Search

CN-121973170-A - Four-limb movement precision control system of humanoid robot

CN121973170ACN 121973170 ACN121973170 ACN 121973170ACN-121973170-A

Abstract

The invention discloses a limb movement precision control system of a humanoid robot, and relates to the technical field of robot control. The system comprises a data acquisition module, a data processing module and a vibration suppression module, wherein the data acquisition module is in butt joint with the data processing module and is used for acquiring multi-dimensional original data of the humanoid robot and sending the multi-dimensional original data to the data processing module, the data processing module is in butt joint with the vibration suppression module and is used for carrying out finite element analysis and neural network processing on the multi-dimensional original data and the motion planning parameters to acquire vibration pre-judgment information and sending the vibration pre-judgment information to the vibration suppression module, and the vibration suppression module is used for generating a vibration suppression mode according to the vibration pre-judgment information and adjusting the motion planning parameters of the humanoid robot according to the vibration suppression mode. According to the invention, the motion planning parameters of the humanoid robot can be restrained in advance according to the vibration prejudgement information, and the high-precision motion control can be completed without depending on subsequent precision compensation.

Inventors

  • CHEN FENG
  • WU PINGZHI

Assignees

  • 合肥中科深谷科技发展有限公司

Dates

Publication Date
20260505
Application Date
20260309

Claims (10)

  1. 1. A humanoid robot extremity movement accuracy control system, comprising: The data acquisition module is in butt joint with the data processing module and is used for acquiring the multi-dimensional original data of the humanoid robot and sending the multi-dimensional original data to the data processing module; The data processing module is in butt joint with the vibration suppression module and is used for carrying out finite element analysis and neural network processing on the multi-dimensional original data and the motion planning parameters, acquiring vibration pre-judging information and sending the vibration pre-judging information to the vibration suppression module; And the vibration suppression module is used for generating a vibration suppression mode according to the vibration pre-judging information and adjusting the motion planning parameters of the humanoid robot according to the vibration suppression mode.
  2. 2. The system of claim 1, wherein the multi-dimensional raw data includes vibration acceleration signals, elastic strain signals, and joint actual rotation angle signals.
  3. 3. The limb movement precision control system of the humanoid robot according to claim 2, wherein the data processing module comprises a data preprocessing unit, an execution scene judging unit, a finite element analysis unit and a neural network processing unit; the data preprocessing unit is in butt joint with the execution scene judging unit, the execution scene judging unit is in butt joint with the finite element analysis unit, and the finite element analysis unit is in butt joint with the neural network processing unit.
  4. 4. The system for controlling the motion precision of the limbs of the humanoid robot according to claim 3, wherein the data preprocessing unit is used for performing filtering, amplifying and noise reduction processing on the multi-dimensional original data to obtain preprocessed multi-dimensional original data.
  5. 5. The system according to claim 4, wherein the execution scene determining unit is configured to determine an execution scene according to the preprocessed multi-dimensional raw data and the motion planning parameter, and the determining logic of the execution scene includes: When the movement speed of the upper limb is more than or equal to 0.3m/s and the load is 0.1-1kg, judging that an upper limb high-speed grabbing scene appears; When the movement speed of the lower limbs is more than or equal to 1.0m/s, judging that a high-speed walking scene appears; and when the rotation angle of the upper limb joint is more than or equal to 180 degrees and the angular speed is more than or equal to 2rad/s, judging that an upper limb rotation operation scene appears.
  6. 6. The system according to claim 5, wherein the finite element analysis unit is configured to construct a stiffness-vibration coupling mathematical model under different execution scene combinations according to the material characteristic parameter and the geometric characteristic parameter, and obtain initial vibration pre-determination information according to the preprocessed multidimensional raw data and the motion planning parameter, where the initial vibration pre-determination information includes stress, strain distribution pre-determination data and vibration response characteristic pre-determination data.
  7. 7. The system for controlling the motion precision of the limbs of the humanoid robot according to claim 6, wherein the neural network processing unit is used for constructing a BP neural network model, taking initial vibration pre-judging information, preprocessed multi-dimensional original data and motion planning parameters as inputs, and outputting final vibration pre-judging information, wherein the vibration pre-judging information comprises a vibration amplitude pre-judging value, a decay time pre-judging value and an angle hysteresis deviation pre-judging value.
  8. 8. The system for controlling the motion precision of limbs of a humanoid robot according to claim 7, wherein the flow of generating the vibration suppression mode by the vibration suppression module according to the vibration pre-judgment information comprises: when the upper limb high-speed grabbing scene appears, based on the initial moment, combining the real-time vibration amplitude in the motion process, dynamically correcting the moment correction coefficient and the speed change rate, so that the real-time vibration amplitude is less than or equal to the allowable maximum vibration amplitude; When a high-speed walking scene appears, determining lower limb joint driving moment according to basic driving moment and compensation moment for vibration suppression of the high-speed walking scene; When an upper limb rotation operation scene appears, determining a real-time correction moment according to the torsion damping moment and the joint rotation basic moment, so that the real-time torsion vibration amplitude is less than or equal to a target threshold value.
  9. 9. The limb movement precision control system of the humanoid robot according to claim 8, wherein when the high-speed upper limb grabbing scene occurs and the vibration amplitude pre-judging value is greater than a preset threshold value, the damping moment of the high-speed upper limb grabbing scene is superimposed on the correction moment.
  10. 10. The system for controlling the motion precision of limbs of a humanoid robot according to claim 8, wherein when the actual rotation angle deviation of the joints in the upper limb rotation operation scene and the upper limb rotation operation scene is greater than a preset threshold, a trimming moment is superimposed on the real-time correction moment, and the calculation formula of the trimming moment is as follows: ; In the formula, In order to fine-tune the moment of force, Compensating coefficient for angle deviation of upper limb rotation operation scene, and , Is the actual rotation angle deviation of the joint.

Description

Four-limb movement precision control system of humanoid robot Technical Field The invention relates to the technical field of robot control, in particular to a limb movement precision control system of a humanoid robot. Background Along with the development of the humanoid robot to the light weight and high flexibility, the four-limb connecting rods and joints of the humanoid robot are made of low-rigidity materials such as carbon fibers and aluminum alloy, and the materials can reduce the dead weight and improve the flexibility of movement of the robot, but under typical working conditions such as high-speed grabbing, rotating operation and walking, the materials are easily influenced by external excitation such as load acting force and ground reaction force, elastic vibration is generated, and further the rotation angle deviation of the joints and the position deviation of the end effector are caused, so that the movement precision is influenced. The control of the limb movement precision of the existing humanoid robot usually adopts a mode of post precision compensation, which has obvious response hysteresis. Disclosure of Invention Aiming at the defects in the prior art, the application solves the technical problem of how to improve the limb movement precision of the humanoid robot by realizing the prejudgment of the vibration parameters in the movement process of the humanoid robot and according to the prejudged vibration parameters. To achieve the above object, an embodiment of the present application provides a motion precision control system for a humanoid robot, including: The data acquisition module is in butt joint with the data processing module and is used for acquiring the multi-dimensional original data of the humanoid robot and sending the multi-dimensional original data to the data processing module; The data processing module is in butt joint with the vibration suppression module and is used for carrying out finite element analysis and neural network processing on the multi-dimensional original data and the motion planning parameters, acquiring vibration pre-judging information and sending the vibration pre-judging information to the vibration suppression module; And the vibration suppression module is used for generating a vibration suppression mode according to the vibration pre-judging information and adjusting the motion planning parameters of the humanoid robot according to the vibration suppression mode. In one embodiment, the multi-dimensional raw data includes vibration acceleration signals, elastic strain signals, and joint actual rotation angle signals. In one embodiment, the data processing module comprises a data preprocessing unit, an execution scene judging unit, a finite element analysis unit and a neural network processing unit; the data preprocessing unit is in butt joint with the execution scene judging unit, the execution scene judging unit is in butt joint with the finite element analysis unit, and the finite element analysis unit is in butt joint with the neural network processing unit. In an embodiment, the data preprocessing unit is configured to perform filtering, amplifying and noise reduction processing on the multi-dimensional original data, and obtain preprocessed multi-dimensional original data. In an embodiment, the execution scenario determination unit is configured to determine an execution scenario according to the preprocessed multi-dimensional raw data and the motion planning parameter, where the determination logic of the execution scenario includes: When the movement speed of the upper limb is more than or equal to 0.3m/s and the load is 0.1-1kg, judging that an upper limb high-speed grabbing scene appears; When the movement speed of the lower limbs is more than or equal to 1.0m/s, judging that a high-speed walking scene appears; and when the rotation angle of the upper limb joint is more than or equal to 180 degrees and the angular speed is more than or equal to 2rad/s, judging that an upper limb rotation operation scene appears. In an embodiment, the finite element analysis unit is configured to construct a stiffness-vibration coupling mathematical model under different execution scene combinations according to the material characteristic parameters and the geometric characteristic parameters, and obtain initial vibration pre-judgment information according to the preprocessed multidimensional raw data and the motion planning parameters, where the initial vibration pre-judgment information includes stress, strain distribution pre-judgment data and vibration response characteristic pre-judgment data. In an embodiment, the neural network processing unit is configured to construct a BP neural network model, and take initial vibration pre-judgment information, preprocessed multidimensional original data and motion planning parameters as inputs to output final vibration pre-judgment information, where the vibration pre-judgment information includes a vibrat