Dexterous hand motion algorithm control method, system and device, storage medium and program product

Abstract

The invention relates to robot motion control, in particular to a dexterous hand motion algorithm control method, system and device, a storage medium and a program product. The dexterous hand motion algorithm control method comprises the following steps: visual detection: acquiring a hand image through a camera, and identifying three-dimensional coordinates of key points of hand joints by adopting a visual detection algorithm; kinematics calculation: based on the coordinates of the joint key points, obtaining target execution angles of a plurality of motors through kinematics forward and inverse calculation; communication distribution: receiving the target execution angle through a master control unit, and distributing the target execution angle to a plurality of slave control units according to a preset distribution rule; and motor control: each slave control unit generates a driving signal by adopting a PID control algorithm based on the received target execution angle, and controls the corresponding motor to move. The dexterous hand motion algorithm control method provided by the invention is high in response speed, high in control precision, stable in communication and accurate in kinematics solution.

Description

Technical Field

[0001] This invention relates to robot motion control, and more particularly to a dexterous hand motion algorithm control method, system, device, storage medium, and program product. Background Technology

[0002] With the development of robotics technology, dexterous hands, as the core execution component for human-computer interaction and precision operation, place extremely high demands on the accuracy, real-time performance, and reliability of motion control. Existing dexterous hand control systems have the following shortcomings: First, the integration of vision detection and motor control is low, resulting in a large response delay when the motor is executed after the detection of key joint points. Second, the communication link redundancy design is insufficient, and data transmission between the host computer and the slave computer is easily interfered with. Third, the motor control algorithm has poor adaptability, making it difficult to achieve high-precision angle tracking on lightweight hollow cup motors; Fourth, the kinematic solution method has a low degree of matching with the actual mechanical structure, and the inverse solution has insufficient accuracy and slow convergence speed.

[0003] In existing technologies, visual inspection often uses independent algorithm modules, lacking a unified communication architecture with motion control modules; lower-level machine control often uses a single-level MCU architecture, which makes it difficult to balance the real-time performance and data processing capabilities of multi-motor control; motor control often uses simple proportional control, which cannot compensate for errors caused by load changes and mechanical backlash; kinematic calculations do not fully incorporate the geometric constraints of the mechanical structure, resulting in low inverse kinematics accuracy and slow convergence. Summary of the Invention

[0004] To address the aforementioned problems of low integration between visual detection and motor control, insufficient communication reliability, low motor control accuracy, and inadequate accuracy and efficiency in kinematic calculation, this invention provides a dexterous hand motion algorithm control method. This method features fast response speed, high control accuracy, stable communication, and precise kinematic calculation. The specific technical solution is as follows: A dexterous hand motion algorithm control method includes the following steps: visual detection: capturing images of a human hand through a camera and using a visual detection algorithm to identify the three-dimensional coordinates of key points of the hand joints; kinematic calculation: based on the coordinates of the key points of the joints, obtaining the target execution angles of multiple motors through forward and inverse kinematic calculations; communication allocation: receiving the target execution angles through a master control unit and distributing the target execution angles to multiple slave control units according to a preset allocation rule; motor control: each slave control unit generates drive signals based on the received target execution angles using a PID control algorithm to control the movement of the corresponding motor.

[0005] Preferably, the kinematic calculation includes: forward kinematic calculation: based on the input joint translations d1 and d2, and the preset link length parameters, the coordinates of each link node are determined using a circular intersection algorithm, and the joint angles q1, q2, and q3 are calculated through vector angle calculation; inverse kinematic calculation: using the Newton iteration method, starting from the initial guessed displacement, the displacements d1 and d2 corresponding to the input target joint angles q1_target and q2_target are iteratively solved.

[0006] Furthermore, in the forward kinematics calculation, the intersection of two circles is calculated using the circleIntersect function to determine the position of the link node, and the angle between the vectors is calculated using the vectorAngle function to obtain the joint angle; the iterative process in the inverse kinematics solution includes: calling the forward kinematics function to calculate the joint angle corresponding to the current displacement, constructing the residual equation F=[q1_fk-q1_target;q2_fk-q2_target], and minimizing the residual using an optimization algorithm until the convergence condition is met.

[0007] Preferably, in the motor control, each slave control unit acquires the real-time angle of the motor through an encoder, calculates the deviation between the target angle and the actual angle, and outputs a PWM signal to drive the motor to move after PID adjustment; the PID adjustment uses a proportional coefficient Kp=1.39, an integral coefficient Ki=0.02, and a derivative coefficient Kd=0.06.

[0008] Preferably, the system debugging also includes: visual calibration: acquiring multiple sets of standard gesture images, calibrating camera intrinsic parameters, and correcting detection coordinate errors; kinematic debugging: testing the angle calculation deviation corresponding to a known displacement, and verifying the convergence and accuracy of the inverse solution under the target angle; communication testing: statistically analyzing the communication delay and packet loss rate between the master controller and the slave controller; motor debugging: testing PID parameters under no-load and load conditions, and verifying angle tracking accuracy; joint debugging verification: recording the joint response time and positioning error of the dexterous hand movements, and optimizing system parameters.

[0009] A dexterous hand motion algorithm control system, used in the aforementioned dexterous hand motion algorithm control method, includes: a host computer layer equipped with a ROS2 platform, used to acquire the coordinates of key points of human hand joints based on a visual detection algorithm and perform kinematic calculations to obtain the target execution angle of the motor; a master control layer, used to communicate with the host computer layer through a first communication interface, receive the target execution angle, and allocate the target execution angle to a slave control layer through a second communication interface; a slave control layer, including multiple slave control units, used to receive the target execution angle and generate drive signals based on a PID control algorithm; and an execution layer, including several motors, used to execute corresponding actions according to the drive signals; wherein the host computer layer, the master control layer, the slave control layer, and the execution layer work collaboratively through a standardized communication link.

[0010] Preferably, the host computer layer includes: a visual detection node, used to acquire images of a human hand through a camera and detect the three-dimensional coordinates of key joint points of the human hand using the MediaPipe algorithm; and a kinematics calculation node, used to perform forward and inverse kinematics calculations on the coordinates of the key joint points based on the geometric constraint model of the dexterous hand, converting the joint coordinates into target execution angles of the motor, and publishing them through a ROS2 topic; wherein, the kinematics calculation node includes: a forward kinematics calculation module, used to determine the coordinates of each link node based on the input joint translations d1 and d2 and preset link length parameters using a circular intersection algorithm, and to obtain the joint angles q1, q2, and q3 through vector angle calculation; and an inverse kinematics calculation module, used to use Newton's iteration method, starting from the initial guessed displacement, to iteratively solve for the displacements d1 and d2 corresponding to the input target joint angles q1_target and q2_target until the preset convergence condition is met.

[0011] A computer device includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the dexterous hand motion algorithm control method.

[0012] A computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps of the dexterous hand motion algorithm control method described above.

[0013] A computer program product, comprising a computer program / instructions, characterized in that, when the computer program / instructions are executed by a processor, they implement the steps of the dexterous hand motion algorithm control method described above.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a dexterous hand motion algorithm control method with fast response speed, high control accuracy, stable communication and accurate kinematic calculation. Attached Figure Description

[0015] Figure 1 is a flowchart of the overall system architecture, showing the composition of each level and the data flow; Figure 2 is a flowchart of the interaction between visual detection and kinematic calculation, illustrating the conversion process from joint coordinates to motor angles; Figure 3 is a flowchart of forward and inverse kinematics calculations, showing the steps of forward kinematics calculation and inverse kinematics solution. Figure 4 is a flowchart of the SPI communication between the master and slave controllers, showing the instruction allocation logic; Figure 5 shows the flowchart of the slave controller and motor PID control, illustrating the closed-loop control principle. Detailed Implementation

[0016] The present invention will now be further described with reference to the accompanying drawings.

[0017] like Figures 1 to 5 As shown, a dexterous hand motion algorithm control method includes the following steps: visual detection: capturing images of a human hand through a camera and using a visual detection algorithm to identify the three-dimensional coordinates of key points of the hand joints; kinematic calculation: obtaining the target execution angles of multiple motors through forward and inverse kinematic calculations based on the coordinates of the key joints; communication allocation: receiving the target execution angles through the master control unit and distributing the target execution angles to multiple slave control units according to a preset allocation rule; motor control: generating drive signals based on the received target execution angles using a PID control algorithm through each slave control unit to control the movement of the corresponding motor.

[0018] The kinematics solution includes: forward kinematics calculation: based on the input joint translations d1 and d2, and the preset link length parameters, the coordinates of each link node are determined using the circular intersection algorithm, and the joint angles q1, q2, and q3 are obtained by calculating the vector angles; inverse kinematics solution: using the Newton-Raphson iteration method, starting from the initial guessed displacement, the displacements d1 and d2 corresponding to the input target joint angles q1_target and q2_target are iteratively solved.

[0019] In the forward kinematics calculation, the intersection of two circles is calculated using the circleIntersect function to determine the position of the link node, and the angle between the vectors is calculated using the vectorAngle function to obtain the joint angle. In the inverse kinematics solution, the iterative process includes: calling the forward kinematics function to calculate the joint angle corresponding to the current displacement, constructing the residual equation F=[q1_fk-q1_target;q2_fk-q2_target], and minimizing the residual through an optimization algorithm until the convergence condition is met.

[0020] In motor control, each slave control unit acquires the real-time angle of the motor through an encoder, calculates the deviation between the target angle and the actual angle, and outputs a PWM signal to drive the motor after PID adjustment. The PID adjustment uses a proportional coefficient Kp=1.39, an integral coefficient Ki=0.02, and a derivative coefficient Kd=0.06.

[0021] This also includes system debugging: Debugging steps: Visual calibration: Acquire 100 sets of standard gesture images, calibrate camera intrinsic parameters, and correct detection coordinate errors; Kinematic adjustment: Test the deviation of angle calculation corresponding to known d1 and d2; verify the convergence and accuracy of the inverse solution under the target angle; Communication test: Statistical analysis of communication latency (≤10ms) and packet loss rate (≤0.1%) between master and slave controllers. Motor debugging: Test PID parameters under no-load and load conditions to verify angle tracking accuracy; Joint debugging and verification: Record the joint response time and positioning error of the dexterous hand's grasping and rotation movements, and optimize system parameters.

[0022] A dexterous hand motion algorithm control system includes a host computer layer, a master control layer, a slave control layer, and an execution layer, with each layer working together through a standardized communication link.

[0023] The upper-level computer layer, based on the Ubuntu operating system and running the ROS2 (Robot Operating System 2) platform, deploys visual inspection nodes and kinematics calculation nodes: The visual inspection node uses the MediaPipe open-source algorithm to capture images of the human hand through a camera and detect the 3D coordinates of 21 key joints of the human hand in real time.

[0024] Based on the pre-defined geometric constraint model of the dexterous hand, the forward and inverse kinematics calculations are performed, and the joint coordinates are converted into the target execution angles of 15 corresponding motors (angle resolution ≤ 0.1°), and published through the ROS2 topic " / cal_vel_l".

[0025] The kinematics solution node includes a forward kinematics calculation module and an inverse kinematics solution module: The forward kinematics calculation module calculates joint angles using geometric constraints, as detailed below: Input the translation amounts d1 and d2 of the first and second joints. Based on preset link length parameters (key length parameters such as L1-L17), the coordinates of each link node are determined using a circular intersection solution algorithm. The link node positions are determined by calculating the intersection of the two circles using the circleIntersect function (based on solving the intersection coordinates through geometric relationships using the simultaneous equations of the two circles). The joint angles q1, q2, and q3 are obtained by calculating the vector angles using the vectorAngle function (the signed angle values ​​are determined by combining the dot product and cross product to ensure the directional nature of the angles).

[0026] Inverse kinematics solution module: The motor displacement is solved using Newton's iteration method, and the specific implementation is as follows: Input the target joint angles q1_target and q2_target. Starting from the initial guessed value d_init=[0.5;0.5], iteratively solve for the displacements d1 and d2 using the fsolve function. During the iteration, the subfunction equationsToSolve calls the forward kinematics function to calculate the joint angles q1_fk and q2_fk corresponding to the current displacement, constructing the residual equation F=[q1_fk-q1_target;q2_fk-q2_target]. The residual is minimized through an optimization algorithm until the convergence condition (function tolerance1e-9, step size tolerance1e-9, optimality tolerance1e-9) is met, obtaining the final displacement solution.

[0027] The main control layer uses a high-performance MCU (taking the STM32F4 series as an example) as the main control unit, and integrates the micro_rosSDK to realize ROS2 communication with the host computer. The main control MCU establishes ROS2 communication with the host computer through the UART interface (baud rate ≥ 1Mbps), subscribes to the " / cal_vel_l" topic, and receives 15 motor target angle information in real time and stores them in the buffer area; The master MCU establishes communication with the slave control layer through the SPI bus (clock frequency ≥ 10MHz). According to the preset address encoding rules, the 15 motor angle information is allocated to 8 slave MCUs (7 slave MCUs control 2 motors, 1 slave MCU controls 1 motor, and the redundancy design avoids single point of failure). Each frame of data includes the target angle value (16-bit precision), check bit and address code.

[0028] In the control layer, each slave MCU (taking the STM32F1 series as an example) is responsible for the closed-loop control of two coreless motors: The MCU receives the target angle command issued by the master MCU through the SPI slave interface, and after verification, it is parsed into the target angles of motor A and motor B. The MCU's built-in PID control algorithm collects the motor's real-time angle through the encoder, calculates the deviation between the target angle and the actual angle, and outputs a PWM signal (frequency ≥ 20kHz) through PID adjustment (proportional coefficient Kp = 1.39f, integral coefficient Ki = 0.02f, derivative coefficient Kd = 0.06f) to drive the motor.

[0029] The actuator layer uses a coreless motor (rated voltage 12V, rated speed 16000rpm) as the actuator. The motor output shaft is connected to the dexterous hand joint through a harmonic reducer (reduction ratio 16:1). The maximum continuous torque is 6.5mN·m, which meets the requirements of precision operation.

[0030] Visual detection accuracy: The MediaPipe algorithm provides high-fidelity tracking of key points of human hand joints.

[0031] Kinematic solution parameters: Positive kinematics: Link length parameter accuracy ≤ 0.01mm, joint angle calculation error ≤ 0.1°.

[0032] Inverse kinematics: Iterative convergence accuracy ≤ 1e-9, solution time ≤ 5ms.

[0033] Communication rate: The communication rate between the host computer and the main control MCU is no less than 5000bps.

[0034] The SPI communication rate between the master MCU and the slave MCU is no less than 5.25Mbps.

[0035] Motor control accuracy: Under the action of PID control algorithm, the deviation between the actual angle of the motor and the target angle does not exceed 1 degree.

[0036] Overall system response time: The overall response time from the movement of the human hand joints to the corresponding joint movements of the dexterous hand does not exceed 25ms.

[0037] System debugging: Visual inspection module debugging: Ensure the camera is installed at the right position and angle to clearly capture images of the hand; adjust MediaPipe algorithm parameters to optimize detection accuracy and real-time performance.

[0038] Kinematics calculation module debugging: The linkage length parameters are calibrated according to the actual mechanical structure of the dexterous hand to ensure the accuracy of the forward kinematics calculation; different initial guess values ​​and iteration parameters are tested to optimize the convergence speed and accuracy of inverse kinematics.

[0039] Communication module debugging: Check the communication stability between the host computer and the main control MCU, and between the main control MCU and the slave control MCU; test the communication delay and packet loss rate; adjust the SPI clock frequency and data format to ensure accurate command transmission.

[0040] Motor control module debugging: Initialize PID parameters using the Ziegler-Nichols method, and optimize the proportional coefficient, integral time, and derivative time based on the actual load; monitor the motor operating status to avoid abnormal vibration and overheating.

[0041] This invention achieves high-precision, high-real-time motion control of dexterous hands through an integrated architecture of "visual detection → kinematic calculation → multi-level communication → closed-loop control".

[0042] A kinematic solution scheme combining geometric constraint method and Newton's iteration method is adopted to improve the accuracy of joint angle and motor displacement conversion. A unified communication framework based on ROS2 and micro_ross is built to solve the problem of integrating vision and control. The SPI bus enables efficient interaction between the master and slave controllers, improving communication reliability. Closed-loop control of the motor based on the PID algorithm ensures precise motor movement.

[0043] The system's modules work together to meet the precision operation needs of dexterous hands, demonstrating significant practical value.

[0044] The dexterous hand motion algorithm control method provided in this application has significant advantages in terms of response speed, control accuracy, communication stability and kinematic solution accuracy. These advantages are achieved through a systematic technical solution.

[0045] In terms of response speed, this application achieves rapid response through multi-level optimization design. The visual inspection stage uses the MediaPipe algorithm to perform real-time detection of 21 key joints of the human hand, processing and outputting more than 30 frames of computational results per second. The kinematics calculation stage employs Newton's iteration method, controlling the average solution time to within 5ms by pre-setting reasonable initial guess values ​​and strict convergence conditions (function tolerance 1e-9). The communication architecture adopts a three-level design of "host computer-master controller-slave controller". The host computer and master controller use UART high-speed communication (≥1Mbps), and the master controller and 8 slave controllers distribute instructions in parallel via SPI bus (≥10MHz), ensuring that the overall system response time from human hand movement to dexterous hand response does not exceed 25ms, which is 2-4 times faster than traditional solutions.

[0046] In terms of control precision, comprehensive and accurate control is achieved. The motor control system employs a high-precision PID algorithm with an encoder resolution of 12 bits (0.088°), optimized PID parameters (Kp=1.39, Ki=0.02, Kd=0.06), and a PWM frequency ≥20kHz, ensuring an actual angle tracking error ≤1°. The mechanical transmission uses a planetary reducer (reduction ratio 16:1) to eliminate gear backlash and provide a maximum continuous torque of 6.5mN·m. The kinematics calculation process ensures a forward kinematics calculation error ≤0.1°, a link length parameter accuracy ≤0.01mm, and an inverse kinematics solution accuracy ≤1e-9, providing a precise kinematic foundation for the system.

[0047] Communication stability is ensured through an innovative system architecture. A redundant design is employed, with seven slave MCUs each controlling two motors, and one slave MCU controlling one motor, ensuring that a single point of failure does not affect the overall system operation. The communication protocol includes a 16-bit checksum, resulting in a packet loss rate of ≤0.1% and a master-slave communication latency of ≤10ms. The standardized interface design includes a ROS2 standard communication interface for the host computer and an industrial-grade SPI protocol between the master and slave controllers, achieving communication rates of ≥5000bps between the host computer and the master controller, and ≥5.25Mbps between the master and slave controllers, providing reliable communication for the system.

[0048] The accuracy of the kinematics solution is achieved through multiple technological innovations. Forward kinematics uses the `circleIntersect` function to accurately calculate the intersection of two circles and the `vectorAngle` function to calculate the angle between signed vectors, fully considering the actual geometric constraints of the mechanical structure. Inverse kinematics is solved using Newton's iteration method combined with mechanical parameter constraints, constructing the residual equation F=[q1_fk-q1_target;q2_fk-q2_target], and ensuring the uniqueness of the solution through multiple convergence conditions. The system also features real-time parameter calibration, including online calibration of link length parameters, dynamic adjustment of iteration initial values, and adaptive convergence thresholds, ensuring consistently reliable solution accuracy.

[0049] The system debugging plan provides comprehensive assurance for system performance. Visual calibration establishes a calibration database by collecting 100 sets of standard hand gestures, dynamically correcting camera intrinsic parameter errors to ensure key point detection errors are ≤0.5mm. Kinematic debugging employs forward and inverse kinematics cross-validation, covering the entire workspace test matrix to ensure a convergence success rate ≥99.9%. Communication testing maintains low latency under pressure, demonstrating excellent anti-interference performance and an automatic retransmission mechanism. Motor debugging utilizes the Ziegler-Nichols parameter tuning method, covering full-state testing under no-load and load conditions, with real-time monitoring of temperature and vibration. Joint debugging and verification include multi-joint coordinated motion testing, dynamic load adaptability verification, and automatic optimization of all system parameters to ensure optimal overall system performance.

[0050] In summary, this application, through systematic technological innovation, achieves a comprehensive improvement in key performance indicators of dexterous hand motion control, providing a reliable technical solution for precise robot operation. The synergistic optimization of each technical component results in overall system performance significantly superior to traditional solutions, demonstrating significant practical value and broad application prospects.

[0051] A computer device includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the dexterous hand motion algorithm control method.

[0052] A computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps of the dexterous hand motion algorithm control method described above.

[0053] A computer program product, comprising a computer program / instructions, characterized in that, when the computer program / instructions are executed by a processor, they implement the steps of the dexterous hand motion algorithm control method described above.

[0054] The aforementioned computer-readable storage media may include, but are not limited to, USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, optical disks, and other media capable of storing computer programs.

[0055] The technical principles of the present invention have been described above with reference to specific embodiments. These descriptions are merely for explaining the principles of the invention and should not be construed as limiting the scope of protection of the invention in any way. Based on this explanation, those skilled in the art can readily conceive of other specific embodiments of the invention without inventive effort, and these embodiments will all fall within the scope of protection of the claims of the present invention.

Claims

1. A dexterous hand motion algorithm control method, characterized in that, Includes the following steps: Visual inspection: Images of the human hand are captured by a camera, and visual inspection algorithms are used to identify the three-dimensional coordinates of key points of the hand joints; Kinematic calculation: Based on the coordinates of the key joint points, the target execution angles of multiple motors are obtained through forward and inverse kinematic calculations; Communication allocation: The main control unit receives the target execution angle and distributes the target execution angle to multiple slave control units according to a preset allocation rule; Motor control: Each slave control unit generates a drive signal based on the received target execution angle using a PID control algorithm to control the movement of the corresponding motor.

2. The dexterous hand motion algorithm control method according to claim 1, characterized in that, The kinematic solution includes: Forward kinematics calculation: Based on the input joint translations d1 and d2, and the preset link length parameters, the coordinates of each link node are determined by the circular intersection algorithm, and the joint angles q1, q2, and q3 are obtained by vector angle calculation; Inverse kinematics solution: Using Newton's iteration method, starting from the initial guessed displacement, the displacements d1 and d2 corresponding to the input target joint angles q1_target and q2_target are iteratively solved.

3. The dexterous hand motion algorithm control method according to claim 2, characterized in that, In the positive kinematics calculation, the intersection of two circles is calculated using the circleIntersect function to determine the position of the link node, and the joint angle is obtained by calculating the vector angle using the vectorAngle function. The iterative process in the inverse kinematics solution includes: calling the forward kinematics function to calculate the joint angle corresponding to the current displacement, constructing the residual equation F=[q1_fk-q1_target;q2_fk-q2_target], and minimizing the residual through an optimization algorithm until the convergence condition is met.

4. The dexterous hand motion algorithm control method according to claim 1, characterized in that, In the motor control, each slave control unit collects the real-time angle of the motor through an encoder, calculates the deviation between the target angle and the actual angle, and outputs a PWM signal to drive the motor to move after PID adjustment. The PID control uses a proportional coefficient Kp=1.39, an integral coefficient Ki=0.02, and a derivative coefficient Kd=0.

06.

5. The dexterous hand motion algorithm control method according to claim 1, characterized in that, This also includes system debugging: Visual calibration step: Acquire multiple sets of standard gesture images, calibrate camera intrinsic parameters, and correct detection coordinate errors; Kinematic adjustment: Test the angle calculation deviation corresponding to the known displacement, and verify the convergence and accuracy of the inverse solution under the target angle; Communication testing: Statistical analysis of communication latency and packet loss rate between the master and slave controllers; Motor debugging: Test PID parameters under no-load and load conditions to verify angle tracking accuracy; Joint debugging and verification: Record the joint response time and positioning error of dexterous hand movements, and optimize system parameters.

6. A dexterous hand motion algorithm control system, used in the dexterous hand motion algorithm control method of claim 1, characterized in that, include: The upper-level computer layer is equipped with the ROS2 platform, which is used to obtain the coordinates of key points of human hand joints based on visual detection algorithms, and perform kinematic calculations to obtain the target execution angle of the motor. The master control layer is used to communicate with the host computer layer through the first communication interface, receive the target execution angle, and allocate the target execution angle to the slave control layer through the second communication interface; The slave control layer includes multiple slave control units, which are used to receive the target execution angle and generate drive signals based on the PID control algorithm; as well as The execution layer includes several motors for performing corresponding actions according to the drive signals; The host computer layer, the master control layer, the slave control layer, and the execution layer work together through a standardized communication link.

7. The dexterous hand motion algorithm control system according to claim 6, characterized in that, The host computer layer includes: A visual inspection node is used to capture images of a human hand via a camera and employs the MediaPipe algorithm to detect the 3D coordinates of key points on the hand joints; and The kinematics solution node is used to perform forward and inverse kinematics solutions on the coordinates of the joint key points based on the geometric constraint model of the dexterous hand, convert the joint coordinates into the target execution angle of the motor, and publish them through the ROS2 topic. The kinematic solution nodes include: The forward kinematics calculation module is used to determine the coordinates of each link node based on the input joint translations d1 and d2 and the preset link length parameters, using the circular intersection algorithm, and to calculate the joint angles q1, q2, and q3 through vector angle calculation. The inverse kinematics solution module is used to iteratively solve for the displacements d1 and d2 corresponding to the input target joint angles q1_target and q2_target using the Newton iteration method, starting from the initial guessed displacement, until the preset convergence condition is met.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the dexterous hand motion algorithm control method according to claim 1.

9. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the steps of the dexterous hand motion algorithm control method as described in claim 1.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the steps of the dexterous hand motion algorithm control method as described in claim 1.

Images