Power maintenance method and device based on hybrid dynamic motion primitives

Abstract

The application discloses a power maintenance method and device based on a hybrid dynamic motion primitive. The method comprises the following steps: obtaining a plurality of demonstration trajectories of an equipotential robot; modeling a dynamic motion primitive for each demonstration trajectory to obtain a corresponding initial hybrid dynamic primitive model; evaluating the fluency of the demonstration trajectory according to an evaluation index to determine the model weight of each initial hybrid dynamic primitive model; determining a hybrid dynamic primitive model according to the initial hybrid dynamic primitive model and the model weight; and writing the hybrid dynamic primitive model into the equipotential robot, so that the equipotential robot generates a maintenance trajectory when receiving a power maintenance task; and controlling the equipotential robot to perform a power maintenance operation based on the maintenance trajectory. The application solves the technical problem that the equipotential robot can only learn a single trajectory in the related art, and that the generated trajectory has too many inflection points in the process of power maintenance, which causes the swing of a boom truck, low maintenance efficiency and poor work quality.

Description

Technical Field

[0001] This invention relates to the field of power maintenance technology, and more specifically, to a power maintenance method and apparatus based on hybrid dynamic motion elements. Background Technology

[0002] In existing technologies, equipotential work robots for power distribution networks are increasingly being used for power maintenance. Traditional trajectory programming methods require pre-setting a series of trajectory points and then controlling the robot to pass through these points in sequence to achieve trajectory planning. This method has two problems in the actual operation of equipotential robots: first, due to the long distance at high altitudes, it is impossible to teach accurately, resulting in low efficiency; second, the trajectory generated by this method has too many inflection points, which will cause the boom truck to shake violently.

[0003] Existing robot trajectory learning methods based on probabilistic motion primitive models can only learn a single trajectory, exhibiting randomness and making them unsuitable for the complex and ever-changing environment of the power industry. Furthermore, the number of Gaussian functions is difficult to determine; too many can lead to overfitting, hindering later trajectory generalization, while too few will result in an inability to properly fit the demonstration trajectory.

[0004] Regarding the aforementioned issues, where the equivalent potential robot can only learn a single trajectory, and the excessive number of inflection points in the trajectory generated during power maintenance causes the boom truck to shake violently, resulting in low maintenance efficiency and poor work quality, no effective solution has yet been proposed. Summary of the Invention

[0005] This invention provides a power maintenance method and apparatus based on hybrid dynamic motion primitives, which at least solves the technical problems in related technologies where isoelectric robots can only learn a single trajectory, and the generated trajectory has too many inflection points during power maintenance, resulting in severe shaking of the boom truck, low maintenance efficiency and poor work quality.

[0006] According to one aspect of the present invention, a power maintenance method based on hybrid dynamic motion primitives is provided, comprising: acquiring multiple demonstration trajectories of an equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance; performing dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory; evaluating the smoothness of each demonstration trajectory according to an evaluation index to obtain an evaluation score; and determining the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score, wherein... The model weights refer to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories based on power maintenance tasks. The hybrid dynamic primitive model is determined based on the initial hybrid dynamic primitive model and the model weights corresponding to each demonstration trajectory. The hybrid dynamic primitive model is written into the equipotential robot so that when the equipotential robot receives the power maintenance task, it uses the hybrid dynamic primitive model to generate the maintenance trajectory based on the maintenance location indicated by the power maintenance task. The equipotential robot is then controlled to perform power maintenance operations based on the maintenance trajectory.

[0007] Optionally, acquiring multiple demonstration trajectories of the equipotential robot includes: acquiring multiple historical maintenance trajectories of the equipotential robot when performing historical maintenance operations within a historical time period, and determining the historical maintenance trajectories as the demonstration trajectories; and / or, acquiring multiple demonstration trajectories by controlling the equipotential robot to simulate the power maintenance process.

[0008] Optionally, dynamic motion primitive modeling is performed on each of the demonstration trajectories to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. This includes: performing dynamic motion primitive modeling on each of the demonstration trajectories based on a proportional-differential controller and a trajectory shape learner to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory, wherein the initial hybrid dynamic primitive model N corresponding to the k-th demonstration trajectory... k The first expression is: τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative of y with respect to time t, and let α represent acceleration. y β represents the proportional parameter in the proportional-derivative controller. yLet f represent the differential parameter in the proportional-differential controller, f represent the nonlinear term in the trajectory shape learner, and g represent the target state of the proportional-differential controller, where the target state refers to the state that the proportional-differential controller needs to achieve. The nonlinear term in the trajectory shape learner is adjusted to obtain the adjusted nonlinear term. Based on the obtained initial hybrid dynamic primitive model, the convergence speed of the demonstration trajectory is adjusted based on the nonlinear term to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory, where the convergence speed refers to the speed at which the demonstration trajectory moves from the current state to the target state. The initial hybrid dynamic primitive model M corresponding to the k-th demonstration trajectory... k The second expression is:

[0009] Optionally, adjusting the nonlinear term in the trajectory shape learner to obtain the adjusted nonlinear term includes: determining the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let σ represent the i-th radial basis function. a c represents the width of the a-th radial basis function. a Let represent the center position of the a-th radial basis function, and x represent the coordinates of a point on the demonstration trajectory. The radial basis functions are used to obtain the nonlinear term through linear superposition. Based on the linear superposition of multiple radial basis functions, the adjusted nonlinear term is determined using a second formula, wherein the second formula is: ω a Let γ represent the weight value of the a-th radial basis function, s represent the phase variable of the first-order system, and γ represent the weight value of the a-th radial basis function. a Let represent the random forgetting factor of the a-th radial basis function.

[0010] Optionally, the evaluation metrics include: a first evaluation metric, a second evaluation metric, and a third evaluation metric. The smoothness of each demonstration trajectory is evaluated based on these metrics to obtain an evaluation score, including at least one of the following: evaluating the smoothness of the demonstration trajectory using a third formula based on the first evaluation metric to obtain a first evaluation score, wherein the first evaluation metric is the first duration of zero speed in each demonstration trajectory, and the third formula is: i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,kLet A(i) represent the velocity at the i-th trajectory point on the k-th demonstration trajectory, and let A(i) represent whether the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero. A(i) = 1 indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is not zero. The smoothness of the demonstration trajectory is evaluated using the fourth formula based on the second evaluation index to obtain a second evaluation score. The second evaluation index is the number of acceleration sign changes in each demonstration trajectory, and the fourth formula is: a represents the second evaluation score of the k-th demonstration trajectory. i,k a represents the acceleration at the i-th point on the k-th demonstration trajectory. i+1,k The acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory is represented by B(i). B(i) indicates whether the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. The smoothness of the demonstration trajectory is evaluated using the fifth formula based on the third evaluation index to obtain a third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds the acceleration threshold in each demonstration trajectory. The absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: The third evaluation score represents the k-th demonstration trajectory, |a i,k The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0011] Optionally, determining the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score includes: fusing the first evaluation score, the second evaluation score, and the third evaluation score using a sixth formula to obtain the evaluation score, wherein the sixth formula is: L kLet n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score; based on the evaluation score, the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory is determined using the seventh formula, wherein the seventh formula is: K represents the total number of the demonstrated trajectories, δ k The model weight represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0012] Optionally, determining the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and the model weights corresponding to each demonstration trajectory includes: determining the hybrid dynamic primitive model using an eighth formula based on the initial hybrid dynamic primitive model and the model weights corresponding to each demonstration trajectory, wherein the eighth formula is: M DMP M represents the hybrid dynamic primitive model. k This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0013] According to another aspect of the present invention, a power maintenance device based on hybrid dynamic motion primitives is also provided, comprising: a first acquisition unit, configured to acquire multiple demonstration trajectories of an equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance; a second acquisition unit, configured to perform dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory; a third acquisition unit, configured to evaluate the smoothness of each demonstration trajectory according to an evaluation index to obtain an evaluation score; and a first determination unit, configured to determine the model of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score. The system comprises: a weighting unit, wherein the model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model, and the hybrid dynamic primitive model is used to generate a maintenance trajectory based on the power maintenance task; a second determining unit, used to determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model corresponding to each demonstration trajectory and the model weight; a generating unit, used to write the hybrid dynamic primitive model into the equipotential robot, so that the equipotential robot can generate the maintenance trajectory based on the maintenance location indicated by the power maintenance task when it receives the power maintenance task; and a control unit, used to control the equipotential robot to perform power maintenance operations based on the maintenance trajectory.

[0014] Optionally, the first acquisition unit includes: a first acquisition module, configured to acquire multiple historical maintenance trajectories of the equipotential robot when performing historical maintenance operations within a historical time period, and determine the historical maintenance trajectory as the demonstration trajectory; and / or, a second acquisition module, configured to acquire multiple demonstration trajectories by controlling the equipotential robot to simulate the power maintenance process.

[0015] Optionally, the second acquisition unit includes: a third acquisition module, configured to perform dynamic motion primitive modeling on each of the demonstration trajectories based on a proportional-differential controller and a trajectory shape learner, to obtain an initial hybrid dynamic primitive model corresponding to each of the demonstration trajectories, wherein the initial hybrid dynamic primitive model N corresponding to the k-th demonstration trajectory... k The first expression is: τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative of y with respect to time t, and let α represent acceleration. y β represents the proportional parameter in the proportional-derivative controller. y The variable term represents the differential parameter in the proportional-differential controller, f represents the nonlinear term in the trajectory shape learner, and g represents the target state of the proportional-differential controller, where the target state refers to the state that the proportional-differential controller needs to achieve. The fourth acquisition module is used to adjust the nonlinear term in the trajectory shape learner to obtain the adjusted nonlinear term. The fifth acquisition module is used to adjust the convergence speed of the demonstration trajectory based on the nonlinear term, based on the obtained initial hybrid dynamic primitive model, to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory, where the convergence speed refers to the speed at which the demonstration trajectory moves from the current state to the target state, and the initial hybrid dynamic primitive model M corresponding to the k-th demonstration trajectory... k The second expression is:

[0016] Optionally, the fourth acquisition module includes: a first determination submodule, used to determine the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let σ represent the i-th radial basis function. a c represents the width of the a-th radial basis function. aLet represent the center position of the a-th radial basis function, and x represent the coordinates of a trajectory point in the demonstration trajectory. The radial basis functions are used to obtain the nonlinear term through linear superposition. The second determining submodule is used to determine the adjusted nonlinear term based on the linear superposition of multiple radial basis functions using a second formula, wherein the second formula is: ω a Let γ represent the weight value of the a-th radial basis function, s represent the phase variable of the first-order system, and γ represent the weight value of the a-th radial basis function. a Let represent the random forgetting factor of the a-th radial basis function.

[0017] Optionally, the evaluation indicators include: a first evaluation indicator, a second evaluation indicator, and a third evaluation indicator. The third acquisition unit includes at least one of the following: a sixth acquisition module, used to evaluate the smoothness of the demonstration trajectory based on the first evaluation indicator using a third formula to obtain a first evaluation score, wherein the first evaluation indicator is the first duration of zero speed in each demonstration trajectory, and the third formula is: i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,k Let A(i) represent the velocity at the i-th trajectory point on the k-th demonstration trajectory, and let A(i) represent whether the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero. A(i) = 1 indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is not zero. The seventh acquisition module is used to evaluate the smoothness of the demonstration trajectory according to the second evaluation index using the fourth formula to obtain a second evaluation score, wherein the second evaluation index is the number of acceleration sign changes in each demonstration trajectory, and the fourth formula is: a represents the second evaluation score of the k-th demonstration trajectory. i,k a represents the acceleration at the i-th point on the k-th demonstration trajectory. i+1,kThe acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory is represented by B(i). B(i) indicates whether the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. The eighth acquisition module is used to evaluate the smoothness of the demonstration trajectory according to the third evaluation index using the fifth formula to obtain a third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds the acceleration threshold in each demonstration trajectory. The absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: The third evaluation score represents the k-th demonstration trajectory, |a i,k The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0018] Optionally, the first determining unit includes: a ninth obtaining module, configured to fuse the first evaluation score, the second evaluation score, and the third evaluation score using a sixth formula to obtain the evaluation score, wherein the sixth formula is: L k Let n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score; the first determining module is used to determine the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score using a seventh formula, wherein the seventh formula is: K represents the total number of the demonstrated trajectories, δ k The model weight represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0019] Optionally, the second determining unit includes: a second determining module, configured to determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model corresponding to each demonstration trajectory and the model weights using an eighth formula, wherein the eighth formula is: M DMP M represents the hybrid dynamic primitive model. k This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0020] According to another aspect of the present invention, a power maintenance system based on hybrid dynamic motion primitives is also provided, wherein the power maintenance system based on hybrid dynamic motion primitives uses any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0021] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein the program executes any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0022] According to another aspect of the present invention, a processor is also provided, the processor being configured to run a program, wherein the program, when running, executes any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0023] According to another aspect of the present invention, a computer program product is also provided, including computer instructions, which, when executed by a processor, perform any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0024] In this embodiment of the invention, multiple demonstration trajectories of an equipotential robot are acquired. The equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance. Dynamic motion primitive modeling is performed on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. The smoothness of each demonstration trajectory is evaluated according to evaluation indicators to obtain an evaluation score. The model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory is determined based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories according to the power maintenance task. The hybrid dynamic primitive model is determined based on the initial hybrid dynamic primitive model and model weight corresponding to each demonstration trajectory. The hybrid dynamic primitive model is written into the equipotential robot so that when the equipotential robot receives a power maintenance task, it uses the hybrid dynamic primitive model to generate a maintenance trajectory according to the maintenance position indicated by the power maintenance task. The equipotential robot is controlled to perform power maintenance operations based on the maintenance trajectory. The above technical solution achieves the goal of obtaining model weights by performing dynamic motion primitive modeling on multiple demonstration trajectories and evaluating the smoothness of each demonstration trajectory, fusing the models corresponding to multiple demonstration trajectories to obtain a hybrid dynamic motion primitive model, and generating maintenance trajectories based on this model according to the maintenance task, so as to control the robot to perform maintenance according to the trajectory. This realizes the technical effect of using the hybrid dynamic primitive model to generate maintenance trajectories for power maintenance by integrating multiple data and factors, thereby improving the smoothness of the trajectory. This improves the maintenance efficiency and quality of equipotential robots in performing power maintenance tasks, and solves the technical problems in related technologies where equipotential robots can only learn a single trajectory, and the generated trajectory has too many inflection points during power maintenance, resulting in severe shaking of the bucket truck, low maintenance efficiency and poor work quality. Attached Figure Description

[0025] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0026] Figure 1 This is a hardware structure block diagram of a mobile terminal for a power maintenance method based on hybrid dynamic motion primitives according to an embodiment of the present invention.

[0027] Figure 2 This is a flowchart of a power maintenance method based on hybrid dynamic motion elements according to an embodiment of the present invention;

[0028] Figure 3 This is a flowchart of a power maintenance method based on hybrid dynamic motion elements according to an embodiment of the present invention;

[0029] Figure 4 This is a schematic diagram of a demonstration trajectory according to an embodiment of the present invention;

[0030] Figure 5 This is a schematic diagram illustrating the effect of the nonlinear term according to an embodiment of the present invention;

[0031] Figure 6 This is a schematic diagram of the visualization of radial basis functions according to an embodiment of the present invention;

[0032] Figure 7 This is a schematic diagram of a power maintenance device based on hybrid dynamic motion elements according to an embodiment of the present invention. Detailed Implementation

[0033] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0034] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] As described in the background section, in related technologies, equal-potential robots can only learn a single trajectory. Furthermore, during power maintenance, the generated trajectory has too many inflection points, causing severe shaking of the boom truck, resulting in low maintenance efficiency and poor work quality. To address these shortcomings, this invention provides a power maintenance method and apparatus based on hybrid dynamic motion primitives.

[0036] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0037] The methods and embodiments provided in this invention can be executed on a mobile terminal, a computer terminal, or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a power maintenance method based on hybrid dynamic motion primitives, according to an embodiment of the present invention. Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0038] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the power maintenance method based on hybrid dynamic motion elements in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0039] According to an embodiment of the present invention, a method embodiment of a power maintenance method based on hybrid dynamic motion primitives is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0040] Figure 2 This is a flowchart of a power maintenance method based on hybrid dynamic motion elements according to an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes the following steps:

[0041] Step S202: Obtain multiple demonstration trajectories of the equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the movement trajectory of the equipotential robot during power maintenance.

[0042] The following is combined Figure 3 The embodiments of the present invention will be described in detail below. Figure 3 This is a flowchart of a power maintenance method based on hybrid dynamic motion elements according to an embodiment of the present invention.

[0043] like Figure 3 As shown, in this embodiment, multiple demonstration trajectories of the equipotential robot can be acquired. Based on the fusion analysis of multiple demonstration trajectories, a hybrid dynamic primitive model can be established. Then, based on the model, a smoother and higher-quality maintenance trajectory can be generated, thereby improving the efficiency and quality of the equipotential robot in performing power maintenance operations.

[0044] It should be noted that, Figure 3 The 'n' shown here is only used as a sequence number to indicate the number of demonstration trajectories, dynamic motion primitives (i.e., the initial hybrid dynamic primitive model) or weights (i.e., model weights). Of course, other letters can also be chosen, such as 'k' in the embodiments of this invention. No specific restrictions are imposed here.

[0045] According to the above embodiments of the present invention, in step S202, acquiring multiple demonstration trajectories of the equipotential robot includes: acquiring multiple historical maintenance trajectories of the equipotential robot when performing historical maintenance operations within a historical time period, and determining the historical maintenance trajectories as demonstration trajectories; and / or, acquiring multiple demonstration trajectories by controlling the equipotential robot to simulate the process of power maintenance.

[0046] The following is combined Figure 4 The embodiments of the present invention will be described in detail below. Figure 4 This is a schematic diagram of a demonstration trajectory according to an embodiment of the present invention.

[0047] Specifically, it is possible to obtain, such as Figure 4The multiple demonstration trajectories shown are assumed to be represented as follows: (The five demonstration trajectories collected are listed below.) in, Let m represent the i-th trajectory point on the k-th demonstration trajectory. k This represents the total number of points on the k-th demonstration trajectory. The demonstration trajectory can be the trajectory recorded by the equipotential robot during power maintenance operations within a historical time period, or it can be obtained by controlling the equipotential robot to simulate the power maintenance process through kinematic teaching or other methods to obtain multiple demonstration trajectories. During kinematic teaching, the operator can directly guide the robot's end effector to experience a series of typical movements in the power maintenance operation process and record these movement trajectories. Of course, other methods can also be used to control the equipotential robot to simulate the power maintenance process to obtain demonstration trajectories, which will not be elaborated here.

[0048] Step S204: Perform dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory.

[0049] As above Figure 3 As shown, in this embodiment, dynamic motion primitives can be modeled based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. Here, the initial hybrid dynamic primitive model is a hybrid dynamic primitive model obtained by modeling dynamic primitives based on each demonstration trajectory. The hybrid dynamic motion primitive (Mixed Dynamic Movement Primitives, Mixed DMPs) model is a custom extension model that combines dynamic motion primitive (DMP) theory with additional innovations (such as random forgetting factors and model weight calculation based on fluency evaluation). In the fields of robotics and control, DMP models have been widely studied and applied for learning and reproducing complex motion trajectories.

[0050] According to the above embodiments of the present invention, in step S204, dynamic motion primitive modeling is performed on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. This includes: performing dynamic motion primitive modeling on each demonstration trajectory based on a proportional-differential controller and a trajectory shape learner to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory, wherein the initial hybrid dynamic primitive model N corresponding to the k-th demonstration trajectory is... k The first expression is: τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative with respect to time t, and let α represent acceleration. yβ represents the proportional parameter in a proportional-derivative controller. y Let f represent the differential parameter in the proportional-differential controller (PDC), g represent the nonlinear term in the trajectory shape learner, and d represent the target state of the PDC, where the target state is the state the PDC needs to achieve. The nonlinear term in the trajectory shape learner is adjusted to obtain the adjusted nonlinear term. Based on the initial hybrid dynamic primitive model, the convergence speed of the demonstration trajectory is adjusted according to the nonlinear term to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory. Here, the convergence speed refers to the speed at which the demonstration trajectory progresses from the current state to the target state. The initial hybrid dynamic primitive model M corresponding to the k-th demonstration trajectory... k The second expression is:

[0051] Specifically, a dynamic motion primitive can be understood as a combination of a PD controller (i.e., a proportional-derivative controller) and a trajectory shape learner; typically, the most commonly used and simplest PD controller can be described by a spring-damped system. Where τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative with respect to time t, and let α represent acceleration. y This represents the proportional parameter (P parameter) in a proportional-derivative controller, β. y Let f represent the derivative parameters (D parameters) in the proportional-derivative controller; assuming f represents the trajectory shape learner, which is a nonlinear function, then the dynamic motion primitives of a demonstration trajectory can be represented as: (First Expression); By modifying the target state g and the nonlinear term f in the first expression, the emphasis and shape of the demonstration trajectory can be adjusted. To obtain the desired trajectory shape, different nonlinear terms f need to be constructed. By adjusting different basis functions and their corresponding weights, a weighted complex trajectory can be obtained, thus yielding the desired trajectory shape. Typically, the complexity of the trajectory is proportional to the number of basis functions. To adjust the convergence speed of the trajectory, the first derivative of the system state representing the velocity needs to be adjusted. By adding a time factor τ, we can obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory:

[0052]

[0053] In the above embodiments of the present invention, adjusting the nonlinear term in the trajectory shape learner to obtain the adjusted nonlinear term includes: determining the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let σ represent the i-th radial basis function. a c represents the width of the a-th radial basis function. a Let represent the center position of the 'a'-th radial basis function, and 'x' represent the coordinates of a point on the demonstration trajectory. Radial basis functions are used to obtain nonlinear terms through linear superposition. Based on the linear superposition of multiple radial basis functions, the adjusted nonlinear terms are determined using the second formula, where the second formula is: ω a Let γ represent the weight of the a-th radial basis function, s represent the phase variable of the first-order system, and γ represent the weight of the a-th radial basis function. a Let represent the random forgetting factor of the a-th radial basis function.

[0054] The following is combined Figure 5 and Figure 6 The embodiments of the present invention will be described in detail below. Figure 5 This is a schematic diagram illustrating the effect of the nonlinear term according to an embodiment of the present invention. Figure 6 This is a schematic diagram of the visualization of radial basis functions according to an embodiment of the present invention.

[0055] To obtain the desired trajectory shape, different nonlinear terms f need to be constructed. This can be achieved by linearly superimposing normalized nonlinear basis functions, and by introducing a random forgetting factor γ. a Acting on basis functions Above, a probability value p is set, and γ is used each time. a The probability of having p is 1, and the probability of having 1-p is 0. Similar to the dropout mechanism in convolutional neural networks, this function can be expressed as: The radial basis function can be expressed as: In the formula, a represents the label of the radial basis function, σ a c represents the width of the a-th radial basis function. a Let x represent the center position of the a-th radial basis function, x represent the coordinates of a point on the demonstration trajectory, and ω represent the center position of the a-th radial basis function. a Let s represent the weight of the a-th radial basis function, and s represent the phase variable of the first-order system, such as... Figure 5As shown, in the DMP model, the weights of each radial basis function (RBF) determine its influence on the overall trajectory. Each RBF is multiplied by its weight, and then all weighted RBF outputs are summed to obtain a weighted summation. This summation reflects the combined influence of multiple shape features and is a key part of generating the final trajectory. RBF activation refers to its ability to respond to specific time points or phases of motion. RBFs typically activate in a specific temporal order throughout the motion, and this activation pattern is determined by the RBF's center position and width parameters. The trajectory is shaped by adjusting the weights and activation order. The magnitude of the weights determines the strength of each RBF's contribution to the trajectory, and the weighted summation combines the outputs of all RBFs into a smooth nonlinear term, such as... Figure 5 The effect of the nonlinear term shown; such as Figure 6 As shown, by adjusting the center position, width, and weight values ​​of radial basis functions, their activation mechanisms can be controlled, thereby optimizing and adjusting the robot's motion trajectory to better match the expected motion target and environmental conditions. In the context of Dynamic Motion Primitives (DMP), Psi usually refers to the phase or time variable in radial basis functions (RBFs), which controls the activation timing of RBFs. Psi activation (phase variable activation or time variable activation) refers to the activation mode of radial basis functions in time or phase as motion progresses in the DMP model.

[0056] Step S206: Evaluate the smoothness of each demonstration trajectory according to the evaluation indicators to obtain an evaluation score.

[0057] Optionally, the above evaluation indicators may include, but are not limited to: the first evaluation indicator, the second evaluation indicator, and the third evaluation indicator.

[0058] In this embodiment, the smoothness of each demonstration trajectory can be evaluated from multiple dimensions based on evaluation indicators to obtain an evaluation score. The smoothness of the demonstration trajectory can be evaluated mainly based on the following three indicators: the duration of zero velocity, the number of acceleration sign changes, and the duration of the absolute value of acceleration exceeding the threshold. Of course, other evaluation indicators can also be selected to evaluate the smoothness of the demonstration trajectory based on the actual situation, and no specific restrictions are imposed here.

[0059] According to the above embodiments of the present invention, in step S206, the smoothness of each demonstration trajectory is evaluated according to an evaluation index to obtain an evaluation score, including at least one of the following: evaluating the smoothness of the demonstration trajectory using a third formula based on a first evaluation index to obtain a first evaluation score, wherein the first evaluation index is the first duration of zero speed in each demonstration trajectory, and the third formula is: i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,k Let A(i) represent the velocity at the i-th point on the k-th demonstration trajectory. A(i) indicates whether the velocity at the i-th point on the k-th demonstration trajectory is zero; A(i) = 1 indicates the velocity at the i-th point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates the velocity at the i-th point on the k-th demonstration trajectory is not zero. The smoothness of the demonstration trajectory is evaluated using the fourth formula based on the second evaluation metric, resulting in a second evaluation score. The second evaluation metric is the number of acceleration sign changes in each demonstration trajectory. The fourth formula is: a represents the second evaluation score of the k-th demonstration trajectory. i,k Let a represent the acceleration at the i-th point on the k-th demonstration trajectory. i+1,k Let B(i) represent the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory. Let B(i) represent the state where the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are the same. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are the same. The smoothness of the demonstration trajectory is evaluated using the fifth formula based on the third evaluation index, resulting in a third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds the acceleration threshold in each demonstration trajectory. Absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: Let |a| represent the third evaluation score of the k-th demonstration trajectory. i,k |The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0060] Specifically, a velocity of zero usually indicates that the presenter may have paused due to their posture or workspace transitions, which is not a characteristic that the trajectory itself should have, and therefore needs to be penalized; this can be achieved using the following formula: Calculate the first evaluation score based on the smoothness of the demonstration trajectory according to the first evaluation index, where i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,k Let A(i) represent the velocity at the i-th point on the k-th demonstration trajectory. A(i) indicates whether the velocity at the i-th point on the k-th demonstration trajectory is zero; A(i) = 1 indicates the velocity at the i-th point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates the velocity at the i-th point on the k-th demonstration trajectory is not zero. To reduce the influence of spikes and noise in the velocity curve, the number of acceleration sign changes is used as a screening criterion to select smoother trajectories. This can be achieved using the following formula: Calculate the second evaluation score, which assesses the smoothness of the demonstration trajectory based on the second evaluation metric, where, a represents the second evaluation score of the k-th demonstration trajectory. i,k Let a represent the acceleration at the i-th point on the k-th demonstration trajectory. i+1,k Let B(i) represent the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory. Let B(i) represent the sign of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. During the demonstration, some motion jitter or motion switching may cause excessive instantaneous acceleration, which is an undesirable phenomenon during the demonstration. This can be addressed using the formula: Calculate the third evaluation score based on the smoothness of the demonstration trajectory according to the third evaluation index, where, Let |a| represent the third evaluation score of the k-th demonstration trajectory. i,k |The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0061] Step S208: Determine the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories based on power maintenance tasks.

[0062] As above Figure 3 As shown, in this embodiment, the evaluation scores obtained from evaluating the smoothness of the demonstration trajectory from multiple evaluation metrics can be fused to obtain the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory.

[0063] According to the above embodiments of the present invention, in step S208, determining the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score includes: fusing the first evaluation score, the second evaluation score, and the third evaluation score using a sixth formula to obtain the evaluation score, wherein the sixth formula is: L k Let n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score. Based on the evaluation scores, the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory are determined using the seventh formula, where the seventh formula is: K represents the total number of demonstration trajectories, δ k This represents the model weight of the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0064] Optionally, the first evaluation score, the second evaluation score, and the third evaluation score mentioned above are evaluation scores obtained by evaluating the smoothness of each demonstration trajectory according to the first evaluation index, the second evaluation index, and the third evaluation index, respectively.

[0065] Specifically, the formula can be used: Calculate the overall evaluation score for smoothness for each demonstration trajectory, where L k Let n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score; then the formula can be used: Calculate the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory, where K represents the total number of demonstration trajectories, and δ k This represents the model weight of the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0066] Step S210: Determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory.

[0067] As above Figure 3 As shown, in this embodiment, the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory obtained in the above steps can be fused to construct a hybrid dynamic primitive model.

[0068] According to the above embodiments of the present invention, in step S210, determining the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory includes: determining the hybrid dynamic primitive model using an eighth formula based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory, wherein the eighth formula is: M DMP M represents a hybrid dynamic primitive model. k This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0069] Specifically, the formula can be used based on the initial hybrid dynamic primitive model and model weights corresponding to multiple demonstration trajectories: They are then fused to obtain the hybrid dynamic primitive model M. DMP .

[0070] Step S212: Write the hybrid dynamic primitive model into the equipotential robot so that when the equipotential robot receives a power maintenance task, it can use the hybrid dynamic primitive model to generate a maintenance trajectory based on the maintenance location indicated by the power maintenance task.

[0071] In this embodiment, the hybrid dynamic primitive model obtained in the above steps can be written into the equipotential robot so that when the equipotential robot receives a power maintenance task, it can use the model to automatically generate a maintenance trajectory for the robot based on the maintenance location indicated by the power maintenance task.

[0072] Step S214: Control the equipotential robot to perform power maintenance operations based on the maintenance trajectory.

[0073] In this embodiment, when performing power maintenance tasks, the equipotential robot can perform power maintenance operations based on the maintenance trajectory generated by the hybrid dynamic primitive model to complete the power maintenance tasks, thereby improving the maintenance efficiency and quality of the equipotential robot in performing power maintenance tasks.

[0074] As described above, the technical solution provided by the above embodiments of the present invention can obtain multiple demonstration trajectories of an equipotential robot. The equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance. Dynamic motion primitive modeling is performed on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. The smoothness of each demonstration trajectory is evaluated according to evaluation indicators to obtain an evaluation score. The model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory is determined based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories according to the power maintenance task. The hybrid dynamic primitive model is determined based on the initial hybrid dynamic primitive model and model weight corresponding to each demonstration trajectory. The model incorporates a hybrid dynamic primitive model into an equipotential robot. Upon receiving a power maintenance task, the robot generates a maintenance trajectory based on the indicated maintenance location. This model controls the equipotential robot to perform power maintenance operations based on the trajectory. The process achieves the goal of using model weights obtained by modeling multiple demonstration trajectories with dynamic motion primitives and evaluating the smoothness of each trajectory, fusing the models corresponding to multiple demonstration trajectories to obtain a hybrid dynamic motion primitive model, and generating a maintenance trajectory based on this model to control the robot to perform maintenance according to the trajectory. This method integrates multiple data and factors to generate power maintenance trajectories using a hybrid dynamic primitive model, improving trajectory smoothness and enhancing the maintenance efficiency and quality of the equipotential robot in performing power maintenance tasks.

[0075] Therefore, the technical solution provided by the above embodiments of the present invention solves the technical problem in the related art that the equal potential robot can only learn a single trajectory, and that the generated trajectory has too many inflection points during the power maintenance process, which causes the bucket truck to shake violently, resulting in low maintenance efficiency and poor work quality.

[0076] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0077] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0078] According to embodiments of the present invention, a power maintenance device based on hybrid dynamic motion primitives is also provided for implementing the above-described power maintenance method based on hybrid dynamic motion primitives. Figure 7 This is a schematic diagram of a power maintenance device based on hybrid dynamic motion elements according to an embodiment of the present invention, as shown below. Figure 7 As shown, the device includes: a first acquisition unit 701, a second acquisition unit 703, a third acquisition unit 705, a first determination unit 707, a second determination unit 709, a generation unit 711, and a control unit 713. The following is a detailed description of this power maintenance device based on hybrid dynamic motion elements.

[0079] The first acquisition unit 701 is used to acquire multiple demonstration trajectories of the equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance.

[0080] The second acquisition unit 703 is used to perform dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory.

[0081] The third acquisition unit 705 is used to evaluate the smoothness of each demonstration trajectory according to the evaluation indicators and obtain an evaluation score.

[0082] The first determining unit 707 is used to determine the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories based on power maintenance tasks.

[0083] The second determining unit 709 is used to determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory.

[0084] The generation unit 711 is used to write the hybrid dynamic primitive model into the equipotential robot, so that when the equipotential robot receives a power maintenance task, it can use the hybrid dynamic primitive model to generate a maintenance trajectory according to the maintenance position indicated by the power maintenance task.

[0085] The control unit 713 is used to control the equipotential robot to perform power maintenance operations based on the maintenance trajectory.

[0086] It should be noted that the first acquisition unit 701, the second acquisition unit 703, the third acquisition unit 705, the first determination unit 707, the second determination unit 709, the generation unit 711 and the control unit 713 mentioned above correspond to steps S202 to S214 in the above embodiments. The seven units and the corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments.

[0087] As can be seen from the above, in the scheme described in the above embodiments of the present invention, multiple demonstration trajectories of the equipotential robot can be acquired using the first acquisition unit. The equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance. Then, the second acquisition unit performs dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. Next, the third acquisition unit evaluates the smoothness of each demonstration trajectory based on evaluation indicators to obtain an evaluation score. Then, the first determination unit determines the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model, which is used to generate maintenance trajectories based on the power maintenance task. Finally, the second determination unit determines the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score. The hybrid dynamic primitive model is determined by model weights. Then, the hybrid dynamic primitive model is written into the equipotential robot using a generation unit. This allows the equipotential robot to generate a maintenance trajectory based on the maintenance location indicated by the power maintenance task when it receives a power maintenance task. Finally, the control unit controls the equipotential robot to perform power maintenance operations based on the maintenance trajectory. This achieves the goal of using model weights obtained by modeling multiple demonstration trajectories with dynamic motion primitives and evaluating the smoothness of each trajectory, fusing the models corresponding to multiple demonstration trajectories to obtain a hybrid dynamic motion primitive model, and generating a maintenance trajectory based on this model to control the robot to perform maintenance according to the trajectory. This achieves the technical effect of integrating multiple data and factors to generate power maintenance trajectories using the hybrid dynamic primitive model, thereby improving trajectory smoothness and enhancing the maintenance efficiency and quality of the equipotential robot in performing power maintenance tasks.

[0088] Therefore, the technical solution provided by the above embodiments of the present invention solves the technical problem in the related art that the equal potential robot can only learn a single trajectory, and that the generated trajectory has too many inflection points during the power maintenance process, which causes the bucket truck to shake violently, resulting in low maintenance efficiency and poor work quality.

[0089] Optionally, the first acquisition unit includes: a first acquisition module, used to acquire multiple historical maintenance trajectories when the equipotential robot performs historical maintenance operations within a historical time period, and determine the historical maintenance trajectories as demonstration trajectories; and / or, a second acquisition module, used to acquire multiple demonstration trajectories by controlling the equipotential robot to simulate the power maintenance process.

[0090] Optionally, the second acquisition unit includes: a third acquisition module, used to perform dynamic motion primitive modeling on each demonstration trajectory based on a proportional-differential controller and a trajectory shape learner, to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory, wherein the initial hybrid dynamic primitive model N corresponding to the k-th demonstration trajectory... k The first expression is: τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative with respect to time t, and let α represent acceleration. y β represents the proportional parameter in a proportional-derivative controller. y The first module represents the differential parameter in the proportional-differential controller (PDC), f represents the nonlinear term in the trajectory shape learner, and g represents the target state of the PDC, which is the state that the PDC needs to achieve. The second module is used to adjust the nonlinear term in the trajectory shape learner to obtain the adjusted nonlinear term. The third module is used to adjust the convergence speed of the demonstration trajectory based on the nonlinear term, based on the obtained initial hybrid dynamic primitive model, to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory. Here, the convergence speed refers to the speed at which the demonstration trajectory moves from the current state to the target state, and the initial hybrid dynamic primitive model M corresponding to the k-th demonstration trajectory is... k The second expression is:

[0091]

[0092] Optionally, the fourth acquisition module includes: a first determination submodule, used to determine the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let σ represent the i-th radial basis function. a c represents the width of the a-th radial basis function. aLet represent the center position of the 'a'-th radial basis function, and 'x' represent the coordinates of a point on the demonstration trajectory. Radial basis functions are used to obtain nonlinear terms through linear superposition. The second determination submodule is used to determine the adjusted nonlinear terms based on the linear superposition of multiple radial basis functions using a second formula, where the second formula is: ω a Let γ represent the weight of the a-th radial basis function, s represent the phase variable of the first-order system, and γ represent the weight of the a-th radial basis function. a Let represent the random forgetting factor of the a-th radial basis function.

[0093] Optionally, the evaluation indicators include: a first evaluation indicator, a second evaluation indicator, and a third evaluation indicator. The third acquisition unit includes at least one of the following: a sixth acquisition module, used to evaluate the smoothness of the demonstration trajectory based on the first evaluation indicator using a third formula to obtain a first evaluation score, wherein the first evaluation indicator is the first duration of zero speed in each demonstration trajectory, and the third formula is: i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,k Let A(i) represent the velocity at the i-th trajectory point on the k-th demonstration trajectory. A(i) indicates whether the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero; A(i) = 1 indicates the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates the velocity at the i-th trajectory point on the k-th demonstration trajectory is not zero. The seventh acquisition module is used to evaluate the smoothness of the demonstration trajectory using the fourth formula based on the second evaluation index, obtaining the second evaluation score. The second evaluation index is the number of acceleration sign changes in each demonstration trajectory, and the fourth formula is: a represents the second evaluation score of the k-th demonstration trajectory. i,k Let a represent the acceleration at the i-th point on the k-th demonstration trajectory. i+1,k B(i) represents the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory. B(i) indicates whether the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory are the same. The eighth acquisition module is used to evaluate the smoothness of the demonstration trajectory according to the third evaluation index and the fifth formula to obtain the third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds the acceleration threshold in each demonstration trajectory. Absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: Let |a| represent the third evaluation score of the k-th demonstration trajectory. i,k |The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0094] Optionally, the first determining unit includes: a ninth obtaining module, used to fuse the first evaluation score, the second evaluation score, and the third evaluation score using a sixth formula to obtain an evaluation score, wherein the sixth formula is: L k Let n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score; the first determining module is used to determine the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation scores using the seventh formula, where the seventh formula is: K represents the total number of demonstration trajectories, δ k This represents the model weight of the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0095] Optionally, the second determining unit includes: a second determining module, used to determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory using the eighth formula, wherein the eighth formula is: M DMP M represents a hybrid dynamic primitive model. k This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0096] According to another aspect of the present invention, a power maintenance system based on hybrid dynamic motion primitives is also provided, wherein the power maintenance system based on hybrid dynamic motion primitives uses any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0097] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein the program executes any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0098] Optionally, in this embodiment, the computer-readable storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any communication device in a group of communication devices.

[0099] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: acquiring multiple demonstration trajectories of an equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the motion trajectory of the equipotential robot during power maintenance; performing dynamic motion primitive modeling based on each demonstration trajectory to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory; evaluating the smoothness of each demonstration trajectory according to evaluation indicators to obtain an evaluation score; determining the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score, wherein the model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model, and the hybrid dynamic primitive model is used to generate maintenance trajectories according to the power maintenance task; determining the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weight corresponding to each demonstration trajectory; writing the hybrid dynamic primitive model into the equipotential robot so that the equipotential robot generates maintenance trajectories according to the maintenance positions indicated by the power maintenance task when receiving a power maintenance task; and controlling the equipotential robot to perform power maintenance operations based on the maintenance trajectories.

[0100] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: acquiring multiple historical maintenance trajectories of the equipotential robot when performing historical maintenance operations within a historical time period, and determining the historical maintenance trajectories as demonstration trajectories; and / or, acquiring multiple demonstration trajectories by controlling the equipotential robot to simulate the process of power maintenance.

[0101] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: performing dynamic motion primitive modeling on each demonstration trajectory according to the proportional-differential controller and the trajectory shape learner to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory, wherein the initial hybrid dynamic primitive model N corresponding to the k-th demonstration trajectory is... k The first expression is: τ represents the time factor, and y represents the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative with respect to time t, and let α represent acceleration. y β represents the proportional parameter in a proportional-derivative controller. yLet f represent the differential parameter in the proportional-differential controller (PDC), g represent the nonlinear term in the trajectory shape learner, and d represent the target state of the PDC, where the target state is the state the PDC needs to achieve. The nonlinear term in the trajectory shape learner is adjusted to obtain the adjusted nonlinear term. Based on the initial hybrid dynamic primitive model, the convergence speed of the demonstration trajectory is adjusted according to the nonlinear term to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory. Here, the convergence speed refers to the speed at which the demonstration trajectory progresses from the current state to the target state. The initial hybrid dynamic primitive model M corresponding to the k-th demonstration trajectory... k The second expression is:

[0102]

[0103] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: determining the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let σ represent the i-th radial basis function. a c represents the width of the a-th radial basis function. a Let represent the center position of the 'a'-th radial basis function, and 'x' represent the coordinates of a point on the demonstration trajectory. Radial basis functions are used to obtain nonlinear terms through linear superposition. Based on the linear superposition of multiple radial basis functions, the adjusted nonlinear terms are determined using the second formula, where the second formula is: ω a Let γ represent the weight of the a-th radial basis function, s represent the phase variable of the first-order system, and γ represent the weight of the a-th radial basis function. a Let represent the random forgetting factor of the a-th radial basis function.

[0104] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: evaluating the smoothness of the demonstration trajectory using a third formula based on a first evaluation index to obtain a first evaluation score, wherein the first evaluation index is the first duration of zero speed in each demonstration trajectory, and the third formula is: i represents the label of the trajectory point on the demonstration trajectory. Let m represent the first evaluation score of the k-th demonstration trajectory. k v represents the total number of trajectory points on the k-th demonstration trajectory. i,kLet A(i) represent the velocity at the i-th point on the k-th demonstration trajectory. A(i) indicates whether the velocity at the i-th point on the k-th demonstration trajectory is zero; A(i) = 1 indicates the velocity at the i-th point on the k-th demonstration trajectory is zero, and A(i) = 0 indicates the velocity at the i-th point on the k-th demonstration trajectory is not zero. The smoothness of the demonstration trajectory is evaluated using the fourth formula based on the second evaluation metric, resulting in a second evaluation score. The second evaluation metric is the number of acceleration sign changes in each demonstration trajectory. The fourth formula is: a represents the second evaluation score of the k-th demonstration trajectory. i,k Let a represent the acceleration at the i-th point on the k-th demonstration trajectory. i+1,k Let B(i) represent the acceleration at the (i+1)th trajectory point on the k-th demonstration trajectory. Let B(i) represent the state where the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are the same. B(i) = 1 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are different, and B(i) = 0 indicates that the signs of the velocity at the ith trajectory point and the acceleration at the (i+1)th trajectory point are the same. The smoothness of the demonstration trajectory is evaluated using the fifth formula based on the third evaluation index, resulting in a third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds the acceleration threshold in each demonstration trajectory. Absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: Let |a| represent the third evaluation score of the k-th demonstration trajectory. i,k |The absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory, where a0 represents the acceleration threshold, and C(i) represents whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. C(i) = 1 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold, and C(i) = 0 indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

[0105] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: fusing the first evaluation score, the second evaluation score, and the third evaluation score using a sixth formula to obtain an evaluation score, wherein the sixth formula is: L k Let n represent the evaluation score of the k-th demonstration trajectory, n1 represent the first weight of the first evaluation score, n2 represent the second weight of the second evaluation score, and n3 represent the third weight of the third evaluation score. Based on the evaluation scores, the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory are determined using the seventh formula, where the seventh formula is: K represents the total number of demonstration trajectories, δ k This represents the model weight of the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0106] Optionally, in this embodiment, the computer-readable storage medium is configured to store program code for performing the following steps: determining the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and model weights corresponding to each demonstration trajectory using an eighth formula, wherein the eighth formula is: M DMP M represents a hybrid dynamic primitive model. k This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

[0107] According to another aspect of the present invention, a processor is also provided, which is used to run a program, wherein the program executes any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0108] According to another aspect of the present invention, a computer program product is also provided, including computer instructions, which, when executed by a processor, perform any of the above-described power maintenance methods based on hybrid dynamic motion primitives.

[0109] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0110] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0112] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0113] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0114] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0115] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A power maintenance method based on hybrid dynamic motion elements, characterized in that, include: Multiple demonstration trajectories of an equipotential robot are obtained, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the movement trajectory of the equipotential robot during power maintenance. Dynamic motion primitives are modeled based on each of the demonstration trajectories to obtain an initial hybrid dynamic primitive model corresponding to each of the demonstration trajectories; The fluency of each demonstration trajectory is evaluated based on the evaluation metrics to obtain an evaluation score; The model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory is determined based on the evaluation score. The model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model. The hybrid dynamic primitive model is used to generate maintenance trajectories based on power maintenance tasks. The hybrid dynamic primitive model is determined based on the initial hybrid dynamic primitive model and the model weights corresponding to each of the demonstration trajectories; The hybrid dynamic primitive model is written into the equipotential robot so that when the equipotential robot receives the power maintenance task, it can use the hybrid dynamic primitive model to generate the maintenance trajectory according to the maintenance location indicated by the power maintenance task. The equipotential robot is controlled to perform power maintenance operations based on the maintenance trajectory. Dynamic motion primitive modeling is performed on each of the demonstration trajectories to obtain an initial hybrid dynamic primitive model corresponding to each demonstration trajectory. This includes: performing dynamic motion primitive modeling on each of the demonstration trajectories based on a proportional-differential controller and a trajectory shape learner to obtain an initial hybrid dynamic primitive model corresponding to each of the demonstration trajectories, wherein the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory... The first expression is: , Let represent the time factor, and y represent the current state of the proportional-derivative controller. Let y be the first derivative with respect to time t, and let represent velocity. Let y be the second derivative of y with respect to time t, and let represent acceleration. This represents the proportional parameter in the proportional-derivative controller. Let f represent the differential parameter in the proportional-differential controller, f represent the nonlinear term in the trajectory shape learner, and g represent the target state of the proportional-differential controller, where the target state refers to the state that the proportional-differential controller needs to achieve. The nonlinear term in the trajectory shape learner is adjusted to obtain the adjusted nonlinear term. Based on the obtained initial hybrid dynamic primitive model, the convergence speed of the demonstration trajectory is adjusted according to the nonlinear term to obtain the initial hybrid dynamic primitive model corresponding to each demonstration trajectory, where the convergence speed refers to the speed at which the demonstration trajectory moves from the current state to the target state. The initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory... The second expression is: ; Adjusting the nonlinear term in the trajectory shape learner to obtain the adjusted nonlinear term includes: determining the radial basis function using a first formula, wherein the first formula is: 'a' represents the label of the radial basis function. Let i represent the i-th radial basis function. This represents the width of the a-th radial basis function. Let represent the center position of the a-th radial basis function, and x represent the coordinates of a point on the demonstration trajectory. The radial basis functions are used to obtain the nonlinear term through linear superposition. Based on the linear superposition of multiple radial basis functions, the adjusted nonlinear term is determined using a second formula, wherein the second formula is: , Let represent the weight value of the a-th radial basis function, and s represent the phase variable of the first-order system. Let represent the random forgetting factor of the a-th radial basis function.

2. The power maintenance method based on hybrid dynamic motion elements according to claim 1, characterized in that, Acquire multiple demonstration trajectories of the equipotential robot, including: Obtain multiple historical maintenance trajectories of the equipotential robot performing historical maintenance operations within a historical time period, and determine the historical maintenance trajectory as the demonstration trajectory; and / or, By controlling the equipotential robot to simulate the power maintenance process, multiple demonstration trajectories are obtained.

3. The power maintenance method based on hybrid dynamic motion elements according to claim 1, characterized in that, The evaluation metrics include: a first evaluation metric, a second evaluation metric, and a third evaluation metric. The smoothness of each demonstration trajectory is evaluated based on the evaluation metrics to obtain an evaluation score, which includes at least one of the following: The smoothness of the demonstration trajectory is evaluated using a third formula based on the first evaluation index to obtain a first evaluation score. The first evaluation index is the first duration of zero speed in each demonstration trajectory, and the third formula is: , , where i represents the label of the trajectory point on the demonstration trajectory. The first evaluation score represents the k-th demonstration trajectory. This represents the total number of trajectory points on the demonstration trajectory described in the k-th example. This represents the velocity at the i-th trajectory point on the k-th demonstration trajectory. This indicates whether the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero. This indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is zero. This indicates that the velocity at the i-th trajectory point on the k-th demonstration trajectory is not zero; The smoothness of the demonstration trajectory is evaluated using the fourth formula based on the second evaluation index to obtain a second evaluation score. The second evaluation index is the number of acceleration sign changes in each demonstration trajectory, and the fourth formula is: , , This represents the second evaluation score for the k-th demonstration trajectory. This represents the acceleration at the i-th point on the k-th demonstration trajectory. This represents the acceleration at the (i+1)th point on the k-th demonstration trajectory. This indicates whether the signs of the velocity at the i-th trajectory point and the acceleration at the (i+1)-th trajectory point on the k-th demonstration trajectory are the same. This indicates that the signs of the velocity at the i-th trajectory point and the acceleration at the (i+1)-th trajectory point on the k-th demonstration trajectory are different. The sign of the velocity at the i-th trajectory point on the k-th demonstration trajectory is the same as that of the acceleration at the (i+1)-th trajectory point; The smoothness of the demonstration trajectory is evaluated using the fifth formula based on the third evaluation index to obtain a third evaluation score. The third evaluation index is the second duration during which the absolute acceleration exceeds an acceleration threshold in each demonstration trajectory. The absolute acceleration refers to the absolute value of the acceleration. The fifth formula is: , , This represents the third evaluation score for the k-th demonstration trajectory. The absolute acceleration at the i-th trajectory point on the demonstration trajectory described in the k-th clause. This represents the acceleration threshold. This indicates whether the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. This indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory exceeds the acceleration threshold. This indicates that the absolute acceleration at the i-th trajectory point on the k-th demonstration trajectory does not exceed the acceleration threshold.

4. The power maintenance method based on hybrid dynamic motion elements according to claim 1, characterized in that, The model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory are determined based on the evaluation scores, including: The first assessment score, the second assessment score, and the third assessment score are combined using a sixth formula to obtain the assessment score, wherein the sixth formula is: , The evaluation score represents the k-th demonstration trajectory. This represents the first weight of the first evaluation score. This represents the second weight of the second evaluation score. This represents the third weight of the third evaluation score; Based on the evaluation score, the model weights of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory are determined using the seventh formula, wherein the seventh formula is: K represents the total number of the demonstrated trajectories. The model weight represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

5. The power maintenance method based on hybrid dynamic motion elements according to claim 4, characterized in that, The hybrid dynamic primitive model is determined based on the initial hybrid dynamic primitive model and the model weights corresponding to each demonstration trajectory, including: The hybrid dynamic primitive model is determined using the eighth formula based on the initial hybrid dynamic primitive model and the model weights corresponding to each demonstration trajectory, wherein the eighth formula is: , This represents the hybrid dynamic primitive model. This represents the initial hybrid dynamic primitive model corresponding to the k-th demonstration trajectory.

6. A power maintenance device based on hybrid dynamic motion elements, characterized in that, The power maintenance method based on hybrid dynamic motion elements, applied to any one of claims 1 to 5, includes: The first acquisition unit is used to acquire multiple demonstration trajectories of the equipotential robot, wherein the equipotential robot is used for power maintenance, and the demonstration trajectory is the movement trajectory of the equipotential robot during power maintenance. The second acquisition unit is used to perform dynamic motion primitive modeling based on each of the demonstration trajectories to obtain an initial hybrid dynamic primitive model corresponding to each of the demonstration trajectories; The third acquisition unit is used to evaluate the smoothness of each demonstration trajectory according to the evaluation index and obtain an evaluation score; The first determining unit is used to determine the model weight of the initial hybrid dynamic primitive model corresponding to each demonstration trajectory based on the evaluation score, wherein the model weight refers to the importance of each initial hybrid dynamic primitive model in the process of constructing the hybrid dynamic primitive model, and the hybrid dynamic primitive model is used to generate maintenance trajectories based on power maintenance tasks. The second determining unit is used to determine the hybrid dynamic primitive model based on the initial hybrid dynamic primitive model and the model weights corresponding to each of the demonstration trajectories; A generation unit is used to write the hybrid dynamic primitive model into the equipotential robot, so that when the equipotential robot receives the power maintenance task, it can use the hybrid dynamic primitive model to generate the maintenance trajectory according to the maintenance location indicated by the power maintenance task. The control unit is used to control the equipotential robot to perform power maintenance operations based on the maintenance trajectory.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein the program executes the power maintenance method based on hybrid dynamic motion primitives as described in any one of claims 1 to 5.

8. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, the power maintenance method based on hybrid dynamic motion primitives as described in any one of claims 1 to 5 is performed.

Images