AI electric butterfly digital twin simulation interaction method based on 3D printing

Abstract

The application relates to the technical field of twin simulation, in particular to an AI electric butterfly digital twin simulation interaction method based on 3D printing, which comprises the following steps: constructing an electric butterfly model, and manufacturing a physical electric butterfly through 3D printing technology to construct a digital twin model corresponding to the physical electric butterfly; obtaining multiple music characteristic information and multiple butterfly actions, obtaining real-time music characteristic information corresponding to an environment where the physical electric butterfly is located, determining corresponding butterfly action information according to the real-time music characteristic information to obtain simulation action data; determining standard action data and physical action data based on the simulation action data; and correcting the physical action data based on the standard action data to obtain target action data, so that the butterfly action can quickly respond to changes in real-time music information, and the continuity of action switching is improved.

Description

Technical Field

[0001] This invention relates to the field of twin simulation technology, specifically to a method for interactive simulation of an AI-powered electric butterfly based on 3D printing. Background Technology

[0002] In recent years, with the rapid development of 3D printing, artificial intelligence, and digital twin technologies, customized and intelligent bionic robots have ushered in new development opportunities. Among them, electric bionic butterflies, which can interact with music, the environment, or the audience, are representative products of art and technology and show broad application prospects in exhibitions, stage performances, and immersive experiences. However, when existing electric bionic butterflies interact with music, they usually use pre-set fixed mapping rules, which cannot accurately match the rhythm of the music. The butterfly's movements cannot quickly respond to changes in real-time music information, affecting the continuity of movement switching and resulting in low efficiency and accuracy in the coordination between the generated movements and the music rhythm.

[0003] To address these issues, we propose a 3D-printed AI-powered electric butterfly digital twin simulation and interaction method. Summary of the Invention

[0004] The purpose of this invention is to provide a 3D-printed AI electric butterfly digital twin simulation interaction method to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a 3D-printed AI-powered electric butterfly digital twin simulation and interaction method, which includes the following steps: Construct an electric butterfly model, and use 3D printing technology to create a physical electric butterfly, and then construct a digital twin model corresponding to the physical electric butterfly; Acquire multiple music feature information and multiple butterfly movements, obtain real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determine the corresponding butterfly movement information based on the real-time music feature information to obtain simulation movement data; Standard motion data and physical motion data are determined based on simulation motion data; target motion data is obtained by correcting the physical motion data based on the standard motion data.

[0006] Preferably, the step of constructing an electric butterfly model and creating a physical electric butterfly using 3D printing technology, and constructing a digital twin model corresponding to the physical electric butterfly, includes: The mechanical structure of an electric butterfly was designed using 3D modeling software, generating a digital 3D model. The digital 3D model is imported into a 3D printer, and the physical structural components of the butterfly are manufactured based on the 3D printing and assembled to generate a physical electric butterfly. A digital twin model is generated by constructing a 3D model based on a physical point electric butterfly device and stored on a control platform. A control terminal is configured for the physical electric butterfly, and a communication channel is established between the control terminal and the digital twin model.

[0007] Preferably, the steps of acquiring multiple music feature information and multiple butterfly motion information, acquiring real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determining the corresponding butterfly motion information based on the real-time music feature information to obtain simulation motion data include: Acquire multiple music data corresponding to the physical electric butterfly, and extract music features from the multiple music data to obtain multiple music feature information; Acquire multiple butterfly motion information of a physical electric butterfly, establish the correlation between multiple music feature information and multiple butterfly motion information, and determine the butterfly motion information of the physical electric butterfly corresponding to the real-time music information and the corresponding motion sequence as simulation motion data based on the correlation.

[0008] Preferably, the step of establishing the association between multiple musical feature information and multiple butterfly motion information includes: Multiple music feature information is obtained, each music feature information is used as a feature node, and multiple feature nodes are connected to each other to obtain a music feature point network; Acquire multiple butterfly motion information, treat each butterfly motion information as an action node, and connect multiple action nodes to obtain a motion information point network; Obtain the feature nodes of music feature information and the action nodes of the corresponding action feature information of music feature information, and map the feature nodes to the action nodes one by one. Differentiation information is configured on the action information point network, wherein the differentiation information includes the main follower point and the point clusters corresponding to the main follower point; Configure movement points on the music feature point network, establish a synchronization chain between the movement points and the master follower point, and use the synchronization chain as the association between multiple music feature information and multiple butterfly motion information.

[0009] Preferably, the step of configuring differentiation information on the action information point network, wherein the differentiation information includes a primary follower point and a cluster of points corresponding to the primary follower point, includes: Obtain each action node and its corresponding associated group on the action information point network. The associated group includes the action node and other action nodes that are associated with the action node. A database is configured for each feature node, and the database stores at least one differentiation information. The differentiation information includes the main follower point set for each action node and the corresponding point cluster. The point cluster includes multiple secondary follower points differentiated from the main follower point, and the main follower point and multiple secondary follower points communicate with each other.

[0010] Preferably, the step of determining the butterfly motion information and corresponding motion sequence of the physical electric butterfly corresponding to the real-time music information as simulation motion data based on the association relationship includes: Real-time acquisition of music information from the environment where the physical electric butterfly is located, extraction of music features and generation of real-time music information containing the sequence of features; Based on real-time music information, determine the target movement point at the current moment, obtain the action node mapped to the target movement point and the corresponding associated group, and distribute the main follow point on the action node to multiple other action nodes in the associated group as secondary follow points. Based on the changing sequence of musical features, from the candidate action nodes where the secondary follower point is located, the target candidate feature node that matches the musical features of the next moment is determined, the secondary follower point on the target candidate action node is taken as the new primary follower point, and a new cluster of points is generated. The sequence of the previous primary follower point and other secondary follower points in its cluster is used as the position transformation order and stored in the database; The differentiation process information of the master follower point corresponding to the target moving point is used as the simulation action data. The differentiation process information includes differentiation information, the position change order of the master follower point and the corresponding action state.

[0011] Preferably, the step of determining standard motion data and entity motion data based on simulation motion data includes: Acquire simulated motion data and input the simulated motion data into the digital twin model; generate standard motion data corresponding to the simulated motion data based on the digital twin model; Based on standard motion data, corresponding motion control commands are generated for the physical electric butterfly; based on the motion control commands, the physical electric butterfly is driven to wave in rhythm with the music to obtain physical motion data.

[0012] Preferably, the step of correcting the entity motion data based on standard motion data to obtain the target motion data includes: Acquire actual motion data and standard motion data, compare the actual motion data and standard motion data in real time, and generate error data; The entity motion data corresponding to the error data that does not meet the preset conditions is adjusted to obtain the actual motion data after adjustment, until the difference between the actual motion data after adjustment and the standard motion data meets the preset conditions, and the actual motion data is used as the target motion data.

[0013] Compared with the prior art, the beneficial effects of the present invention are: By differentiating secondary follower points from the primary follower point at the current moment, i.e., from subsequent possible action nodes in its associated cluster, an action preparation is performed. When the music features change and the movement point moves to the next feature node, there is no need to temporarily calculate which action corresponds to it. Instead, the secondary follower points that have been pre-deployed on the corresponding action node are directly collected and promoted to new primary follower points. By placing the search and preparation process of the action path before the music changes, the latency of data analysis and conversion is reduced, enabling the butterfly motion to respond quickly to changes in real-time music information and improving the coherence of action switching. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the motion information network of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0017] Example: Please refer to Figures 1 to 2 This invention provides a technical solution for a 3D-printed AI electric butterfly digital twin simulation and interaction method: the 3D-printed AI electric butterfly digital twin simulation and interaction method includes the following steps: S1: Construct an electric butterfly model and use 3D printing technology to create a physical electric butterfly, and construct a digital twin model corresponding to the physical electric butterfly; The steps for constructing an electric butterfly model and then creating a physical electric butterfly using 3D printing technology, along with the corresponding digital twin model, include: designing the mechanical structure of the electric butterfly using 3D modeling software to generate a digital 3D model; importing the digital 3D model into a 3D printer to manufacture the physical structural components of the butterfly and assembling them to generate a physical electric butterfly; constructing a digital twin model based on the physical electric butterfly device and storing the digital twin model on a control platform; configuring a control terminal for the physical electric butterfly and establishing a communication channel between the control terminal and the digital twin model, with the control terminal used to transmit the actual motion data of the physical electric butterfly to the digital twin model via the communication channel to determine the standardization of the motion; Specifically, using 3D modeling software, the mechanical structure of the electric butterfly is designed, generating a digital 3D model that includes movable wings, an internal cavity, and a fixed structure. The internal cavity houses the drive mechanism and control unit. The digital 3D model is imported into a 3D printer, and the physical structural components of the butterfly are manufactured using 3D printing technology, followed by post-processing. Subsequently, the micro-drive mechanism, control unit, and sensor unit are integrated and installed inside the printed physical structure to assemble a physical electric butterfly. In a digital twin development environment, a visual virtual model with geometric and physical properties consistent with the physical electric butterfly is created, and dynamic parameters are configured for the virtual model. Numerical and kinematic constraints are applied; a communication protocol and interface are configured between the control unit of the physical electric butterfly and the digital twin model to establish a bidirectional, real-time data exchange channel between the two; through the bidirectional data exchange channel, the digital twin model can send motion control commands to the physical electric butterfly, and the physical electric butterfly can feed back its own state data to the digital twin model. The physical electric butterfly includes a 3D printed main structure, wing components, an electromagnetic induction drive module, a control module, and a power supply module. The digital twin model is constructed based on the structural parameters of the physical electric butterfly and includes a virtual twin model and a data interaction module to achieve real-time data synchronization with the physical device.

[0018] S2: Acquire multiple music feature information and multiple butterfly movements, acquire real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determine the corresponding butterfly movement information based on the real-time music feature information to obtain simulation movement data; The steps for obtaining simulation motion data include: acquiring multiple music feature information and multiple butterfly motion information, acquiring real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determining the corresponding butterfly motion information based on the real-time music feature information; acquiring multiple butterfly motion information of the physical electric butterfly, establishing the correlation between the multiple music feature information and the multiple butterfly motion information, and determining the butterfly motion information of the physical electric butterfly corresponding to the real-time music information and the corresponding motion sequence as simulation motion data based on the correlation. Specifically, multiple music files matching the expected performance scenarios of the physical electric butterflies are acquired to form an initial, standardized music dataset. The music data can come from local storage, streaming platform API interfaces, or recordings in specific environments. For each music file in the dataset, a comprehensive digital signal analysis is performed to extract its global statistical features and temporal local features, thereby obtaining multi-dimensional feature information of the music. Musical feature information can include low-level features (such as pitch and melody, rhythm and beat) and high-level features (such as style and emotion). The steps for extracting low-level features are as follows: For pitch and melody, a fundamental tone extraction algorithm (such as autocorrelation method, YIN algorithm) is used to identify the tonic pitch, and the note sequence is obtained by combining MIDI analysis. For example, the pitch contour is extracted by using the Librosa library, and the note start time and pitch value are marked. For rhythm and beat, the spectrum is analyzed by using short-time Fourier transform (STFT) or constant Q transform (CQT), and the beat position is detected by combining dynamic programming algorithm (such as DBN). For example, the accent position of each measure is extracted as rhythm feature point. For volume and dynamics, the RMS energy or peak amplitude of each frame of audio is calculated, and the volume change points (such as sudden increase / decrease) are marked as dynamic feature points. The high-level feature extraction steps are as follows: For genre classification, Mel frequency cepstral coefficients (MFCC) are used as timbre features, combined with support vector machines (SVM) or deep learning models (such as LSTM) to classify genres (e.g., classical, jazz, electronic); for emotion recognition, features such as MFCC, pitch range, and rhythm complexity are extracted and input into a pre-trained emotion classification model (e.g., the eGeMAPS feature set extracted by OpenSMILE tool), outputting emotion labels (e.g., cheerful, sad); the extracted features are transformed into a feature node network and associated with butterfly movements, defining time nodes, using the music timeline as a reference, and marking feature change points; feature type nodes are defined, layered according to feature categories, such as "high pitch layer," "rhythm layer," and "emotion layer," each layer containing specific feature values ​​(e.g., pitch = C4, rhythm = allegro), connecting different feature nodes at the same time point (e.g., pitch C4 and allegro rhythm appearing simultaneously). Similar features across time points are connected (e.g., consecutive occurrences of the same pitch or rhythmic pattern), and weights are assigned according to feature importance (e.g., emotion features have higher weights than pitch features because they have a greater impact on the movement). The butterfly's movements are broken down into quantifiable parameters and mapped to musical characteristics: for example, body posture: wing angle (0°-180°), body tilt (-90° to 90°). Movement trajectory: flight path type (straight, spiral, hovering), speed (0-10 units / second).Movement frequency: Wing flapping frequency (1-20Hz), body sway amplitude (0-100%); A movement information network is constructed using movement parameters as nodes, defining posture nodes such as "wings 90°" and "body horizontal"; defining trajectory nodes such as "spiral flight" and "speed 5 units / second"; defining frequency nodes such as "flapping frequency 10Hz" and "sway amplitude 50%"; movement continuity: the ending state of the previous movement connects to the starting state of the next movement (e.g., "wings 90°" → "wings 120°"); movement combination: different parameters... Combinations of numbers form complex movements (e.g., "wings 120° + spiral flight + flapping frequency 15Hz"), establishing a mapping relationship between musical characteristics and movement parameters. For pitch → posture, for example, high register (above C5) → wings spread (120°-180°), body raised (30°-90°); low register (below C2) → wings folded (0°-60°), body lowered (-30° to 0°). For rhythm → trajectory, for example, fast rhythm (>120 BPM) → spiral flight, speed 8-10 units / second; slow rhythm (<60 BPM) → straight flight, speed 1-3 units / second. For emotion → frequency: cheerful emotion → wing flapping frequency 15-20Hz, swing amplitude 70%-100%. Sadness → Wing flapping frequency 1-5Hz, swing amplitude 10%-30%. Traverse the music feature point network to find the feature corresponding to the current time point (e.g., pitch = G4, rhythm = allegro, mood = cheerful). Calculate the target motion parameters (e.g., wing angle = 150°, flight speed = 9 units / second, flapping frequency = 18Hz) according to the mapping rules.

[0019] The musical features in the music feature point network can be any one of pitch, rhythm, and volume; the motion information in the motion information network can be any one of posture, trajectory, and frequency. Among them, the feature nodes of pitch correspond one-to-one with the motion nodes of posture, the feature nodes of rhythm correspond one-to-one with the motion nodes of trajectory, and the feature nodes of emotion correspond one-to-one with the motion nodes of frequency. When constructing the feature node network, it is assumed that the music feature point network corresponds to pitch features, including the middle, high, and low registers. Each register serves as a feature node in the music feature point network, corresponding one-to-one with the posture action nodes in the action information network, recording the specific pitch value of each register. In a piece of music, assuming the current moment corresponds to the feature node of the low register, the middle and high registers are connected to the low register respectively. The feature nodes corresponding to the middle and high registers form association groups for the feature nodes of the low register. Multiple feature nodes in each association group are then connected to the feature nodes of the low register. The system connects lines and labels interval information. Based on the mapping relationship, it determines the association of the action nodes of the corresponding bass region's feature nodes, and configures movement points for these feature nodes. These movement points are implemented through a data acquisition port to collect music change information. When a music feature is detected to shift from the bass region to other regions, the movement point is moved from the current feature node to the feature node corresponding to the next region, simultaneously causing the master follower point to change position. For example, when the music begins to shift from the bass region to the treble region or to the middle region, the feature nodes corresponding to the treble and middle regions respectively become the corresponding bass regions. The system identifies the association groups of feature nodes in a region, obtains the feature nodes of the current pitch range of the moving point, and determines the association groups corresponding to the action nodes based on the association groups of the feature nodes. In the association groups, one action node corresponds to one secondary follower point. For example, if the current pitch range is high, it may shift to low or middle range, meaning the next pitch range may be low or middle, thus corresponding to two different feature nodes. These two different feature nodes are used as the association groups corresponding to the current feature node. When the moving point starts moving from the previous pitch range to the current pitch range, the secondary follower point on the action node corresponding to the feature node of the current pitch range is upgraded to a new primary follower point and marked. At the same time, the association group corresponding to the action node where the new primary follower point is located is obtained as the center. New secondary follower points are distributed to other action nodes in the association group through the new primary follower point. The primary follower point corresponding to the previous pitch range and the unupgraded secondary follower points are stored in the database of the corresponding action nodes. The primary follower point at the previous moment and the new primary follower point are connected to form the action sequence corresponding to the change in musical features. The action sequence and the action state at different times are used for subsequent control of the physical electric butterfly.

[0020] Action nodes corresponding to feature nodes are extracted, and butterfly movements are matched with musical feature information. Each feature node corresponds to an action node. The actions of a real butterfly include, but are not limited to, wing flapping, hovering, forward ascent, descent, and swaying. Each action corresponds to a musical feature. The duration of the movement point on the corresponding feature node is determined based on the interval information to control the duration of the real butterfly's movements. Musical feature nodes are mapped to action information networks. For example, high-pitched notes are associated with a butterfly's fully extended wings and upward-raised body posture. When a high-pitched melody appears in the music, the butterfly makes this movement, expressing excitement and activity. Low-pitched notes are mapped to a butterfly's folded wings and downward-sinking body posture. Mapping body posture to musical features, low-pitched music usually gives a sense of stability and weight, and this butterfly movement can correspond to it. A fast rhythm can correspond to the butterfly's winding and fast flight path. When the music rhythm speeds up, the butterfly quickly changes its flight direction in space, increasing the complexity and dynamism of the movement, while a slow rhythm matches the butterfly's straight flight or slow hovering path. This mapping allows the butterfly's movements to harmonize with the rhythm of the music, creating a soothing and tranquil atmosphere. Based on the previously established mapping relationship between characteristic changes and butterfly movements, a correspondence is established between the corresponding characteristic change nodes and butterfly movement nodes. For example, the high-pitched note change nodes correspond to the butterfly movement nodes of wings unfolding and body rising, reflecting the relationship between musical characteristic changes and butterfly movements.

[0021] The steps for establishing the association between multiple musical feature information and multiple butterfly motion information include: acquiring multiple musical feature information, each musical feature information being treated as a feature node, and connecting multiple feature nodes to obtain a musical feature point network; acquiring multiple butterfly motion information, each butterfly motion information being treated as an action node, and connecting multiple action nodes to obtain an action information point network; acquiring the feature nodes of the musical feature information and the action nodes of the corresponding action feature information, and establishing a one-to-one correspondence between feature nodes and action nodes; configuring differentiation information on the action information point network, wherein the differentiation information includes a master follower point and a cluster of points corresponding to the master follower point; configuring movement points on the musical feature point network, establishing a synchronization chain between the movement points and the master follower point, and using the synchronization chain as the association between multiple musical feature information and multiple butterfly motion information; Specifically, multiple musical feature information is obtained, each musical feature information is used as a feature node, and multiple feature nodes are connected to each other to obtain a musical feature point network. The musical feature information can be similarity or continuity such as pitch, rhythm or intensity, and multiple butterfly action information, such as hovering, flapping wings, gliding, turning, etc. Based on other feature nodes in the associated group corresponding to the feature node, the main follower point will differentiate into multiple secondary follower points and distribute them to the action nodes corresponding to other feature nodes in the associated group. Then, based on the changes in music features, the secondary follower points that need to be stored are determined. The secondary follower points on the action nodes corresponding to the next music feature are taken as the main follower points, and multiple secondary follower points are further differentiated to the action nodes corresponding to the feature nodes in the associated group of that feature node until the music in the current environment is completed. Real-time music feature information and the corresponding feature order are obtained. The movement path of the main follower point on the action node network is synchronously driven by the real-time music information where the movement point is located, thereby determining the action information and action order corresponding to the real-time music information as simulation action data, which can improve the efficiency and accuracy of the coordination between action and real-time music. The steps for configuring differentiation information on the action information point network, wherein the differentiation information includes a primary follower point and a point cluster corresponding to the primary follower point, include: obtaining each action node and its corresponding associated cluster on the action information point network, wherein the associated cluster includes the action node and other action nodes associated with the action node; configuring a database for each feature node, wherein the database stores at least one piece of differentiation information, wherein the differentiation information includes the primary follower point set for each action node and the corresponding point cluster, and the point cluster includes multiple secondary follower points differentiated from the primary follower point, and the primary follower point and the multiple secondary follower points are connected for communication; The steps for determining the butterfly motion information and corresponding motion sequence of the physical electric butterfly corresponding to real-time music information as simulation motion data based on the correlation relationship include: acquiring music information of the environment where the physical electric butterfly is located in real time, extracting music features and generating real-time music information containing feature sequence; determining the target movement point at the current moment based on the real-time music information, acquiring the action node mapped by the target movement point and the corresponding association cluster, and distributing secondary follower points from the main follower point on the action node to multiple other action nodes in the association cluster; determining the target candidate feature node that matches the music feature at the next moment from the candidate action nodes where the secondary follower point is located based on the change sequence of music features, taking the secondary follower point on the target candidate action node as the new main follower point, and generating a new point cluster; storing the sequence of the previous main follower point and other secondary follower points in its point cluster as the position transformation sequence in the database; and using the differentiation process information of the main follower point corresponding to the target movement point as simulation motion data, wherein the differentiation process information includes differentiation information, the position transformation sequence of the main follower point, and the corresponding motion state. It should be noted that by taking the secondary follower point on the target candidate action node as the new primary follower point and generating a new cluster of points, the primary follower point on the target action node can be copied. During copying, the corresponding cluster of points must also be copied. The primary follower point and its corresponding cluster are then distributed to other action nodes in the associated cluster. The distributed primary follower point and its corresponding cluster are treated as a whole, belonging to the secondary follower point distributed from the cluster by the previous primary follower point. The movement point, primary follower point, and secondary follower point all serve as virtual ports for collecting relevant information from the corresponding nodes. The movement point collects the music features at the current moment and determines the movement point to move to the corresponding music feature based on the changes in the music features at the next moment. During the movement of the movement point, the primary follower point is synchronously driven to differentiate and change positions in the action node network through the synchronization chain. The differentiation information and the order of position changes are used as simulation action data. The simulation action data is input into the digital twin model for simulation and serves as standard action data for subsequent judgment of whether the real-time action of the physical electric butterfly is abnormal, facilitating timely correction. Specifically, the primary follower point is a virtual marker located on the current action node, representing the current dominant butterfly action state. The cluster of points is a set of multiple secondary follower points derived from the primary follower point. Each secondary follower point is distributed to other action nodes associated with the current action node (i.e., the "association cluster" of that action node). The primary follower point maintains a communication connection with all secondary follower points for state synchronization. A mobile point is configured on the music feature point network, which moves within the network based on the real-time input music feature sequence. A synchronization chain is established between this mobile point and the primary follower point on the action information point network. This synchronization chain determines how changes in music features drive the migration path of the primary follower point in the action node network, thus forming a dynamic association between music features and butterfly action information. Based on the real-time music features at the current moment, the target mobile point (i.e., the feature node corresponding to the current music feature) is determined on the music feature point network. Through the synchronization chain, the action node mapped to the target mobile point (called the current action node) and its association cluster are located. The primary follower point on the current action node is triggered to distribute secondary follower points to multiple other action nodes in its association cluster. These secondary follower points represent the possible action states that may transition to in the next musical moment. Based on the music information flow, the music characteristics of the next moment are identified. According to these characteristics, target candidate action nodes that match the next music characteristics are determined from the candidate action nodes that have been assigned secondary follower points. The secondary follower points on the target candidate action node are promoted to new primary follower points. Subsequently, the new primary follower point generates a new cluster of points on its new action node (i.e., assigns new secondary follower points to the new associated cluster). The information of the previous primary follower point and other secondary follower points in its cluster is stored in the database as the position transformation order to record the action evolution process until the music ends. Throughout the process, the movement path of the moving point drives the migration of the primary follower point on the action information point network in real time through the synchronization chain. Finally, the position transformation order of the primary follower point and its differentiation process information (i.e., differentiation information and position transformation order) are output as simulation action data. This data contains the butterfly motion information sequence and its execution order, used to control the physical motorized butterfly to perform. The main follow point, secondary follow point, and main movement point use virtual ports to collect and record the changing order of motion nodes following music nodes. Based on real-time changes in the music, it can determine the next step in possible motion transition paths in advance, allowing the main follow point to enter the next feature node ahead of time. It also determines which secondary follow point among multiple feature nodes can be upgraded to the main follow point based on real-time changes in the music. This allows motion changes to respond quickly to changes in music features, improving the efficiency and accuracy of the coordination between motion and real-time music, resulting in smoother generated butterfly motion.

[0022] Specifically, other related feature nodes refer to the probability, logic, or artistic patterns of transitioning from one state (current node) to another state (related node) during the music's progression. This belongs to a temporal and dynamic relationship. Based on extensive music data analysis, the probability of the music transitioning to feature node B, feature node C, feature node D, etc., when it is in feature node A, has been statistically determined. After "Exhilarating Allegro" (feature node A), there is a 70% probability of transitioning to "Soothing Adagio" (feature node B), and a 30% probability of maintaining the exhilarating tempo but slightly decreasing (feature node C). Feature nodes B, C, and D are the related nodes of feature node A. The set of feature nodes B, C, and D belongs to the related group corresponding to feature node A. Feature node A corresponds to action node A1, and feature node B corresponds to... Action node B1 corresponds to action node C1, feature node D corresponds to action node D1, and so on, mapping feature nodes on the music feature network to action nodes on the action information network one-to-one. Based on the necessary or common progression relationships determined by music theory (such as chord progressions and musical form), for each feature node, its "association group" is encapsulated into a data structure. The first one belongs to the initial movement point, and so on up to the nth movement point. The butterfly action information corresponding to the order and distribution of the movement points is used as simulation action data, which can quickly find the node corresponding to the next music feature and perform pre-differentiation. When needed, the data is directly used. The secondary follower points on the corresponding action node are collected and promoted to the primary follower point at the current moment. The primary follower point at the previous moment is connected to the primary follower point at the current moment. The connection path obtained by connecting the primary follower points at different moments in the playback of a piece of music information is used as the position transformation order. Since the primary follower point will follow the movement point when the movement point moves to the feature node corresponding to the current moment, the secondary follower point on the action node of the feature node corresponding to the movement point is used as the primary follower point to complete one position transformation of the primary follower point. At the same time as the position transformation is completed, the new primary follower point differentiates into new secondary follower points from other action nodes in the associated group corresponding to the action node, which are used for the next step. The position of the primary follower point changes during each movement, allowing for advance preparation for the movement point to move to the next feature node following the music features. This reduces the latency of data analysis and conversion, thereby improving the smoothness of action switching. Multiple secondary follower points and the primary follower point within the point cluster are all interconnected. When a movement point moves to the action node corresponding to any feature node, the secondary follower point on that action node is immediately designated as the primary follower point at that moment. Simultaneously, a connection path is formed with the previous primary follower point, which is stored in the corresponding database. This improves the efficiency of the action node in following changes in the music feature nodes, thereby increasing the efficiency of matching action information with music information.

[0023] S3: Determine standard motion data and entity motion data based on simulation motion data; correct the entity motion data based on the standard motion data to obtain the target motion data; The steps for determining standard motion data and physical motion data based on simulated motion data include: acquiring simulated motion data and inputting it into a digital twin model; generating standard motion data corresponding to the simulated motion data based on the digital twin model; generating motion control commands for the corresponding physical electric butterfly based on the standard motion data; and driving the physical electric butterfly to wave according to the rhythm of the music based on the motion control commands to obtain physical motion data. The steps for correcting entity motion data based on standard motion data to obtain target motion data include: acquiring actual motion data and standard motion data; comparing the actual motion data and standard motion data in real time to generate error data; adjusting the entity motion data corresponding to the error data that does not meet the preset conditions to obtain adjusted actual motion data, until the difference between the adjusted actual motion data and the standard motion data meets the preset conditions, and using the actual motion data as the target motion data.

[0024] Specifically, simulated motion data generated from music feature information through methods such as dynamic dot mesh differentiation is acquired. This simulated motion data is an idealized sequence of movements, without considering the physical limitations of the physical mechanism. The simulated motion data is input into a pre-constructed digital twin model. This digital twin model is a virtual mapping of the physical electric butterfly, accurately simulating its mechanical structure (such as links and joints), dynamic characteristics (such as mass, inertia, and friction), and motion constraints (such as servo angle range and maximum angular velocity). The digital twin model verifies and optimizes the feasibility of the input simulated motion data, outputting a set of standard motion data. This standard motion data is theoretically the optimal sequence of motion commands that can be executed safely, smoothly, and accurately on the physical mechanism. Specific motion control commands (such as PWM signals) are generated based on the standard motion data. These motion control commands are sent to the drive mechanism (such as a servo motor) of the physical electric butterfly, driving it to wave according to the rhythm of the music, thereby obtaining the physical motion data (i.e., the actual executed motion). The physical motion data is collected in real time by sensors installed on the physical electric butterfly (such as an IMU consisting of a gyroscope, accelerometer, and magnetometer, or a vision-based motion capture system). Real-time comparison and feedback correction generate target motion data. Actual motion data and benchmark standard motion data are acquired in real time. The actual data and benchmark data at the same moment are compared to calculate error data. This error data may include position error, angle error, speed error, etc. It is determined whether the error data meets preset conditions (e.g., whether the absolute value of the error is less than a certain threshold ε). These preset conditions are the criteria for judging whether the motion is accurate and qualified. For entity motion data corresponding to error data that does not meet the preset conditions, a control mechanism is activated. Specifically, the control mechanism generates control instructions (such as an incremental PID control signal) based on the magnitude and direction of the error data. The control instructions are combined with the original benchmark motion data to generate new, corrected motion control instructions. The new motion control instructions are used to drive the physical electric butterfly, and the entity motion data is collected again at this time; the entity motion data that meets the conditions at this time is marked as target motion data. This target motion data represents the optimal performance effect achieved after correction in a real-world environment. The system performs repeated comparisons to form a closed-loop feedback control loop until the difference between the current physical action data and the standard action data (i.e., the error generated by the new round of comparison) meets the preset conditions. This allows the system to actively sense and reduce deviations during execution, increase the consistency between the electric butterfly's performance movements and the music rhythm, and improve the accuracy of controlling the physical electric butterfly.

[0025] By differentiating secondary follower points from the primary follower point at the current moment, i.e., from subsequent possible action nodes in its associated cluster, an action preparation is performed. When the music features change and the movement point moves to the next feature node, there is no need to temporarily calculate which action corresponds to it. Instead, the secondary follower points that have been pre-deployed on the corresponding action node are directly collected and promoted to new primary follower points. The process of searching and preparing the action path is placed before the music changes, reducing the delay in data analysis and conversion. This allows the butterfly motion to quickly respond to changes in real-time music information and improves the coherence of action switching.

[0026] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for interactive simulation of an AI-powered electric butterfly based on 3D printing, characterized in that, Includes the following steps: Construct an electric butterfly model, and use 3D printing technology to create a physical electric butterfly, and then construct a digital twin model corresponding to the physical electric butterfly; Acquire multiple music feature information and multiple butterfly movements, obtain real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determine the corresponding butterfly movement information based on the real-time music feature information to obtain simulation movement data; Determine standard motion data and physical motion data based on simulation motion data; The target motion data is obtained by correcting the entity motion data based on the standard motion data.

2. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 1, characterized in that: The steps of constructing an electric butterfly model, creating a physical electric butterfly using 3D printing technology, and constructing a digital twin model corresponding to the physical electric butterfly include: The mechanical structure of an electric butterfly was designed using 3D modeling software, generating a digital 3D model. The digital 3D model is imported into a 3D printer, and the physical structural components of the butterfly are manufactured based on the 3D printing and assembled to generate a physical electric butterfly. A digital twin model is generated by constructing a 3D model based on a physical point electric butterfly device and stored on a control platform. A control terminal is configured for the physical electric butterfly, and a communication channel is established between the control terminal and the digital twin model.

3. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 1, characterized in that: The steps of acquiring multiple music feature information and multiple butterfly motion information, acquiring real-time music feature information corresponding to the environment where the physical electric butterfly is located, and determining the corresponding butterfly motion information based on the real-time music feature information to obtain simulation motion data include: Acquire multiple music data corresponding to the physical electric butterfly, and extract music features from the multiple music data to obtain multiple music feature information; Acquire multiple butterfly motion information of a physical electric butterfly, establish the correlation between multiple music feature information and multiple butterfly motion information, and determine the butterfly motion information of the physical electric butterfly corresponding to the real-time music information and the corresponding motion sequence as simulation motion data based on the correlation.

4. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 3, characterized in that: The steps for establishing the association between multiple musical feature information and multiple butterfly motion information include: Multiple music feature information is obtained, each music feature information is used as a feature node, and multiple feature nodes are connected to each other to obtain a music feature point network; Acquire multiple butterfly motion information, treat each butterfly motion information as an action node, and connect multiple action nodes to obtain a motion information point network; Obtain the feature nodes of music feature information and the action nodes of the corresponding action feature information of music feature information, and map the feature nodes to the action nodes one by one. Differentiation information is configured on the action information point network, wherein the differentiation information includes the main follower point and the point clusters corresponding to the main follower point; Configure movement points on the music feature point network, establish a synchronization chain between the movement points and the master follower point, and use the synchronization chain as the association between multiple music feature information and multiple butterfly motion information.

5. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 4, characterized in that: The step of configuring differentiation information on the action information point network, wherein the differentiation information includes the main follower point and the point cluster corresponding to the main follower point, includes: Obtain each action node and its corresponding associated group on the action information point network. The associated group includes the action node and other action nodes that are associated with the action node. A database is configured for each feature node, and the database stores at least one differentiation information. The differentiation information includes the main follower point set for each action node and the corresponding point cluster. The point cluster includes multiple secondary follower points differentiated from the main follower point, and the main follower point and multiple secondary follower points communicate with each other.

6. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 3, characterized in that: The step of determining the butterfly motion information and corresponding motion sequence of the physical electric butterfly corresponding to the real-time music information as simulation motion data based on the association relationship includes: Real-time acquisition of music information from the environment where the physical electric butterfly is located, extraction of music features and generation of real-time music information containing the sequence of features; Based on real-time music information, determine the target movement point at the current moment, obtain the action node mapped to the target movement point and the corresponding associated group, and distribute the main follow point on the action node to multiple other action nodes in the associated group as secondary follow points. Based on the changing sequence of musical features, from the candidate action nodes where the secondary follower point is located, the target candidate feature node that matches the musical features of the next moment is determined, the secondary follower point on the target candidate action node is taken as the new primary follower point, and a new cluster of points is generated. Store the sequence of the previous primary follower point and other secondary follower points in its cluster as the position transformation order in the database; The differentiation process information of the master follower point corresponding to the target moving point is used as the simulation action data. The differentiation process information includes differentiation information, the position change order of the master follower point and the corresponding action state.

7. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 1, characterized in that: The steps for determining standard motion data and entity motion data based on simulated motion data include: Acquire simulated motion data and input it into a digital twin model; generate standard motion data corresponding to the simulated motion data based on the digital twin model; Based on standard motion data, corresponding motion control commands are generated for the physical electric butterfly; based on the motion control commands, the physical electric butterfly is driven to wave in rhythm with the music to obtain physical motion data.

8. The AI-powered electric butterfly digital twin simulation and interaction method based on 3D printing according to claim 1, characterized in that: The step of correcting the entity motion data based on standard motion data to obtain the target motion data includes: Acquire actual motion data and standard motion data, compare the actual motion data and standard motion data in real time, and generate error data; The entity motion data corresponding to the error data that does not meet the preset conditions is adjusted to obtain the actual motion data after adjustment, until the difference between the actual motion data after adjustment and the standard motion data meets the preset conditions, and the actual motion data is used as the target motion data.

Images