection matrix to obtain multiple second regression matrices respectively corresponding to the h joints at the multiple time points, and establishing a first correlation relationship among the second regression matrices, the dynamic parameters and the moment vectors of the h joints; determining the dynamic parameter vector of the mechanical arm based on the multiple second regression matrices, the moment vectors of the h joints respectively corresponding to the multiple time points and the first correlation relationship.

[0005] The embodiments of the present application provide a data processing device of a mechanical arm, which comprises: The mechanical arm parameter acquisition module is configured to control the mechanical arm to move along a preset excitation track, and to acquire motion data vectors of n joints of the mechanical arm at multiple time points and torque vectors of h joints of the mechanical arm at the multiple time points during the movement, where n is a total number of joints of the mechanical arm, and h is a positive integer less than n. The dynamics parameter calculation module is configured to generate a plurality of first regression matrices respectively corresponding to the n joints at the multiple time points based on the motion data vectors of the n joints at the multiple time points, modify the plurality of first regression matrices by using a preset selection matrix to obtain a plurality of second regression matrices respectively corresponding to the h joints at the multiple time points, and establish a first correlation relationship among the second regression matrices, the dynamics parameters, and the torque vectors of the h joints. The dynamics parameter vector of the mechanical arm is determined based on the plurality of second regression matrices, the torque vectors of the h joints at the multiple time points, and the first correlation relationship.

[0006] In the above scheme, the mechanical arm parameter acquisition module is further configured to acquire position values of the n joints of the mechanical arm at the multiple time points, respectively combine the positions of the n joints at the multiple time points to obtain a position vector of the n joints of the mechanical arm, perform central difference calculation on positions of a previous time point and a next time point of each time point among the multiple time points to obtain velocity values of the n joints of the mechanical arm at the multiple time points, respectively combine the velocity values of the n joints at the multiple time points to obtain a velocity vector of the n joints of the mechanical arm, perform central difference calculation on velocities of the previous time point and the next time point of each time point among the multiple time points to obtain acceleration values of the n joints of the mechanical arm at the multiple time points, respectively combine the acceleration values of the n joints at the multiple time points to obtain an acceleration vector of the n joints of the mechanical arm, acquire torque values of the h joints of the mechanical arm at the multiple time points, and respectively combine the torque values of the h joints at the multiple time points to obtain a torque vector of the h joints of the mechanical arm.

[0007] In the above scheme, the mechanical arm parameter acquisition module is further configured to perform central difference calculation on velocity values of a previous time point and a next time point of each time point to obtain initial acceleration values of the n joints of the mechanical arm at the multiple time points, and perform low-pass filtering processing on the initial acceleration values of the n joints at the multiple time points to obtain the acceleration values of the n joints of the mechanical arm at the multiple time points.

[0008] In the scheme, the dynamics parameter calculation module is further configured to multiply the preset selection matrix with the plurality of first regression matrices respectively to obtain a plurality of second regression matrices corresponding to the h joints at the plurality of time points respectively; and establish the first correlation relationship based on the preset selection matrix and a preset second correlation relationship, wherein the preset second correlation relationship is a relationship among the first regression matrices, the dynamics parameters and the torque vector of the n joints.

[0009] In the scheme, the dynamics parameter calculation module is further configured to obtain a zero matrix with h rows and n columns; obtain a joint serial number j of the n joints corresponding to an i th joint of the h joints for the i th joint i , and set a value of an element in an i th row and j th column of the zero matrix to 1 to obtain the preset selection matrix. i

[0010] In the scheme, the dynamics parameter calculation module is further configured to multiply the preset selection matrix with the torque vector of the n joints to obtain a torque vector of the h joints; replace the torque vector of the n joints in the second correlation relationship with the torque vector of the h joints, and replace the first regression matrix in the second correlation relationship with the second regression matrix to establish the first correlation relationship In the scheme, the dynamics parameter calculation module is further configured to substitute the plurality of second regression matrices and the torque vectors of the h joints at the plurality of time points into the first correlation relationship respectively to obtain an equation group conforming to the first correlation relationship; integrate the plurality of second regression matrices in the equation group conforming to the first correlation relationship into a comprehensive regression matrix, integrate the torque vectors of the h joints at the plurality of time points in the equation group conforming to the first correlation relationship into a comprehensive torque matrix, obtain a relationship among the comprehensive regression matrix, the dynamics parameters and the comprehensive torque matrix; and solve the relationship by using a least square method to obtain the dynamics parameter vector.

[0011] An electronic device is provided in an embodiment of the present application, and the electronic device comprises: a memory configured to store computer executable instructions or computer programs; a processor configured to execute the computer executable instructions or computer programs stored in the memory to implement the data processing method of the mechanical arm provided in the embodiments of the present application.

[0012] A computer readable storage medium is provided in an embodiment of the present application, and the computer readable storage medium stores computer programs or computer executable instructions, and is configured to implement the data processing method of the mechanical arm provided in the embodiments of the present application when executed by a processor.

[0013] ​The embodiment of the present application provides a computer program product, comprising a computer program or computer executable instructions, which, when executed by a processor, implement the data processing method of the mechanical arm provided by the embodiment of the present application.

[0014] The embodiment of the present application has the following beneficial effects: By selectively collecting the moment, the moment collection device (such as a torque sensor) does not need to be provided for all joints, thereby reducing the hardware complexity and cost of the identification of the mechanical arm dynamics parameters. Moreover, by introducing the selection matrix, the complete first regression matrix based on the motion data vector of all n joints is accurately trimmed, and the second regression matrix strictly corresponding to the h target joint moments is directly constructed, and then the redundant high-dimensional problem of describing the dynamics of all joints of the mechanical arm is losslessly converted into the simplified low-dimensional first correlation relationship of the h joints of the mechanical arm. In addition, since the generated first correlation relationship is a linear equation relationship with smaller scale and higher calculation efficiency, it not only retains the global coupling effect, but also focuses on the effective target information, thereby significantly improving the robustness and calculation efficiency of the solution of the dynamics parameter vector. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 FIG. 1 is a structural schematic diagram of a data processing system architecture of a mechanical arm provided by the embodiment of the present application; Figure 2 FIG. 2 is a structural schematic diagram of an electronic device provided by the embodiment of the present application; Figure 3 FIG. 3 is a first flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 4 FIG. 4 is a second flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 5 FIG. 5 is a third flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 6 FIG. 6 is a fourth flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 7 FIG. 7 is a fifth flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 8 FIG. 8 is a sixth flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 9 FIG. 9 is a seventh flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 10 FIG. 10 is an eighth flowchart of a data processing method of a mechanical arm provided by the embodiment of the present application; Figure 11is a structural schematic diagram of a mechanical arm provided by an embodiment of the present application. Figure 12 is a ninth flowchart of a data processing method of a mechanical arm provided by an embodiment of the present application.

[0016] It should be noted that the above-mentioned "first", "second" are only used to distinguish different schemes, and do not represent the advantages or disadvantages of the schemes or the priority in the implementation process. DETAILED DESCRIPTION

[0017] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be described in further detail below with reference to the drawings, and the described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present application.

[0018] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0019] In the following description, the terms "first", "second", "third" are only used to distinguish similar objects, and do not represent a specific order of the objects. It can be understood that "first", "second", "third" can be interchanged in a specific order or sequence as allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0020] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined target, and can be implemented entirely or partially by using software, hardware (such as a processing circuit or a memory) or a combination thereof. Similarly, one processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of an overall module or unit that includes the functions of the module or unit.

[0021] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meanings as those commonly understood by those skilled in the art. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0022] In the embodiments of the present application, the relevant data collection and processing should strictly comply with the requirements of relevant laws and regulations, obtain the informed consent or separate consent of the personal information subject, and within the scope of authorization of laws and regulations and the personal information subject, carry out subsequent data use and processing.

[0023] Before further detailing the embodiments of the present application, the terms and phrases involved in the embodiments of the present application are explained, and the terms and phrases involved in the embodiments of the present application are applicable to the following explanations.

[0024] 1) Selection matrix: The selection matrix is a special matrix mainly composed of 0 and 1, which functions to "select" or "extract" specific elements of interest from a larger vector or matrix and arrange them into a new, smaller vector or matrix. It realizes the screening and reorganization of data through matrix multiplication operation.

[0025] 2) Manipulator: The manipulator is a machine device composed of a series of rigid links connected by multiple joints (such as rotary joints or moving joints), which can imitate the function of human arms to perform tasks such as grasping, carrying, assembling or machining in three-dimensional space.

[0026] 3) Kinetic parameters: Kinetic parameters are quantitative values of a series of inherent physical properties that describe how an object or system (such as a manipulator) moves under the action of external force. Kinetic parameters are the mathematical basis for establishing accurate kinetic models, determining the relationship between force / torque and acceleration. For a certain link of a manipulator, its kinetic parameters usually refer to: the mass of the link (which determines its inertia), the center of mass position (which determines the point of action of gravity and the distribution of inertia), and the inertia tensor (which describes the difficulty of rotating the object around its center of mass).

[0027] 4) Excitation trajectory: The excitation trajectory is a specially designed motion path for the manipulator to execute, which aims to fully "excite" or "reveal" the influence of all to-be-identified kinetic parameters on the dynamic behavior of the system during the motion, so as to ensure that the collected data can uniquely and accurately solve these parameters.

[0028] 5) Low-pass filtering: Low-pass filtering is a signal processing technique that allows low-frequency components of a signal to pass through while attenuating or suppressing high-frequency components. In kinetic parameter identification, it is mainly used to filter out high-frequency noise in the original data collected from sensors (such as motor encoders) to extract smooth and true motion information.

[0029] 6) Regression matrix: The regression matrix is a numerical matrix calculated according to the motion state of the manipulator (such as joint position, velocity, acceleration), which functions to establish a mathematical bridge between unknown linear kinetic parameters and measurable joint torques. Each column of the regression matrix corresponds to a to-be-identified kinetic parameter, and the values in the column represent the contribution coefficient of the parameter to the total torque.

[0030] 7) Least squares: Least squares is a mathematical optimization technique used to find the best parameters of a mathematical model from a set of data that contains measurement errors. Its core principle is to find a set of parameters that minimize the sum of the squares of the errors between the predicted values calculated by the model and the actual observations.

[0031] 8) Generalized inverse matrix: Generalized inverse matrix (also known as pseudo-inverse) is a generalization of traditional matrix inverse operation, which makes non-square or non-invertible square matrices also have a "inverse" with similar properties. In solving linear equations Yπ=T, if Y does not have a traditional inverse, the generalized inverse matrix Y + can provide a "best" solution, i.e. the least squares solution.

[0032] The dynamics parameters of the robot arm are the data basis for the robot arm to realize high-performance control (such as admittance control, impedance control, etc. compliant control). The traditional dynamics parameter identification method usually needs to install torque sensors at each joint of the robot arm, so as to obtain complete torque information of all joints during the movement of the robot arm. However, in actual application, considering the cost control, most robot arms do not configure torque sensors on all joints, but only on some key joints. For this kind of robot arm with only partial joint torque information measurable, the traditional identification method is no longer applicable because it cannot obtain the complete torque vector of all joints, which leads to the difficulty in accurately identifying the complete dynamics parameters. Therefore, it is urgent to propose a new identification method to solve the dynamics parameter identification problem of the robot arm with torque sensors installed only on partial joints.

[0033] The embodiments of the present application provide a data processing method, device and equipment of a robot arm, a computer readable storage medium and a computer program product, which can accurately obtain complete dynamics parameters by using only partial joint torque vectors. The following describes an exemplary application of the electronic device provided by the embodiments of the present application. The electronic device provided by the embodiments of the present application can be implemented as a notebook computer, a tablet computer, a desktop computer, a set-top box, a smart phone, a smart speaker, a smart watch, a smart television, a vehicle-mounted terminal, and various types of terminals. It can also be implemented as a server. The electronic device can also be integrated inside a robot and directly controlled by a controller of the robot. The electronic device can also be connected to the robot as an independent external control device and controlled by an external control.

[0034] The following describes an exemplary application of the electronic device implemented as a terminal or a server.

[0035] Referring to Figure 1 , Figure 1 is an architecture schematic diagram of the data processing system 100 of the robot arm provided by the embodiments of the present application, Figure 1The application relates to a server 200, a network 300 and a terminal 400. The terminal 400 (an example of a graphic interface 410) is connected to the server 200 through the network 300, which can be a wide area network or a local area network, or a combination of the two.

[0036] In some embodiments, the application can be implemented by the server and the terminal. For example, the terminal 400 controls the movement of the mechanical arm along the preset excitation track, and collects the motion data vectors of n joints of the mechanical arm at multiple time points and the moment vectors of h joints of the mechanical arm at multiple time points during the movement, wherein n is the total number of joints of the mechanical arm, and h is a positive integer less than n; the terminal 400 sends the motion data vectors of n joints at multiple time points and the moment vectors of h joints at multiple time points to the server 200; the server 200 generates a plurality of first regression matrices corresponding to n joints at multiple time points respectively based on the motion data vectors of n joints at multiple time points; the server 200 modifies the plurality of first regression matrices by using a preset selection matrix to obtain a plurality of second regression matrices corresponding to h joints at multiple time points respectively, and establishes a first correlation relationship among the second regression matrices, the dynamic parameters and the moment vectors of h joints; and the server 200 determines the dynamic parameter vector of the mechanical arm based on the plurality of second regression matrices, the moment vectors of h joints at multiple time points respectively and the first correlation relationship.

[0037] In some embodiments, the application can be implemented by the terminal alone. For example, the terminal 400 controls the movement of the mechanical arm along the preset excitation track, and collects the motion data vectors of n joints of the mechanical arm at multiple time points and the moment vectors of h joints of the mechanical arm at multiple time points during the movement, wherein n is the total number of joints of the mechanical arm, and h is a positive integer less than n; the terminal 400 generates a plurality of first regression matrices corresponding to n joints at multiple time points respectively based on the motion data vectors of n joints at multiple time points; the terminal 400 modifies the plurality of first regression matrices by using a preset selection matrix to obtain a plurality of second regression matrices corresponding to h joints at multiple time points respectively, and establishes a first correlation relationship among the second regression matrices, the dynamic parameters and the moment vectors of h joints; and the terminal 400 determines the dynamic parameter vector of the mechanical arm based on the plurality of second regression matrices, the moment vectors of h joints at multiple time points respectively and the first correlation relationship.

[0038] In some embodiments, the server 200 can be a stand-alone physical server, a server cluster or a distributed system composed of multiple physical servers, a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and basic cloud computing services such as big data and artificial intelligence platforms. The terminal and the server can be connected directly or indirectly through wired or wireless communication, which is not limited in the embodiments of the present application. The data processing method of the mechanical arm provided in the embodiments of the present application can be used in the scene of compliant control algorithm such as development admittance control and impedance control of the mechanical arm equipped with a torque sensor on part of the joints, for example: 1) Industrial motor end cover precision assembly scene: In the assembly process of the motor stator and the end cover, the mechanical arm needs to realize "light touch and accurate alignment" operation based on admittance control to avoid part deformation caused by rigid collision. During operation, the mechanical arm synchronously collects motion data such as positions and angular velocities of all joints, and torque data of two key joints of the wrist and the elbow; the data processing method of the mechanical arm provided in the embodiments of the present application is used to generate a first regression matrix corresponding to all joints, a second regression matrix matching part of the torque is obtained through trimming and optimization, and finally the association between the matrix and the dynamic parameters is established, and a complete parameter vector is solved. Based on the admittance control algorithm optimized based on the parameters, the assembly force can be adjusted in real time, and the cost of the traditional full-sensor solution can be significantly reduced.

[0039] 2) Service robot household fragile product carrying scene: When the household robot carries glass tableware, it needs to rely on impedance control to buffer the ground bump or obstacle interference to prevent the tableware from breaking. During the carrying process, the robot collects motion data such as accelerations and displacements of all joints, and torque feedback of two load-bearing joints of the shoulder and the hand. The data processing method of the mechanical arm provided in the embodiments of the present application is used to generate a first regression matrix corresponding to all joints, and a second regression matrix adapted to part of the torque is obtained through selective modification, and the dynamic parameters are efficiently solved. The optimized impedance control algorithm can quickly offset the external force impact, and solves the problem of difficult landing in household scenarios caused by the high cost of full-joint sensors.

[0040] 3) Medical rehabilitation robot upper limb training scene: when the rehabilitation robot assists the stroke patient to carry out the upper limb lifting training, the "follows the force" needs to be realized based on the admittance control, and the secondary damage to the patient is avoided. During the training, the robot collects all joint motion angle, speed data of the upper limb of the patient driving the mechanical arm, and torque data of two core joints of the upper arm and forearm; the construction and simplification of the regression matrix are completed through the data processing method of the mechanical arm provided in the application, and the dynamic parameter vector is accurately obtained. The admittance control based on the parameter can match the patient muscle strength adjustment auxiliary degree, while reducing the hardware cost, the safety and effectiveness of the rehabilitation training are guaranteed.

[0041] Referring to Figure 2 , Figure 2 is a structural schematic diagram of an electronic device 500 provided by an embodiment of the application, Figure 2 The electronic device 500 shown in the figure includes at least one processor 510, a memory 550, at least one network interface 520, and a user interface 530. Various components in the electronic device 500 are coupled together through a bus system 540. It can be understood that the bus system 540 is used to realize the connection communication between the components. In addition to including a data bus, the bus system 540 also includes a power bus, a control bus, and a status signal bus. However, for the purpose of clear illustration, all kinds of buses are marked as the bus system 540 in Figure 2 .

[0042] The processor 510 can be an integrated circuit chip with a signal processing capability, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0043] The user interface 530 includes one or more output devices 531 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 530 also includes one or more input devices 532 that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0044] The memory 550 can be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard drives, optical drives, etc. The memory 550 optionally includes one or more storage devices physically located in proximity to the processor 510.

[0045] The memory 550 includes volatile memory or nonvolatile memory, and can include both volatile and nonvolatile memory. The nonvolatile memory can be read only memory (ROM), and the volatile memory can be random access memory (RAM). The memory 550 described in the embodiments of the present application is intended to include any suitable type of memory.

[0046] In some embodiments, the memory 550 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, which are exemplarily illustrated below.

[0047] The operating system 551 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks. The network communication module 552 is used to communicate with other electronic devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 including Bluetooth, wireless compatibility certification (WiFi), and universal serial bus (USB), and the like. The presentation module 553 is used to enable the presentation of information via one or more output devices 531 associated with the user interface 530 (e.g., a display screen, a speaker, and the like) (e.g., a user interface for operating peripheral devices and displaying content and information). The input processing module 554 is used to detect and interpret one or more user inputs or interactions from one or more input devices 532.

[0048] In some embodiments, the device provided by the embodiments of the present application can be implemented in software, Figure 2 A data processing device 555 of a robot arm stored in the memory 550 is shown, which can be software in the form of programs and plug-ins, including the following software modules: a robot arm parameter acquisition module 5551 and a dynamics parameter calculation module 5552, which are logical, and thus can be combined or further split according to the functions implemented. The functions of each module will be described below.

[0049] In some embodiments, the apparatus provided by the embodiments of the present application can be implemented in a hardware manner. For example, the apparatus provided by the embodiments of the present application can be a processor in the form of a hardware decoding processor programmed to perform the data processing method of the mechanical arm provided by the embodiments of the present application. For example, the processor in the form of a hardware decoding processor can be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Programmable Logic Devices (PLDs), Complex Programmable Logic Devices (CPLDs), Field-Programmable Gate Arrays (FPGAs), or other electronic elements.

[0050] The data processing method of the mechanical arm provided by the embodiments of the present application will be described below. As described above, the electronic device implementing the data processing method of the mechanical arm according to the embodiments of the present application can be a terminal, a server, or a combination of the two. Therefore, the execution subject of each step will not be described again below.

[0051] Referring to Figure 3 , Figure 3 FIG. 1 is a first flowchart of the data processing method of the mechanical arm provided by the embodiments of the present application. The steps shown in Figure 3 will be described below. Figure 3 The subject of the step is an electronic device.

[0052] In step 101, the mechanical arm is controlled to move along a preset excitation trajectory, and motion data vectors of n joints of the mechanical arm at multiple time points and moment vectors of h joints of the mechanical arm at multiple time points are collected during the movement, where n is the total number of joints of the mechanical arm, and h is a positive integer less than n.

[0053] Here, the excitation trajectory refers to one or more movement paths specially designed in the field of robot dynamics and control for identifying the dynamic parameters (such as mass, inertia, friction coefficient, etc.) of the system. The excitation trajectory is used to let the mechanical arm perform a series of multi-dimensional characteristic information movements, so as to sufficiently “excite” or “stimulate” its inherent dynamic characteristics, so that the sensors (such as encoders, torque sensors) can measure sufficient effective data.

[0054] Here, the motion data vector refers to a set of kinematic states of all n joints of the robot arm during the motion process. The motion data vector is a set of time series data that changes over time, which collectively constitutes the necessary input for dynamic parameter identification.

[0055] In embodiments of the present application, by controlling the robot arm to execute a specific excitation trajectory, it is ensured that the collected data can fully reflect the dynamic characteristics of the robot arm. The collected data accurately records the kinematic states of all joints of the robot arm at multiple time instants, as well as the dynamic responses (i.e., torques) of some joints at these same time instants, providing a complete and reliable empirical input and data basis for subsequently establishing a mathematical correlation between the motion data vector and the torque.

[0056] In some embodiments, the multiple time instants can be set as a series of consecutive time instants, wherein the time difference between any two adjacent time instants is the same control period.

[0057] In some embodiments, the motion data vector of the n joints at the multiple time instants can be embodied by multiple sets of motion data vectors of the n joints. Each time instant corresponds to a set of motion data vectors.

[0058] In some embodiments, the torque vector of the h joints at the multiple time instants can be embodied by multiple torque vectors. Specifically, each time instant corresponds to a 1xh torque vector, and each torque vector contains h elements corresponding to the h joints, each element representing the torque value of the corresponding joint at that time instant.

[0059] In some embodiments, the motion data vector includes a position vector, a velocity vector, and an acceleration vector. In step 101, the motion data vector of the n joints of the robot arm at the multiple time instants during the motion process can be collected by the following methods: collecting the positions of the n joints of the robot arm at the multiple time instants through an encoder, respectively combining the positions of the n joints at the multiple time instants to obtain the position vector of the n joints of the robot arm; collecting the velocity values of the n joints of the robot arm at the multiple time instants through a tachometer, respectively combining the velocity values of the n joints at the multiple time instants to obtain the velocity vector of the n joints of the robot arm; and collecting the acceleration values of the n joints of the robot arm at the multiple time instants through an acceleration sensor, respectively combining the acceleration values of the n joints at the multiple time instants to obtain the acceleration vector of the n joints of the robot arm.

[0060] In some embodiments, the positions of the n joints of the robot arm at the multiple time instants can be collected by the encoder in the following manner: an encoder is installed on each joint of the robot arm, and the values of the n encoders are read synchronously at each sampling time instant to obtain the accurate positions of the n joints at each sampling time instant. By way of example, the encoder can be an absolute value encoder or an incremental encoder.

[0061] In some embodiments, n encoders may be mounted on the motor of each joint.

[0062] In some embodiments, the speed values ​​of n joints of a robotic arm at multiple moments can be acquired by using a tachogenerator. This can be achieved by coaxially mounting a tachogenerator on the motor shaft of each joint, with the tachogenerator outputting a voltage signal proportional to the motor speed. By acquiring these voltage signals and performing calibration conversion, the speed values ​​of the n joints at each moment can be obtained.

[0063] In some embodiments, acquiring the acceleration values ​​of n joints of a robotic arm at multiple moments using accelerometers can be achieved by installing accelerometers at key locations near each link or joint of the robotic arm, and directly measuring the linear or angular acceleration values ​​of the components using the accelerometers. For example, the accelerometer can be a microelectromechanical system (MEMS) accelerometer.

[0064] In other embodiments, the motion data vector includes a position vector, a velocity vector, and an acceleration vector, see [link to relevant documentation]. Figure 4 , Figure 4 This is a second flowchart illustrating the data processing method for the robotic arm provided in this application embodiment. Figure 4 The step 101, "collecting motion data vectors of n joints of the robotic arm at multiple moments and torque vectors of h joints during the motion process," can be achieved through the following steps 1011 to 1014: In step 1011, the positions of the n joints of the robotic arm at multiple times are collected, and the positions of the n joints at multiple times are combined to obtain the position vectors of the n joints of the robotic arm.

[0065] In some embodiments, the positions of n joints of the robotic arm at multiple times can be acquired by mounting encoders on the motors of each joint. For example, the encoders can be absolute encoders and incremental encoders.

[0066] Taking the acquisition of the positions of the six joints of the robotic arm at time k as an example, the position of each joint is denoted as... ,get , , , , , Where i represents the i-th joint of the robotic arm. This represents the position of the i-th joint of the robotic arm at the k-th moment.

[0067] In some embodiments, the positions of the n joints of the robot arm at the plurality of time points are combined respectively to obtain position vectors of the n joints of the robot arm, which can be achieved by the following manner: for each time point, the positions of the n joints of the robot arm are sequentially written into the position vector according to the serial numbers of the joints, to obtain the position vectors of the n joints of the robot arm corresponding to the plurality of time points respectively.

[0068] Here, the joint sequence of the robot arm refers to a specific arrangement manner of the joints thereof, which can be from right to left or from left to right.

[0069] Taking the above example, the position vector of the 6 joints of the robot arm at the kth time point is: (1) wherein, represents the position vector of the 6 joints of the robot arm at the kth time point, represents the position of the 1st joint at the kth time point, represents the position of the 2nd joint at the kth time point, represents the position of the 6th joint at the kth time point.

[0070] In step 1012, for each time point of the n joints at the plurality of time points, the positions of the previous time point and the next time point of each time point are used to perform central difference calculation, to obtain velocity values of the n joints of the robot arm at the plurality of time points, and the velocity values of the n joints at the plurality of time points are combined respectively to obtain velocity vectors of the n joints of the robot arm.

[0071] In some embodiments, the positions of the previous time point and the next time point of each time point are used to perform central difference calculation to obtain velocity values of the n joints of the robot arm at the plurality of time points, which can be calculated by the following formula (2): ( ) / 2Δt (2) wherein, represents the velocity value of the i th joint of the robot arm at the k th time point, represents the position of the i th joint of the robot arm at the k+1 th time point, represents the position of the i th joint of the robot arm at the k-1 th time point, and Δt represents the time interval of acquisition.

[0072] Taking the above example, the velocity vector of the 6 joints of the robot arm at the kth time point is: (3) wherein, represents the velocity vector of the 6 joints of the robot arm at the kth time point, represents the velocity value of the 1st joint at the kth time point, denotes the velocity value of the 2nd joint at the kth moment, denotes the velocity value of the 6th joint at the kth moment.

[0073] In step 1013, for each of the plurality of moments, the acceleration values of the n joints of the robot arm at the plurality of moments are calculated by using the central difference calculation of the velocity of the previous moment and the next moment of each moment, and the acceleration values of the n joints at the plurality of moments are combined respectively to obtain the acceleration vector of the n joints of the robot arm.

[0074] In some embodiments, the central difference calculation of the velocity of the previous moment and the next moment of each moment can be calculated by formula (4): ( ) / 2Δt (4) wherein, denotes the acceleration value of the i th joint of the robot arm at the k th moment, denotes the velocity value of the i th joint of the robot arm at the k+1 th moment, denotes the velocity value of the i th joint of the robot arm at the k-1 th moment, and Δt denotes the time interval of acquisition.

[0075] Based on the above example, the acceleration vector of the 6 joints of the robot arm at the k th moment is: (5) wherein, denotes the acceleration vector of the 6 joints of the robot arm at the k th moment, denotes the acceleration value of the 1st joint at the k th moment, denotes the acceleration value of the 2nd joint at the k th moment, denotes the acceleration value of the 6th joint at the k th moment.

[0076] In some embodiments, Figure 5 The third flowchart of the data processing method of the robot arm provided by the embodiments of the present application, Figure 5 The step 1013 of "calculating the acceleration values of the n joints of the robot arm at the plurality of moments by using the central difference calculation of the velocity of the previous moment and the next moment of each moment" can be realized by the following steps 10131 to 10132, which will be described in detail below.

[0077] In step 10131, the initial acceleration values of the n joints of the robot arm at the plurality of moments are calculated by using the central difference calculation of the velocity value of the previous moment and the next moment of each moment.

[0078] In some embodiments, the initial acceleration values of the n joints of the robot arm at the plurality of time points are obtained by using the central difference calculation of the velocity values of the previous time point and the next time point of each time point, which can be calculated by the following formula (6): ( ) / 2Δt (6) wherein, represents the initial acceleration value of the i th joint of the robot arm at the k th time point, represents the velocity value of the i th joint of the robot arm at the k+1 th time point, represents the velocity value of the i th joint of the robot arm at the k-1 th time point, and Δt represents the time interval of acquisition.

[0079] In step 10132, the initial acceleration values of the n joints at the plurality of time points are subjected to low-pass filtering processing to obtain the acceleration values of the n joints of the robot arm at the plurality of time points.

[0080] In the embodiments of the present application, since the initial acceleration is calculated by the central difference method of the velocity of the previous time point and the next time point, it is inevitable to introduce and amplify the high-frequency noise in the initial acceleration value. By subjecting the initial acceleration value to low-pass filtering processing, the high-frequency noise in the initial acceleration value is effectively filtered out, and the acceleration value reflecting the true motion trend is accurately extracted.

[0081] In some embodiments, the initial acceleration values of the n joints at the plurality of time points can be subjected to low-pass filtering processing by Kalman filtering or Butterworth low-pass filtering.

[0082] In step 1014, the torque values of the h joints of the robot arm at the plurality of time points are acquired, and the torque values of the h joints at the plurality of time points are combined respectively to obtain the torque vector of the h joints of the robot arm at the plurality of time points.

[0083] In the embodiments of the present application, the velocity and acceleration data are derived from the directly measured position data by using the central difference method, which avoids equipping each joint with a special velocity sensor or acceleration sensor, thereby simplifying the hardware configuration of the robot arm and reducing the hardware cost. At the same time, by this calculation method, all data (i.e. position, velocity, acceleration, torque value) are acquired or calculated based on the same set of time reference, and this strict time alignment ensures that the motion state at any time point accurately corresponds to the dynamic response at the time point in the subsequent analysis, which provides a basic guarantee for the accuracy of parameter identification.

[0084] In some embodiments, the torque vectors of the h joints of the robot arm at the plurality of time points can be obtained by combining the torque values of the h joints at the plurality of time points respectively, which can be achieved by the following manner: for each time point, the torque values of the h joints of the robot arm are sequentially written into a position vector according to the joint number, to obtain the torque vectors of the h joints of the robot arm corresponding to the plurality of time points respectively.

[0085] Taking the above example, assuming that only the torques of the first, second, fourth and fifth joints of the robot arm are collected, the torque vectors of the four joints of the robot arm are: (7) wherein, represents the torque vector of the four joints of the robot arm at the kth time point, represents the torque value of the first joint at the kth time point, represents the torque value of the second joint at the kth time point, represents the torque value of the fourth joint at the kth time point, represents the torque value of the fifth joint at the kth time point.

[0086] In some embodiments, the torque values of the h joints of the robot arm at the plurality of time points can be collected by the torque sensor, specifically, the torque sensor can be installed on the joint to be measured of the robot arm, and the sensitive element inside the torque sensor, such as a strain gauge, can be used to convert the physical deformation of the sensitive element caused by bearing torque into an electrical signal (such as voltage or current signal) that is proportional to the physical deformation.

[0087] In some preferred embodiments, the torque sensor can be installed between the output end of the joint reducer (such as a harmonic reducer) of the joint to be measured and the connecting rod driven by the joint to be measured. Such installation can ensure that the torque sensor measures the net output torque after deceleration and torque increase, which is finally used to drive the subsequent connecting rod and overcome all loads such as gravity and inertia, thereby improving the accuracy of torque collection.

[0088] In step 102, based on the motion data vectors of the n joints at the plurality of time points, a plurality of first regression matrices corresponding to the n joints at the plurality of time points respectively are generated.

[0089] In some embodiments, step 102 can be achieved by the following manner: establishing a standard dynamic equation of the robot arm; performing parameter linearization processing on the standard dynamic equation to obtain a theoretical form of the first regression matrix; substituting the motion data vectors at the plurality of time points into the theoretical form of the first regression matrix to obtain the plurality of first regression matrices corresponding to the n joints at the plurality of time points respectively.

[0090] In some embodiments, the parameter linearization processing of the standard dynamics equation can be realized by the following manner: expanding the torque vector of each joint in the standard dynamics equation to obtain the expanded torque equation of each joint; traversing the expanded torque equation of each joint to identify the constant physical parameters to be identified as the standard dynamics parameters; and algebraically reorganizing the expanded standard dynamics equation so that each standard dynamics parameter is multiplied by a function item composed of only the motion data vector, and the function item combination is taken as the theoretical form of the first regression matrix.

[0091] In some embodiments, the torque vector of each joint in the standard dynamics equation can be expanded by using the Lagrange method or the Newton-Euler method to obtain the expanded torque equation of each joint.

[0092] In step 103, the plurality of first regression matrices are modified by using the preset selection matrix to obtain a plurality of second regression matrices corresponding to the h joints at a plurality of time points respectively, and a first association relationship between the second regression matrix, the dynamics parameter, and the torque vector of the h joints is established.

[0093] Here, the selection matrix refers to a special matrix composed of 0 and 1 in advance. The selection matrix is used to accurately and sequentially extract the specified row or column from a larger matrix or vector to construct a new smaller matrix or vector.

[0094] In some embodiments, Figure 6 is a fourth flowchart of the data processing method of the mechanical arm provided by the embodiments of the present application, Figure 6 It is shown that step 103 can be realized by step 1031 to step 1032, which will be explained in detail below.

[0095] In step 1031, the preset selection matrix is multiplied by the plurality of first regression matrices to obtain a plurality of second regression matrices corresponding to the h joints at a plurality of time points respectively: (8) Wherein, D is the selection matrix, is the first regression matrix at the k th time point, is the second regression matrix at the k th time point.

[0096] In some embodiments, the preset selection matrix is an h x n matrix, the plurality of first regression matrices are n x p matrices, and the h x n selection matrix is multiplied by the plurality of n x p first regression matrices to obtain a plurality of h x p second regression matrices.

[0097] In step 1032, the first correlation relationship is established based on the preset selection matrix and the preset second correlation relationship, where the preset second correlation relationship is a relationship among the first regression matrix, the kinetic parameters, and the torque vector of the n joints.

[0098] In the embodiments of the present application, by multiplying the preset selection matrix and the first regression matrix corresponding to the n joints, information related only to the h measurable torque joints can be accurately extracted from the first regression matrix, and a dimension-adapted hxp second regression matrix is generated. This processing process effectively filters out the redundant interference caused by the unknown torque joints, and makes the dimension of the second regression matrix completely match the actually measured h joint torque vector, thereby solving the problem that the original n-dimensional regression matrix and the h-dimensional torque vector cannot be directly adapted and processed. In addition, with the aid of the preset selection matrix, the second correlation relationship among the n-dimensional second regression matrix, the n-dimensional torque vector and the kinetic parameter vector can also be successfully reduced to the first correlation relationship among the h-dimensional second regression matrix, the h-dimensional torque vector and the kinetic parameter vector. This dimension reduction processing avoids the problem that the kinetic parameters cannot be identified due to the unobtainable part of the joint torque in the n-dimensional torque vector, and finally realizes the accurate identification of the kinetic parameters under the condition that only part of the joint torque is known.

[0099] In some embodiments, the preset second correlation relationship can be an equation relationship that the product of the first regression matrix and the kinetic parameters is equal to the torque vector of the n joints, which conforms to the following formula (9): (9) wherein, is the first regression matrix at the kth moment, is the kinetic parameter vector, is the torque vector of the n joints at the kth moment.

[0100] In some embodiments, Figure 7 is a fifth flowchart of the data processing method provided by the embodiments of the present application, Figure 7 It is shown that step 1032 can be implemented by steps 10321 to 10322, which will be described in detail below.

[0101] In step 10321, the preset selection matrix is multiplied by the torque vector of the n joints to obtain the torque vector of the h joints.

[0102] For ease of understanding, the relationship among the torque vector of the h joints, the preset selection matrix, and the torque vector of the n joints conforms to the following formula (10): (10) Here, D is the selection matrix, a torque vector of the h joints at the kth moment, a torque vector of the n joints at the kth moment.

[0103] In step 10322, the torque vector of the n joints in the second correlation relationship is replaced by the torque vector of the h joints, and the first regression matrix in the second correlation relationship is replaced by the second regression matrix, to establish the first correlation relationship.

[0104] In the embodiments of the present application, by synchronously replacing the torque vector of the n joints and the first regression matrix corresponding to the n joints in the second correlation relationship with the torque vector of the h joints and the second regression matrix corresponding to the h joints, the second correlation relationship describing the global dynamics of the robot arm is converted into the first correlation relationship describing only the target subsystem of the robot arm. Specifically, by adapting the h-dimensional target torque vector with the second regression matrix of h rows, the closure and effectiveness of the newly established first correlation relationship in mathematical form are ensured. In addition, the reconstructed first correlation relationship is an equation of smaller size (h

[0105] In some embodiments, replacing the torque vector of the n joints in the second correlation relationship with the torque vector of the h joints can be understood as: replacing the torque vector of the h joints at the kth moment in formula (10) in formula (9) , replacing in formula (8) in formula (9) , to obtain the following formula (11): (11) wherein, is the second regression matrix at the kth moment, is the dynamics parameter vector, is the torque vector of the h joints at the kth moment.

[0106] In other embodiments, replacing the torque vector of the n joints in the second correlation relationship with the torque vector of the h joints can also be understood as: multiplying the selection matrix D on both sides of the equation of formula (9) to obtain formula (11): (9) (12) (11) wherein, is the first regression matrix at the kth moment, a dynamics parameter vector of the robot arm is determined based on the plurality of second regression matrices, the torque vectors corresponding to the h joints at the plurality of time instants respectively, and the first correlation relationship. is a torque vector of n joints at a kth time instant, and D is a selection matrix. , .

[0107] In step 104, a dynamics parameter vector of the robot arm is determined based on the plurality of second regression matrices, the torque vectors corresponding to the h joints at the plurality of time instants respectively, and the first correlation relationship.

[0108] In the embodiments of the present application, by selectively collecting torque, it is not necessary to equip torque collection devices (such as torque sensors) for all joints, thereby reducing the hardware complexity and cost of identification of the dynamics parameters of the robot arm. Moreover, by introducing a selection matrix, the complete first regression matrix based on the motion data vectors of all n joints is accurately pruned, and the second regression matrix strictly corresponding to the torque of the h target joints is directly constructed, and then the redundant high-dimensional problem describing the dynamics of all joints of the robot arm is losslessly converted into the simplified low-dimensional first correlation relationship describing the h joints of the robot arm. In addition, since the generated first correlation relationship is a linear equation relationship with smaller scale and higher calculation efficiency, it not only retains the global coupling effect, but also focuses on effective target information, thereby significantly improving the robustness and calculation efficiency of the solution of the dynamics parameter vector.

[0109] In some embodiments, Figure 8 is a sixth flowchart of the data processing method of the robot arm provided by the embodiments of the present application, Figure 8 It is shown that step 104 can be implemented by step 1041 to step 1043, which will be described in detail below.

[0110] In step 1041, the plurality of second regression matrices and the torque vectors corresponding to the h joints at the plurality of time instants respectively are substituted into the first correlation relationship respectively to obtain an equation group conforming to the first correlation relationship.

[0111] It is easy to understand that the equation group conforming to the first correlation relationship is composed of a plurality of equations conforming to the first correlation relationship, and each equation conforming to the first correlation relationship represents the relationship equation between the second regression matrix, the dynamics parameter vector and the torque vector of the corresponding time instant.

[0112] In some embodiments, the first correlation relationship can be that the product of the second regression matrix and the dynamics parameter vector is equal to the torque vector, and the equation group of the first correlation relationship can be represented by the following formula (13): (13) Wherein, represents the second regression matrix at the first time instant, represents the second regression matrix at the second time instant, a second regression matrix representing the Kth moment, a dynamics parameter vector, an h-moment matrix representing the 1st moment, an h-moment matrix representing the 2nd moment, an h-moment matrix representing the Kth moment.

[0113] In step 1042, the plurality of second regression matrices in the equation group conforming to the first correlation relationship are integrated into a comprehensive regression matrix, and the h joint moment vectors corresponding to the plurality of moments of the h joints in the equation group conforming to the first correlation relationship are integrated into a comprehensive joint moment vector, to obtain a relationship between the comprehensive regression matrix, the dynamics parameter, and the comprehensive joint moment vector.

[0114] In some embodiments, the plurality of second regression matrices in the equation group conforming to the first correlation relationship can be integrated into a comprehensive regression matrix in a time sequence order of the plurality of moments, and the comprehensive regression matrix conforms to the following formula (14): (14) wherein, the comprehensive regression matrix, a second regression matrix representing the 1st moment, a second regression matrix representing the 2nd moment, a second regression matrix representing the Kth moment.

[0115] In some embodiments, the h joint moment vectors corresponding to the plurality of moments of the h joints in the equation group conforming to the first correlation relationship can be integrated into a comprehensive joint moment vector in a moment sequence order of the plurality of moments, and the comprehensive joint moment vector conforms to the following formula (15): (15) wherein, an h joint moment vector of the 1st moment, an h joint moment vector of the 2nd moment, an h joint moment vector of the Kth moment, the comprehensive joint moment vector.

[0116] In step 1043, the relationship is solved by using the least square method to obtain the dynamics parameter vector.

[0117] In the embodiments of the present application, by substituting the measurement data (i.e., the second regression matrix and the moment vector) of each sampling time into the first correlation relationship in batches, the abstract theoretical relationship can be instantiated as a series of independent linear equations with actual measurement values - each equation is an effective snapshot of the dynamic behavior of the robot arm at that time. Specifically, by systematically integrating multiple independent equations, the second regression matrix and the moment vector in all equations are combined to obtain a single global relationship. This operation equivalently converts the complex problem of multiple time-independent equations into a single high-dimensional matrix equation. Finally, by solving the global relationship using the least squares method, the unique optimal solution that minimizes the sum of squared errors of all time equations can be accurately locked. At the same time, since the global relationship has considered all time motion data and moments, the accidental deviation of single-time data is effectively avoided, ensuring that the solved dynamic parameter vector has both universality and practical value.

[0118] In some embodiments, the relationship is that the product of the comprehensive regression matrix and the dynamic parameter vector is equal to the comprehensive moment vector, Figure 9 FIG. 7 is a seventh flowchart of a data processing method of a robot arm according to an embodiment of the present application, Figure 9 It is shown that step 1043 can be implemented by step 10431 to step 10433, which will be described in detail below.

[0119] In step 10431, the comprehensive moment vector in the relationship is subtracted from the product of the comprehensive regression matrix and the dynamic parameter vector, and the calculation result is taken as an error vector.

[0120] For ease of understanding, the comprehensive moment vector in the relationship is subtracted from the product of the comprehensive regression matrix and the dynamic parameter vector, and the calculation result is taken as an error vector, which can be represented by the following formula (16): - (16) Here, Y represents the comprehensive regression matrix, e represents the error vector, and Y represents the comprehensive regression matrix, represents the dynamic parameter vector.

[0121] In step 10432, the square of the Euclidean norm of the error vector is taken as an objective function.

[0122] For ease of understanding, the square of the Euclidean norm (i.e., the L2 norm) of the error vector is taken as an objective function, which can be represented by the following formula (17): (17) Here, represents the objective function, Y represents a comprehensive regression matrix, e represents an error vector, and Y represents a comprehensive regression matrix, Y represents a comprehensive regression matrix, e represents an error vector, and Y represents a comprehensive regression matrix,

[0123] In step 10433, the minimum value of the objective function is calculated, and the estimated value of the kinetic parameter vector when the objective function reaches the minimum value is taken as the kinetic parameter vector.

[0124] In some embodiments, the minimum value of the objective function is calculated, and the estimated value of the kinetic parameter vector when the objective function reaches the minimum value is taken as the kinetic parameter vector, which can be realized by solving the minimum value of the objective function by applying the generalized inverse matrix to obtain the estimated value of the kinetic parameter vector.

[0125] It should be noted that according to the linear algebra theory, the generalized inverse matrix of a matrix and the matrix are multiplied to obtain the unit matrix. Therefore, solving the minimum value of the objective function is mathematically equivalent to multiplying the comprehensive regression matrix by the generalized inverse matrix on the left of the comprehensive force matrix.

[0126] For ease of understanding, the above relationship can be represented by the following formula (18): (18) Wherein, Y represents a comprehensive regression matrix, e represents an error vector, and Y represents a comprehensive regression matrix, Y represents a comprehensive regression matrix, e represents an error vector, and Y represents a comprehensive regression matrix, Y represents a comprehensive regression matrix, e represents an error vector, and Y represents a comprehensive regression matrix,

[0127] In some embodiments, Figure 10 is the eighth flowchart of the data processing method of the mechanical arm provided in the embodiments of the application, Figure 10 It is shown that before step 103, steps 105 to 106 are further included, which will be explained in detail below.

[0128] In step 105, a zero matrix with h rows and n columns is obtained.

[0129] In step 106, for the i th joint of the h joints, the joint number j i of the n joints corresponding to the i th joint is obtained, and the value of the element in the i th row and the j i th column of the zero matrix is set to 1, to obtain a preset selection matrix.

[0130] In the embodiments of the present application, through the construction logic of "zero matrix initialization-target joint coordinate set 1", the selection matrix naturally has clear joint correlation attributes and can accurately map the corresponding relationship between the h measurable torque joints and the n full joints, ensuring that each "1" element in the matrix is directly pointed to the specific measurable joint dynamics information channel. At the same time, the selection matrix constructed in this way has a fixed dimension of h x n, and the element distribution forms a clear "effective information identification-ineffective information shielding" structure. This structural feature makes the subsequent operation process with the first regression matrix have a clear logical direction, and the directional filtering of information can be completed without additional complex judgments, providing a standardized operation carrier for the entire data processing process. In addition, the binary element characteristics of "0" and "1" in the selection matrix also enable the subsequent operation with the first regression matrix to form a rigid information filtering rule, which not only can stably shield the information interference of non-target joints and effectively avoid the distortion of effective data caused by joint information confusion, but also provides a stable pre-protection for the generation of the second regression matrix, further improving the consistency and reliability of the dynamics parameter identification result.

[0131] Take a mechanical arm with 7 joints as an example, wherein the torque values of the 1st, 2nd, 4th and 5th joints are known. A zero matrix with 4 rows and 7 columns is obtained. For the 1st joint of the 4 joints, the joint number 1 of the 7 joints corresponding to the 1st joint is obtained, and the value of the element in the 1st row and the 1st column of the zero matrix is set to 1. For the 2nd joint of the 4 joints, the joint number 2 of the 7 joints corresponding to the 2nd joint is obtained, and the value of the element in the 2nd row and the 2nd column of the zero matrix is set to 1. For the 3rd joint of the 4 joints, the joint number 4 of the 7 joints corresponding to the 3rd joint is obtained, and the value of the element in the 3rd row and the 4th column of the zero matrix is set to 1. For the 4th joint of the 4 joints, the joint number 5 of the 7 joints corresponding to the 4th joint is obtained, and the value of the element in the 4th row and the 5th column of the zero matrix is set to 1. Finally, the selection matrix as shown in formula (19) is obtained.

[0132] (19) In the following, an exemplary application of the embodiments of the present application in an actual application scenario will be described.

[0133] In the related art, in order to realize high-performance motion control of a mechanical arm, an accurate dynamics model must be established, and the key to obtaining the model is to accurately identify its dynamics parameters. The mainstream parameter identification method relies on establishing a linear equation set associated with joint driving torque, joint motion state and to-be-solved dynamics parameters, and solving the equation set by collecting multiple groups of data of the mechanical arm under a specific excitation trajectory.

[0134] However, the existing dynamic parameter identification technology scheme has a major defect: they must rely on each joint equipped with expensive and complex installation torque sensor to directly measure the driving torque of all joints. This premise seriously restricts the application breadth and economy of the technology. The specific defects are: 1) High hardware cost: high-precision torque sensors are expensive, and configuring each joint of the robot will result in a sharp increase in overall hardware costs, which is unacceptable for cost-sensitive industrial robots and consumer robots.

[0135] 2) Complex system integration: integrating torque sensors into each joint of the robot increases the complexity, weight, and potential failure points of the mechanical structure, while also requiring higher wiring and signal acquisition requirements for the control system.

[0136] 3) Limited application scenarios: for many robots that have been manufactured or are compact in structure, it is almost impossible to install torque sensors later. This makes it impossible for a large number of existing robots that do not have full-joint torque sensors to use these advanced identification methods to improve their performance.

[0137] These defects make precise dynamic parameter identification technology a high-end configuration with high barriers to entry, making it difficult to popularize among ordinary industrial robots and collaborative robots, and preventing its performance potential from being fully exploited.

[0138] To address the problem of existing dynamic parameter identification methods relying heavily on full-joint torque sensors, resulting in high cost, integration difficulties, and limited application, the present embodiment provides a data processing method for a robot. This method cleverly eliminates the direct dependence on unmeasured joint torque data by reconstructing and separating the variables of the overall robot dynamics equation, and only uses the torque information of part of the measurable joints (such as the base joint or the end joint) and the kinematics information of all joints to construct an effective solving equation.

[0139] This application breaks the rigid constraint that "every joint must have a torque sensor," requiring only sensors to be installed on a few (or even one) key joints, or torque to be estimated directly using motor current. This significantly lowers the hardware threshold and overall cost for achieving accurate dynamic identification. Simultaneously, the reduced number of required sensors simplifies the mechanical structure and electrical system of the robotic arm, reducing system complexity, weight, and failure rate, and improving overall reliability and maintainability. This application makes it possible to accurately identify dynamic parameters and upgrade the performance of a large number of existing robotic arms that lack full-joint torque sensors, greatly expanding the applicability and commercial value of this advanced technology and contributing to the widespread adoption and application of high-precision, high-dynamic-response robot control technology in a broader range of industrial automation fields.

[0140] The following describes the specific implementation process of the data processing method for the robotic arm provided in the embodiments of this application.

[0141] Figure 11 This is the ninth flowchart of the data processing method for the robotic arm provided in the embodiments of this application. Figure 12 This is a schematic diagram of the robotic arm provided in an embodiment of this application. Figure 12 As shown, the robotic arm consists of 7 joints, with torque sensors installed on the 1st, 2nd, 4th, and 5th joints.

[0142] The following is combined with Figure 12 and Figure 11 The steps shown illustrate the specific implementation process of the data processing method for the robotic arm provided in the embodiments of this application.

[0143] In step 201, the position vector, velocity vector, acceleration vector of all joints and the torque vector of some joints are obtained.

[0144] Design several smooth, continuous acceleration trajectories as excitation trajectories. The robotic arm moves sequentially along these trajectories, recording the positions of the seven joints at each control cycle, resulting in multiple time-stamped positions of the seven joints. Center difference calculation is performed on the positions of the seven joints at multiple time-stamped times to obtain their velocity values. Center difference calculation is then performed on the velocity values ​​of the seven joints at multiple time-stamped times to obtain their initial acceleration values. Finally, low-pass filtering is applied to the initial acceleration values ​​of the seven joints at multiple time-stamped times to obtain their final acceleration values.

[0145] The positions, velocities, and accelerations of the seven joints at multiple times are combined in the order of joints 1 to 7 to obtain position vectors, velocity vectors, and acceleration vectors at multiple times.

[0146] The torque values of the first, second, fourth and fifth joints at multiple time points are obtained by using the torque sensors on the first, second, fourth and fifth joints. The torque values of the first, second, fourth and fifth joints at each time point are combined in the order of 1 to 7 respectively to obtain the torque vectors of the corresponding four joints at multiple time points.

[0147] In step 202, the first regression matrix corresponding to multiple time points is generated by using the position vectors, velocity vectors and acceleration vectors of all joints at multiple time points.

[0148] The standard dynamic equation of the robot arm is established. The torque vector of each joint in the standard dynamic equation is expanded by using the Lagrange method to obtain the expanded torque equation of each joint. The constant physical parameters to be identified are identified by traversing the expanded torque equation of each joint, which are used as the standard dynamic parameters. The expanded standard dynamic equation is algebraically reorganized so that each standard dynamic parameter is multiplied by a function item composed of only the motion data vector. The function item is combined as the theoretical form of the first regression matrix. The motion data vectors at multiple time points are substituted into the theoretical form of the theoretical form of the first regression matrix to obtain multiple first regression matrices corresponding to multiple time points respectively.

[0149] The product of the first regression matrix and the dynamic parameter vector is equal to the torque vector containing the torque values of the seven joints, which is used as the second correlation relationship.

[0150] In step 203, the first correlation relationship is obtained by modifying the second correlation relationship.

[0151] A 4*7 zero matrix is obtained. The element values of the first row and the first column, the element values of the second row and the second column, the element values of the third row and the fourth column, and the element values of the fourth row and the fifth column in the zero matrix are set to 1 respectively to obtain a selection matrix D.

[0152] The selection matrix is multiplied by the left side of the first correlation relationship, i.e. the product of the selection matrix and the first regression matrix and the dynamic parameter vector and the torque vector containing the torque values of the seven joints are multiplied by the left side of the first correlation relationship at the same time to obtain the first correlation relationship. The product of the selection matrix and the first regression matrix is used as the second regression matrix. The first correlation relationship is expressed as the product of the second regression matrix and the dynamic parameter vector being equal to the torque vector containing the torque values of the four joints.

[0153] In step 204, the data at all time points are integrated to obtain the relationship between the comprehensive regression matrix, the dynamic parameters and the comprehensive torque matrix.

[0154] The plurality of second regression matrices and the plurality of moment vectors containing 7 joint moment values are respectively substituted into the first correlation relationship to obtain an equation group conforming to the first correlation relationship. The plurality of second regression matrices in the equation group is integrated according to a time sequence to obtain a comprehensive regression matrix; the plurality of moment vectors containing 7 joint moment values in the equation group is integrated according to a time sequence to obtain a relationship between the comprehensive regression matrix, the dynamic parameter and a comprehensive moment matrix.

[0155] In step 205, the relationship between the comprehensive regression matrix, the dynamic parameter and the comprehensive moment matrix is solved to obtain a dynamic parameter vector.

[0156] The comprehensive moment vector in the relationship is subtracted by the product of the comprehensive regression matrix and the dynamic parameter vector, the calculation result is taken as an error vector, and the square of the L2 norm of the error vector is taken as an objective function. The generalized inverse matrix is used to solve when the objective function reaches a minimum value, that is, the product of the generalized inverse matrix of the comprehensive regression matrix and the comprehensive moment matrix is taken as the dynamic parameter vector.

[0157] The data processing method of the mechanical arm provided by the embodiment has the following beneficial effects: 1) Breakthrough of full joint sensing constraint: the embodiment innovatively reconstructs the overall dynamics equation of the mechanical arm to mathematically separate the measurable joint moment from the unmeasured joint moment. Through variable substitution and equation elimination, the original dynamics equation requiring full joint moment information is converted into a new form of linear regression equation only depending on partial measurable joint moment information and the kinematics information of all joints. The reconstructed equation still maintains the linear relationship of the to-be-solved dynamic parameter vector, so that the subsequent identification algorithm can be directly applied. This directly solves the fundamental constraint of the prior art that each joint must have a moment sensor, and provides a new feasible path for low-cost and high-precision parameter identification.

[0158] 2) Simplify data acquisition requirements and ensure the integrity of the identification process: after the effective solving equation is constructed by the above method, only the joint moment data of a small number of installed sensors and the position, speed and acceleration of all joints need to be collected, and then the least squares method and other mature optimization tools can be used for solving. This process completely bypasses the need for unmeasured joint moment data at the algorithm level, thereby ensuring the integrity and feasibility of the entire identification process. This ensures that even if the hardware configuration is greatly simplified, a complete, reliable and efficient parameter identification process can still be performed.

[0159] 3) Transfer precision depends on the dimension, reduce the technical application threshold: the embodiment of the present application makes the identification of the dynamics parameters of the mechanical arm no longer a technology limited by high hardware cost. By transferring the identification accuracy dependence on hardware to the clever design of the algorithm, a large number of inventory or low-cost mechanical arms without full-joint torque sensors can also obtain accurate dynamic models through this method. Ultimately, it helps to promote the popularization of high-performance control algorithms in a wider range of industrial and commercial robots, and improve the automation level and production efficiency of the entire robot industry.

[0160] In summary, the embodiment of the present application solves the many shortcomings of the prior art caused by the dependence on full-joint torque sensors, such as high cost, high integration difficulty, and limited application range. More importantly, it builds a low-cost, easy-to-implement, and high-precision core dynamics foundation for the general mechanical arm. This will enable a large number of existing robots to achieve a huge leap in performance through software upgrades and minor hardware modifications, providing more accurate trajectory tracking, more stable high-speed operation, and safer physical interaction capabilities, and exhibiting the evolution potential from "usable" to "good use". Thus, it fundamentally improves the popularity, economic benefits, and application depth of high-performance robot technology, and promotes its transformation from a high-end laboratory technology to a truly popular industrial productivity.

[0161] The following continues to illustrate an exemplary structure of the data processing device 555 of the mechanical arm provided by the embodiment of the present application as a software module. In some embodiments, as shown in Figure 2 The software module stored in the data processing device 555 of the mechanical arm of the storage 550 can include: A mechanical arm parameter acquisition module 5551 for controlling the mechanical arm to move along a preset excitation trajectory and acquiring motion data vectors of n joints of the mechanical arm at multiple time points and torque vectors of h joints at multiple time points during the movement, wherein n is the total number of joints of the mechanical arm, and h is a positive integer less than n; A dynamics parameter calculation module 5552 for generating a plurality of first regression matrices respectively corresponding to the n joints at multiple time points based on the motion data vectors of the n joints at multiple time points; modifying the plurality of first regression matrices using a preset selection matrix to obtain a plurality of second regression matrices respectively corresponding to the h joints at multiple time points, and establishing a first correlation relationship between the second regression matrices, the dynamics parameters, and the torque vectors of the h joints; determining the dynamics parameter vector of the mechanical arm based on the plurality of second regression matrices, the torque vectors of the h joints at multiple time points, and the first correlation relationship.

[0162] In some embodiments, the robot parameter acquisition module 5551 is further configured to acquire position values of the n joints of the robot at multiple time points, combine the position values of the n joints at the multiple time points respectively to obtain position vectors of the n joints of the robot; for each time point of the n joints at the multiple time points, perform central difference calculation on the positions of the previous time point and the next time point of each time point to obtain velocity values of the n joints of the robot at the multiple time points, combine the velocity values of the n joints at the multiple time points respectively to obtain velocity vectors of the n joints of the robot; for each time point of the n joints at the multiple time points, perform central difference calculation on the velocities of the previous time point and the next time point of each time point to obtain acceleration values of the n joints of the robot at the multiple time points, combine the acceleration values of the n joints at the multiple time points respectively to obtain acceleration vectors of the n joints of the robot; and acquire torque values of the h joints of the robot at the multiple time points, combine the torque values of the h joints at the multiple time points respectively to obtain a torque vector of the h joints of the robot.

[0163] In some embodiments, the robot parameter acquisition module 5551 is further configured to perform central difference calculation on the velocity values of the previous time point and the next time point of each time point to obtain initial acceleration values of the n joints of the robot at the multiple time points; and perform low-pass filtering processing on the initial acceleration values of the n joints at the multiple time points to obtain the acceleration values of the n joints of the robot at the multiple time points.

[0164] In some embodiments, the dynamics parameter calculation module 5552 is further configured to multiply the preset selection matrix and the multiple first regression matrices respectively to obtain multiple second regression matrices corresponding to the h joints at the multiple time points respectively; and establish a first correlation relationship based on the preset selection matrix and a preset second correlation relationship, wherein the preset second correlation relationship is a relationship among the first regression matrices, the dynamics parameters, and the torque vector of the n joints.

[0165] In some embodiments, the dynamics parameter calculation module 5552 is further configured to obtain a zero matrix with h rows and n columns; for an i th joint of the h joints, obtain a joint serial number j of the n joints corresponding to the i th joint i , and set a value of an element in an i th row and a j th column of the zero matrix to 1 to obtain the preset selection matrix. i

[0166] In some embodiments, the dynamics parameter calculation module 5552 is further configured to multiply the preset selection matrix and the torque vector of the n joints to obtain a torque vector of the h joints; replace the torque vector of the n joints in the second correlation relationship with the torque vector of the h joints, and replace the first regression matrix in the second correlation relationship with the second regression matrix to establish the first correlation relationship. ​In some embodiments, the dynamics parameter calculation module 5552 is further configured to substitute the plurality of second regression matrices and the moment vectors of the h joints at the plurality of time points into the first correlation relationship respectively to obtain an equation group conforming to the first correlation relationship; integrate the plurality of second regression matrices in the equation group conforming to the first correlation relationship into a comprehensive regression matrix, and integrate the moment vectors of the h joints at the plurality of time points in the equation group conforming to the first correlation relationship into a comprehensive moment matrix, to obtain a relationship between the comprehensive regression matrix, the dynamics parameters and the comprehensive moment matrix; and solve the relationship by using a least square method to obtain the dynamics parameter vector.

[0167] The embodiment of the present application provides a computer program product, which comprises a computer program or computer executable instructions stored in a computer readable storage medium. A processor of an electronic device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the electronic device executes the data processing method of the mechanical arm provided in the embodiment of the present application.

[0168] The embodiment of the present application provides a computer readable storage medium, which stores computer executable instructions or computer programs. When the computer executable instructions or computer programs are executed by a processor, the processor executes the data processing method of the mechanical arm provided in the embodiment of the present application, for example, the data processing method of the mechanical arm shown in the embodiment of the present application. Figure 3 The embodiment of the present application provides a computer readable storage medium, which stores computer executable instructions or computer programs. When the computer executable instructions or computer programs are executed by a processor, the processor executes the data processing method of the mechanical arm provided in the embodiment of the present application, for example, the data processing method of the mechanical arm shown in the embodiment of the present application.

[0169] In some embodiments, the computer readable storage medium can be RAM, ROM, flash memory, magnetic surface memory, optical disc, or CD-ROM, etc. storage; or can be various devices including one or any combination of the above storage.

[0170] In some embodiments, the computer executable instructions can be in the form of programs, software, software modules, scripts or codes, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and can be deployed in any form, including being deployed as independent programs or being deployed as modules, components, subroutines or other units suitable for use in a computing environment.

[0171] As an example, the computer executable instructions can but not necessarily correspond to files in a file system, can be stored in a part of a file storing other programs or data, for example, stored in one or more scripts in a Hyper Text Markup Language (HTML) document, stored in a single file dedicated to the program in question, or stored in multiple cooperative files (for example, files storing one or more modules, subroutines or code parts).

[0172] By way of example, computer-executable instructions can be deployed to be executed on one electronic device or on multiple electronic devices that are located at one site or distributed across multiple sites and interconnected by a communication network.

[0173] In summary, the embodiment of the present application provides a systematic, low-cost and high-robustness mechanical arm dynamics parameter identification technical solution, which fundamentally solves the three core problems of high hardware dependence, high noise sensitivity and unstable solving process in the actual application of the prior art. Overall, the present application breaks the high dependence on full-joint torque sensors through innovative equation reconstruction technology, significantly reducing the hardware cost and integration complexity of the technology landing; through systematic excitation trajectory design and low-pass filtering processing, the effectiveness and purity of the input data are guaranteed, and the anti-interference ability of the identification process to actual physical noise is significantly improved; finally, with the help of global least squares and generalized inverse matrix solving, the uniqueness, stability and optimality of the parameter solution are ensured. The present application transforms the originally complex, expensive and uncertain engineering problem into an economic, reliable and standardized technical process. This not only directly and significantly improves the motion control precision, running efficiency and operation safety of the mechanical arm, but also clears the key bottom model barrier for a series of advanced robot technologies such as load identification, collision detection, compliant control, etc., and thus promotes the development of the entire robot and automation industry to a higher performance and higher intelligence stage.

[0174] The above only describes the embodiments of the present application and is not intended to limit the protection scope of the present application. Any modification, equivalent replacement and improvement made within the spirit and scope of the present application shall be included in the protection scope of the present application.

Claims

1. A data processing method of a robot arm, characterized by, The method comprises: controlling the mechanical arm to move along a preset excitation track, and collecting motion data vectors of n joints of the mechanical arm at multiple time points and torque vectors of h joints at multiple time points during the movement, wherein n is the total number of joints of the mechanical arm, and h is a positive integer less than n; generating multiple first regression matrices respectively corresponding to the n joints at the multiple time points based on the motion data vectors of the n joints at the multiple time points; modifying the multiple first regression matrices by using a preset selection matrix to obtain multiple second regression matrices respectively corresponding to the h joints at the multiple time points, and establishing a first correlation relationship among the second regression matrices, the dynamic parameters and the torque vectors of the h joints; determining a dynamic parameter vector of the mechanical arm based on the multiple second regression matrices, the torque vectors of the h joints at the multiple time points and the first correlation relationship.

2. The method of claim 1, wherein, The motion data vector comprises a position vector, a velocity vector and an acceleration vector, and the collecting of the motion data vectors of the n joints of the mechanical arm at multiple time points and the torque vectors of the h joints during the movement comprises: collecting position values of the n joints of the mechanical arm at the multiple time points, and combining the position values of the n joints at the multiple time points respectively to obtain a position vector of the n joints of the mechanical arm; for each time point of the n joints at the multiple time points, performing central difference calculation by using the positions of the previous time point and the next time point of the each time point to obtain velocity values of the n joints of the mechanical arm at the multiple time points, and combining the velocity values of the n joints at the multiple time points respectively to obtain a velocity vector of the n joints of the mechanical arm; for each time point of the n joints at the multiple time points, performing central difference calculation by using the velocities of the previous time point and the next time point of the each time point to obtain acceleration values of the n joints of the mechanical arm at the multiple time points, and combining the acceleration values of the n joints at the multiple time points respectively to obtain an acceleration vector of the n joints of the mechanical arm; collecting torque values of the h joints of the mechanical arm at the multiple time points, and combining the torque values of the h joints at the multiple time points respectively to obtain a torque vector of the h joints of the mechanical arm.

3. The method of claim 2, wherein, The central difference calculation by using the velocities of the previous time point and the next time point of the each time point to obtain the acceleration values of the n joints of the mechanical arm at the multiple time points comprises: performing central difference calculation by using the velocity values of the previous time point and the next time point of the each time point to obtain initial acceleration values of the n joints of the mechanical arm at the multiple time points; performing low-pass filtering processing on the initial acceleration values of the n joints at the multiple time points to obtain the acceleration values of the n joints of the mechanical arm at the multiple time points.

4. The method of claim 1, wherein, The step of modifying the plurality of first regression matrices using a preset selection matrix to obtain a plurality of second regression matrices corresponding to the h joints at the plurality of times, and establishing a first correlation relationship between the second regression matrices, the dynamic parameters, and the torque vectors of the h joints, includes: The preset selection matrix is ​​multiplied by the plurality of first regression matrices respectively to obtain the plurality of second regression matrices corresponding to the h joints at the plurality of times; Based on the preset selection matrix and the preset second correlation, the first correlation is established, wherein the preset second correlation is the relationship between the first regression matrix, the dynamic parameters, and the torque vectors of the n joints.

5. The method of claim 4, wherein, Before modifying the plurality of first regression matrices using a preset selection matrix to obtain the plurality of second regression matrices corresponding to the h joints at the plurality of times, and before establishing the first correlation relationship between the second regression matrices, the dynamic parameters, and the torque vectors of the h joints, the method further includes: Obtain the zero matrix with h rows and n columns; For the i-th joint among the h joints, obtain the joint number j among the n joints corresponding to the i-th joint. i and the i-th row and j-th row of the zero matrix i The values ​​of the elements in the column are set to 1 to obtain the preset selection matrix.

6. The method of claim 4, wherein, The step of establishing the first association relationship based on the preset selection matrix and the second association relationship includes: Multiply the preset selection matrix by the torque vectors of the n joints to obtain the torque vectors of the h joints; The torque vectors of the n joints in the second association relationship are replaced with the torque vectors of the h joints, and the first regression matrix in the second association relationship is replaced with the second regression matrix to establish the first association relationship.

7. The method of claim 1, wherein, The determination of the dynamic parameter vector of the robotic arm based on the plurality of second regression matrices, the torque vectors corresponding to the h joints at the plurality of times, and the first correlation relationship includes: Substituting the plurality of second regression matrices and the torque vectors of the h joints at multiple times into the first correlation relationship, respectively, yields a set of equations that conform to the first correlation relationship; The multiple second regression matrices in the equation system that conforms to the first correlation relationship are integrated into a comprehensive regression matrix, and the torque vectors of the h joints in the equation system that conforms to the first correlation relationship at multiple times are integrated into a comprehensive torque matrix, so as to obtain the relationship between the comprehensive regression matrix, dynamic parameters and comprehensive torque matrix; The dynamic parameter vector is obtained by solving the relation using the least squares method.

8. A data processing device of a robot arm, characterized by The device includes: The robotic arm parameter acquisition module is used to control the robotic arm to move along a preset excitation trajectory, and to acquire motion data vectors of n joints of the robotic arm at multiple times and torque vectors of h joints at multiple times during the movement, where n is the total number of joints of the robotic arm and h is a positive integer less than n. The dynamic parameter calculation module is used to generate multiple first regression matrices corresponding to the n joints at multiple times based on the motion data vectors of the n joints at multiple times; modify the multiple first regression matrices using a preset selection matrix to obtain multiple second regression matrices corresponding to the h joints at multiple times, and establish a first correlation relationship between the second regression matrices, the dynamic parameters, and the torque vectors of the h joints; and determine the dynamic parameter vector of the robotic arm based on the multiple second regression matrices, the torque vectors of the h joints at multiple times, and the first correlation relationship.

9. An electronic device, comprising: The electronic device includes: Memory is used to store executable instructions or computer programs. A processor, when executing computer-executable instructions or computer programs stored in the memory, implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, they implement the method described in any one of claims 1 to 7.

11. A computer program product comprising computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, they implement the method according to any one of claims 1 to 7.

Images