A swimming control system for a biomimetic robotic ray

Abstract

This invention provides a swimming control system for a biomimetic robotic ray, comprising a shell, left and right pectoral fins located on opposite sides of the shell, and a caudal fin located at the tail of the shell. The left pectoral fin includes multiple first waterproof servos interconnected by connectors, serving as multiple joints of the left pectoral fin; the right pectoral fin includes multiple second waterproof servos interconnected by connectors, serving as multiple joints of the right pectoral fin; the caudal fin includes a third waterproof servo fixedly connected to the shell, serving as a joint of the caudal fin; wherein each joint is rotated by a two-level CPG control system. The two-level CPG network topology enables the biomimetic robotic ray to adapt to more complex rhythmic movements and reduces the time to reach the target phase and synchronization time during neuron coupling, improving the robustness of the biomimetic robotic ray's rhythmic movements and making it more realistically simulate the various movement modes of biological rays.

Description

Technical Field

[0001] This invention belongs to the field of biomimetic robotic fish motion control technology, and particularly relates to a swimming control system for a biomimetic robotic ray. Background Technology

[0002] With the continuous development of biomimetic robotic fish control technology, motion control methods based on Central Pattern Generators (CPGs) have become an important technical means widely used in the field of biomimetic robotic fish motion control. This method can achieve motion control of robotic fish by mimicking the rhythmic swimming of fish in nature, and has stable periodic signal output and good robustness. In the CPG-based motion control method, the robotic fish's motion pattern is realized by forming a stable periodic signal. The CPG network, as a central pattern generator, can drive the robotic fish to generate rhythmic swimming. By adjusting the phase relationship and frequency between neurons, CPG can control various rhythmic motion patterns of the robotic fish, such as body swaying, fin flapping, and swimming speed. This allows the robotic fish to more realistically simulate the movement behavior of real fish and has better motion performance and interaction capabilities.

[0003] However, CPG motion controllers in biomimetic robotic rays generally employ a traditional semi-centralized model, using only a single-layer neuron model to simulate movement rhythms, neglecting the optimization and improvement of the CPG network structure. The structure of a CPG network consists of neurons and the connections between them; different structures have varying impacts on the robotic ray's motion control capabilities. Adjusting the network structure can further improve the robotic ray's motion efficiency and performance. For example, changing the topology between neurons or introducing adaptive mechanisms holds promise for optimizing the network's motion control capabilities. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies by providing a swimming control system for a biomimetic robotic ray.

[0005] The present invention provides a swimming control system for a biomimetic robotic ray, comprising a shell, a left pectoral fin and a right pectoral fin located on both sides of the shell, and a tail fin located at the tail of the shell;

[0006] The left pectoral fin includes multiple first waterproof servos interconnected by connectors to serve as multiple joints of the left pectoral fin; the right pectoral fin includes multiple second waterproof servos interconnected by connectors to serve as multiple joints of the right pectoral fin; the caudal fin includes a third waterproof servo fixedly connected to the shell to serve as a joint of the caudal fin; wherein each joint is rotated by a dual-level CPG.

[0007] The dual-level CPG includes a rhythm layer and a pattern layer; the CPG neurons in the rhythm layer are used to control the phase between the joints of the left pectoral fin, the right pectoral fin, and the caudal fin; the CPG neurons in the pattern layer are used to control the phase difference between the joints of the left pectoral fin, the joints of the right pectoral fin, and the caudal fin; the rhythm signals generated by the dual-level CPG are used as the rotation angles of the first waterproof servo, the second waterproof servo, and the third waterproof servo, respectively.

[0008] Furthermore, all CPG neurons are coupled in a symmetrical bidirectional manner; the rhythm layer CPG neurons of each joint of the left and right pectoral fins are all in a chain-like topology; and the CPG neurons of the pattern layer are in a network topology.

[0009] Furthermore, the Kuramoto oscillator was used to model the neurons in the pattern layer:

[0010]

[0011] in, Let θ be the frequency of the phase change of the i-th oscillator; i ω is the phase angle of the i-th oscillator; i ω is the phase change angular frequency of the i-th oscillator; N is the total number of oscillators; w ij θ represents the coupling strength between the i-th oscillator and the j-th oscillator. j Let be the phase angle of the j-th oscillator; Let be the set phase difference between the i-th oscillator and the j-th oscillator; denoted as , where is the difference in phase angle between the i-th oscillator and the (i-th)-th oscillator; e is the natural constant; τ is the time constant of the output signal shielding function; σ is the center position of the output signal shielding function; t is the time variable; A is the amplitude constant; and γ is the phase difference constant.

[0012] Furthermore, when the biomimetic robotic ray swims in a straight line, the input parameters of the CPG network are set to have the same amplitude of waveform parameters for the left and right pectoral fins, the same frequency of waveform parameters for the left and right pectoral fins, and the phase difference between the amplitude and frequency of waveform parameters for the left and right pectoral fins are both π.

[0013] Furthermore, when the bionic robotic ray turns left, the input parameters for the CPG network are set such that the amplitude of the left pectoral fin waveform parameter is less than that of the right pectoral fin waveform parameter, and the frequency of the left pectoral fin waveform parameter is less than that of the right pectoral fin waveform parameter; when the bionic robotic ray turns right, the input parameters for the CPG network are set such that the amplitude of the left pectoral fin waveform parameter is greater than that of the right pectoral fin waveform parameter, and the frequency of the left pectoral fin waveform parameter is greater than that of the right pectoral fin waveform parameter.

[0014] Furthermore, when the biomimetic robotic ray dives, the input parameters of the CPG network are set such that the waveform parameters of the left and right pectoral fins have the same amplitude, the waveform parameters of the left and right pectoral fins have the same frequency, and the phase difference between the amplitude and frequency of the waveform parameters of the left and right pectoral fins are both π. Among these, the frequency of the biomimetic robotic ray when diving is greater than the frequency when swimming in a straight line, and the amplitude of the upward swing of the left and right pectoral fins is greater than the amplitude of the downward swing.

[0015] Furthermore, the rhythm layer comprises three CPG neurons; the pattern layer comprises two groups of neurons and a single neuron.

[0016] This invention provides a swimming control system for a biomimetic robotic ray, which adopts a two-level CPG network topology, enabling the biomimetic robotic ray to adapt to more complex rhythmic movements and reducing the time for neurons to reach the target phase and the synchronization time during coupling. This improves the robustness of the biomimetic robotic ray's rhythmic movements and makes it more realistically simulate the various movement modes of biological rays. Attached Figure Description

[0017] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the swimming control system for a biomimetic robotic ray provided in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the control structure of the biomimetic robotic ray provided in an embodiment of the present invention;

[0020] Figure 3 The output waveform diagram of the linear forward swimming algorithm of the biomimetic robotic ray provided in the embodiment of the present invention;

[0021] Figure 4 The output waveform diagram of the turning algorithm of the biomimetic robotic ray provided in the embodiment of the present invention;

[0022] Figure 5 The waveform diagram is the output of the diving algorithm of the biomimetic robotic ray provided in the embodiment of the present invention. Detailed Implementation

[0023] 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.

[0024] like Figure 1 As shown, this embodiment of the invention provides a swimming control system for a biomimetic robotic ray, including a shell 1, a left pectoral fin 2 and a right pectoral fin 3 located on both sides of the shell 1, and a tail fin 4 located at the tail of the shell 1.

[0025] The left pectoral fin 2 includes multiple first waterproof servos 22 interconnected by connectors 21, serving as multiple joints of the left pectoral fin 2; the right pectoral fin 3 includes multiple second waterproof servos 31 interconnected by connectors 21, serving as multiple joints of the right pectoral fin 3; the tail fin 4 includes a third waterproof servo 41 fixedly connected to the housing 1, serving as a joint of the tail fin 4; wherein each joint is rotated by a dual-level CPG. Exemplarily, the connector 21 is a U-shaped bracket, the first waterproof servos 22 are located in the middle of the concave side of the U-shaped bracket, and the output shaft of the first waterproof servo 22 is fixedly connected to the side of the U-shaped bracket. The rotation angle range of each servo is 0 degrees to 180 degrees.

[0026] The dual-level CPG includes a rhythm layer and a pattern layer. The CPG neurons in the rhythm layer are used to control the phase between the left pectoral fin joint 2, the right pectoral fin joint 3, and the caudal fin joint 4. The CPG neurons in the pattern layer are used to control the phase difference between the joints of the left pectoral fin 2, the joints of the right pectoral fin 3, and the caudal fin joint 4. The rhythm signals generated by the dual-level CPG are used as the rotation angles of the first waterproof servo motor 22, the second waterproof servo motor 31, and the third waterproof servo motor 41, respectively.

[0027] In the control process of the biomimetic robotic ray, the rhythmic movement patterns of a biological ray can be simulated by inputting corresponding rotation angle values ​​to the first waterproof servo 22, the second waterproof servo 31, and the third waterproof servo 41. Therefore, it is necessary to introduce CPG network control, using the rhythmic signals generated by the CPG as rotation angle inputs to the servos. The angle adjustment of these servos will enable the movement of the pectoral and caudal fins of the biomimetic robotic ray to simulate the swimming movements of a biological ray. By precisely controlling the movement of the servos, the biomimetic ray can achieve smooth transitions between different modes, thus exhibiting diverse movement characteristics in the underwater environment.

[0028] The rhythmic movements of organisms are actually subject to complex regulation by multi-level neural networks. These multi-level neural networks cooperate to form a highly complex system, with different levels of the network undertaking different motion control tasks. For the modal simulation and transformation of biomimetic manta rays, choosing a suitable CPG network topology is particularly important. Different neuron coupling methods and topologies affect the coupling transition, synchronization time, and adaptability to complex rhythmic movements of the CPG network. Therefore, establishing a CPG network with a multi-level neuron topology has a positive impact on achieving more diverse and realistic movement behaviors in biomimetic manta rays. Such a network model can better simulate the motion control methods of biological manta rays, improve the motion performance and adaptability of biomimetic manta rays, and enable them to achieve more complex rhythmic movements.

[0029] For example, such as Figure 2 As shown, when constructing a two-layer CPG network, the rhythm layer includes three CPG neurons. Figure 2 The darker areas (in the middle) are responsible for controlling the phase between the joint groups of the left and right pectoral fins and caudal fins, respectively. Each pectoral fin joint group contains 3 joints, while the caudal fin contains only 1 joint. The pattern layer includes 2 groups of neurons and 1 single neuron ( Figure 2 The light-colored portion corresponds one-to-one with the joint groups in the rhythm layer. To ensure the effectiveness of the double-layer CPG network, all CPG neurons are coupled symmetrically in both directions. The rhythm layer CPG neurons in the left pectoral fin joint 2 and the right pectoral fin joint 3 are chain-like topologies. The CPG neurons in the pattern layer are mesh-like topologies, which better mimic the function of neurons in living organisms and achieve precise control of the movement of the biomimetic robotic ray.

[0030] To mimic the function of neurons in living organisms, oscillators need to be introduced into CPG motion controllers to replace the function of biological neurons and generate periodic and controllable control signals. The Kuramoto oscillator has become a focus of research due to its excellent practicality. This oscillator can effectively describe the phase coupling relationship between various oscillators in the system, its parameters are easy to control, and the model can be flexibly adjusted to adapt to different research needs.

[0031] For example, the mathematical model of the rhythm layer neurons based on the Kuramoto oscillator is as follows:

[0032]

[0033] Modeling of pattern layer neurons using the Kuramoto oscillator:

[0034]

[0035] in, For θi The derivative with respect to time is the frequency of the phase change of the i-th oscillator; θ i ω is the phase angle of the i-th oscillator; i ω is the phase change angular frequency of the i-th oscillator; N is the total number of oscillators; w ij θ represents the coupling strength between the i-th oscillator and the j-th oscillator. j Let be the phase angle of the j-th oscillator; Let be the set phase difference between the i-th oscillator and the j-th oscillator; τ is the difference in phase angle between the i-th oscillator and the (i-1)-th oscillator; e is the natural constant; τ is the time constant of the output signal shielding function; σ is the center position of the output signal shielding function, determined according to the synchronization time of the output signal of the two-level CPG network; t is the time variable; A is the amplitude constant; γ is the phase difference constant.

[0036] Model parameter coupling strength w ij The angular frequency ω determines the synchronous convergence speed of CPG neurons; a larger coupling strength value will accelerate the convergence speed. i and phase difference These factors determine the frequency and phase relationship of the CPG neuron's output signal, respectively.

[0037] To simplify the model, the coupling strength w of each neuron can be... ij All are set to a uniform value and satisfy a specific phase difference condition. Then, the angular frequency and phase difference are modified according to different motion modes.

[0038] At this point, the phase angle waveform will initially generate a turbulent signal due to the CPG network not yet being synchronized. Without constraint, this may cause issues when used as a servo control signal due to θ. i The sudden change caused the servo to malfunction. To solve this problem, a sigmoid masking function needs to be added to the output signal for constraint. The mathematical model of the masking function is S(t) = (1 + e^(-t / t)). -τ(t-σ) ) -1 .

[0039] like Figure 3As shown, when controlling the biomimetic manta ray to swim in a straight line, considering that the amplitude and frequency of the left and right pectoral fin swings are basically the same, the input parameters of the CPG network are set to have the same amplitude and frequency for the waveform parameters of the left and right pectoral fins 2 and 3, respectively, and the phase difference between the amplitude and frequency of the waveform parameters of the left and right pectoral fins 2 and 3 are both π. Meanwhile, the main function of the caudal fin 4 is to maintain the balance of the fish body, so the amplitude and frequency of the caudal fin 4 are set to be relatively small. During the debugging process, the algorithm for the biomimetic manta ray to swim in a straight line was successfully implemented by stabilizing the output waveform. Furthermore, due to the left-right symmetry of the pectoral fins, the output waveform exhibits a symmetrical distribution along the x-axis.

[0040] like Figure 4 As shown, when controlling the biomimetic manta ray to turn, the amplitude and frequency of the left and right pectoral fins are different. When the manta ray turns to the left, the amplitude and frequency of the right pectoral fin's swing are larger, while the opposite is true when turning to the right. To make the turning effect more obvious, the effect of one side of the pectoral fin needs to be reduced; therefore, the amplitude and frequency of the swing of that side's pectoral fin are decreased. At the same time, the main function of the caudal fin 4 is to maintain the fish's balance; therefore, the amplitude and frequency of the caudal fin 4 are set to be smaller to ensure stability during the turning process.

[0041] When the bionic robotic ray turns left, the input parameters of the CPG network are set as follows: the amplitude of waveform parameter 2 of the left pectoral fin is less than the amplitude of waveform parameter 3 of the right pectoral fin, and the frequency of waveform parameter 2 of the left pectoral fin is less than the frequency of waveform parameter 3 of the right pectoral fin. When the bionic robotic ray turns right, the input parameters of the CPG network are set as follows: the amplitude of waveform parameter 2 of the left pectoral fin is greater than the amplitude of waveform parameter 3 of the right pectoral fin, and the frequency of waveform parameter 2 of the left pectoral fin is greater than the frequency of waveform parameter 3 of the right pectoral fin.

[0042] like Figure 5 As shown, when controlling the biomimetic robotic ray to dive, the amplitude and frequency of the swinging of the left and right pectoral fins are basically the same, but the frequency is higher when diving compared to swimming in a straight line. Therefore, the input parameters of the CPG network are set such that the waveform parameters of the left pectoral fin 2 and right pectoral fin 3 have the same amplitude, the waveform parameters of the left pectoral fin 2 and right pectoral fin 3 have the same frequency, and the phase difference between the amplitude and frequency of the waveform parameters of the left pectoral fin 2 and right pectoral fin 3 are both π. The frequency of the biomimetic robotic ray diving is greater than the frequency during straight-line swimming, and the amplitude of the upward swinging of the left pectoral fin 2 and right pectoral fin 3 is greater than the amplitude of the downward swinging. After the output waveform stabilizes, the waveform of one pectoral fin shifts to one side of the x-axis, making the amplitude of the upward swinging of the robotic ray's pectoral fin greater than the amplitude of the downward swing. This design allows the biomimetic robotic ray to sink quickly.

[0043] The final output waveform regarding the phase angle can be converted into PWM as the input of the corresponding servo motor through mapping or a control library.

[0044] The present invention has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present invention. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and embodiments of the present invention without departing from the spirit and scope of the invention, and all such modifications and improvements fall within the scope of the present invention. The scope of protection of the present invention is defined by the appended claims.

Claims

1. A biomimetic robotic ray swimming control system, comprising a shell (1), a left pectoral fin (2) and a right pectoral fin (3) respectively located on both sides of the shell (1), and a caudal fin (4) located at the tail of the shell (1), characterized in that, The left pectoral fin (2) includes multiple first waterproof servos (22) interconnected by connectors (21) to serve as multiple joints of the left pectoral fin (2); the right pectoral fin (3) includes multiple second waterproof servos (31) interconnected by connectors (21) to serve as multiple joints of the right pectoral fin (3); the tail fin (4) includes a third waterproof servo (41) fixedly connected to the shell (1) to serve as a joint of the tail fin (4); wherein each joint is rotated by a dual-level CPG. The dual-level CPG includes a rhythm layer and a pattern layer; the CPG neurons in the rhythm layer are used to control the phase between the joints of the left pectoral fin (2), the right pectoral fin (3), and the caudal fin (4); the CPG neurons in the pattern layer are used to control the phase difference between the joints of the left pectoral fin (2), the joints of the right pectoral fin (3), and the caudal fin (4); the rhythm signals generated by the dual-level CPG are used as the rotation angles of the first waterproof servo (22), the second waterproof servo (31), and the third waterproof servo (41), respectively. All CPG neurons are coupled in a symmetrical bidirectional manner; the rhythm layer CPG neurons of each joint of the left pectoral fin (2) and each joint of the right pectoral fin (3) are all chain-like topological structures; the CPG neurons of the pattern layer are network-like topological structures. CPG neurons in the pattern layer were modeled using the Kuramoto oscillator: ； in, For the first i The frequency of the phase change of each oscillator; θ i For the first i Phase angle of each oscillator; ω i For the first The phase change angular frequency of each oscillator; N This represents the total number of oscillators. w ij For the first i The oscillator and the first j The coupling strength between the oscillators; θ j For the first j Phase angle of each oscillator; φ ij For the first i The oscillator and the first j The phase difference between the groups of oscillators; φ i For the first i The oscillator and the first i -1 is the difference in phase angle between the oscillators; e It is a natural constant; τ The time constant of the output signal shielding function; σ The center position of the output signal shielding function; t It is a time variable; A It is the amplitude constant; γ It is the phase difference constant; The rhythm layer consists of three CPG neurons; the pattern layer consists of two groups of neurons and a single neuron.

2. The swimming control system for the biomimetic robotic ray according to claim 1, characterized in that, When the biomimetic robotic ray swims in a straight line, the input parameters of the CPG network are set to have the same amplitude of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3), the same frequency of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3), and the phase difference between the amplitude of waveform parameters for the left pectoral fin (2) and the phase difference between the frequency of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3). π .

3. The swimming control system for the biomimetic robotic ray according to claim 1, characterized in that, When the bionic robotic ray turns left, the input parameters of the CPG network are set as follows: the amplitude of the waveform parameter of the left pectoral fin (2) is less than the amplitude of the waveform parameter of the right pectoral fin (3), and the frequency of the waveform parameter of the left pectoral fin (2) is less than the frequency of the waveform parameter of the right pectoral fin (3); when the bionic robotic ray turns right, the input parameters of the CPG network are set as follows: the amplitude of the waveform parameter of the left pectoral fin (2) is greater than the amplitude of the waveform parameter of the right pectoral fin (3), and the frequency of the waveform parameter of the left pectoral fin (2) is greater than the frequency of the waveform parameter of the right pectoral fin (3).

4. The swimming control system for the biomimetic robotic ray according to claim 2, characterized in that, When the biomimetic robotic ray dives, the input parameters of the CPG network are set to have the same amplitude of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3), the same frequency of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3), and the phase difference between the amplitude of waveform parameters for the left pectoral fin (2) and the phase difference between the frequency of waveform parameters for the left pectoral fin (2) and the right pectoral fin (3). π When the bionic robotic ray dives, the swing frequency of the left pectoral fin (2) and right pectoral fin (3) is greater than that when it swims in a straight line. The amplitude of the upward swing of the left pectoral fin (2) and right pectoral fin (3) is greater than that of the downward swing.

Images