Dance formation self-organization control method based on robot group

Abstract

The invention discloses a dance formation self-organization control method based on a robot group. The method comprises the following steps: S1, obtaining dynamic changes of light frequency, color and intensity through a spectrum camera; s2, judging an effective light source and an invalid light source, and removing the invalid light source to form a target light signal; s3, converting the optical signal into a digital signal through a photoelectric converter and a digital signal processor, and extracting the characteristics of the digital signal to form a characteristic instruction; s4, comparing the feature instruction with a predefined action matching database in real time, and outputting a robot action instruction; s5, analyzing the characteristics of the received digital signal, predicting an action instruction at the next moment, and generating the next action instruction in advance; s6, generating a robot action track by using a path planning algorithm; and S7, adopting a distributed control strategy, dynamically adjusting the motion track, and completing dance arrangement. According to the invention, dance arrangement of the robot group is controlled through light, and diversity of dance of the robot group is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of robot dance, and in particular to a dance formation self-organization control method based on a robot group. Background Art

[0002] With the development of science and technology, robotics has been widely used, especially in the field of entertainment, where robot group performances have become a new art form. However, existing robot group performances mostly rely on pre-programmed action sequences, lack flexibility and interactivity, and cannot adapt well to changes in the on-site environment, such as the influence of factors such as lighting and music. In addition, traditional methods make it difficult to achieve efficient coordination of large-scale robot groups, resulting in limited performance effects. Therefore, it is particularly important to develop a method that enables robot groups to self-organize and control dance formations according to environmental changes.

[0003] Traditional robot group dance formation control methods mainly rely on the leader for control. The leader sets the dance program, and other robots follow, resulting in a single dance phenomenon. At the same time, some robot central control systems guide the robot's actions through pre-set motion trajectories, accurately control the actions of each robot, and ensure the coordination of the entire group. However, there are obvious limitations, such as poor adaptability to environmental changes. Once encountering external interference, such as sudden changes in lighting, it may cause chaos in the entire formation and affect the performance effect. In addition, since all decisions are made by the central system, for large-scale robot groups, the central system's computational burden is too heavy and easily becomes a performance bottleneck.

[0004] In recent years, with the development of sensor technology and artificial intelligence, some studies have begun to explore robot swarm self-organization control methods based on environmental perception, which can improve the flexibility and adaptability of robot swarms to a certain extent. However, existing technologies often focus on a single environmental factor, such as sound or visual signals, while ignoring the combined effects of multiple environmental factors, especially the important role of lighting on dance formations. In addition, existing methods still have shortcomings in multi-robot collaborative control, making it difficult to achieve efficient path planning and collision avoidance.

[0005] This technical solution proposes a self-organizing control method for dance formation based on robot groups. It uses a spectral camera to obtain changes in ambient light, and combines a deep learning model to predict the robot's next action, achieving rapid response and adaptation to complex environmental changes. Compared with the existing technology, this technology not only improves the adaptability of the robot group to environmental changes, but also optimizes the collaborative work between multiple robots through a distributed control strategy, reduces the computing burden of the central system, and improves the stability and efficiency of the overall system. In addition, by dynamically adjusting the movement trajectory in real time, the coordination of the robot group in the dance formation process is ensured, greatly enriching the artistic effect of the performance. Summary of the invention

[0006] The present invention provides a dance formation self-organizing control method based on a robot group, comprising:

[0007] S1. Obtain the dynamic changes of light frequency, color and intensity through the spectral camera to form the original light signal;

[0008] S2, using a spectrum recognition algorithm to determine effective light sources and invalid light sources, and removing invalid light sources through an adaptive filter to form a target light signal;

[0009] S3, the target light signal is converted into an electrical signal by a photoelectric converter, the electrical signal is converted into a digital signal by a digital signal processor, and the digital signal feature is extracted by Fourier transform to form a feature instruction;

[0010] S4, through the feature instruction parsing mechanism, the feature instruction is compared with the predefined action matching database in real time, and the robot action instruction is output;

[0011] S5. Analyze the characteristics of the received digital signal through a deep learning model, predict the action instruction at the next moment, and generate the next action instruction in advance;

[0012] S6. The predicted next action instruction is assigned to each robot through the multi-robot collaborative control mechanism, and the path planning algorithm is used to enable each robot to generate an action trajectory according to the received instruction;

[0013] S7, using distributed control strategy to dynamically adjust the movement trajectory in real time, so that the robot can complete the dance choreography under the guidance of lights.

[0014] Preferably, the spectral camera in S1 is located on the top of the robot, and forms an original light signal by acquiring the dynamic changes of light frequency, color and intensity; the light signal color C, intensity, and frequency f are obtained by the spectral camera; the light signal color C formula is: Where E(λ) is the spectral radiance, X(λ), Y(λ) and Z(λ) are the CIE standard observer color matching functions, λ1 and λ2 represent the minimum and maximum response wavelengths of the spectral camera, and λ is the wavelength; the intensity I formula is: Where R(λ) is the response function of the spectral camera; the frequency f formula is: where c is the speed of light.

[0015] Preferably, S2 specifically includes: using a spectrum recognition algorithm to determine effective light sources and invalid light sources, and determining the effectiveness of the light sources by calculating the spectrum similarity S, the formula is: Wherein, E(λ) is the spectral radiance, T(λ) is the preset effective light source spectrum template, λ1 and λ2 represent the minimum and maximum response wavelengths of the spectral camera; the spectral similarity S is calculated, and when S exceeds the preset threshold S th. , the light source is determined to be a valid light source; otherwise, it is determined to be an invalid light source.

[0016] Preferably, the invalid light source is removed by an adaptive filter to form a target light signal, and the adaptive filter performs adaptive filtering in an infinite approach manner, and the formula is: Wherein, x(n) is the original light signal, u(n) is the invalid light source signal, w(k) represents the weight coefficient of the adaptive filter, k is the weight coefficient index of the adaptive filter, and N represents the order of the filter;

[0017] The adaptive filter iteratively updates the weight coefficient w(k) so that the filtered signal y(n) is as close to the target light signal as possible. The updating formula of the weight coefficient is: w(k+1)=w(k)+μ·e(n)·u(n), wherein e(n)=d(n)-y(n), e(n) is the error signal, d(n) is the desired pure light signal, μ represents the step size factor, and the adaptive filter removes invalid light source signals to form a pure target light signal G.

[0018] Preferably, in S3, the target optical signal is converted into an electrical signal by a photoelectric converter, and then the electrical signal is converted into a digital signal by a digital signal processor, specifically including: the photoelectric converter converts the target optical signal G into an electrical signal V, the formula is: V=A·G+B, wherein G represents the target optical signal, A represents the gain coefficient of the photoelectric converter, and B represents the bias coefficient; the electrical signal V is converted into a digital signal D by the digital signal processor, the formula is: D=ADC(V), wherein ADC represents the analog-to-digital conversion function, which converts the continuous electrical signal into a discrete digital signal.

[0019] Preferably, the step of extracting digital signal features through Fourier transform of discrete digital signals to form feature instructions specifically includes converting digital signals into frequency domain signals through Fourier transform, and the formula is: Among them, D(t) is a digital signal, ω represents the angular frequency, j is an imaginary unit, and t is time. By analyzing the frequency domain signal F(ω), the frequency, intensity and color characteristics of the signal are extracted to form the feature instruction F f , the feature instruction F f For feature extraction, the formula is: F f =(|F(ω1)|, |F(ω2)|,..., |F(ω n )|), where, ω1, ω2,...,ω n Indicates the frequency point, F(ω i ) indicates that at the frequency point ω i The signal amplitude at .

[0020] Preferably, in S4, the feature instruction F(ω) is compared with the predefined action matching database in real time through the feature instruction parsing mechanism to calculate the feature instruction F f The similarity Ssim with the predefined action pattern: Among them, F f (i) represents the amplitude of the characteristic instruction at the i-th frequency point, A pre (i) represents the amplitude of the predefined action pattern at the i-th frequency point, n represents the number of frequency points, and the similarity S is calculated. sim , when S sim Exceeding the preset threshold S th When the most similar action mode is selected as the current robot action instruction A cmd ,The robot action instructions are output through the embedded control mechanism, and the corresponding actions are performed according to the feature instructions.

[0021] Preferably, the deep learning model in S5 is a long short-term memory network model, which uses the long short-term memory network model to analyze the digital signal feature F f The deep learning model is used to analyze the characteristics of the received digital signal, predict the action command at the next moment, and generate the next action command in advance. The formula is: A n =LSTM(F f ,h t-1 ), where F f is the characteristic instruction, h t-1 is the hidden state at the previous moment, A n The predicted action instruction for the next moment.

[0022] Preferably, in S6, the predicted action instruction A is transmitted through a multi-robot collaborative control mechanism. n The distributed control strategy is assigned to each robot, and the path planning algorithm is used to generate the motion trajectory P of each robot. i (t), the path planning formula is: Among them, P i (t) represents the position of the ith robot at time t, P i (0) represents the initial position, v i (t) represents the speed of the i-th robot at time t; the path planning algorithm ensures that each robot will not collide when executing action instructions through collision avoidance and coordination of the robot group; the multi-robot collaborative control mechanism can effectively assign the predicted next action instruction to each robot and generate their respective action trajectories to ensure the coordination of the dance formation of the entire robot group.

[0023] Preferably, the distributed control strategy is assigned to each robot, and the motion trajectory is adjusted dynamically in real time through the distributed control strategy, so that the robot can complete the dance choreography under the guidance of the light. The robot monitors its own position and speed in real time through the local controller, and exchanges information with other robots to make real-time dynamic adjustments. The adjustment formula is: P i (t+Δt)=P i (t)+v i (t)·Δt+α·∑ j≠i (P j (t)-P i (t)), where Δt is the time step, α is the adjustment coefficient, and P j (t) represents the position of other robots at time t; by adjusting the coefficient α, each robot can dynamically adjust its own motion trajectory according to the positions of other robots and its own position to ensure the coordination of the entire robot group; each robot also includes real-time updating of motion instructions according to the received light signal to ensure that the predetermined dance choreography is completed under the guidance of the light.

[0024] Compared with the prior art, the technical solution of this application has the following technical effects:

[0025] The present invention controls the dance program of the robot through dynamic changes in the frequency, color and intensity of the light, so that the robot can display different dance groups under different light intensities, thereby achieving the diversity of the dance of the robot group.

[0026] The present invention uses a spectral camera to obtain the dynamic changes of light frequency, color and intensity and form an original light signal, thus solving the problem that traditional technology cannot effectively capture changes in ambient light. This enables the robot to adaptively adjust dance movements according to changes in ambient light, thereby enhancing the flexibility and interactivity of robot group performances and greatly improving the audience's viewing experience.

[0027] The present invention determines effective light sources and invalid light sources through a spectral recognition algorithm, and uses an adaptive filter to remove invalid light sources to form a target light signal, thereby solving the problem of inaccurate signals caused by ambient light interference in traditional technologies, ensuring that the robot can receive clear, interference-free light signal instructions, and improving the accuracy and reliability of the robot group dance formation.

[0028] The present invention uses a feature instruction parsing mechanism to compare feature instructions with a predefined action matching database in real time, thereby solving the problem of robot actions being out of sync with environmental changes in traditional technologies. This enables robots to quickly and accurately perform actions that match environmental changes, thereby enhancing the coherence and natural fluency of robot group performances.

[0029] The present invention uses a deep learning model to predict the action instructions at the next moment and generate the next action instructions in advance, which solves the problems of robot response delay and incoherent action in traditional technologies, improves the predictability and reaction speed of the robot group to environmental changes, ensures the smooth transition and overall coordination of the dance formation, and significantly improves the artistic effect and viewing value of the performance.

[0030] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application so that it can be implemented in accordance with the contents of the specification, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following is a detailed description of the preferred embodiments of the present application in conjunction with the accompanying drawings as follows.

[0031] Based on the detailed description of the specific embodiments of the present application in combination with the accompanying drawings below, those skilled in the art will become more aware of the above and other objects, advantages and features of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings without creative work. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0033] Figure 1 A schematic diagram of a dance formation self-organizing control method based on a robot group according to the present invention;

[0034] Figure 2 This is a flow chart of the dance formation self-organization control method based on robot groups of the present invention

[0035] Figure 3 It is a single schematic diagram of the dance formation self-organization control method based on robot groups of the present invention. DETAILED DESCRIPTION

[0036] To make the purpose, technical scheme and advantages of the embodiment of the present application clearer, the technical scheme in the embodiment of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiment of the present application. Obviously, the described embodiment is a part of the embodiment of the present application, rather than all of the embodiments. In the following description, specific details such as specific configuration and components are provided only to help fully understand the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, for clarity and brevity, the description of known functions and structures is omitted in the embodiment.

[0037] It should be understood that the references to "one embodiment" or "this embodiment" throughout the specification mean that the specific features, structures, or characteristics associated with the embodiment are included in at least one embodiment of the present application. Therefore, the references to "one embodiment" or "this embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0038] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity, and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0039] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" in this article describes another type of association object relationship, indicating that there can be two relationships. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the previous and next associated objects are in an "or" relationship.

[0040] The term "at least one" in this article is merely a description of the association relationship of associated objects, indicating that there may be three relationships. For example, at least one of A and B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0041] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusions.

[0042] Example 1

[0043] This embodiment mainly describes a dance formation self-organizing control method based on a robot group. Figure 1 As shown in the figure, the spectral camera is used to obtain changes in ambient light, and combined with the deep learning model to predict the robot's next action, it achieves rapid response and adaptation to complex environmental changes;

[0044] like Figure 2 As shown, the specific steps include:

[0045] S1. Obtain the dynamic changes of light frequency, color and intensity through the spectral camera to form the original light signal;

[0046] Furthermore, the spectral camera is located on the top of the robot, and the original light signal is formed by acquiring the dynamic changes of the light frequency, color and intensity. The light signal color C, intensity, and frequency f are obtained through the spectral camera. The formula of the light signal color C is: Where E(λ) is the spectral radiance, X(λ), Y(λ) and Z(λ) are the CIE standard observer color matching functions, λ1 and λ2 represent the minimum and maximum response wavelengths of the spectral camera, and λ is the wavelength; intensity, the formula is: Where R(λ) is the response function of the spectral camera; the frequency f formula is: where c is the speed of light.

[0047] S2, using a spectrum recognition algorithm to determine effective light sources and invalid light sources, and removing invalid light sources through an adaptive filter to form a target light signal;

[0048] Furthermore, the spectrum recognition algorithm is used to determine the effective light source and the invalid light source, and the effectiveness of the light source is determined by calculating the spectrum similarity S. The formula is: Among them, E(λ) is the spectral radiance, T(λ) is the preset effective light source spectrum template, λ1 and λ2 represent the lowest and highest response wavelengths of the spectral camera; by calculating the spectral similarity S, when S exceeds the preset threshold S th. When , the light source is determined to be a valid light source; otherwise, it is determined to be an invalid light source;

[0049] Furthermore, the invalid light source is removed by an adaptive filter to form a target light signal. The adaptive filter performs adaptive filtering in an infinite approach, and the formula is: Wherein, x(n) is the original light signal, u(n) is the invalid light source signal, w(k) represents the weight coefficient of the adaptive filter, k is the weight coefficient index of the adaptive filter, and N represents the order of the filter;

[0050] The adaptive filter iteratively updates the weight coefficient w(k) to make the filtered signal y(n) as close to the target light signal as possible. The updating formula of the weight coefficient is: w(k+1)=w(k)+μ·e(n)·u(n), where e(n)=d(n)-y(n), e(n) is the error signal, d(n) is the desired pure light signal, and μ represents the step size factor. The adaptive filter removes invalid light source signals to form a pure target light signal G.

[0051] S3, the target light signal is converted into an electrical signal by a photoelectric converter, the electrical signal is converted into a digital signal by a digital signal processor, and the digital signal feature is extracted by Fourier transform to form a feature instruction;

[0052] Furthermore, the target optical signal is converted into an electrical signal by a photoelectric converter, and then the electrical signal is converted into a digital signal by a digital signal processor, specifically including: the photoelectric converter converts the target optical signal G into an electrical signal V, the formula is: V=A·G+B, where G represents the target optical signal, A represents the gain coefficient of the photoelectric converter, and B represents the bias coefficient; the electrical signal V is converted into a digital signal D by the digital signal processor, the formula is: D=ADC(V), where ADC represents the analog-to-digital conversion function, which converts the continuous electrical signal into a discrete digital signal.

[0053] Furthermore, the discrete digital signal is subjected to Fourier transform to extract digital signal features and form feature instructions, which specifically includes converting the digital signal into a frequency domain signal using Fourier transform, and the formula is: Among them, D(t) is a digital signal, ω represents the angular frequency, j is an imaginary unit, and t is time. By analyzing the frequency domain signal F(ω), the frequency, intensity and color characteristics of the signal are extracted to form the feature instruction F f , for the feature instruction F f For feature extraction, the formula is: F f=(|F(ω1)|, |F(ω2)|,..., |F(ω n )|), where, ω1, ω2,...,ω n Indicates the frequency point, F(ω i ) indicates that at the frequency point ω i The signal amplitude at .

[0054] S4, through the feature instruction parsing mechanism, the feature instruction is compared with the predefined action matching database in real time, and the robot action instruction is output;

[0055] Furthermore, through the feature instruction parsing mechanism, the feature instruction F(ω) is compared with the predefined action matching database in real time to calculate the feature instruction F f The similarity S between the predefined action patterns sim : Among them, F f (i) represents the amplitude of the characteristic instruction at the i-th frequency point, A pre (i) represents the amplitude of the predefined action pattern at the i-th frequency point, n represents the number of frequency points, and the similarity S is calculated. sim , when S sim Exceeding the preset threshold S th When the most similar action mode is selected as the current robot action instruction A cmd , output robot action instructions through the embedded control mechanism, and execute corresponding actions according to the feature instructions;

[0056] S5. Analyze the characteristics of the received digital signal through a deep learning model, predict the action instruction at the next moment, and generate the next action instruction in advance;

[0057] Furthermore, the deep learning model is a long short-term memory network model, which uses the long short-term memory network model to analyze the digital signal feature F f The deep learning model is used to analyze the characteristics of the received digital signal, predict the action command at the next moment, and generate the next action command in advance. The formula is: A n =LSTM(F f ,h t-1 ), where F f is the characteristic instruction, h t-1 is the hidden state at the previous moment, A n The predicted action instruction for the next moment.

[0058] S6. The predicted next action instruction is assigned to each robot through the multi-robot collaborative control mechanism, and the path planning algorithm is used to enable each robot to generate an action trajectory according to the received instruction;

[0059] Furthermore, through the multi-robot collaborative control mechanism, the predicted action command A n The distributed control strategy is assigned to each robot, and the path planning algorithm is used to generate the motion trajectory P of each robot. i (t), the path planning formula is: Among them, P i (t) represents the position of the ith robot at time t, P i (0) represents the initial position, v i (t) represents the speed of the i-th robot at time t; the path planning algorithm ensures that each robot will not collide when executing action instructions through collision avoidance and coordination of the robot group; the multi-robot collaborative control mechanism can effectively assign the predicted next action instruction to each robot and generate their respective action trajectories to ensure the coordination of the dance formation of the entire robot group.

[0060] S7, using distributed control strategy to dynamically adjust the movement trajectory in real time, so that the robot can complete the dance choreography under the guidance of the light;

[0061] Furthermore, a distributed control strategy is used to allocate to each robot. The distributed control strategy is used to dynamically adjust the motion trajectory in real time, so that the robot can complete the dance choreography under the guidance of the light. The robot monitors its own position and speed in real time through the local controller, and exchanges information with other robots to make real-time dynamic adjustments. The adjustment formula is: P i (t+Δt)=P i (t)+v i (t)·Δt+α·∑ j≠i (P j (t)-P i (t)), where Δt is the time step, α is the adjustment coefficient, and P j (t) represents the position of other robots at time t; by adjusting the coefficient α, each robot can dynamically adjust its own motion trajectory according to the positions of other robots and its own position to ensure the coordination of the entire robot group; each robot also includes real-time updating of motion instructions based on the received light signals to ensure that the predetermined dance choreography is completed under the guidance of the light.

[0062] This embodiment solves the problems of poor environmental adaptability, serious signal interference, slow motion response and low collaborative work efficiency in the control of traditional robot group dance formations by integrating spectral cameras, spectral recognition algorithms, adaptive filters, digital signal processing, deep learning models and multi-robot collaborative control mechanisms. It achieves rapid response and adaptive adjustment of robots to changes in ambient light, improves the accuracy, continuity and viewing experience of dance formations, and optimizes the collaboration between multiple robots through distributed control strategies, significantly enhancing the work efficiency and stability of the entire system.

[0063] Example 2

[0064] This embodiment is based on Embodiment 1, and describes in detail that due to the different lighting areas on the stage, the robot receives different dynamic changes in frequency, color and intensity of different lighting areas, and performs different dances;

[0065] On the stage, by setting up multiple different lighting zones, the dynamic changes of the lighting frequency, color and intensity in each zone are different. For example, one side of the stage may be set to soft blue light with low frequency and moderate intensity, while the other side may be set to high-frequency strong light with red tone. Each robot is equipped with a highly sensitive spectral camera that can capture the light signals of the area in real time. These spectral cameras can not only detect the color of light, but also perceive the frequency and intensity of light, ensuring that the robot can accurately obtain the lighting characteristics of the area in which it is located.

[0066] When the robot moves to different lighting areas, the light signals captured by the spectral camera will change accordingly. For example, when the robot moves from a soft blue lighting area to a high-frequency red strong light area, the light signal detected by the spectral camera will change from low-frequency, low-intensity blue light to high-frequency, high-intensity red light. These different light signals will be converted into electrical signals and further processed into digital signals for subsequent signal processing and action command generation. Therefore, each robot receives different light signals according to the lighting characteristics of the area it is in, thereby triggering different action commands.

[0067] In this way, robots in different areas can show different dance moves. For example, in a soft light area with a blue tone, the robot may perform slow and graceful movements, while in a high-frequency strong light area with a red tone, the robot may perform fast and powerful movements. This method of adjusting dance moves according to the changes in light signals in different lighting areas makes the dance performances on the entire stage more colorful. Robots in each area can show a unique dance style according to the guidance of the light, thereby realizing diversified dance choreography.

[0068] This embodiment sets up multiple different lighting areas on the stage, and the dynamic changes of the lighting frequency, color and intensity of each area are different. The high-sensitivity spectral camera equipped with the robot can capture and analyze the light signals of the area in real time. When the robot moves to different lighting areas, the light signals detected by the spectral camera change, thereby triggering different action instructions, so that the robot can show different dance movements according to the lighting characteristics of the area. This technology not only realizes the diversification of dance, but also ensures the coordination of the dance movements of the entire group, and enhances the visual effect and viewing of the performance.

[0069] Example 3

[0070] This embodiment is based on embodiment 2. Figure 3 As shown in the figure, if the robot group wants to achieve the same dance, the middle robot of the group dancing robot will receive the dynamic changes of the frequency, color and intensity of the light area by locking the middle robot of the group dancing robot. The other robots will turn off the spectral camera and connect the middle robot through the multi-robot collaborative control mechanism to achieve the same dance of the robots due to the different light areas.

[0071] When a group of robots needs to perform the same dance, this can be achieved by locking the middle robot in the group. The middle robot is equipped with a highly sensitive spectral camera to capture and analyze the dynamic changes in the frequency, color and intensity of light in the area in real time. The middle robot serves as the "eyes" of the entire group, responsible for receiving and processing light signals to ensure that all robots can perform the same dance moves according to the same set of light instructions.

[0072] Other robots will turn off their spectral cameras and no longer directly receive and process light signals. These robots will communicate with the middle robot through a multi-robot collaborative control mechanism. The middle robot will convert the received light signals into digital signals and extract feature instructions, and then transmit these instructions to other robots through wireless communication or other means. In this way, all robots can receive the same action instructions, ensuring the consistency and coordination of the actions of the entire group.

[0073] Through this centralized control method, even if the robots are distributed in different lighting areas, they can perform the same dance moves synchronously because all robots receive the same instructions from the middle robot; the middle robot is not only responsible for receiving and processing light signals, but also undertakes the task of command distribution, ensuring that each robot can execute action instructions in a timely and accurate manner, thereby achieving synchronous dance of the entire group.

[0074] In order to further improve the reliability and robustness of the system, real-time feedback and dynamic adjustment functions can be added to the multi-robot collaborative control mechanism. If a robot deviates when performing an action, it can be corrected in real time by the intermediate robot to ensure that the dance movements of the entire group are always consistent. This technical solution not only simplifies the complexity of the system, but also improves the coordination and viewing of the dance performance, allowing the robot group to perform the same dance movements in different lighting areas.

[0075] This embodiment locks the middle robot in the robot group to receive and process the dynamic changes of the frequency, color and intensity of the light area, and the other robots turn off the spectral camera and connect to the middle robot through the multi-robot collaborative control mechanism to receive the same action instructions. This technical solution ensures that all robots can synchronously perform the same dance movements even in different light areas, achieving the unity and coordination of the dance and improving the visual effect and viewing experience of the performance.

[0076] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any changes, modifications, replacements, integrations and parameter changes to these embodiments within the spirit and principles of the present invention through conventional substitutions or without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. A dance formation self-organizing control method based on robot groups, characterized in that: include: S1. Obtain the dynamic changes of light frequency, color and intensity through the spectral camera to form the original light signal; S2, using a spectrum recognition algorithm to determine effective light sources and invalid light sources, and removing invalid light sources through an adaptive filter to form a target light signal; S3, the target light signal is converted into an electrical signal by a photoelectric converter, the electrical signal is converted into a digital signal by a digital signal processor, and the digital signal feature is extracted by Fourier transform to form a feature instruction; S4, through the feature instruction parsing mechanism, the feature instruction is compared with the predefined action matching database in real time, and the robot action instruction is output; S5. Analyze the characteristics of the received digital signal through a deep learning model, predict the action instruction at the next moment, and generate the next action instruction in advance; S6. The predicted next action instruction is assigned to each robot through the multi-robot collaborative control mechanism, and the path planning algorithm is used to enable each robot to generate an action trajectory according to the received instruction; S7. Use a distributed control strategy to dynamically adjust the movement trajectory in real time, so that the robot can complete the dance choreography under the guidance of the light.

2. The dance formation self-organizing control method based on robot groups according to claim 1 is characterized in that: The spectral camera in S1 is located on the top of the robot, and forms the original light signal by acquiring the dynamic changes of the light frequency, color and intensity; the light signal color C, intensity I and frequency f are obtained by the spectral camera; the light signal color C formula is: Where E(λ) is the spectral radiance, X(λ), Y(λ) and Z(λ) are the CIE standard observer color matching functions, λ1 and λ2 represent the minimum and maximum response wavelengths of the spectral camera, and λ is the wavelength; the intensity I formula is: Where R(λ) is the response function of the spectral camera; the frequency f formula is: where c is the speed of light.

3. The dance formation self-organizing control method based on robot groups according to claim 1 or 2, characterized in that: The S2 specifically includes: using a spectrum recognition algorithm to determine effective light sources and invalid light sources, and determining the effectiveness of the light sources by calculating the spectrum similarity S, the formula is: Wherein, E(λ) is the spectral radiance, T(λ) is the preset effective light source spectrum template, λ1 and λ2 represent the minimum and maximum response wavelengths of the spectral camera; the spectral similarity S is calculated, and when S exceeds the preset threshold S th. , the light source is determined to be a valid light source; otherwise, it is determined to be an invalid light source.

4. The dance formation self-organizing control method based on robot groups according to claim 1 is characterized in that: The invalid light source is removed by the adaptive filter to form a target light signal. The adaptive filter performs adaptive filtering in an infinite approach manner, and the formula is: Wherein, x(n) is the original light signal, u(n) is the invalid light source signal, w(k) represents the weight coefficient of the adaptive filter, k is the weight coefficient index of the adaptive filter, and N represents the order of the filter; The adaptive filter iteratively updates the weight coefficient w(k) so that the filtered signal y(n) is as close to the target light signal as possible. The updating formula of the weight coefficient is: w(k+1)=w(k)+μ·e(n)·u(n), wherein e(n)=d(n)-y(n), e(n) is the error signal, d(n) is the desired pure light signal, μ represents the step size factor, and the adaptive filter removes invalid light source signals to form a pure target light signal G.

5. The dance formation self-organizing control method based on robot groups according to claim 1 is characterized in that: In S3, the target optical signal is converted into an electrical signal by a photoelectric converter, and then the electrical signal is converted into a digital signal by a digital signal processor, which specifically includes: the photoelectric converter converts the target optical signal G into an electrical signal V, and the formula is: V=A·G+B, where G represents the target optical signal, A represents the gain coefficient of the photoelectric converter, and B represents the bias coefficient; the electrical signal V is converted into a digital signal D by the digital signal processor through analog-to-digital conversion, and the formula is: D=ADC(V), where ADC represents the analog-to-digital conversion function, which converts the continuous electrical signal into a discrete digital signal.

6. The dance formation self-organizing control method based on robot groups according to claim 5 is characterized in that: The method of extracting digital signal features through Fourier transform of discrete digital signals to form feature instructions specifically includes converting digital signals into frequency domain signals using Fourier transform, and the formula is: Among them, D(t) is a digital signal, ω represents the angular frequency, j is an imaginary unit, and t is time. By analyzing the frequency domain signal F(ω), the frequency, intensity and color characteristics of the signal are extracted to form the feature instruction F f , the feature instruction F f For feature extraction, the formula is: F f =(|F(ω1)|,|F(ω2)|,…,|F(ω n )|), where ω1,ω2,…,ω n Indicates the frequency point, F(ω i ) indicates that at the frequency point ω i The signal amplitude at .

7. The dance formation self-organizing control method based on robot groups according to claim 1 or 6, characterized in that: In S4, the feature instruction F(ω) is compared with the predefined action matching database in real time through the feature instruction parsing mechanism to calculate the feature instruction F f The similarity S between the predefined action patterns sim : Among them, F f (i) represents the amplitude of the characteristic instruction at the i-th frequency point, A pre (i) represents the amplitude of the predefined action pattern at the i-th frequency point, n represents the number of frequency points, and the similarity S is calculated. sim , when S sim Exceeding the preset threshold S th When the most similar action mode is selected as the current robot action instruction A cmd ,The robot action instructions are output through the embedded control mechanism, and the corresponding actions are performed according to the feature instructions.

8. The dance formation self-organizing control method based on robot groups according to claim 1 or 6, characterized in that: The deep learning model in S5 is a long short-term memory network model. The long short-term memory network model is used to analyze the digital signal feature F f The deep learning model is used to analyze the characteristics of the received digital signal, predict the action instructions at the next moment, and generate the next action instructions in advance. The formula is: A n =LSTM(F f ,h t-1 ), where F f is the characteristic instruction, h t-1 is the hidden state at the previous moment, A n The predicted action instruction for the next moment.

9. The dance formation self-organizing control method based on robot groups according to claim 1, characterized in that: In S6, the predicted action instruction A is transmitted through the multi-robot collaborative control mechanism. n The distributed control strategy is assigned to each robot, and the path planning algorithm is used to generate the motion trajectory P of each robot. i (t), the path planning formula is: Among them, P i (t) represents the position of the ith robot at time t, P i (0) represents the initial position, v i (t) represents the speed of the i-th robot at time t; the path planning algorithm ensures that each robot will not collide when executing action instructions through collision avoidance and coordination of the robot group; the multi-robot collaborative control mechanism can effectively assign the predicted next action instruction to each robot and generate their respective action trajectories to ensure the coordination of the dance formation of the entire robot group.

10. The robot group-based dance formation self-organizing control method according to claim 9, characterized in that: The distributed control strategy is assigned to each robot, and the motion trajectory is adjusted dynamically in real time through the distributed control strategy, so that the robot can complete the dance choreography under the guidance of the light. The robot monitors its own position and speed in real time through the local controller, and exchanges information with other robots to make real-time dynamic adjustments. The adjustment formula is: P i (t+Δt)=P i (t)+v i (t)·Δt+α·∑ j≠i (P j (t)-P i (t)), where Δt is the time step, α is the adjustment coefficient, and P j (t) represents the position of other robots at time t; by adjusting the coefficient α, each robot can dynamically adjust its own motion trajectory according to the positions of other robots and its own position to ensure the coordination of the entire robot group; each robot also includes real-time updating of motion instructions according to the received light signal to ensure that the predetermined dance choreography is completed under the guidance of the light.

Images