Search

CN-122008199-A - Robot remote interaction method and system based on tactile feedback

CN122008199ACN 122008199 ACN122008199 ACN 122008199ACN-122008199-A

Abstract

The invention relates to the technical field of robot remote interaction, and discloses a robot remote interaction method and a system based on tactile feedback, wherein the method comprises the steps of obtaining force data, position data and material friction data of remote robot interaction with environment; the method comprises the steps of obtaining an initial contact signal through time domain registration and differential processing, extracting force components after correction to generate a regular contact signal, distinguishing a dynamic contact type through super-rigidity threshold, constructing a multidimensional input vector by combining contact dynamic parameters and friction data, substituting the multidimensional input vector into a pre-training model to generate predicted contact force, adjusting friction characteristic weight through the super-deviation threshold to obtain a correction feedback signal, decomposing high-frequency texture and low-frequency contact force characteristics to generate a composite touch driving signal, and driving an executing mechanism to realize high-fidelity force touch output. The method can realize high-fidelity force tactile feedback of robot remote interaction and meet the strict requirements on the accuracy and safety of remote operation.

Inventors

  • YANG ZIHONG

Assignees

  • 广州宏和网络科技有限公司

Dates

Publication Date
20260512
Application Date
20260127

Claims (9)

  1. 1. A method of robot remote interaction based on haptic feedback, comprising: Acquiring force data, position data and material friction data of interaction between a remote robot and an environment; performing time domain registration and differential processing on the force data and the position data to obtain an initial contact signal; constructing a state vector according to the initial contact signal, substituting a preset state transition matrix to obtain a predicted state vector, collecting an observed value of real-time force, correcting the predicted state vector according to the observed value, and extracting force components of the corrected state vector to obtain a regular contact signal; extracting a rigidity change value in the regular contact signal, analyzing the amplitude and the frequency of the regular contact signal if the rigidity change value exceeds a preset rigidity judgment threshold value, comparing the amplitude and the frequency with a contact type template acquired in advance, and distinguishing a dynamic contact type; Collecting current interactive contact dynamic parameters, constructing a multidimensional model input vector by combining the dynamic contact type and the material friction data, substituting the multidimensional model input vector into a pre-trained friction prediction model, and generating a predicted contact force; Calculating the deviation between the predicted contact force and the actual contact force, if the deviation exceeds a preset deviation judging threshold value, adjusting the friction characteristic weight of the friction prediction model, regenerating the predicted contact force, and performing deviation verification again until the deviation meets the requirement, and obtaining a correction feedback signal; and carrying out characteristic decomposition on the correction feedback signal, generating a composite touch driving signal according to the decomposed high-frequency texture characteristics and low-frequency contact force characteristics, and executing the composite touch driving signal to obtain force touch output consistent with actual interaction.
  2. 2. The method for remotely interacting with a robot based on haptic feedback of claim 1, wherein the obtaining force data, position data, material friction data of the remote robot interacting with the environment comprises: collecting force data of interaction between the remote robot and the environment through the force sensor; collecting position data of the remote robot through a position encoder; And retrieving pre-stored material friction data, wherein the material friction data comprises basic friction factors and surface texture information of different materials.
  3. 3. The method of claim 1, wherein performing time domain registration and differentiation on the force data and the position data to obtain an initial contact signal comprises: aligning the force data with the position data according to the time stamp to generate a synchronous data frame; performing differential processing on the synchronous data frame, and calculating a force variation gradient; if the force variation gradient exceeds a preset gradient judgment threshold value, locking the corresponding synchronous data frame as instantaneous contact data; and extracting the instantaneous contact data, and combining the position data to generate an initial contact signal.
  4. 4. The method according to claim 1, wherein the constructing a state vector according to the initial contact signal and substituting a preset state transition matrix to obtain a predicted state vector, collecting an observed value of a real-time force and correcting the predicted state vector according to the observed value, extracting force components of the corrected state vector to obtain a regular contact signal, comprises: Constructing a state vector according to the initial contact signal, substituting a preset state transition matrix, and deducting to obtain a predicted state vector; Collecting an observation value of real-time force, extracting a time stamp of the observation value and a deduction time stamp of the prediction state vector, calculating a time difference, and generating a delay compensation factor according to the time difference; Correcting the observed value according to the delay compensation factor to obtain aligned observed data; Based on a preset noise suppression coefficient, carrying out weighted fusion on the aligned observation data and the predicted state vector, and filtering sensor noise through signal smoothing processing to obtain a corrected state vector; And extracting force component values from the corrected state vector to obtain a regular contact signal.
  5. 5. The method for remote interaction of a robot based on haptic feedback according to claim 1, wherein the extracting the stiffness variation value of the regular contact signal, analyzing the amplitude and frequency of the regular contact signal if the stiffness variation value exceeds a preset stiffness determination threshold, comparing with a pre-acquired contact type template, and distinguishing the dynamic contact type, comprises: Calculating a stiffness variation value by a ratio of a force variation in the regular contact signal to a synchronous displacement variation in the position data; If the rigidity change value exceeds a preset rigidity judgment threshold value, carrying out frequency analysis on the regular contact signal to obtain a frequency distribution spectrum; and calculating vibration energy density according to the frequency distribution spectrum, comparing the vibration energy density with a pre-acquired contact type template, and distinguishing the dynamic contact type.
  6. 6. The method for remote interaction of a robot based on haptic feedback of claim 1, wherein the collecting the current interaction contact dynamic parameters, combining the dynamic contact type and the material friction data, constructing a multidimensional model input vector, and substituting the multidimensional model input vector into a pre-trained friction prediction model, generating a predicted contact force, comprises: Collecting the relative sliding speed and the forward load of the current interaction as contact dynamic parameters; Searching corresponding friction characteristic parameters in the material friction data according to the dynamic contact type; based on the contact dynamic parameter and the friction characteristic parameter, weighting and calculating a thermal coupling correction value caused by friction heat generation according to a preset weight coefficient; Fusing the contact dynamic parameter, the friction characteristic parameter and the thermal coupling correction value to form a multidimensional model input vector; Substituting the multi-dimensional model input vector into a pre-trained friction prediction model to generate a predicted contact force.
  7. 7. The method according to claim 1, wherein calculating the deviation between the predicted contact force and the actual contact force, if the deviation exceeds a preset deviation determination threshold, adjusting the friction characteristic weight of the friction prediction model, regenerating the predicted contact force, and performing deviation verification again until the deviation meets the requirement, and obtaining a correction feedback signal, includes: Performing discretization sampling on the predicted contact force, and obtaining a time sequence friction data sequence after the discretization sampling; extracting the change trend characteristics of the friction data sequence, and calculating the deviation between the predicted contact force and the actual contact force; if the deviation does not exceed a preset deviation judging threshold value, taking the current predicted contact force as a correction feedback signal; if the deviation exceeds the deviation judging threshold value, converting the deviation into a deviation characteristic vector; And calculating a weight adjustment value of the corresponding dimension of the friction prediction model according to the deviation feature vector, updating friction characteristic weight, reconstructing a model output signal, performing deviation verification again, and outputting the latest model as a correction feedback signal after the deviation meets the requirement.
  8. 8. The method of claim 1, wherein the performing feature decomposition on the correction feedback signal, generating a composite haptic driving signal according to the decomposed high frequency texture feature and low frequency contact force feature, and performing the same to obtain a haptic output consistent with the actual interaction, comprises: performing multi-scale decomposition on the correction feedback signal to obtain high-frequency texture characteristics and low-frequency contact force characteristics; generating a micro-vibration driving waveform according to the high-frequency texture characteristics, and generating a moment resisting instruction according to the low-frequency contact force characteristics; Superposing the micro-vibration driving waveform and the moment resistance instruction to construct a composite touch driving signal; And converting the composite touch driving signal into a driving electric signal matched with the executing mechanism, transmitting the driving electric signal to the touch executing mechanism, and driving the touch executing mechanism to generate physical displacement and reverse resistance to obtain force touch output.
  9. 9. A robot remote interaction system based on haptic feedback, comprising: the data acquisition module is used for acquiring force data, position data and material friction data of interaction between the remote robot and the environment; The contact signal module is used for performing time domain registration and differential processing on the force data and the position data to obtain an initial contact signal; The signal processing module is used for constructing a state vector according to the initial contact signal, substituting a preset state transition matrix to obtain a predicted state vector, collecting an observed value of real-time force, correcting the predicted state vector according to the observed value, and extracting force components of the corrected state vector to obtain a regular contact signal; The contact judging module is used for extracting a rigidity change value in the regular contact signal, analyzing the amplitude and the frequency of the regular contact signal if the rigidity change value exceeds a preset rigidity judging threshold value, comparing the amplitude and the frequency with a pre-acquired contact type template, and distinguishing a dynamic contact type; The contact force prediction module is used for collecting the contact dynamic parameters of current interaction, combining the dynamic contact type and the material friction data, constructing a multidimensional model input vector, substituting the multidimensional model input vector into a pre-trained friction prediction model, and generating a predicted contact force; the signal correction module is used for calculating the deviation between the predicted contact force and the actual contact force, if the deviation exceeds a preset deviation judgment threshold value, the friction characteristic weight of the friction prediction model is adjusted, the predicted contact force is regenerated, deviation verification is carried out again, and a correction feedback signal is obtained after the deviation meets the requirement; And the haptic output module is used for carrying out characteristic decomposition on the correction feedback signal, generating a composite haptic driving signal according to the decomposed high-frequency texture characteristic and low-frequency contact force characteristic, and executing the composite haptic driving signal to obtain force haptic output consistent with actual interaction.

Description

Robot remote interaction method and system based on tactile feedback Technical Field The invention relates to the technical field of robot remote interaction, in particular to a robot remote interaction method and system based on tactile feedback. Background At present, in the technical field of robot remote interaction, along with continuous expansion of remote operation scenes and continuous improvement of precision operation demands, the robot remote interaction based on haptic feedback is used as a core technical support, the authenticity and the accuracy of the feedback are directly related to the safety and the operation precision of the operation, and the intelligent chip can provide support for realizing the goal aiming at high-efficiency analysis and processing capacity of multi-source data. In the prior art, the remote interactive tactile feedback method of the robot mainly relies on fixed parameter mapping or simple force value transmission, for example, remote force data is directly amplified and then output, or a preset friction prediction model is adopted to ignore dynamic change of contact characteristics, or effective processing of signal noise and delay is lacked. However, this approach is clearly inadequate in complex interaction scenarios. The feedback force distortion and the direction deviation are caused by the fact that the fixed parameters cannot be adapted to dynamic changes of different materials and contact angles, the noise and the transmission delay of the sensor are difficult to filter by simple force value transmission, the operation judgment is affected, the real-time tracking and model iteration of the contact stiffness and friction characteristics are lacked, excessive force or operation failure is easy to cause particularly in scenes such as precise medical treatment and dangerous environment operation, and the high-precision interaction requirement is difficult to meet. To sum up, the prior art is difficult to realize high-fidelity force tactile feedback of robot remote interaction, and cannot meet the strict requirements on the accuracy and safety of remote operation. Disclosure of Invention The invention provides a robot remote interaction method and a system based on haptic feedback, which are used for realizing high-fidelity force haptic feedback of robot remote interaction and meeting the strict requirements on the accuracy and safety of remote operation. In order to solve the above technical problems, the present invention provides a robot remote interaction method based on haptic feedback, including: Acquiring force data, position data and material friction data of interaction between a remote robot and an environment; performing time domain registration and differential processing on the force data and the position data to obtain an initial contact signal; constructing a state vector according to the initial contact signal, substituting a preset state transition matrix to obtain a predicted state vector, collecting an observed value of real-time force, correcting the predicted state vector according to the observed value, and extracting force components of the corrected state vector to obtain a regular contact signal; extracting a rigidity change value in the regular contact signal, analyzing the amplitude and the frequency of the regular contact signal if the rigidity change value exceeds a preset rigidity judgment threshold value, comparing the amplitude and the frequency with a contact type template acquired in advance, and distinguishing a dynamic contact type; Collecting current interactive contact dynamic parameters, constructing a multidimensional model input vector by combining the dynamic contact type and the material friction data, substituting the multidimensional model input vector into a pre-trained friction prediction model, and generating a predicted contact force; Calculating the deviation between the predicted contact force and the actual contact force, if the deviation exceeds a preset deviation judging threshold value, adjusting the friction characteristic weight of the friction prediction model, regenerating the predicted contact force, and performing deviation verification again until the deviation meets the requirement, and obtaining a correction feedback signal; and carrying out characteristic decomposition on the correction feedback signal, generating a composite touch driving signal according to the decomposed high-frequency texture characteristics and low-frequency contact force characteristics, and executing the composite touch driving signal to obtain force touch output consistent with actual interaction. In a second aspect, the present invention provides a robot remote interaction system based on haptic feedback, comprising: the data acquisition module is used for acquiring force data, position data and material friction data of interaction between the remote robot and the environment; The contact signal module is used for performing t