Methods, devices, equipment, media, and products for transferring robot operational skills

By acquiring and adjusting the interaction characteristics of heterogeneous actuators, operation control parameters are generated, solving the adaptation problem of robot operation skills after migration between heterogeneous actuators, and achieving higher contact stability and task completion requirements.

CN123103680APending Publication Date: 2026-11-17ASTRIBOT CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611544595.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-30
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

After robot operation skills are transferred between heterogeneous actuators, the degree of adaptation with the actuator receiving the skills is low, resulting in insufficient contact stability.

Method used

By acquiring the interaction characteristics of the original actuator, adjusting the interaction characteristics of the target actuator to adapt to its capability limitations, and generating operation control parameters, the target actuator is ensured to be compatible with heterogeneous actuators.

Benefits of technology

This improves the adaptability of robot operation skills after they are transferred between heterogeneous actuators, ensuring the stability of contact interaction and the completion of task requirements.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123103680A_ABST
    Figure CN123103680A_ABST
Patent Text Reader

Abstract

This disclosure provides a method, apparatus, device, medium, and product for transferring robot operation skills, relating to the field of robotics. The method includes acquiring the original interaction features of each original task stage, wherein the original interaction features characterize the response relationship between motion changes and contact interaction changes; adjusting the interaction features to be adjusted in each target task stage to the corresponding original interaction features under the capability limitations of the target actuator, thereby obtaining target interaction features, wherein the corresponding target task stage and the original task stage have the same stage completion requirements; and generating operation control parameters for the target actuator based on the target interaction features and stage completion requirements. This method solves the technical problem of low adaptability between robot operation skills transferred between heterogeneous actuators and the actuator receiving the skill, achieving the technical effect of improving the adaptability between robot operation skills transferred between heterogeneous actuators and the actuator receiving the skill.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure relates to the field of robotics technology, and more specifically, to a method, apparatus, device, medium, and product for transferring robot operational skills. Background Technology

[0002] In robotic manipulation tasks, relevant technologies can acquire motion trajectories and stiffness profiles through teaching, and based on this, develop operational skills to control actuators to perform tasks such as grasping. When reusing existing operational skills, motion trajectories can be redirected and impedance control parameters adjusted to adapt to changes in task execution conditions.

[0003] However, heterogeneous actuators differ in their actuation methods, degrees of freedom, and contact structures. Even when using the same or redirected motion trajectory, the forces and motion exhibited when in contact with the manipulated object may vary. Therefore, when skill transfer is primarily based on motion trajectory and stiffness profile, it is difficult to fully adapt to the differences in contact interaction between actuators. This results in a low degree of compatibility between the robot's operational skills after they have been transferred between heterogeneous actuators and the actuator receiving the skills. Summary of the Invention

[0004] This disclosure provides a method, apparatus, device, medium, and product for transferring robot operation skills, in order to at least solve the technical problem in the related art that the degree of adaptation between robot operation skills and the receiving actuator is low after the skills are transferred between heterogeneous actuators.

[0005] According to one aspect of the present disclosure, a method for transferring robot operational skills is provided, comprising: Receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; In response to the skill transfer request, the original interaction features of each original task stage during the execution of the target task by the original executor through the operation skill are obtained, wherein the original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when interacting with the operation object during the execution of the target task. For each target task stage corresponding to each original task stage when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, the interaction features to be adjusted in each target task stage are adjusted in the direction of the original interaction features of the corresponding original task stage to obtain the target interaction features of each target task stage, wherein the corresponding target task stage and the original task stage have the same stage completion requirements. Based on the target interaction characteristics and the stage completion requirements of each target task stage, the operation control parameters of the target executor in each target task stage are generated to obtain the migrated operation skills adapted to the target executor.

[0006] According to another aspect of the present disclosure, a device for transferring robot operational skills is also provided, comprising: A receiving module is configured to receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; The first acquisition module is used to respond to the skill transfer request and acquire the original interaction features of each original task stage during the process of the original executor performing the target task through the operation skill. The original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when it interacts with the operation object during the execution of the target task. An adjustment module is used to adjust the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, so as to adjust the interaction features of each of the target task stages in the direction of the original interaction features of the corresponding original task stages, thereby obtaining the target interaction features of each of the target task stages, wherein the corresponding target task stages and the original task stages have the same stage completion requirements. The generation module is used to generate operation control parameters for the target executor in each of the target task stages based on the target interaction features and the stage completion requirements, so as to obtain the migrated operation skills adapted to the target executor.

[0007] Optionally, the first acquisition module includes: The first acquisition unit is used to acquire force signals and velocity signals during the execution of the target task by the original actuator, wherein the force signal represents the change of contact force received by the original actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement velocity of the original actuator over time. The first analysis unit is used to perform spectrum analysis on the corresponding force signal and velocity signal for each of the original task stages, and obtain force signal components and velocity signal components at multiple reference frequencies, wherein the multiple reference frequencies correspond to the variation components of different periods in the force signal and velocity signal. The comparison unit is used to compare the force signal component with the velocity signal component with the same reference frequency for each original task stage, determine the amplitude relationship and phase relationship of the force signal component relative to the velocity signal component, and combine the multiple reference frequencies with the corresponding amplitude relationship and phase relationship to form the original interaction feature of the original task stage.

[0008] Optionally, the first acquisition unit is also used for: The detection excitation signal is superimposed as an additional control quantity onto the initial task command used to control the original actuator to execute the target task, thereby obtaining the adjusted target task command. The detection excitation signal is used to control the original actuator to apply additional detection actions during the execution of the target task, and the frequency of the detection excitation signal is within a preset frequency range. During the execution of the target task by the original actuator, the displacement change amplitude of the original actuator caused by the additional detection action and the contact force change amplitude between the original actuator and the operating object are respectively limited to the preset upper limit of displacement amplitude and the upper limit of force amplitude. The mechanical energy input by the original actuator to the contact interaction process between the original actuator and the operating object through the additional detection action is limited to the preset upper limit of energy, and the force signal and the velocity signal are collected.

[0009] Optionally, the first analysis unit is also used for: For each of the original task stages, spectral analysis is performed on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple candidate frequencies, and the coherence coefficient of the force signal and the velocity signal at each of the candidate frequencies is calculated, wherein the coherence coefficient is used to indicate the degree of correlation between the force signal and the velocity signal at the corresponding candidate frequency. The candidate frequencies with a coherence coefficient not less than a preset threshold are determined as a plurality of reference frequencies, and the force signal components and velocity signal components at the plurality of reference frequencies are retained.

[0010] Optional adjustment modules include: The second acquisition unit is used to acquire the confidence intervals of each of the original interaction features and the set of reachable response relationships of the target actuator. The original interaction features include the amplitude and phase relationships of the contact force received by the original actuator relative to the movement speed of the original actuator at multiple reference frequencies when the original actuator interacts with the operation object. The confidence intervals indicate the estimated uncertainty range of the amplitude and phase relationships of the original interaction features at each of the reference frequencies. The set of reachable response relationships includes multiple response relationships that the target actuator can reach. Each reachable response relationship is a response relationship that the target actuator can achieve under the capability limitation, expressed by the amplitude and phase relationships of the contact force received by the target actuator relative to the movement speed of the target actuator at multiple reference frequencies. The filtering unit is configured to, for each of the target task stages, filter from the set of achievable response relationships that the amplitude relationship values ​​at each of the reference frequencies are within the range of amplitude relationship values ​​defined by the confidence interval of the corresponding original task stage, and the phase relationship values ​​are within the range of phase relationship values ​​defined by the confidence interval, to obtain at least one target response relationship; The first determining unit is configured to, for each of the target task stages, at each of the reference frequencies, determine the allowable range of the amplitude relationship and the phase relationship of the target response relationship at the reference frequencies, with the minimum and maximum values ​​of the amplitude relationship and the phase relationship of each target response relationship as the lower limit and the upper limit, respectively; An adjustment unit is configured to, for each of the target task stages, under the constraints of the corresponding allowable range and the set of achievable response relationships, adjust the amplitude relationship and the phase relationship in the interaction feature to be adjusted with the goal of reducing the difference between the interaction feature to be adjusted and the corresponding original interaction feature in each of the target task stages, thereby obtaining the target interaction feature of the target task stage.

[0011] Optionally, the device further includes: The second acquisition module is used to acquire, before adjusting the interaction features to be adjusted in each of the original task stages and each of the target task stages to the direction of the original interaction features corresponding to the original task stages when the target task is executed for the target actuator, under the limitation of the target actuator's ability in terms of motion and contact interaction, to obtain the target interaction features of each of the target task stages, and before obtaining the target interaction features of each of the original task stages, each of the target task stages acquires the entry contact event, contact mode and energy flow direction of each of the original task stages and each of the target task stages. The entry contact event is a contact event used to identify the start of the corresponding task stage. The contact mode refers to the contact interaction mode between the original actuator or the target actuator and the operation object in the corresponding task stage. The energy flow direction indicates the direction of mechanical energy transfer between the original actuator or the target actuator and the operation object in the corresponding task stage. A module is established to establish a stage correspondence between the original task stage and the target task stage, where the stage completion requirements are the same, the energy flow direction is the same, and the original actuator and the target actuator are respectively allowed to achieve the stage completion requirements through their respective contact methods. The existing sequential relationship between the entry contact events of the original task stage remains unchanged between the corresponding target task stages.

[0012] Optionally, the device further includes: The third acquisition module is used to, after generating operation control parameters for the target actuator in each target task stage based on the target interaction characteristics and stage completion requirements of each target task stage, and obtaining the transferred operation skills adapted to the target actuator, acquire the actual interaction characteristics of the currently executed target task stage during the process of controlling the target actuator to execute the target task according to the operation control parameters, wherein the actual interaction characteristics characterize the response relationship between the motion changes of the target actuator when it interacts with the operation object and the changes in the contact action it receives; A determining module is used to determine the deviation between the actual interaction features and the target interaction features of the currently executed target task stage, and to determine the parameter adjustment amount for reducing the deviation based on the deviation and the sensitivity of the deviation to the operation control parameters, wherein the sensitivity characterizes the degree of change of the deviation caused by the change of the operation control parameters; The fourth acquisition module is used to acquire the available energy value for this parameter adjustment, and to determine the additional mechanical energy that the target actuator is expected to input into its contact interaction process with the operation object when updating the operation control parameters according to the parameter adjustment amount, and to obtain the additional mechanical energy value, wherein the available energy value represents the upper limit of the additional mechanical energy allowed for this parameter adjustment; The update module is used to update the operation control parameters of the currently executed target task phase according to the parameter adjustment amount, provided that the additional mechanical energy value does not exceed the available energy value. The reduction module is used to reduce the magnitude of the parameter adjustment amount when the additional mechanical energy value exceeds the available energy value, to obtain a limited parameter adjustment amount, and update the operation control parameters of the currently executed target task stage according to the limited parameter adjustment amount, or cancel the current parameter adjustment and retain the operation control parameters, wherein when updating the operation control parameters according to the limited parameter adjustment amount, the additional mechanical energy expected to be input by the target actuator to its contact interaction process with the operation object does not exceed the available energy value; The control module is used to control the target actuator to continue executing the currently executing target task stage according to the updated or retained operation control parameters.

[0013] Optional, the fourth acquisition module includes: The second determining unit is used to determine the energy update period by taking the time of the last update of the available energy value as the start time and the current time before the current parameter adjustment as the end time, and to obtain the available energy value updated at the start time. The third acquisition unit is used to acquire the dissipated power and additional input power during the energy update period, wherein the dissipated power characterizes the rate of mechanical energy loss during the contact interaction process between the target actuator and the operated object, and the additional input power characterizes the rate at which the target actuator inputs additional mechanical energy into the contact interaction process due to the implemented parameter adjustments; The subtraction unit is used to take the difference between the dissipated power and the additional input power as the rate of change of the available energy value; An integration unit is used to integrate the rate of change of the available energy value over time during the energy update period to obtain the energy change amount. The sum of the available energy value updated at the starting time and the energy change amount is determined as the available energy value used for this parameter adjustment.

[0014] Optional, the third acquisition module includes: The superposition unit is used to superimpose the verification excitation signal as an additional control quantity onto the initial control command generated according to the operation control parameters of the currently executed target task stage to obtain the adjusted target control command, wherein the verification excitation signal is used to control the target actuator to apply additional detection actions within a preset verification period. The acquisition unit is used to control the target actuator to perform the motion and contact interaction corresponding to the current target task stage according to the target control command during the preset verification period, limit the displacement change amplitude of the target actuator caused by the additional detection action to within a preset displacement amplitude upper limit, and acquire force signals and velocity signals, wherein the force signal represents the change of contact force received by the target actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement speed of the target actuator over time; The second analysis unit is used to perform spectral analysis on the force signal and the velocity signal, compare the force signal component and the velocity signal component at the same frequency, determine the amplitude relationship and phase relationship of the contact force relative to the motion velocity, and obtain the actual interaction characteristics. The actual interaction characteristics and the target interaction characteristics are respectively represented by their amplitude relationship and phase relationship at the same frequency.

[0015] Optionally, the device further includes: The fifth acquisition module is used to acquire, after determining the deviation between the actual interaction feature and the target interaction feature in the currently executed target task stage, the deviation between the actual interaction feature and the target interaction feature determined in multiple consecutive steps during the execution of the current target task stage by the target executor. The reduction module is used to reduce the control bandwidth of the target actuator during the contact interaction process when the deviation is determined to increase sequentially in multiple consecutive iterations. The reduced control bandwidth is used to perform impedance control on the target actuator, and the execution flow of the target task is rolled back to the starting point of the currently executed target task stage. The impedance control is used to adjust the response relationship between the motion change of the target actuator and the change in the contact action it receives. The control bandwidth characterizes the upper limit of the frequency at which the target actuator can effectively follow the changes in control commands under corresponding control.

[0016] According to another aspect of the present disclosure, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory, wherein the processor is configured to execute the computer program to implement the steps of any of the above-described robot operation skill transfer methods.

[0017] According to another aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above-described robot operation skill transfer methods.

[0018] According to one aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of any of the above-described robot operation skill transfer methods.

[0019] The technical solution provided in this disclosure proposes a method for transferring robot operation skills. Using the original interaction features of the original task stage as a reference, the method adjusts the interaction features to be adjusted for the corresponding target task stage under the capability limitations of the target actuator, and generates operation control parameters based on the target interaction features and stage completion requirements. Since the original interaction features characterize the response relationship between motion changes and changes in the received contact action, and the corresponding task stages have the same stage completion requirements, skill transfer can be based on the contact interaction performance when completing the corresponding task stage. Simultaneously, the capability limitations of the target actuator constrain the adjustment range of the interaction features, and the generated operation control parameters can take into account both the task completion requirements and the actual capabilities of the target actuator. Therefore, this method can solve the technical problem in related technologies where the adaptability of robot operation skills after transfer between heterogeneous actuators is low with the actuator receiving the skill, achieving the technical effect of improving the adaptability of robot operation skills after transfer between heterogeneous actuators with the actuator receiving the skill. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments of this disclosure will be briefly introduced below.

[0021] Figure 1 This is a schematic diagram of an application environment for an optional robot operation skill transfer method according to an embodiment of the present disclosure; Figure 2 This is a flowchart illustrating an optional method for transferring robot operating skills according to an embodiment of the present disclosure; Figure 3 This is a schematic diagram illustrating an optional stage correspondence between the original task stage and the target task stage according to an embodiment of the present disclosure; Figure 4 This is a schematic diagram of an optional robot operation skill transfer and execution feedback process according to an embodiment of the present disclosure; Figure 5 This is a schematic diagram of an optional heterogeneous actuator operation skill migration process according to an embodiment of the present disclosure; Figure 6 This is a schematic diagram of an optional skill transfer comparison verification according to an embodiment of the present disclosure; Figure 7 A schematic diagram of the structure of a robot operation skill transfer device provided in an embodiment of this disclosure; Figure 8This is a schematic diagram of the structure of an optional electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0022] The embodiments of this disclosure are described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of this disclosure, and do not constitute a limitation on the technical solutions of the embodiments of this disclosure.

[0023] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the terms “comprising” and “including” as used in embodiments of this disclosure mean that the corresponding feature can be implemented as the presented feature, information, data, step, operation, element, and / or component, but do not exclude implementation as other features, information, data, step, operation, element, component, and / or combinations thereof supported by the art. It should be understood that when we say that an element is “connected” or “coupled” to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element are connected through an intermediate element. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The term “and / or” as used herein indicates at least one of the items defined by the term, for example, “A and / or B” or “A, B” indicates implementation as “A,” or implementation as “B,” or implementation as “A and B.”

[0024] To make the objectives, technical solutions, and advantages of this disclosure clearer, the embodiments of this disclosure will be described in further detail below with reference to the accompanying drawings.

[0025] First, the technical terms used in this disclosure will be introduced and explained: End effector: refers to a device installed at the operating end of a robot for direct interaction with the object being manipulated, such as a gripper, dexterous hand, suction cup, or tool end.

[0026] Degrees of freedom: These refer to the number of independent variables required to describe the independent motion of a mechanical mechanism. For example, an independent translation along a direction or an independent rotation about an axis each corresponds to one degree of freedom.

[0027] Stiffness refers to the ability of an object or mechanical structure to resist deformation. Under the same stress conditions, the greater the stiffness, the smaller the deformation will generally be.

[0028] Damping: refers to the force that opposes relative motion and dissipates mechanical energy during motion, and can be used to suppress vibration. In the viscous damping model, the damping force is proportional to the relative velocity and its direction is opposite to the direction of relative motion.

[0029] Frequency response refers to the relationship between the amplitude ratio and phase difference of the output relative to the input as a function of frequency under periodic inputs of different frequencies. The amplitude reflects the magnitude of the periodic change, while the phase difference reflects the relative lead or lag of two periodic changes of the same frequency within one cycle.

[0030] Mechanical impedance refers to the dynamic relationship between force and velocity in mechanical interaction. In the frequency domain, it can be represented by the amplitude ratio and phase difference of force and velocity at the same frequency, reflecting the combined influence of factors such as inertia, stiffness, and damping on the motion response.

[0031] Impedance control refers to a control method that regulates the contact interaction between an actuator and the manipulated object by adjusting the dynamic relationship between motion and force. For example, by setting stiffness and damping parameters, the motion response of the actuator when subjected to contact force can be adjusted.

[0032] Control bandwidth: refers to the frequency range within which an actuator can effectively follow changes in control commands under corresponding control. In this disclosure, the upper limit of this frequency range is used to represent the control bandwidth, which is different from the transmission frequency of control commands.

[0033] This disclosure proposes a method for transferring robot operational skills. Before introducing the optional embodiments of this disclosure, relevant technologies will be described first to facilitate understanding of the technical concept of this disclosure: After replacing a robot's gripper, dexterous hand, or other end effector, it is usually necessary to apply existing operational skills to the new actuator. Related technologies can record motion trajectories, force changes, or stiffness parameters during the teaching process, and achieve skill reuse by adjusting the trajectory and control parameters. For example, when tightening a bottle cap with an elastic sealing ring, the rotation trajectory and gripping parameters of the original actuator can be used as a reference for another actuator to complete the tightening operation. However, actuators differ in structure, driving capability, and contact method; similar motion commands may not produce similar contact performance, potentially leading to unstable gripping, insufficient clamping, or excessive contact force. Different actuators may also require different actions and durations to complete the same operation. Directly executing according to the original action sequence and duration may result in premature switching before the conditions for entering the next operation are met. Furthermore, changes in object friction or contact state may cause the execution process to deviate from expectations, while improper adjustment of control parameters may cause vibration or contact instability. Therefore, when transferring robot operational skills between heterogeneous actuators, there are still problems such as low compatibility between the transferred skills and the receiving actuator, and insufficient contact stability during actual execution.

[0034] In view of at least one of the above-mentioned technical problems or areas that need improvement in the related technologies, this disclosure proposes a method for transferring robot operation skills, which can solve the technical problem in the related technologies that the degree of adaptation between robot operation skills and the receiving actuator after the transfer between heterogeneous actuators is low, and achieve the technical effect of improving the degree of adaptation between robot operation skills and the receiving actuator after the transfer between heterogeneous actuators.

[0035] The technical solutions and their effects according to the embodiments of this disclosure are described below through several exemplary implementation methods. It should be noted that the following implementation methods can be referenced, learned from, or combined with each other, and the same terms, similar features, and similar implementation steps in different implementation methods will not be described again.

[0036] According to one aspect of the embodiments of this disclosure, a method for transferring robot operational skills is provided. Optionally, in this embodiment, Figure 1 This is a schematic diagram illustrating an application environment for an optional robot operation skill transfer method according to an embodiment of the present disclosure. The robot operation skill transfer method described above can be applied to, but is not limited to, situations such as... Figure 1 The hardware environment shown includes terminal device 102 and server 104. Server 104 can be connected to terminal device 102 via a network and can be used to provide services (e.g., application 107, etc.) to terminal device 102 or clients installed on terminal device 102. Database 105 can be set up on or independently of server 104 to provide data storage services for server 104.

[0037] The aforementioned network may include, but is not limited to, at least one of the following: wired network and wireless network. The aforementioned wired network may include, but is not limited to, at least one of the following: wide area network (WAN), metropolitan area network (MAN), and local area network (LAN). The aforementioned wireless network may include, but is not limited to, at least one of the following: Wireless Fidelity (WIFI) and Bluetooth. Terminal device 102 may be, but is not limited to, a personal computer (PC), mobile phone, tablet computer, etc. Server 104 may be, but is not limited to, a cloud server, server cluster, or other server types.

[0038] The robot operation skill transfer method of this disclosure embodiment can be executed by server 104, terminal device 102, or jointly by server 104 and terminal device 102. Alternatively, the robot operation skill transfer method of this disclosure embodiment can be executed by a client installed on the terminal device 102.

[0039] Taking the robot operation skill transfer method in this embodiment, executed by terminal device 102 (server 104), as an example, Figure 2 This is a flowchart illustrating an optional method for transferring robot operating skills according to an embodiment of the present disclosure, as shown below. Figure 2 As shown, the process of this method may include the following steps: Step S11: Receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; Step S12: In response to the skill transfer request, obtain the original interaction features of each original task stage during the process of the original executor performing the target task through the operation skill, wherein the original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when interacting with the operation object during the execution of the target task. Step S13: For each target task stage corresponding to each original task stage when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, the interaction features to be adjusted in each target task stage are adjusted in the direction of the original interaction features corresponding to the original task stage to obtain the target interaction features of each target task stage, wherein the corresponding target task stage and the original task stage have the same stage completion requirements. Step S14: Based on the target interaction features and the stage completion requirements of each target task stage, generate the operation control parameters of the target executor in each target task stage, and obtain the migrated operation skills adapted to the target executor.

[0040] In step S11, the original actuator is the actuator that provides the operational skill to be transferred, and the target actuator is the actuator that receives the operational skill. The heterogeneity between the original actuator and the target actuator can be manifested in at least one difference in drive type, degrees of freedom, gripping structure, sensor configuration, or control interface. For example, the original actuator can be a three-finger dexterous hand, and the target actuator can be a parallel gripper.

[0041] Optionally, in this embodiment, the target task can be a grasping, twisting, or assembly task involving contact interaction. The process of the original actuator performing the target task can be divided into multiple original task stages according to the changes in the contact state, such as the initial contact, constrained movement, and locking stages. Different original task stages correspond to different motion and force manifestations. By acquiring the original interaction characteristics of each original task stage, the interaction characteristics of the operation skill under different contact states can be preserved.

[0042] In step S12, the original interaction characteristics of each original task stage can be obtained based on the motion and force information recorded by the original actuator when performing the target task. These original interaction characteristics reflect how the contact action changes when the motion changes, and the relative lag between these changes. For example, when the motion speed changes periodically, the magnitude of the change in contact force and its lag relative to the speed change can be used to describe the response relationship of the corresponding task stage. This response relationship reflects the dynamic characteristics such as stiffness, damping, and resonance when the original actuator contacts the operated object.

[0043] As an optional implementation, the original interaction features can be represented using an interaction impedance spectrum, which describes the interaction behavior through the amplitude and phase relationships between contact force and motion velocity at multiple frequencies. For tasks involving rotational motion, the response relationship between torque and angular velocity can also be described. The response relationships corresponding to translational and rotational motions can be represented separately or combined into a matrix form containing three translational directions and three rotational directions. Thus, the original interaction features can retain the interaction behavior in different motion directions related to the target task, for reference when subsequently determining the target interaction features.

[0044] In step S13, the correspondence between the original task stage and the target task stage is first determined, and task stages with the same stage completion requirements are associated. The stage completion requirements are used to determine whether the corresponding operation is completed, such as whether stable contact has been established or whether locking has been completed. The original actuator and the target actuator can use their respective applicable contact methods and motion methods to achieve the same stage completion requirements, and the corresponding target task stage is adjusted based on the original interaction characteristics of the original task stage.

[0045] Furthermore, the limitations of the target actuator in terms of motion and contact interaction can be determined based on its structure, rated power, force and displacement range, effective mass, control bandwidth, and available contact methods. Motion limitations can include the direction and range of motion, while contact interaction limitations can include the direction and magnitude of the applicable contact force, as well as the achievable contact methods and dynamic response range. For actuators with different sizes, masses, or rated power, the interaction characteristics used for comparison can be converted to a comparable scale based on their respective equipment parameters before subsequent adjustments are made.

[0046] The interaction features to be adjusted represent the response relationship that the target actuator intends to achieve in the corresponding target task stage. For each target task stage, within the range of response relationships allowed by the target actuator's capabilities, the differences between different candidate response relationships and the original response relationship are compared with the corresponding original interaction features as a reference. The values ​​of the interaction features to be adjusted are then adjusted to obtain target interaction features that meet capability constraints and are close to the original interaction features. When using interactive impedance spectroscopy representation, the above adjustments can be reflected in adjusting the amplitude and phase relationships at different frequencies. The target interaction features can also include the allowable range of variation of these relationships for use in generating operation control parameters.

[0047] In step S14, the target interaction features are taken as the interaction requirements that the target actuator needs to achieve in the corresponding task stage. Combined with the stage completion requirements, the operation control parameters required for motion and contact interaction are jointly determined. The operation control parameters may include motion path, motion speed, clamping or adsorption commands, stiffness parameters, damping parameters, and stage switching conditions. The motion path may include path components along the normal and tangential directions of the contact surface, corresponding to motion perpendicular to and along the contact surface, respectively. The clamping or adsorption commands are used to establish and maintain the corresponding contact. The stiffness and damping parameters are used to adjust the response relationship between changes in motion and changes in contact action. The stage switching conditions are determined according to the stage completion requirements.

[0048] Optionally, in this embodiment, the aforementioned operation control parameters can be used as variables in the joint solution, with the interaction requirements and stage completion requirements of each target task stage as the solution basis, to adjust and compare the candidate operation control parameters. Within the joint range of motion, contact force range, control bandwidth, and the allowable range of the manipulated object of the target actuator, the difference between the interaction performance corresponding to the candidate operation control parameters and the target interaction characteristics, as well as the difference between the completion status of the task stage and the stage completion requirements, is reduced, while also taking into account the smoothness of the motion trajectory and energy consumption. For operations that require friction to maintain contact, the direction and magnitude of the contact force can also be limited to a range that can maintain the corresponding contact state. Through the above joint solution, the mutually cooperating motion parameters and contact interaction parameters are determined.

[0049] For example, when transferring the skill of tightening a bottle cap with an elastic sealing ring using a three-finger dexterity hand to a parallel gripper, the normal gripping force, wrist rotation speed, stiffness, and damping parameters of the gripper can be jointly determined based on the target interaction characteristics of the corresponding task stage. In the sealing ring tightening stage, the parameter combination is determined based on the interaction performance corresponding to the above parameters approximating the target interaction characteristics and meeting the tightening completion requirements. Simultaneously, the stage switching conditions are determined based on the interaction changes characterizing the tightening completion, thus forming the operation content suitable for the parallel gripper to perform.

[0050] By combining the operational control parameters of each target task stage with the corresponding stage switching conditions, a migrated operational skill suitable for the target actuator is obtained. As an optional implementation method, the generated operational skill can also be simulated or executed at low speed to verify whether the corresponding motion and contact interaction meet the target interaction characteristics and stage completion requirements.

[0051] By combining the interactive characteristics of the original actuator at each task stage with the actual capabilities of the target actuator in the above manner, and generating corresponding motion and contact interaction parameters, we can adapt to the differences in structure, motion capabilities and contact methods of heterogeneous actuators while retaining the task completion requirements, thereby improving the adaptability of the transferred operating skills to the target actuator.

[0052] As an optional approach, obtaining the original interaction features of each original task stage during the process of the original executor performing the target task through the operational skill further includes: S21, acquire the force signal and velocity signal during the execution of the target task by the original actuator, wherein the force signal represents the change of the contact force received by the original actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement velocity of the original actuator over time; S22, for each of the original task stages, perform spectrum analysis on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple reference frequencies, wherein the multiple reference frequencies correspond to the variation components of different periods in the force signal and velocity signal. S23, for each of the original task stages, the force signal component with the same reference frequency is compared with the velocity signal component to determine the amplitude relationship and phase relationship of the force signal component relative to the velocity signal component, and the multiple reference frequencies are combined with the corresponding amplitude relationship and phase relationship to form the original interaction feature of the original task stage.

[0053] Optionally, in this embodiment, during the execution of the target task by the original actuator, signals of contact force and motion velocity changes over time are recorded, and a temporal correspondence between the two types of signals is established based on the acquisition time. For signals in different motion directions, the correspondence between the contact force component and the velocity component in the corresponding direction is retained to analyze the force and motion changes within the same time range and in the same motion direction. During task execution, the clamping state, the motion of the manipulated object, and the task completion signal can also be recorded as auxiliary information for identifying the original task stage.

[0054] In step S21, the changes used to form the force and velocity signals can originate from detection actions superimposed during the task motion, or from natural task disturbances or passive environmental stimuli. For example, when the original actuator contacts the manipulated object and relative motion occurs, changes in the contact state can cause changes in force and velocity. When analyzing the corresponding signals, the signals need to contain changing components that can be used to estimate the response relationship within the required frequency range, and satisfy the safety energy constraints of the contact interaction process.

[0055] In step S22, force and velocity signals within the time range corresponding to the original task phase are selected for spectral analysis, decomposing the time-varying signals into components of different frequencies. Each frequency component of the force and velocity signals has amplitude and phase, respectively. The amplitude describes the magnitude of the corresponding periodic change, and the phase determines the position of the change within one period. By analyzing the two types of signals with a time correspondence using the same time reference, the magnitudes of the two types of changes at the same frequency and their relative lead or lag can be compared.

[0056] The boundaries of the original task stages can be determined by combining the above analysis results. Specifically, the changes in the cross-impedance spectrum within adjacent time windows can be compared, and contact events detected visually or tactilely can be used to identify state changes such as initial contact, entry guidance, sliding, or locking. The corresponding change positions are used as the basis for stage division. For each original task stage obtained, the response relationship is estimated using signal segments with relatively stable contact states within the corresponding stage, thereby correlating the frequency analysis results with the original task stages.

[0057] In step S23, for the same reference frequency, the amplitude of the force signal component is compared with the amplitude of the velocity signal component to obtain the amplitude relationship, and the phases of the two are compared to obtain the phase relationship. Optionally, the above relationship can be estimated by the cross spectrum of the force signal and the velocity signal, and the autospectrum of the velocity signal. The cross spectrum describes the common variation and phase relationship of the two types of signals at the same frequency, while the autospectrum describes the intensity distribution of the velocity signal at each frequency. The cross-impedance spectrum can be calculated according to the following relationship:

[0058] in, Indicates angular frequency; This represents the interaction impedance at the corresponding angular frequency; Represents the cross spectrum of force and velocity signals; The autospectrum representing the velocity signal; This represents the inverse of the velocity self-spectrum matrix, and in the case of a single channel, it represents the reciprocal of the velocity self-spectrum. For multi-channel signals, the cross-spectrum and self-spectrum are in matrix form, with the rows and columns of the matrix corresponding to the signal channels involved in the analysis. The calculation is applicable to cases where the velocity self-spectrum matrix is ​​invertible.

[0059] For the complex interactive impedance obtained from single-channel estimation, its magnitude represents the amplitude ratio of the contact force to the motion velocity, and its argument represents the phase difference between the contact force and the motion velocity. For multi-channel estimation results, the amplitude and phase relationships between the corresponding channels can be preserved. By associating and recording each reference frequency with its corresponding amplitude and phase relationships, the original interactive characteristics of the initial task phase can be formed, which can be used by the target actuator when performing skill adaptation in the corresponding task phase.

[0060] For example, when tightening a bottle cap with an elastic sealing ring, the cross-impedance spectrum during the initial contact phase can show an increase in low-frequency stiffness, while the cross-impedance spectrum after thread engagement can show phase lag during torsional motion. By retaining the frequency and amplitude-phase relationships of these phases, the force and motion responses under different contact states can be described. As an optional implementation, features such as low-frequency equivalent stiffness, damping variation characteristics, resonant peak frequency, and peak gain can be extracted from the cross-impedance spectrum, and the dissipated energy can be determined by combining this with the force and motion records within one motion cycle. This information can then be used as a supplement to the original cross-impedance characteristics.

[0061] By using the above method, the force and motion changes of the original actuator when performing the target task are converted into multi-frequency amplitude-phase relationships of each original task stage. This can preserve information such as stiffness, damping, resonance, and phase lag in the contact interaction, improve the ability to characterize the differences in dynamic response under different contact states, and provide a reference corresponding to the task stage for adjusting the interaction characteristics of the target actuator.

[0062] As an optional solution, the acquisition of force and velocity signals during the execution of the target task by the original actuator further includes: S31, the detection excitation signal is superimposed as an additional control quantity onto the initial task command used to control the original actuator to execute the target task, thereby obtaining the adjusted target task command, wherein the detection excitation signal is used to control the original actuator to apply additional detection actions during the execution of the target task, and the frequency of the detection excitation signal is within a preset frequency range; S32, during the process of the original actuator executing the target task command, the displacement change amplitude of the original actuator caused by the additional detection action and the contact force change amplitude between the original actuator and the operating object are respectively limited to the preset upper limit of displacement amplitude and the upper limit of force amplitude. The mechanical energy input by the original actuator to the contact interaction process between the original actuator and the operating object through the additional detection action is limited to the preset upper limit of energy, and the force signal and the velocity signal are collected.

[0063] Optionally, in this embodiment, during the execution of the target task by the original actuator, an additional control quantity for detecting contact dynamic characteristics is superimposed on the initial task command. The initial task command is used to drive the original actuator to complete the motion and contact operations in the target task. The additional control quantity is provided by the detection excitation signal and is used to introduce restricted motion changes based on the task motion, and obtain the corresponding contact force and velocity changes. In this way, signals for estimating the original interaction characteristics can be obtained during task execution.

[0064] In step S31, a preset frequency range can be determined based on the contact dynamic characteristics to be analyzed, and the frequency of the detection excitation signal can be selected within this range. The preset frequency range corresponds to the frequency band where the response relationship needs to be identified. When selecting the frequency band, the dynamic characteristics of the original actuator and the contact changes that the operated object can withstand are taken into account, and frequency bands with well-defined dynamic characteristics and suitable for safe detection are preferred. For example, the identification frequency band can be selected from 0.5 Hz to 20 Hz to cover the frequency range corresponding to quasi-static stiffness and main contact dynamics. The selection of the detection excitation signal frequency is linked to the subsequent spectrum analysis, providing excitation for the estimation of the response relationship at the corresponding frequency.

[0065] As an optional implementation, the detection excitation signal can be a multi-sine signal, combining multiple sinusoidal components of different frequencies into an additional control quantity, the expression of which is:

[0066] in, Indicates at time The applied detection excitation signal; The number represents the sine component, and the summation range is the sum of all sine components involved in the combination; , and These represent the amplitude, angular frequency, and initial phase of the corresponding sinusoidal component, respectively. The angular frequency of each component corresponds to a frequency within a preset frequency range. The combined signal is superimposed on the initial task command as an additional control quantity to obtain the target task command.

[0067] The excitation signal can also be a linear sweep frequency signal, a logarithmic sweep frequency signal, or a pseudo-random binary sequence. Linear and logarithmic sweep frequency signals change the excitation frequency according to linear and logarithmic laws, respectively, to sequentially detect the response within a selected frequency range. The pseudo-random binary sequence provides the detected change by taking the values ​​of two sequentially changing signals. When using these signals, their rate of change and spectral distribution can be set in conjunction with a preset frequency range, and additional control quantities can be determined according to displacement, contact force, and mechanical energy limitations.

[0068] In step S32, the original actuator executes the target task according to the target task command and generates an additional detection action in addition to the task motion. The displacement change amplitude is the magnitude of the displacement change of the original actuator relative to the initial task motion caused by the additional detection action, and the contact force change amplitude is the magnitude of the contact force change caused by this additional action. Both types of amplitudes are constrained by their corresponding upper limits, and the amplitude of the detection excitation signal needs to be determined by combining the relationship between the actual control quantity and the motion and force. For example, when the additional displacement is used as the detection control quantity, the force change is also related to the contact stiffness of the manipulated object, and the applicable additional displacement needs to be determined by combining the limit of the contact force change amplitude.

[0069] The preset upper limit of displacement amplitude can be determined by considering the allowable motion offset, and the preset upper limit of force amplitude can be determined by considering the strength of the object being manipulated. For example, the displacement amplitude of a minute detection action can be selected within the range of 0.05 mm to 0.5 mm, the contact force change amplitude can be selected within the range of 0.2 N to 5 N, and the detection duration of a single initial task phase can be selected within the range of 0.3 s to 2 s. The above values ​​are determined by the specific actuator and the object being manipulated, and satisfy the corresponding upper limit of amplitude and energy constraints.

[0070] Regarding mechanical energy constraints, the driving action of the original actuator transmits mechanical energy to the manipulated object through contact. A detection excitation signal is used to specify additional actions, and the mechanical energy is provided by the actuator when performing these additional actions. The mechanical energy input into the contact interaction process during the detection period can be evaluated by combining the force and motion conditions of the additional detection action. By constraining the amplitude and duration of the additional action, the input mechanical energy can be controlled within a preset energy upper limit. Displacement and contact force amplitude constraints limit the magnitude of the additional changes, while mechanical energy constraints further limit the cumulative effect of the additional action during its application.

[0071] During the task execution process that satisfies the above constraints, force and velocity signals from the original actuator are collected, and the correspondence between these two types of signals and the detection period is retained. The collected signals are used to analyze the response of the original actuator to the additional detection action in the current contact state, thereby determining the original interaction characteristics of the corresponding original task stage.

[0072] By introducing additional probing actions with selected frequency components during the execution of the target task in the above manner, the required motion and force changes can be provided for interactive response estimation. At the same time, by limiting the additional displacement, contact force changes and input mechanical energy, the influence of the probing actions on the task motion and the manipulated object can be controlled, taking into account both the effectiveness of interactive feature acquisition and the safety of contact interaction.

[0073] As an optional approach, the step of performing spectral analysis on the corresponding force signal and velocity signal for each of the original task stages to obtain force signal components and velocity signal components at multiple reference frequencies further includes: S41, for each of the original task stages, perform spectrum analysis on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple candidate frequencies, and calculate the coherence coefficient of the force signal and the velocity signal at each of the candidate frequencies, wherein the coherence coefficient is used to indicate the degree of correlation between the force signal and the velocity signal at the corresponding candidate frequency. S42, the candidate frequencies with coherence coefficients not less than a preset threshold are determined as a plurality of reference frequencies, and the force signal components and velocity signal components at the plurality of reference frequencies are retained.

[0074] Optionally, in this embodiment, the frequencies obtained from spectral analysis are used as candidate frequencies, and a reference frequency for forming the original interaction characteristics is determined based on the correlation between the force signal and the velocity signal at each candidate frequency. The candidate frequencies correspond to the signal components participating in the screening, and the reference frequencies correspond to the screened signal components used for subsequent amplitude and phase relationship analysis.

[0075] In step S41, for each initial task stage, the coherence coefficients of the force and velocity signals at each candidate frequency are calculated, and the calculation results are correlated with the corresponding candidate frequencies. The coherence coefficient is used to evaluate the degree of linear correlation between the two types of signals at the corresponding frequency. When using a coherence coefficient with a value ranging from 0 to 1, the closer the value is to 1, the higher the degree of linear correlation between the two types of signals at the corresponding frequency, which can provide a more reliable basis for estimating the interaction response relationship.

[0076] For example, when the actuator comes into contact with the object being manipulated, the change in velocity and the change in contact force can have a certain amplitude ratio and phase difference. A high coherence coefficient indicates that the two types of signals can be related by a relatively stable linear relationship at the corresponding frequency, while the amplitude ratio and phase difference are used to describe the specific manifestation of this relationship. Therefore, even if there is a phase lag between the change in contact force and the change in velocity, the two types of signals can still have a high coherence coefficient.

[0077] In step S42, the coherence coefficients corresponding to each candidate frequency are compared one by one with a preset threshold, and candidate frequencies whose coherence coefficients reach the preset threshold are retained. For example, the preset threshold can be set to 0.8, and candidate frequencies with coherence coefficients greater than or equal to 0.8 are determined as reference frequencies, and the force signal components and velocity signal components corresponding to these reference frequencies are retained. Subsequently, the correspondence between the two types of signal components is established according to the reference frequencies, the amplitude relationship and phase relationship at each reference frequency are determined, and these are used to form the original interaction features of the corresponding original task stage.

[0078] By using the above method, the signal components used for interaction feature extraction can be selected according to the degree of correlation between the force signal and the velocity signal at different frequencies. This reduces the impact of frequency components with low correlation on response relationship estimation, improves the reliability of the original interaction features, and provides a basis for subsequent adjustment of target interaction features.

[0079] As an optional solution, when the target actuator performs the target task, the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages are adjusted in the direction of the original interaction features corresponding to the original task stages, under the limitations of the target actuator's ability in terms of motion and contact interaction, to obtain the target interaction features of each target task stage, further comprising: S51, obtain the confidence intervals of each of the original interaction features and the set of reachable response relationships of the target actuator, wherein the original interaction features include the amplitude and phase relationships of the contact force received by the original actuator relative to the movement speed of the original actuator at multiple reference frequencies when the original actuator interacts with the operation object, the confidence intervals indicate the estimated uncertainty range of the amplitude and phase relationships of the original interaction features at each of the reference frequencies, and the set of reachable response relationships includes multiple response relationships reachable by the target actuator, each reachable response relationship being a response relationship that the target actuator can achieve under the capability limitation, expressed by the amplitude and phase relationships of the contact force received by the target actuator relative to the movement speed of the target actuator at multiple reference frequencies; S52, for each of the target task stages, from the set of achievable response relationships, select response relationships whose amplitude relationship values ​​at each of the reference frequencies are within the range of amplitude relationship values ​​defined by the confidence interval of the corresponding original task stage, and whose phase relationship values ​​are within the range of phase relationship values ​​defined by the confidence interval, to obtain at least one target response relationship; S53, for each of the target task stages, at each of the reference frequencies, the minimum and maximum values ​​of the amplitude relationship and the phase relationship of each of the target response relationships are used as the lower limit and upper limit, respectively, to determine the allowable range of the amplitude relationship and the phase relationship of the interaction feature to be adjusted in the target task stage at the reference frequency; S54, for each target task stage, under the constraints of the corresponding allowable range and the set of achievable response relationships, with the goal of reducing the difference between the interaction features to be adjusted and the corresponding original interaction features in each target task stage, the amplitude relationship and the phase relationship in the interaction features to be adjusted are adjusted to obtain the target interaction features of the target task stage.

[0080] Optionally, in this embodiment, the adjustment range of the target interaction features is determined by combining the estimated uncertainty range of the original interaction features with the response relationships that the target actuator can achieve. The confidence interval of the original interaction features is used to reflect the uncertainty of the response relationship estimation results, and the set of achievable response relationships is used to reflect the actual capability of the target actuator; both participate in the determination of the target interaction features.

[0081] In step S51, a sliding time window can be used to estimate the interaction response within the same original task phase, and the estimation results of the amplitude and phase relationships at each reference frequency are recorded to form corresponding means and confidence intervals. Different reference frequencies correspond to their respective amplitude and phase relationship confidence intervals. During comparison, the amplitude and phase relationships are represented in a consistent manner, and the phase intervals and the phase values ​​to be compared use a consistent phase representation benchmark. The confidence interval reflects the uncertainty range of the estimation results, and the allowable range subsequently determined based on the target actuator capability is used to constrain the interaction characteristics to be adjusted.

[0082] The set of reachable response relationships for a target actuator can be determined based on its structure, rated power, force and displacement range, effective mass, control bandwidth, and available contact methods. A single response relationship within this set comprises interrelated amplitude and phase relationships at multiple reference frequencies, representing a complete set of interactive responses that the target actuator can form within its capability limitations. When represented using an interactive impedance spectrum, the set of reachable response relationships corresponds to the set of reachable impedances of the target actuator.

[0083] To compare the response relationships of actuators of different sizes, masses, or rated power, the force, speed, and frequency can be normalized based on the equipment parameters of both the original and target actuators before screening, and the response relationships and corresponding confidence intervals can be calculated simultaneously. For each actuator, its force variable can be calculated according to the following relationship:

[0084] in, Represents the normalized force variables; Indicates contact force; This represents the reference force for the corresponding actuator. The reference force is a positive value, and the force variable is converted to a relative value according to the corresponding reference force.

[0085] Velocity variables can be converted according to the following relationship:

[0086] in, This represents the normalized velocity variable; Indicates the speed of movement; This represents the reference velocity of the corresponding actuator. The reference velocity is a positive value, and the normalized force and velocity are used to form the response relationship at a relative scale.

[0087] Frequency variables can be converted according to the following relationship:

[0088] in, This represents the normalized angular frequency; Indicates the angular frequency involved in the analysis; This represents the upper limit of the angular frequency corresponding to the control bandwidth of the respective actuator, and uses the same unit as the angular frequencies involved in the analysis. When using normalized frequency comparisons, the reference frequency corresponds to this relative frequency scale.

[0089] The reference force, reference velocity, and reference power satisfy the following relationship:

[0090] in, and These represent the reference force and reference velocity, respectively. This indicates the reference power. The reference power can be selected based on the actuator's rated power, for example, choosing an applicable value within the range of 5 watts to 100 watts, and then determining the reference force and reference speed that satisfy the above relationship.

[0091] After the above conversion, the response relationship between the original actuator and the target actuator is expressed as a comparable relative value according to their respective reference values. When rotational and translational channels are involved, channel conversion can also be performed by combining characteristic length and effective mass. For actuators using contact methods such as adsorption, the corresponding achievable response relationship is determined by combining the contact action direction they can provide, while retaining the limitations of the contact method on the interaction capability.

[0092] In step S52, for each target task stage, a confidence interval for the corresponding original task stage is selected, and candidate response relationships in the reachable response relationship set are checked one by one. At each reference frequency, it is determined whether the amplitude relationship value and the phase relationship value of the candidate response relationship are within the amplitude relationship confidence interval and the phase relationship confidence interval, respectively. When a candidate response relationship satisfies the above conditions at all reference frequencies involved in the comparison, this complete response relationship is retained as the target response relationship. This screening is equivalent to finding response relationships that jointly satisfy the conditions between the confidence interval constraints of the original interaction features and the reachable response relationship set of the target actuator.

[0093] If at least one target response relationship is obtained, proceed to step S53. For each reference frequency, first compare the amplitude values ​​of the retained target response relationships at this frequency, and determine the lower and upper limits of the allowable amplitude relationship range using the minimum and maximum values, respectively; then compare the phase relationship values, and determine the lower and upper limits of the allowable phase relationship range using the minimum and maximum values, respectively. Correlating the allowable amplitude relationship ranges at different reference frequencies forms the target spectrum envelope, and the allowable phase relationship range forms the corresponding phase band. When there is only one target response relationship, the upper and lower limits of the corresponding allowable range can be the same. The target spectrum envelope can be expressed as:

[0094] in, This represents the amplitude envelope of the target task phase at the corresponding angular frequency; and These represent the lower and upper limits of the permissible range of amplitude relationships, respectively. This represents the corresponding angular frequency. When using normalized representation, the envelope and its upper and lower limits are based on the normalized comparison scale.

[0095] In step S54, using the original interaction characteristics corresponding to the original task stage as a reference, the amplitude and phase relationships of the interaction characteristics to be adjusted at each reference frequency are adjusted. During adjustment, it is required that the values ​​at each frequency are within the corresponding allowable range, and that the complete response relationship formed by these values ​​belongs to the set of achievable response relationships. Under these constraints, the estimated mean of the original interaction characteristics can be used as a comparison basis to reduce the difference between the interaction characteristics to be adjusted and the original interaction characteristics. This preserves the achievable relationships between different frequencies and amplitude / phase values, obtaining the target interaction characteristics that the target actuator can achieve, which are then used to generate subsequent operation control parameters.

[0096] As an alternative implementation, the mean and confidence interval of the original interaction features can be projected onto the reachable impedance set of the target actuator to determine the target spectral envelope and phase band. When the bandwidth of the target actuator is insufficient to cover the original analysis frequency band, the action bandwidth of the target skill can be reduced and the target spectral envelope can be redefined. For cases of insufficient reachability, environmental support, dual-arm cooperation, or adjustment of the contact method can be selected based on the missing contact degrees of freedom. If the migration requirements still cannot be met, the migration is terminated and the reasons such as missing force direction, insufficient bandwidth, or insufficient energy capacity are output. The above alternative routes are applicable to the range determination process after obtaining at least one target response relationship.

[0097] By combining the estimation uncertainty of the original interaction features with the capability limitations of the target actuator to determine the adjustment range and maintaining the feasibility of the complete response relationship during the adjustment process, the matching degree between the target interaction features and the original interaction features can be improved, interaction requirements that exceed the capability of the target actuator can be reduced, and a basis for generating suitable operation control parameters can be provided.

[0098] As an optional approach, before adjusting the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages during the execution of the target task for the target actuator, under the limitations of the target actuator's capabilities in terms of motion and contact interaction, in order to obtain the target interaction features of each target task stage, the method further includes: S61, obtain the entry contact event, contact mode and energy flow direction of each of the original task stages and each of the target task stages, wherein the entry contact event is a contact event used to identify the start of the corresponding task stage, the contact mode refers to the contact interaction mode between the original actuator or the target actuator and the operation object in the corresponding task stage, and the energy flow direction indicates the direction of mechanical energy transfer between the original actuator or the target actuator and the operation object in the corresponding task stage; S62, between the original task stage and the target task stage, where the stage completion requirements are the same, the energy flow direction is the same, and the original actuator and the target actuator are respectively allowed to achieve the stage completion requirements through their respective contact methods, a stage correspondence is established, wherein the existing sequential relationship between the entry contact events of the original task stage remains unchanged between the corresponding target task stages.

[0099] Optionally, in this embodiment, before adjusting the interaction features to be adjusted in the target task stage, a stage correspondence between the original task stage and the target task stage is established based on the contact conditions, completion requirements, and mechanical energy transfer direction of the task stage. Through this correspondence, the original task stage that each target task stage should refer to is determined, and the corresponding original interaction features are obtained as the basis for subsequent adjustments.

[0100] Figure 3 This is a schematic diagram illustrating an optional stage correspondence between the original task stage and the target task stage according to an embodiment of this disclosure. Figure 3 As shown, the top and bottom rows represent the phased flow of the original actuator and the target actuator when performing the target task, including free movement, initial contact and guidance, and stable bearing. Arrows within the same row indicate the sequence of phase progression, and the dotted lines between the two rows indicate the correspondence between the original task phase and the target task phase. For the contact interaction phase, the correspondence between phases is determined based on the entry contact event, contact method, energy flow direction, and phase completion requirements.

[0101] In step S61, the entry contact event is used to determine the contact conditions for entering the corresponding task stage. For example, detecting the initial contact between the actuator and the object being operated on can be used as the entry contact event for the initial contact stage; detecting the entry into the guiding contact state can be used as the entry contact event for the corresponding guiding stage. The entry contact event for the original task stage can be determined based on the contact changes in the task execution record, while the target task stage determines the contact conditions that need to be met to enter the corresponding stage by combining the contact method that the target actuator intends to adopt.

[0102] The contact method describes the form of contact through which the actuator applies force to the manipulated object. For example, a gripper holds the manipulated object by clamping, while a suction cup holds it by adsorption. The energy flow direction describes the direction of mechanical energy transfer during the corresponding contact phase. For example, the actuator transfers mechanical energy to the manipulated object when driving it, or the actuator absorbs mechanical energy from the object's reaction. When comparing energy flow directions, the actuator and the manipulated object are used as the two corresponding ends to determine whether their mechanical energy transfer directions are consistent.

[0103] In step S62, candidate corresponding stages can be determined first based on the stage completion requirements, and then the energy flow direction and contact method of the candidate stages can be checked. For two stages with the same stage completion requirements, the same mechanical energy transfer direction, and where both the original actuator and the target actuator can achieve the corresponding completion requirements through their respective contact methods, a stage correspondence is established. Here, the stage completion requirements describe the physical state that needs to be achieved, such as forming a stable load-bearing state that meets the requirements of subsequent tightening operations; the contact method describes the specific way to achieve this state.

[0104] For example, a two-finger gripper establishes a stable load through clamping before performing rotation, and a suction cup establishes a stable load through adsorption before performing rotation. If both can achieve the required stable load stage through their respective contact methods, and the energy flow direction is the same at each stage, a correspondence can be established between their stable load stages. In this way, the target actuator can adopt contact actions suitable for its own structure, while using the interaction performance of the original actuator at the corresponding stage as a reference for skill adaptation.

[0105] When establishing the phase correspondence, the existing sequential relationship between entry contact events is also maintained. For example, if the original task first enters the initial contact and guidance phase and then the stable bearing phase, the entry contact events corresponding to these two phases in the target task will still occur in this order. The sequential relationship constrains the order of occurrence of related contact events, and the duration of each phase can be determined by combining the movement and contact capabilities of the target actuator.

[0106] As an optional implementation, a task phase graph can be used to represent the aforementioned stages and their transition relationships. Here, a task phase corresponds to a task stage, each task stage is represented as a node in the graph, and the transition relationships between stages triggered by contact events are represented as directed edges. The entry contact event, stage completion requirements, contact method, and energy flow direction are associated with the corresponding nodes. The stage correspondence between the original task stage and the target task stage can be represented as follows:

[0107] in, This represents the task phase diagram when the original executor executes the target task; This represents the task phase diagram when the target executor performs the target task; This represents the mapping established between two task phase diagrams. The node correspondence is used to represent the phase correspondence between the original task phase and the target task phase. The mapping maintains the completion requirements of the corresponding phases and the sequence of key contact events.

[0108] The above mappings can be obtained using graph isomorphism, graph edit distance, dynamic programming, optimal transport, or graph neural networks, with stage completion requirements, energy flow, feasibility of contact methods, and the sequence of contact events serving as mapping constraints. For cases where the corresponding operation can be completed through a combination of partial stages, a subgraph mapping that maintains stage completion requirements can be used to establish a correspondence between selected stages and their relationships. After establishing the correspondence, the target task stages are associated with the original interaction features according to the mapping results for subsequent use in determining the target interaction features.

[0109] By using the above method, a phase correspondence can be established based on the physical state and contact interaction conditions required for each phase of the task, and the order of occurrence of key contact events can be preserved. This allows for the adaptation to different actuator contact methods while determining a reference that is compatible with the task process for adjusting the target interaction characteristics, thereby reducing the impact of phase misalignment on skill transfer.

[0110] As an optional approach, after generating the operation control parameters of the target executor in each of the target task stages based on the target interaction characteristics and stage completion requirements, and obtaining the transferred operation skills adapted to the target executor, the method further includes: S71, during the process of controlling the target actuator to execute the target task according to the operation control parameters, the actual interaction characteristics of the currently executed target task stage are obtained, wherein the actual interaction characteristics characterize the response relationship between the motion change of the target actuator when it comes into contact with the operation object and the change of the contact action it receives; S72, determine the deviation between the actual interaction feature and the target interaction feature of the currently executed target task stage, and determine the parameter adjustment amount for reducing the deviation based on the deviation and the sensitivity of the deviation to the operation control parameter, wherein the sensitivity characterizes the degree of change of the deviation caused by the change of the operation control parameter; S73, obtain the available energy value for this parameter adjustment, and determine the additional mechanical energy that the target actuator is expected to input to its contact interaction process with the operation object when updating the operation control parameters according to the parameter adjustment amount, and obtain the additional mechanical energy value, wherein the available energy value represents the upper limit of the additional mechanical energy allowed for this parameter adjustment; S74, if the additional mechanical energy value does not exceed the available energy value, update the operation control parameters of the currently executed target task phase according to the parameter adjustment amount; S75, if the additional mechanical energy value exceeds the available energy value, reduce the range of the parameter adjustment amount to obtain a limited parameter adjustment amount, and update the operation control parameters of the currently executed target task stage according to the limited parameter adjustment amount, or cancel the current parameter adjustment and retain the operation control parameters, wherein, when updating the operation control parameters according to the limited parameter adjustment amount, the additional mechanical energy expected to be input by the target actuator to its contact interaction process with the operation object does not exceed the available energy value; S76, according to the updated or retained operation control parameters, control the target actuator to continue executing the currently executing target task phase.

[0111] Optionally, in this embodiment, when the target actuator performs the target task according to the transferred operational skills, the operation control parameters are corrected online based on the actual performance of the current contact interaction. The online correction applies to the currently executed target task stage, adopts the corresponding stage completion requirements, and adjusts the motion and contact interaction parameters in conjunction with changes in the contact state.

[0112] In step S71, the motion and force information of the target actuator in the current task stage is acquired to form actual interaction characteristics. The actual interaction characteristics and the target interaction characteristics use corresponding motion directions, frequencies, and representation scales to compare their corresponding response relationships. When using cross-impedance spectrum representation, the actual interaction characteristics correspond to the current measured cross-impedance spectrum of the target actuator, and the target interaction characteristics correspond to the interaction response requirements to be achieved in the corresponding stage.

[0113] In step S72, the amplitude and phase relationships of the actual interaction features and the target interaction features at various frequencies can be compared to determine the amplitude and phase deviations. When the target interaction feature is represented using an allowable range, the deviation is determined based on the positional relationship between the actual value and the corresponding allowable range; values ​​within the allowable range meet the corresponding interaction requirements, while values ​​outside the allowable range correspond to deviations that need to be corrected. For interaction features that also contain dissipation energy information, the difference in dissipation energy can be combined to form a comprehensive deviation used to evaluate the difference between the actual interaction performance and the target interaction requirements.

[0114] Furthermore, based on the aforementioned deviations and their sensitivity to the operating control parameters, the direction and magnitude of parameter adjustment are determined. Sensitivity reflects the direction and degree of deviation change when a certain operating control parameter changes. For example, for the same parameter change magnitude, different parameters may cause different degrees of amplitude or phase deviation changes. Based on this, the parameter change direction that can reduce the current deviation can be selected, and the corresponding adjustment amount can be determined. Adjustable parameters may include stiffness, damping, motion speed, and clamping force, and may also include the normal or tangential components of the motion trajectory. When determining the adjustment amount, the adjusted parameter values ​​are considered in conjunction with the feasible range constraints of the target actuator.

[0115] In step S73, before actually applying the parameter adjustment amount, the available energy value corresponding to this adjustment is obtained, and the expected additional mechanical energy input to the contact interaction process when using this adjustment amount is assessed. The available energy value can be determined by recording the state of the energy tank, which represents the energy budget available for parameter adjustment. The additional mechanical energy value corresponds to the active mechanical energy input caused by this parameter change, and is distinguished from the mechanical energy exchange when the target actuator continues to perform the task according to the original parameters. By comparing the two, it is determined whether this parameter adjustment is within the allowable energy budget.

[0116] In step S74, if the expected additional mechanical energy input does not exceed the available energy value, the parameter adjustment amount is applied to the current operating control parameters to obtain updated parameters. In step S75, if the expected additional mechanical energy input exceeds the available energy value, the parameter adjustment amount is reduced based on the available energy budget, and the additional mechanical energy corresponding to the reduced adjustment amount is reassessed. The adjustment amount whose expected additional mechanical energy does not exceed the available energy value is determined as the restricted parameter adjustment amount, and the operating control parameters are updated accordingly. The reduction in adjustment magnitude, combined with the mechanical energy assessment, can balance reducing interaction deviations and meeting energy constraints.

[0117] As an alternative implementation, if the expected additional mechanical energy exceeds the available energy value, the current parameter adjustment is cancelled and the current operating control parameters are retained. Then, following step S76, the target actuator continues to execute the current task phase using the updated or retained parameters, and actual interaction characteristics are continuously acquired during subsequent execution, forming an online correction process consisting of interactive feedback, deviation judgment, parameter adjustment, and energy constraints. For example, the online correction period can be selected within the range of 5 to 20 milliseconds based on the actuator's control conditions.

[0118] The aforementioned interactive response correction can be achieved through impedance control, or through admittance control, model predictive control, or hybrid force-potential control. Admittance control adjusts the motion response based on the force conditions; model predictive control determines the control action by combining predicted motion and contact responses; and hybrid force-potential control controls the contact force and motion position in the corresponding directions. When using different control methods, the adjustment amount is determined by considering the influence of the corresponding parameters on the interactive deviation, and the additional mechanical energy input caused by parameter adjustment is constrained by the available energy value.

[0119] In this way, the control parameters can be corrected based on the deviation between the interaction performance and the target requirements when the target actuator is actually performing the task, and the additional mechanical energy input can be checked before applying the parameter adjustment. This improves the adaptability of the execution process to changes in contact state and reduces the risk of contact vibration or instability caused by parameter adjustment.

[0120] As an optional approach, obtaining the available energy value for this parameter adjustment further includes: S81, taking the time of the last update of the available energy value as the start time and the current time before this parameter adjustment as the end time, determine the energy update period and obtain the available energy value updated at the start time; S82, obtain the dissipated power and additional input power during the energy update period, wherein the dissipated power characterizes the rate of mechanical energy loss during the contact interaction process between the target actuator and the operated object, and the additional input power characterizes the rate at which the target actuator inputs additional mechanical energy into the contact interaction process due to the implemented parameter adjustments; S83, the difference between the dissipated power and the additional input power is used as the rate of change of the available energy value; S84, the rate of change of the available energy value is integrated over time during the energy update period to obtain the energy change amount, and the sum of the available energy value updated at the starting time and the energy change amount is determined as the available energy value used for this parameter adjustment.

[0121] Optionally, in this embodiment, the available energy value for subsequent parameter adjustments is updated by continuously recording the mechanical energy dissipation during contact interaction and the additional mechanical energy input caused by parameter adjustments. The available energy value represents the budget for additional mechanical energy input allowed by the control process, and the energy tank records changes in this budget through state variables. An increase in the budget corresponding to mechanical energy dissipation is reflected in a change in the constraint on subsequent active energy input.

[0122] In step S81, the time when the available energy value was last updated is taken as the start time of the current energy update period, and the current time before the current parameter adjustment is taken as the end time. The available energy value already updated at the start time is read. Therefore, this update is based on the result of the last update and only accumulates the energy changes that occurred between these two times. After this update is completed, the current time and its corresponding available energy value can be used as the starting point for the next update.

[0123] In step S82, the dissipated power and additional input power during the energy update period are obtained. Dissipated power describes the rate at which mechanical energy is dissipated due to damping, friction, etc., during the contact interaction process. Additional input power describes the rate of active mechanical energy input caused by the implemented parameter adjustments. Both types of power record the rate of change of mechanical energy at corresponding moments. After accumulation over time, they correspond to the dissipated mechanical energy and additional input mechanical energy during this period, respectively. The input recorded here is the input generated by the implemented adjustments; the expected additional mechanical energy corresponding to the parameter adjustments yet to be implemented is used in subsequent energy budget judgments.

[0124] In step S83, the rate of change of the available energy value is determined using the difference between the dissipated power and the additional input power. The energy tank status can be updated according to the following relationship:

[0125] in, It represents the energy state recorded by the energy tank and is used to characterize the available energy budget; This represents the rate of change of this energy state over time; Indicates power dissipation; This represents the additional input power resulting from the implemented parameter adjustments. When the energy state is expressed in Joules, both types of power and the rate of change of energy state are expressed in watts.

[0126] The aforementioned update relationship includes dissipated mechanical energy in the budget increase and additional mechanical energy input in the budget decrease. Physically, this means that dissipation during the contact process reduces the amount of mechanical energy available for further motion exchange, thus allowing the energy budget to supplement mechanical energy under constrained conditions. When parameter adjustments have already input additional mechanical energy, the amount that can be input subsequently is reduced accordingly. Therefore, the budget increase with dissipation reflects the update of control constraints; the actual dissipated mechanical energy is still converted into other forms of energy according to the physical laws of the contact process.

[0127] In step S84, the rate of change of available energy value during the energy update period is integrated over time to obtain the total energy change for the entire period. This change is then added to the available energy value at the start time to obtain the available energy value used for this parameter adjustment. When the dissipated mechanical energy during this period is greater than the additional input mechanical energy, the energy change is positive, and the available energy value increases; when the dissipated mechanical energy is less than the additional input mechanical energy, the energy change is negative, and the available energy value decreases; when the two are equal, the available energy value remains unchanged.

[0128] As an optional implementation, a lower limit can also be set for the energy tank state, and the following relationship can be satisfied as a protection requirement for the energy state:

[0129] in, Indicates the status of the energy tank; This indicates the preset energy lower limit, which can be set to, for example, 0.02 joules.

[0130] Before applying any parameter adjustment, check whether the adjusted energy state meets the preset energy lower limit, taking into account the expected additional mechanical energy input. If the expected adjusted energy state is lower than the preset energy lower limit, reduce the parameter adjustment amount and recheck, or cancel the current parameter adjustment. When using the preset energy lower limit, the available energy value of this parameter adjustment corresponds to the allowable additional mechanical energy input after maintaining the energy state at its lower limit.

[0131] The updated available energy value is compared with the additional mechanical energy expected to be input during this parameter adjustment to determine whether to use the original parameter adjustment amount, the constrained parameter adjustment amount, or retain the current operating control parameters. In this way, the energy changes that have already occurred participate in the next adjustment decision, and subsequent parameter adjustments are constrained by the updated budget, forming an energy management process that is connected to the actual contact and interaction process.

[0132] In this way, the available energy budget can be continuously updated based on the actual dissipation in the contact interaction and the energy input of the parameters that have been adjusted. This provides energy constraints that change with the execution state for subsequent parameter adjustments, reduces the excessive input of active mechanical energy caused by continuous adjustments, and helps to maintain the stability of the contact interaction.

[0133] As an optional approach, the step of obtaining the actual interaction characteristics of the currently executed target task stage during the process of controlling the target executor to execute the target task according to the operation control parameters further includes: S91, the verification excitation signal is superimposed as an additional control quantity onto the initial control command generated according to the operation control parameters of the currently executed target task stage to obtain the adjusted target control command, wherein the verification excitation signal is used to control the target actuator to apply additional detection actions within a preset verification period. S92, during the preset verification period, the target actuator is controlled to perform the motion and contact interaction corresponding to the current target task stage according to the target control command, the displacement change amplitude of the target actuator caused by the additional detection action is limited to within the preset displacement amplitude upper limit, and force signal and velocity signal are collected, wherein the force signal represents the change of contact force received by the target actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement speed of the target actuator over time; S93, perform spectrum analysis on the force signal and the velocity signal, compare the force signal component and the velocity signal component at the same frequency, determine the amplitude relationship and phase relationship of the contact force relative to the motion velocity, and obtain the actual interaction feature, wherein the actual interaction feature and the target interaction feature are respectively represented by their amplitude relationship and phase relationship at the same frequency.

[0134] Optionally, in this embodiment, when the target actuator performs the migrated operational skill, the actual interactive performance of the current target task stage is obtained through short-term verification probing. Verification probing is arranged within the task stage where additional probing actions are permitted, and a preset verification period corresponds to a probing time segment within this stage. The selection of the verification period combines the current contact state and task motion requirements to maintain the correspondence between the acquired signals and the current target task stage.

[0135] In step S91, an initial control command is generated based on the operation control parameters of the current target task stage, and a verification excitation signal is superimposed to obtain the target control command. The initial control command is used to implement the motion and contact interaction of the corresponding stage, while the verification excitation signal is used to introduce identifiable additional changes. The verification excitation can use the signal form of the aforementioned limited detection, and the detection amplitude is determined by combining the capability of the target actuator and the contact conditions of the current stage. As an optional implementation, when using the same additional control quantity and the amplitude is represented by a comparable scale, the amplitude of the verification excitation can be set to be smaller than the excitation amplitude of the original actuator detection, so as to reduce the impact of verification detection on the execution of the target task.

[0136] In step S92, within a preset verification period, the target actuator is driven according to the target control command to perform the motion and contact interaction corresponding to the current task stage, while limiting the displacement change amplitude caused by the additional detection action. Here, the displacement change amplitude corresponds to the additional displacement relative to the motion determined by the initial control command, and the preset upper limit of the displacement amplitude limits the magnitude of this additional change. For example, if stable contact has been established in the current stage, a small detection change can be applied based on the motion maintaining the corresponding contact to obtain force and velocity feedback of the target actuator in the current contact state.

[0137] Force and velocity signals are acquired during the verification period, and the temporal correspondence between the two types of signals is retained. In step S93, the aforementioned spectrum analysis and response relationship estimation methods can be used to obtain the corresponding force and velocity signal components at the same frequency, and then the amplitude and phase relationships of the contact force relative to the motion velocity are determined. The analysis results are correlated with the current target task stage to form the actual interaction characteristics of this stage.

[0138] The actual interaction features and the target interaction features are correlated according to the same frequency position, corresponding direction of motion, and consistent representation scale. For each frequency involved in the comparison, the actual interaction features provide the currently measured amplitude and phase relationships, while the target interaction features provide the corresponding target values ​​or allowable ranges. When using normalized representation, both adopt a consistent normalization benchmark. Thus, the deviation of the actual interaction performance from the target requirements can be determined frequency by frequency, and the obtained deviations can be used to adjust the operational control parameters for the current target task phase.

[0139] In this way, actual interaction features corresponding to the current contact state can be obtained during the execution of the target task. The influence of the probe on the task motion can be verified by controlling the displacement amplitude limit, providing a test basis for online parameter correction and improving the adaptability of the transferred operation skills to actual contact conditions.

[0140] As an optional approach, after determining the deviation between the actual interaction characteristics and the target interaction characteristics of the currently executed target task phase, the method further includes: S101, during the process of the target executor executing the current target task phase, the deviation between the actual interaction feature and the target interaction feature is obtained by the target executor in multiple consecutive determinations; S102, when the deviation is determined to increase sequentially in multiple consecutive steps, the control bandwidth of the target actuator during the contact interaction process is reduced, the target actuator is subjected to impedance control using the reduced control bandwidth, and the execution flow of the target task is rolled back to the starting point of the currently executed target task stage. The impedance control is used to adjust the response relationship between the motion change of the target actuator and the change of the contact action it receives. The control bandwidth characterizes the upper limit of the frequency at which the target actuator can effectively follow the changes in the control command under the corresponding control.

[0141] Optionally, in this embodiment, after determining the deviation based on the actual interaction characteristics and the target interaction characteristics, it is further determined whether to reduce the control bandwidth and roll back the task flow based on the change of the deviation during the execution process. The above determination is performed on the current target task stage, and each deviation uses a consistent comparison object and calculation method to determine the continuous trend of the deviation.

[0142] In step S101, the deviations determined consecutively within the current target task phase are recorded in chronological order of acquisition time. Each deviation reflects the degree of deviation of the actual interaction performance at the corresponding moment from the target requirements of the current phase. In step S102, adjacent deviations are compared. When each subsequent deviation is greater than the previous one in the consecutive deviations involved in the judgment, the deviation is determined to increase sequentially. This change indicates that the actual contact interaction performance is continuously deviating from the target requirements of the current phase, which can serve as a trigger condition for switching to a conservative execution mode.

[0143] When the above conditions are met, the contact interaction control of the target actuator is switched to an impedance control configuration with a bandwidth lower than before the switch. Reducing the control bandwidth lowers the upper limit of the frequency range that can effectively follow changes in control commands, allowing the target actuator to perform motion and contact interaction within a lower effective response frequency range. Impedance control continues to adjust the response relationship between changes in motion and changes in contact action, and uses control parameters adapted to the reduced bandwidth to weaken the response to rapid changes and reduce the risk of oscillations caused by dynamic changes in contact.

[0144] Simultaneously, the execution flow of the target task is rolled back to the starting point of the current target task stage. This starting point corresponds to the task stage at the time of entering the current stage, and is associated with the corresponding entry contact event and contact conditions. After the flow rollback, the execution of the current stage is reorganized from this point, and corresponding motion and contact interactions are performed under reduced control bandwidth. Thus, adjusting the control method in conjunction with the task flow rollback provides conditions for re-establishing a contact state that meets the task requirements in the current stage.

[0145] For example, during the stable bearing phase, if the continuously acquired interaction deviations increase sequentially, impedance control with a lower bandwidth can be used, and the execution flow can be rolled back to the beginning of the stable bearing phase, reorganizing the execution based on the contact conditions of this phase. Subsequent execution will still be judged based on the completion requirements of the stable bearing phase, thus maintaining the connection between the phase rollback and the original task objectives.

[0146] As alternative triggering conditions, the aforementioned lower bandwidth impedance control and phase backoff method can also be used when resonance phenomena not included in the original model occur, or when the coherence coefficient of the force and velocity signals drops sharply. Resonance phenomena may manifest as a significant enhancement of the interaction response near a specific frequency, while a sharp drop in the coherence coefficient indicates a decrease in the correlation between the corresponding signals used for response estimation. These situations can be used to trigger conservative execution for contact interaction anomalies.

[0147] By using the above methods, when the actual interaction performance continues to deviate from the target requirements, the control bandwidth of the contact interaction can be reduced in a timely manner, and the execution process can be rolled back to the starting point of the current stage. This reduces the instability risk caused by continuing to advance the task when the deviation continues to increase, and improves the execution reliability of the transferred operational skills.

[0148] To better understand and illustrate the solutions provided by the embodiments of this disclosure and their beneficial effects, the following description, in conjunction with a specific scenario embodiment, will illustrate the solutions provided by the embodiments of this disclosure.

[0149] In this scenario, the objective is to tighten a bottle cap with a resilient sealing ring. The original actuator is a three-finger dexterous hand, and the target actuator is a parallel gripper. Both hold the bottle cap through their respective gripping structures and work together in a rotational motion to complete the thread engagement and sealing ring tightening. Skill transfer is based on the contact interaction performance and stage completion requirements in each operation step, forming corresponding operation control parameters for the parallel gripper.

[0150] Figure 4 This is a schematic diagram of an optional robot operation skill transfer and execution feedback process according to an embodiment of this disclosure. Figure 4As shown, the original actuator's task execution and restricted detection, frequency response identification, normalization and stage correspondence, target spectrum envelope generation and skill synthesis, and execution feedback correction are sequentially linked. Among them, frequency response identification is used to obtain the original interaction features of each original task stage, the target spectrum envelope is used to describe the allowable range of target interaction features, and execution feedback correction corresponds to adjusting the operation control parameters according to the deviation between the actual interaction features and the target interaction features.

[0151] When a three-finger dexterous hand performs a tightening task, its posture, velocity, clamping state, bottle cap movement, and task completion signal are recorded, along with six-dimensional force information including forces and torques in three directions. During the permissible contact phase, restricted micro-excitations are applied, and interactive characteristics for each phase are formed as described above. Upon initial contact, the interactive characteristics may manifest as an increase in low-frequency equivalent stiffness; after thread engagement, phase lag may occur during torsional motion; and when the sealing ring is tightened, mechanical energy dissipation increases. These changes are respectively associated with the original task phases, providing a basis for identifying task progress and forming transfer references.

[0152] When determining stage boundaries, candidate stage switching positions can be identified by combining spectral differences between adjacent time windows, changes in contact relationships, and short-term signal changes accompanying slippage or locking. For example, when using a corresponding spectral distance metric, a spectral distance of 0.25 or higher can be used as a condition to trigger a candidate stage switching position, further combined with visual or tactile contact events to determine the task stage. In addition to the aforementioned amplitude and phase relationships, information such as damping slope, resonant frequency, peak gain, and phase margin can also be recorded. The damping slope describes the change of relevant damping characteristics with frequency, the resonant frequency and peak gain describe the location and degree of response enhancement, and the phase margin describes the stability margin of the corresponding control loop. This information is used to further describe the dynamic interaction characteristics of each stage.

[0153] When determining the capabilities of a parallel gripper, robot structural information recorded in the Unified Robot Description Format (URDF) can be read, and combined with the gripper's rated power, effective mass, force and displacement range, control bandwidth, and available contact methods, its set of reachable response relationships can be determined. When comparing the interaction characteristics of two types of actuators, an appropriate representation scale is selected by combining rated power, characteristic length, effective mass, and sensor range, and then the target interaction characteristics are determined according to the aforementioned stage correspondence and reachability constraints.

[0154] Figure 5 This is a schematic diagram of an optional heterogeneous actuator operation skill migration process according to an embodiment of this disclosure. Figure 5As shown, the results of original skill acquisition, constrained micro-stimulus, stage interaction feature extraction, normalization, and stage correspondence are used together to synthesize the target skill. The phase isomorphism in the figure corresponds to the mapping between the original task stages and the target task stages, maintaining the stage completion requirements and the sequence of key events. For this scenario, stage correspondence is established according to physical states such as initial contact, thread engagement, and clamping completion, and the normal clamping force of the parallel gripper, wrist rotation speed, stiffness, and damping parameters are jointly determined.

[0155] During skill synthesis, accessibility can be evaluated based on the feasibility of the target interaction requirements within the set of reachable response relationships, and the evaluation result is expressed as a spectral accessibility score. When the score is below a preset threshold, depending on insufficient capabilities in areas such as the required force direction, control bandwidth, or energy capacity, appropriate treatment methods are selected, including environmental support, two-arm collaboration, alternative contact methods, or termination of migration. For cases where the target skill can be formed, its motion and contact performance is verified through simulation or low-speed execution before feedback correction during actual execution.

[0156] When the parallel grippers perform a tightening task, the abrupt change in the torsional interaction response can be used as one of the criteria for judging the completion of the sealing ring clamping, and combined with the corresponding stage completion requirements to advance the task. During execution, differences in amplitude, phase, and dissipated energy can be combined according to their respective weights to form an interaction deviation, which is used for the correction of the aforementioned parameters. The evaluation threshold for spectral matching can be set according to the task requirements; for example, the upper limit of amplitude envelope error can be set to 3 dB, and the upper limit of phase envelope error can be set to 20 degrees. These thresholds are used to evaluate the matching of the interaction response, and whether a stage is completed is judged according to the corresponding physical completion requirements.

[0157] In other alternative implementations, the interaction response can also be represented using admittance, transmissibility, wave variables, or time-frequency wavelet spectra. Admittance describes the interaction relationship from the perspective of the motion response caused by force changes; transmissibility describes the transmission relationship between selected interaction quantities; wave variables describe the energy exchange at the interaction ports through the combination of force and velocity; and time-frequency wavelet spectra describe the change in response with time and frequency. When using corresponding representations, the original interaction characteristics, target interaction characteristics, and actual interaction characteristics are represented in a way that allows for comparison, while maintaining the connection between stage correspondence, capability constraints, and the generation of operational control parameters. Online correction can also employ reinforcement learning residual control, where the residual control output provides correction to existing control actions, combined with energy tanks or power limit constraints on active energy input.

[0158] Figure 6 This is a schematic diagram of an optional skill transfer comparison verification according to an embodiment of this disclosure. Figure 6As shown, trajectory scaling, force profile transfer, and stiffness profile transfer can be set as comparison schemes to be compared and verified with the transfer method used in this scenario. During verification, the gripper mass, control bandwidth, bottle cap friction, and sealing ring hardness are changed, and the task success rate, peak torque, object damage, and spectral envelope error are statistically analyzed. The task stage recognition error rate and energy constraint trigger count can also be statistically analyzed. Through these indicators, the task completion, contact action intensity, stage recognition, and interaction response matching are evaluated respectively.

[0159] The above process can also be extended to multi-finger grippers, suction cups, magnetic tools, flexible hands, dual-arm combinations, or leg-foot contact ends, depending on the operational task. For different actuators, their contact methods, capability limitations, and task stage completion requirements are determined separately, and then applicable interaction characteristics and operational control parameters are formed.

[0160] Through the skill transfer process in this scenario, the dynamic contact characteristics of the original actuator when completing the task can be transformed into interactive requirements suitable for the target actuator. Operation control parameters can be generated by combining the actual capabilities of the target actuator and making restrictive corrections based on execution feedback. This improves the adaptability and execution stability of operation skills after being transferred between heterogeneous actuators.

[0161] This disclosure provides an apparatus for transferring robot operational skills. Figure 7 This is a schematic diagram of the structure of a robot operation skill transfer device provided in an embodiment of this disclosure, as shown below. Figure 7 As shown, the robot operation skill transfer device 70 may include: The receiving module 701 is configured to receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; The first acquisition module 702 is used to respond to the skill transfer request and acquire the original interaction features of each original task stage during the process of the original executor performing the target task through the operation skill. The original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when it interacts with the operation object during the execution of the target task. The adjustment module 703 is used to adjust the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, so as to adjust the interaction features of each of the target task stages in the direction of the original interaction features corresponding to the original task stages, thereby obtaining the target interaction features of each of the target task stages, wherein the corresponding target task stages and the original task stages have the same stage completion requirements. The generation module 704 is used to generate operation control parameters of the target executor in each of the target task stages based on the target interaction features and the stage completion requirements, so as to obtain the migrated operation skills adapted to the target executor.

[0162] In one exemplary embodiment, the first acquisition module includes: The first acquisition unit is used to acquire force signals and velocity signals during the execution of the target task by the original actuator, wherein the force signal represents the change of contact force received by the original actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement velocity of the original actuator over time. The first analysis unit is used to perform spectrum analysis on the corresponding force signal and velocity signal for each of the original task stages, and obtain force signal components and velocity signal components at multiple reference frequencies, wherein the multiple reference frequencies correspond to the variation components of different periods in the force signal and velocity signal. The comparison unit is used to compare the force signal component with the velocity signal component with the same reference frequency for each original task stage, determine the amplitude relationship and phase relationship of the force signal component relative to the velocity signal component, and combine the multiple reference frequencies with the corresponding amplitude relationship and phase relationship to form the original interaction feature of the original task stage.

[0163] In one exemplary embodiment, the first acquiring unit is further configured to: The detection excitation signal is superimposed as an additional control quantity onto the initial task command used to control the original actuator to execute the target task, thereby obtaining the adjusted target task command. The detection excitation signal is used to control the original actuator to apply additional detection actions during the execution of the target task, and the frequency of the detection excitation signal is within a preset frequency range. During the execution of the target task by the original actuator, the displacement change amplitude of the original actuator caused by the additional detection action and the contact force change amplitude between the original actuator and the operating object are respectively limited to the preset upper limit of displacement amplitude and the upper limit of force amplitude. The mechanical energy input by the original actuator to the contact interaction process between the original actuator and the operating object through the additional detection action is limited to the preset upper limit of energy, and the force signal and the velocity signal are collected.

[0164] In an exemplary embodiment, the first analysis unit is further configured to: For each of the original task stages, spectral analysis is performed on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple candidate frequencies, and the coherence coefficient of the force signal and the velocity signal at each of the candidate frequencies is calculated, wherein the coherence coefficient is used to indicate the degree of correlation between the force signal and the velocity signal at the corresponding candidate frequency. The candidate frequencies with a coherence coefficient not less than a preset threshold are determined as a plurality of reference frequencies, and the force signal components and velocity signal components at the plurality of reference frequencies are retained.

[0165] In one exemplary embodiment, the adjustment module includes: The second acquisition unit is used to acquire the confidence intervals of each of the original interaction features and the set of reachable response relationships of the target actuator. The original interaction features include the amplitude and phase relationships of the contact force received by the original actuator relative to the movement speed of the original actuator at multiple reference frequencies when the original actuator interacts with the operation object. The confidence intervals indicate the estimated uncertainty range of the amplitude and phase relationships of the original interaction features at each of the reference frequencies. The set of reachable response relationships includes multiple response relationships that the target actuator can reach. Each reachable response relationship is a response relationship that the target actuator can achieve under the capability limitation, expressed by the amplitude and phase relationships of the contact force received by the target actuator relative to the movement speed of the target actuator at multiple reference frequencies. The filtering unit is configured to, for each of the target task stages, filter from the set of achievable response relationships that the amplitude relationship values ​​at each of the reference frequencies are within the range of amplitude relationship values ​​defined by the confidence interval of the corresponding original task stage, and the phase relationship values ​​are within the range of phase relationship values ​​defined by the confidence interval, to obtain at least one target response relationship; The first determining unit is configured to, for each of the target task stages, at each of the reference frequencies, determine the allowable range of the amplitude relationship and the phase relationship of the target response relationship at the reference frequencies, with the minimum and maximum values ​​of the amplitude relationship and the phase relationship of each target response relationship as the lower limit and the upper limit, respectively; An adjustment unit is configured to, for each of the target task stages, under the constraints of the corresponding allowable range and the set of achievable response relationships, adjust the amplitude relationship and the phase relationship in the interaction feature to be adjusted with the goal of reducing the difference between the interaction feature to be adjusted and the corresponding original interaction feature in each of the target task stages, thereby obtaining the target interaction feature of the target task stage.

[0166] In one exemplary embodiment, the apparatus further includes: The second acquisition module is used to acquire, before adjusting the interaction features to be adjusted in each of the original task stages and each of the target task stages to the direction of the original interaction features corresponding to the original task stages when the target task is executed for the target actuator, under the limitation of the target actuator's ability in terms of motion and contact interaction, to obtain the target interaction features of each of the target task stages, and before obtaining the target interaction features of each of the original task stages, each of the target task stages acquires the entry contact event, contact mode and energy flow direction of each of the original task stages and each of the target task stages. The entry contact event is a contact event used to identify the start of the corresponding task stage. The contact mode refers to the contact interaction mode between the original actuator or the target actuator and the operation object in the corresponding task stage. The energy flow direction indicates the direction of mechanical energy transfer between the original actuator or the target actuator and the operation object in the corresponding task stage. A module is established to establish a stage correspondence between the original task stage and the target task stage, where the stage completion requirements are the same, the energy flow direction is the same, and the original actuator and the target actuator are respectively allowed to achieve the stage completion requirements through their respective contact methods. The existing sequential relationship between the entry contact events of the original task stage remains unchanged between the corresponding target task stages.

[0167] In one exemplary embodiment, the apparatus further includes: The third acquisition module is used to, after generating operation control parameters for the target actuator in each target task stage based on the target interaction characteristics and stage completion requirements of each target task stage, and obtaining the transferred operation skills adapted to the target actuator, acquire the actual interaction characteristics of the currently executed target task stage during the process of controlling the target actuator to execute the target task according to the operation control parameters, wherein the actual interaction characteristics characterize the response relationship between the motion changes of the target actuator when it interacts with the operation object and the changes in the contact action it receives; A determining module is used to determine the deviation between the actual interaction features and the target interaction features of the currently executed target task stage, and to determine the parameter adjustment amount for reducing the deviation based on the deviation and the sensitivity of the deviation to the operation control parameters, wherein the sensitivity characterizes the degree of change of the deviation caused by the change of the operation control parameters; The fourth acquisition module is used to acquire the available energy value for this parameter adjustment, and to determine the additional mechanical energy that the target actuator is expected to input into its contact interaction process with the operation object when updating the operation control parameters according to the parameter adjustment amount, and to obtain the additional mechanical energy value, wherein the available energy value represents the upper limit of the additional mechanical energy allowed for this parameter adjustment; The update module is used to update the operation control parameters of the currently executed target task phase according to the parameter adjustment amount, provided that the additional mechanical energy value does not exceed the available energy value. The reduction module is used to reduce the magnitude of the parameter adjustment amount when the additional mechanical energy value exceeds the available energy value, to obtain a limited parameter adjustment amount, and update the operation control parameters of the currently executed target task stage according to the limited parameter adjustment amount, or cancel the current parameter adjustment and retain the operation control parameters, wherein when updating the operation control parameters according to the limited parameter adjustment amount, the additional mechanical energy expected to be input by the target actuator to its contact interaction process with the operation object does not exceed the available energy value; The control module is used to control the target actuator to continue executing the currently executing target task stage according to the updated or retained operation control parameters.

[0168] In one exemplary embodiment, the fourth acquisition module includes: The second determining unit is used to determine the energy update period by taking the time of the last update of the available energy value as the start time and the current time before the current parameter adjustment as the end time, and to obtain the available energy value updated at the start time. The third acquisition unit is used to acquire the dissipated power and additional input power during the energy update period, wherein the dissipated power characterizes the rate of mechanical energy loss during the contact interaction process between the target actuator and the operated object, and the additional input power characterizes the rate at which the target actuator inputs additional mechanical energy into the contact interaction process due to the implemented parameter adjustments; The subtraction unit is used to take the difference between the dissipated power and the additional input power as the rate of change of the available energy value; An integration unit is used to integrate the rate of change of the available energy value over time during the energy update period to obtain the energy change amount. The sum of the available energy value updated at the starting time and the energy change amount is determined as the available energy value used for this parameter adjustment.

[0169] In one exemplary embodiment, the third acquisition module includes: The superposition unit is used to superimpose the verification excitation signal as an additional control quantity onto the initial control command generated according to the operation control parameters of the currently executed target task stage to obtain the adjusted target control command, wherein the verification excitation signal is used to control the target actuator to apply additional detection actions within a preset verification period. The acquisition unit is used to control the target actuator to perform the motion and contact interaction corresponding to the current target task stage according to the target control command during the preset verification period, limit the displacement change amplitude of the target actuator caused by the additional detection action to within a preset displacement amplitude upper limit, and acquire force signals and velocity signals, wherein the force signal represents the change of contact force received by the target actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement speed of the target actuator over time; The second analysis unit is used to perform spectral analysis on the force signal and the velocity signal, compare the force signal component and the velocity signal component at the same frequency, determine the amplitude relationship and phase relationship of the contact force relative to the motion velocity, and obtain the actual interaction characteristics. The actual interaction characteristics and the target interaction characteristics are respectively represented by their amplitude relationship and phase relationship at the same frequency.

[0170] In one exemplary embodiment, the apparatus further includes: The fifth acquisition module is used to acquire, after determining the deviation between the actual interaction feature and the target interaction feature in the currently executed target task stage, the deviation between the actual interaction feature and the target interaction feature determined in multiple consecutive steps during the execution of the current target task stage by the target executor. The reduction module is used to reduce the control bandwidth of the target actuator during the contact interaction process when the deviation is determined to increase sequentially in multiple consecutive iterations. The reduced control bandwidth is used to perform impedance control on the target actuator, and the execution flow of the target task is rolled back to the starting point of the currently executed target task stage. The impedance control is used to adjust the response relationship between the motion change of the target actuator and the change in the contact action it receives. The control bandwidth characterizes the upper limit of the frequency at which the target actuator can effectively follow the changes in control commands under corresponding control.

[0171] The apparatus of this disclosure embodiment can execute the method provided in this disclosure embodiment, and its implementation principle is similar, and it has corresponding technical effects. The actions performed by each module in the apparatus of each embodiment of this disclosure correspond to the steps in the method of each embodiment of this disclosure. For a detailed functional description of each module of the apparatus, please refer to the description in the corresponding method shown above, and it will not be repeated here.

[0172] This disclosure provides an electronic device (computer apparatus / device / system) including 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 method provided in any optional embodiment of this disclosure.

[0173] In one alternative embodiment, an electronic device is provided. Figure 8 The following is a schematic diagram of the structure of an optional electronic device provided in an embodiment of this disclosure, such as... Figure 8 As shown, Figure 8 The illustrated electronic device 4000 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of this disclosure.

[0174] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0175] Bus 4002 may include a pathway for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 8 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0176] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium capable of carrying or storing computer programs and capable of being read by a computer, without limitation herein.

[0177] The memory 4003 is used to store computer programs that execute embodiments of the present disclosure, and is controlled by the processor 4001 to execute them. The processor 4001 is used to execute the computer programs stored in the memory 4003 to implement the steps shown in the foregoing method embodiments.

[0178] This disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0179] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, can implement the steps and corresponding content of the aforementioned method embodiments.

[0180] It should be understood that although arrows indicate various operation steps in the flowcharts of the embodiments of this disclosure, the order in which these steps are implemented is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of the embodiments of this disclosure, the implementation steps in each flowchart can be executed in other orders as required. Furthermore, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages based on the actual implementation scenario. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage can also be executed at different times. In scenarios where execution times differ, the execution order of these sub-steps or stages can be flexibly configured as required, and the embodiments of this disclosure do not limit this.

[0181] The above description is only an optional implementation method for some implementation scenarios of this disclosure. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this disclosure without departing from the technical concept of this disclosure also fall within the protection scope of the embodiments of this disclosure.

Claims

1. A method for transferring robot operation skills, characterized in that, include: Receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; In response to the skill transfer request, the original interaction features of each original task stage during the execution of the target task by the original executor through the operation skill are obtained, wherein the original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when interacting with the operation object during the execution of the target task. For each target task stage corresponding to each original task stage when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, the interaction features to be adjusted in each target task stage are adjusted in the direction of the original interaction features of the corresponding original task stage to obtain the target interaction features of each target task stage, wherein the corresponding target task stage and the original task stage have the same stage completion requirements. Based on the target interaction characteristics and the stage completion requirements of each target task stage, the operation control parameters of the target executor in each target task stage are generated to obtain the migrated operation skills adapted to the target executor.

2. The method for transferring robot operation skills according to claim 1, characterized in that, The acquisition of the original interaction features of each original task stage during the process of the original executor performing the target task through the operational skill includes: The force signal and velocity signal are acquired during the execution of the target task by the original actuator, wherein the force signal represents the change of the contact force received by the original actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement velocity of the original actuator over time. For each of the original task stages, spectral analysis is performed on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple reference frequencies, wherein the multiple reference frequencies correspond to the variation components of different periods in the force signal and velocity signal. For each of the original task stages, the force signal component and the velocity signal component with the same reference frequency are compared to determine the amplitude and phase relationship of the force signal component relative to the velocity signal component. The multiple reference frequencies are combined with the corresponding amplitude and phase relationships to form the original interaction features of the original task stage.

3. The method for transferring robot operation skills according to claim 2, characterized in that, The acquisition of force and velocity signals during the execution of the target task by the original actuator includes: The detection excitation signal is superimposed as an additional control quantity onto the initial task command used to control the original actuator to execute the target task, thereby obtaining the adjusted target task command. The detection excitation signal is used to control the original actuator to apply additional detection actions during the execution of the target task, and the frequency of the detection excitation signal is within a preset frequency range. During the execution of the target task by the original actuator, the displacement change amplitude of the original actuator caused by the additional detection action and the contact force change amplitude between the original actuator and the operating object are respectively limited to the preset upper limit of displacement amplitude and the upper limit of force amplitude. The mechanical energy input by the original actuator to the contact interaction process between the original actuator and the operating object through the additional detection action is limited to the preset upper limit of energy, and the force signal and the velocity signal are collected.

4. The method for transferring robot operation skills according to claim 2, characterized in that, For each of the original task stages, spectral analysis is performed on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple reference frequencies, including: For each of the original task stages, spectral analysis is performed on the corresponding force signal and velocity signal to obtain force signal components and velocity signal components at multiple candidate frequencies, and the coherence coefficient of the force signal and the velocity signal at each of the candidate frequencies is calculated, wherein the coherence coefficient is used to indicate the degree of correlation between the force signal and the velocity signal at the corresponding candidate frequency. The candidate frequencies with a coherence coefficient not less than a preset threshold are determined as a plurality of reference frequencies, and the force signal components and velocity signal components at the plurality of reference frequencies are retained.

5. The method for transferring robot operation skills according to claim 1, characterized in that, When the target actuator performs the target task, the target task stages corresponding to each of the original task stages are adjusted in the direction of the original interaction features corresponding to the original task stages, under the limitations of the target actuator's ability in terms of motion and contact interaction, to obtain the target interaction features of each target task stage, including: The confidence intervals of each of the original interaction features and the set of reachable response relationships of the target actuator are obtained. The original interaction features include the amplitude and phase relationships of the contact force experienced by the original actuator relative to the motion speed of the original actuator at multiple reference frequencies when the original actuator interacts with the operation object. The confidence intervals indicate the estimated uncertainty range of the amplitude and phase relationships of the original interaction features at each of the reference frequencies. The set of reachable response relationships includes multiple response relationships that the target actuator can reach. Each reachable response relationship is a response relationship that the target actuator can achieve under the capability limit, expressed by the amplitude and phase relationships of the contact force experienced by the target actuator relative to the motion speed of the target actuator at multiple reference frequencies. For each target task stage, from the set of achievable response relationships, at least one target response relationship is obtained by selecting response relationships whose amplitude relationship values ​​at each reference frequency are within the range of amplitude relationship values ​​defined by the confidence interval of the corresponding original task stage, and whose phase relationship values ​​are within the range of phase relationship values ​​defined by the confidence interval. For each target task stage, at each reference frequency, the minimum and maximum values ​​of the amplitude relationship and the phase relationship of each target response relationship are used as the lower and upper limits, respectively, to determine the allowable range of the amplitude relationship and the phase relationship of the interaction feature to be adjusted in the target task stage at the reference frequency; For each target task stage, under the constraints of the corresponding allowable range and the set of reachable response relationships, with the goal of reducing the difference between the interaction features to be adjusted and the corresponding original interaction features in each target task stage, the amplitude relationship and the phase relationship in the interaction features to be adjusted are adjusted to obtain the target interaction features of the target task stage.

6. The method for transferring robot operation skills according to claim 1, characterized in that, Before adjusting the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages when the target actuator performs the target task for the target actuator, under the limitations of the target actuator's ability in terms of motion and contact interaction, in order to obtain the target interaction features of each target task stage, the method further includes: The entry contact event, contact mode, and energy flow direction of each of the original task stages and the target task stages are obtained. The entry contact event is a contact event used to identify the start of the corresponding task stage. The contact mode refers to the contact interaction mode between the original actuator or the target actuator and the operation object in the corresponding task stage. The energy flow direction indicates the direction of mechanical energy transfer between the original actuator or the target actuator and the operation object in the corresponding task stage. A phase correspondence is established between the original task phase and the target task phase, where the phase completion requirements are the same, the energy flow direction is the same, and the original actuator and the target actuator are respectively allowed to achieve the phase completion requirements through their respective contact methods. The existing sequential relationship between the entry contact events of the original task phase remains unchanged between the corresponding target task phases.

7. The method for transferring robot operation skills according to claim 1, characterized in that, After generating the operation control parameters of the target actuator in each of the target task stages based on the target interaction characteristics and stage completion requirements, and obtaining the transferred operation skills adapted to the target actuator, the method further includes: During the process of controlling the target actuator to execute the target task according to the operation control parameters, the actual interaction characteristics of the currently executed target task stage are obtained, wherein the actual interaction characteristics characterize the response relationship between the motion changes of the target actuator when it comes into contact with the operation object and the changes in the contact action it receives; Determine the deviation between the actual interaction feature and the target interaction feature of the currently executed target task stage, and determine the parameter adjustment amount to reduce the deviation based on the deviation and the sensitivity of the deviation to the operation control parameters, wherein the sensitivity characterizes the degree of change of the deviation caused by the change of the operation control parameters; The available energy value for this parameter adjustment is obtained, and the additional mechanical energy expected to be input by the target actuator to its contact interaction process with the operated object when the operation control parameters are updated according to the parameter adjustment amount is determined, and the additional mechanical energy value is obtained, wherein the available energy value represents the upper limit of the additional mechanical energy allowed for this parameter adjustment; If the additional mechanical energy value does not exceed the available energy value, the operation control parameters of the currently executed target task phase are updated according to the parameter adjustment amount; If the additional mechanical energy value exceeds the available energy value, the range of the parameter adjustment amount is reduced to obtain a limited parameter adjustment amount, and the operation control parameters of the currently executed target task stage are updated according to the limited parameter adjustment amount, or the current parameter adjustment is canceled and the operation control parameters are retained. When the operation control parameters are updated according to the limited parameter adjustment amount, the additional mechanical energy expected to be input by the target actuator to its contact interaction process with the operation object does not exceed the available energy value. According to the updated or retained operation control parameters, the target executor is controlled to continue executing the currently executing target task phase.

8. The method for transferring robot operation skills according to claim 7, characterized in that, The process of obtaining the available energy value for this parameter adjustment includes: The energy update period is determined by taking the last update of the available energy value as the start time and the current time before this parameter adjustment as the end time, and the available energy value updated at the start time is obtained. The dissipated power and additional input power during the energy update period are obtained, wherein the dissipated power characterizes the rate of mechanical energy loss during the contact interaction process between the target actuator and the operated object, and the additional input power characterizes the rate at which the target actuator inputs additional mechanical energy into the contact interaction process due to the implemented parameter adjustments; The difference between the dissipated power and the additional input power is used as the rate of change of the available energy value; The rate of change of the available energy value is integrated over time during the energy update period to obtain the energy change amount. The sum of the available energy value updated at the starting time and the energy change amount is determined as the available energy value used for this parameter adjustment.

9. The method for transferring robot operation skills according to claim 7, characterized in that, The step of obtaining the actual interaction characteristics of the currently executed target task stage during the process of controlling the target executor to execute the target task according to the operation control parameters includes: The verification stimulus signal is superimposed as an additional control quantity onto the initial control command generated based on the operation control parameters of the currently executed target task stage to obtain the adjusted target control command, wherein the verification stimulus signal is used to control the target actuator to apply additional detection actions within a preset verification period. During the preset verification period, the target actuator is controlled to perform the motion and contact interaction corresponding to the current target task stage according to the target control command. The displacement change amplitude of the target actuator caused by the additional detection action is limited to within the preset upper limit of displacement amplitude, and force signal and velocity signal are collected. The force signal represents the change of contact force received by the target actuator when it interacts with the operation object over time, and the velocity signal represents the change of the movement speed of the target actuator over time. Spectral analysis is performed on the force signal and the velocity signal. The force signal component and the velocity signal component at the same frequency are compared to determine the amplitude and phase relationship of the contact force relative to the motion velocity, thereby obtaining the actual interaction characteristics. The actual interaction characteristics and the target interaction characteristics are respectively represented by their amplitude and phase relationships at the same frequency.

10. The method for transferring robot operation skills according to claim 7, characterized in that, After determining the deviation between the actual interaction features and the target interaction features of the currently executed target task phase, the method further includes: During the execution of the current target task phase by the target executor, the deviation between the actual interaction feature and the target interaction feature is obtained multiple times in succession. When the deviation is determined to increase sequentially in multiple consecutive iterations, the control bandwidth of the target actuator during the contact interaction process is reduced. The reduced control bandwidth is used to perform impedance control on the target actuator, and the execution flow of the target task is rolled back to the starting point of the currently executed target task stage. The impedance control is used to adjust the response relationship between the motion change of the target actuator and the change in the contact action it receives. The control bandwidth characterizes the upper limit of the frequency at which the target actuator can effectively follow the changes in the control command under the corresponding control.

11. A device for transferring robot operational skills, characterized in that, include: A receiving module is configured to receive a skill transfer request for a target actuator, wherein the target actuator is heterogeneous with the robot's original actuator, and the skill transfer request instructs the transfer of the operating skills of the original actuator to the target actuator; The first acquisition module is used to respond to the skill transfer request and acquire the original interaction features of each original task stage during the process of the original executor performing the target task through the operation skill. The original interaction features characterize the response relationship between the motion changes and the changes in contact action received by the original executor when it interacts with the operation object during the execution of the target task. An adjustment module is used to adjust the interaction features to be adjusted in each of the target task stages corresponding to each of the original task stages when the target actuator performs the target task, under the limitation of the target actuator's ability in terms of motion and contact interaction, so as to adjust the interaction features of each of the target task stages in the direction of the original interaction features of the corresponding original task stages, thereby obtaining the target interaction features of each of the target task stages, wherein the corresponding target task stages and the original task stages have the same stage completion requirements. The generation module is used to generate operation control parameters for the target executor in each of the target task stages based on the target interaction features and the stage completion requirements, so as to obtain the migrated operation skills adapted to the target executor.

12. An electronic 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 method of any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 10.

14. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 10.