A fin drive type bionic chest fin spanwise and chordwise cooperative motion control method, device and medium

Abstract

The application discloses a fin strip driving type bionic chest fin spanwise and chordwise cooperative motion control method and device and medium, and the method comprises the following steps: the bionic chest fin is composed of fin roots, fin strips and fin surfaces; the first fin root spanwise fluctuation, the second and third fin root spanwise fluctuation and adduction-abduction coupled motion; three-dimensional cooperative motion of the chest fin in the spanwise direction, chordwise direction and adduction / abduction is realized through a rhythm signal generator; the motion state of the chest fin is obtained by using a state sensing module; a double-layer optimization strategy of inner-outer nesting is constructed based on a deep deterministic policy gradient algorithm; the local structure stiffness is optimized in the inner layer, and the motion parameters are optimized in the outer layer; the trajectory tracking controller is used to enable the chest fin driving mechanism to realize fast and stable target angle tracking, so that a sensing-decision-control framework is constructed; and the method can improve the propulsion performance and enhance the stability of the robotic fish in complex water areas, and solves the technical problems of efficient propulsion and stable control of the fin strip driving type bionic chest fin under unstructured working conditions.

Description

Technical Field

[0001] This invention relates to the field of biomimetic robot control, and in particular to a fin-driven biomimetic pectoral fin spanwise and chordwise coordinated motion control method, device and medium. Background Technology

[0002] In the field of underwater biomimetic robots, fish pectoral fins are widely regarded as an important propulsion structure for achieving high maneuverability and complex attitude adjustment. Compared with traditional propeller-driven methods, biomimetic pectoral fins have lower disturbance, higher fluid adaptability and more degrees of freedom, and therefore have broad application prospects in underwater exploration, environmental monitoring and intelligent inspection. However, most common biomimetic pectoral fins currently use a single fin structure with a relatively simple motion mode, making it difficult to achieve stable attitude control while ensuring propulsion efficiency, especially under unstructured flow fields and complex working conditions.

[0003] Existing research indicates that, in order to improve the hydrodynamic performance and control capability of pectoral fins, researchers have proposed a multi-fin coordinated wave mechanism, which achieves spanwise and chordwise wave action on the fin surface through the phase difference between multiple fins, thereby improving thrust direction control and attitude response speed. However, current control methods still face the following challenges: (1) how to construct a composite motion model that combines spanwise and chordwise wave action; (2) how to achieve coordinated control and real-time optimization among multiple fins in complex environments; and (3) how to ensure that the control output has sufficient smoothness and robustness to avoid the impact of signal jumps on the actuator.

[0004] Furthermore, existing research lacks a modeling and optimization mechanism for the coupling relationship between the structural stiffness and control parameters of biomimetic pectoral fins. As a flexible driving component, the stiffness variation of the fin rays directly affects the feasibility of the motion trajectory and the stability of the force output. Therefore, there is an urgent need for a method that can combine motion control parameters and structural stiffness parameters for joint optimization, along with real-time state perception and intelligent control algorithms, to achieve a more efficient and adaptable biomimetic pectoral fin motion control scheme. Summary of the Invention

[0005] To address the problems of insufficient control precision, low coordination efficiency, and unstable structural response in existing technologies for pectoral fins, this invention proposes a fin-driven biomimetic pectoral fin spanwise and chordwise coordinated motion control method, device, and medium. This method achieves efficient spatiotemporal coordinated oscillation of the fin surface by precisely controlling the spanwise undulation, chordwise phase, and inward and outward movements of the fin root, thereby significantly improving the propulsion performance and attitude stability of the multi-fin driven pectoral fin, and ultimately enhancing the overall swimming performance and environmental adaptability of the aircraft.

[0006] The purpose of this invention is to provide a fin-ray driven bionic pectoral fin spanwise and chordwise coordinated motion control method, device and medium. The method includes the following steps: (1) Based on the skeletal structure of fish pectoral fins, a fin-ray driven bionic pectoral fin is designed, consisting of independently driven fin roots, elastic fin rays and fin surfaces. The spanwise undulation of the fin surface is formed by the active undulation of the fin roots and the passive undulation of the fin rays. The phase difference of the undulation between the fin roots forms the chordwise undulation of the fin surface. The deflection phase difference between the fin roots forms the adduction and abduction motion of the fin surface. At the same time, based on the pectoral fin gait characteristic data, a fin root kinematic model is constructed, and c(t) is defined as the fin root motion control parameter; (2) In order to make the fin root kinematic model continuously output a smooth, continuous and rhythmic control signal, an improved central pattern generator based on a Hopf oscillator is used. The system uses CPGs (procedure generators) as rhythm signal generators. These generators are matched with the input parameters of the fin root kinematic model. By changing the fin root motion control parameter c(t), the rhythm signal generators process the output of the fin root kinematic model to output a smooth fin root angular displacement. This achieves the three-dimensional coordinated motion of the fin-ray driven biomimetic pectoral fin in the spanwise, chordwise, adduction, and abduction directions. The smooth fin root angular displacement is defined as... (3) During the movement of the pectoral fin, the combined thrust and torque of the pectoral fin are collected in real time by mounting the pectoral fin on a six-dimensional force sensor, and the local fluid load P is obtained by using a strain gauge array. i rays An angle encoder is used to obtain the actual deflection angle displacement of the fin root during motion. Using a bending sensor to collect the actual angular displacement of the fin root during motion and fin passive wave angular displacement The actual deflection angle displacement of the fin root and the actual undulation angle displacement of the fin root are used to measure these parameters. and the passive wave angular displacement of the fin rays The linear displacement and equivalent stiffness K of the fins were calculated. i The actual angular displacement of the fin root is defined as... (4) Construct inner and outer optimization strategies for reinforcement learning based on the deep deterministic policy gradient (DDPG) algorithm; the outer optimization strategy adopts a weighted combination of maximizing propulsion performance and maximizing attitude stability as the optimization objective, and optimizes to obtain the optimal fin root motion control parameters. To achieve a balance between propulsion performance and attitude stability, and enhance the system's adaptability to environmental changes, the inner-layer optimization strategy establishes an equivalent stiffness adjustment model based on the structural response characteristics of each fin, and dynamically adjusts the equivalent stiffness K through nested optimization strategies. i, so that the fin strip structure responds more closely to the optimal dynamic deformation state, thereby improving structural compliance; (5) To ensure accurate tracking of the fin root expected angular displacement generated by the outer layer optimization strategy, a trajectory tracking controller based on model predictive control (MPC) method is introduced, which is constructed on the basis of outer layer control trajectory and inner layer equivalent stiffness adjustment, based on trajectory prediction and rolling optimization mechanism, receives the error between the current fin root actual angular displacement and the fin root expected angular displacement , predicts future state evolution, and solves the optimal control input , so as to realize high-precision, low-overshoot and strong robustness trajectory tracking execution.

[0007] As a further technical solution, the fin strip driving bionic pectoral fin is composed of independently driven fin roots, passively deformed elastic fin strips and passively deformed flexible fin surfaces, wherein the first fin root realizes spanwise fluctuation, the second and third fin roots further have the coupling motion ability of adduction and abduction on the basis of spanwise fluctuation, the fin root active fluctuation and the fin strip passive fluctuation realize the spanwise fluctuation of the whole fin surface, the fluctuation phase difference realizes the chordwise fluctuation of the whole fin surface, and the deflection phase difference realizes the adduction and abduction motion of the whole fin surface.

[0008] The equation of the fin root kinematic model is:

[0009]

[0010]

[0011] In the formula, α i (t) is the fluctuation angular displacement of the i-th fin root at t, is the amplitude of the fluctuation angular displacement of the i-th fin root, is the bias of the fluctuation angular displacement of the i-th fin root, r i 1 ,r i 2 ,r i 3 ,r i 4 is the fin root fluctuation motion period ratio, which divides the period into four parts, r i 1 +r i 2 +r i 3 +r i 4 =1, T α is the fluctuation motion period; β i(t) is the i-th fin root deflection angle displacement at time t, is the amplitude of the i-th fin root deflection angle displacement, is the bias of the i-th fin root deflection angle displacement, r i is the fin root deflection motion period ratio, dividing the period into two parts, 0 < r i < 1, T β is the deflection motion period, t is time; the fin root motion control parameters i takes the values 1, 2, 3, the elastic fin strip relies on the elastic properties of its own material to generate fluctuations under the active drive of the fin root, and drives the flexible fin surface to achieve passive spanwise fluctuations.

[0012] As a further technical solution, the rhythm signal generator is constructed by the improved central pattern generator and the fin root kinematics model, and the equation of the rhythm signal generator is as follows:

[0013]

[0014] In the formula, is the α i (t) output, is the output of the rhythm signal generator about α i (t), is the β i (t) output after smoothing by the rhythm signal generator, is the output of the rhythm signal generator about β i (t), α k β is the response frequency (the greater the response, the faster the response), is the damping ratio (controls the degree of smoothing), t is time, and the smoothed fin root angle displacement is

[0015] The rhythm signal generator not only has good approximation ability to the original control signal, but also has the ability to maintain the rhythm, smoothness and differentiability of the trajectory.

[0016] As a further technical solution, the collection and processing process of the pectoral fin thrust and moment, the local fluid load, the fin root actual fluctuation angle displacement, the fin strip passive fluctuation angle displacement, the fin root actual deflection angle displacement and the fin strip motion line displacement are as follows: first, through the six-dimensional force sensor installed above the pectoral fin base, the pectoral fin surface in three-axis direction is collected in real time. The pectoral fin thrust and moment are recorded as τ global = [F x , F y , F z , M x , My M z ], F x ,F y ,F z M represents the thrust of the pectoral fin surface in the x, y, and z directions, respectively. x M y M z These are the torques on the pectoral fin surfaces about the x, y, and z axes, respectively.

[0017] Simultaneously, strain gauge arrays were deployed at the fin root, middle section of the fin, and fin tip to measure the distribution of the local fluid load as P. i rays =[P i root ,P i mid ,P i end ], P i root ,P i mid ,P i end Let be the fluid loads at the root, middle, and end of the i-th fin, respectively. Using quadratic interpolation, the fin fluid pressure function along the fin length is constructed as follows:

[0018] f i (l)=al 2 +bl+c,l∈[0,L i ];

[0019] In the formula, L i Let be the length of the fin installed at the i-th fin root, where a, b, and c are constants. This length is obtained using fluid load data from the fin root, the middle section of the fin, and the end of the fin. The fluid pressure at the end of the i-th fin is... The fluid torque acting on the end of the i-th fin is

[0020] Secondly, the actual deflection angle displacement of the fin root is collected using an angle encoder. The actual angular displacement of the fin root is obtained by a flexible bending sensor. and the passive wave angular displacement of the fin rays Based on the relationship between the geometric model and the material response, the linear displacement of the fin tip is calculated as follows:

[0021]

[0022] In the formula, Let be the end displacement of the i-th fin in the normal direction at time t. Let L be the tangential end displacement of the i-th fin at time t. iLength of the i-th fin strip installed on the fin root;

[0023] The equivalent stiffness K i The calculation formula is

[0024]

[0025] In the formula, K i The equivalent stiffness of the i-th fin strip (the greater the more difficult to bend), The passive wave angle displacement of the i-th fin strip after being subjected to fluid torque, i takes the value of 1, 2, 3;

[0026] Finally, the above data are uniformly sampled, filtered, normalized and feature fusion processed by the state perception module to form a state vector:

[0027]

[0028] In the formula,

[0029] The state vector is used as the environmental observation input of the outer optimization strategy for evaluating the propulsion performance and attitude stability of the system; meanwhile, wherein P i rays The data are also synchronously transmitted to the inner optimization strategy for dynamically correcting the equivalent stiffness of the fin strip.

[0030] As a further technical solution, the outer optimization strategy maximizes the propulsion performance and the attitude stability as the outer optimization target, and learns the optimal fin root motion control parameter The superscript "~" only represents the optimal value of the motion control parameter, without changing the meaning represented by the letter;

[0031] The equation of the outer optimization target is:

[0032]

[0033] In the formula, η thrust The effective thrust generated by the unit power consumption, P is the power consumption, ξ yaw Indicates the attitude disturbance of the pectoral fin in the yaw degree of freedom, ξ pitch Indicates the attitude disturbance of the pectoral fin in the pitch degree of freedom, Var() indicates the variance function, w1, w2, w3 are weight coefficients, when ξ yaw , ξ pitch , the attitude stability is better;

[0034] The inner layer optimization strategy maximizes the structural compliance in the fin root motion process under the premise of keeping the optimal fin root motion control parameters unchanged, adjusts the equivalent stiffness K i The actual deformation trajectory under fluid load has good structural compliance;

[0035] The equation of the inner layer optimization target is:

[0036]

[0037] In the formula, the first term measures the passive bending response of the fin under the action of fluid, controls the compliance deformation degree, and the second term measures the smoothness of the fin motion curve (limits the adjustment of equivalent stiffness too frequent or too intense to avoid material fatigue or uncontrollable deformation); δ i (t, K i ) is the passive fluctuation angular displacement of the i-th fin when the equivalent stiffness is K i is the normal end displacement of the i-th fin when the equivalent stiffness is K i , w4 and w5 are weighting coefficients of the two parts, T is the motion period of the pectoral fin, and t is time;

[0038] The outer layer optimization strategy realizes high-level decision control of the fin root motion behavior target, and the inner layer optimization strategy ensures the compliance of the fin dynamic response by adjusting the equivalent stiffness. The two strategies work together to effectively improve the environmental adaptability, posture stability and propulsion performance of the bionic pectoral fin control system under dynamic working conditions.

[0039] As a further technical solution, the trajectory tracking controller uses a model predictive control method to ensure accurate tracking of the fin root desired angular displacement Based on the trajectory prediction and rolling optimization mechanism, the fin root desired angular displacement and the fin root actual angular displacement are used to calculate the displacement error, construct a dynamic evolution model within the prediction window, minimize the sum of squares of displacement errors and minimize the weighted sum of control inputs, solve the problem in real time, output a series of control inputs in the future, and obtain the optimal control input in each control period to achieve accurate tracking of the target trajectory;

[0040] The equation of the trajectory tracking controller is:

[0041]

[0042] In the formula, is the i-th fin root desired angular displacement at the prediction step k at time t, ​ is the actual angular displacement of the i-th fin root at the prediction step k, is the optimal control input of the i-th fin root at the prediction step k, the optimal control input is obtained in each control period The superscript '' only represents the optimal value, and does not change the meaning represented by the letter; Q is an error weight matrix, R is a control cost matrix, N p is the prediction step; the trajectory tracking controller is used to enable the driving mechanism to achieve fast, smooth and no overshoot target angle tracking response, thereby constructing a complete perception-decision-control closed loop path.

[0043] As a further technical solution, the fin strip driving type bionic pectoral fin is made of a rigid fin bone, a rigid fin root, an elastic fin strip and a flexible fin membrane; the fin bone and the fin root are integrally formed through a 3D printing process and are made of a material with good structural strength and forming precision; the fin strip is selected from a material with high elasticity and light weight characteristics to ensure good deformation response capability under fluid load; and the fin membrane is made of a flexible and deformable material to enhance its adaptability and fluctuation performance in response to changes in hydrodynamic force.

[0044] The application also provides a fin strip driving type bionic pectoral fin spanwise and chordwise cooperative motion control device, characterized by comprising: (1) a driving control module for outputting control signals of the rhythm signal generator and driving the actuator at the fin root to perform corresponding motion; (2) a state sensing module for collecting state vectors at the first time in pectoral fin motion through sensors, wherein the state variables include the pectoral fin combined thrust and torque, the local fluid load, the fin root actual fluctuation angular displacement, the fin strip passive fluctuation angular displacement and the fin root actual deflection angular displacement, for real-time calling by a feedback adjustment module; (3) a signal processing module for generating fin root motion control parameters through calculation of an outer optimization strategy, an inner optimization strategy and a trajectory tracking controller; (4) a feedback adjustment module for comparing data results of the state sensing module with current control parameters, generating dynamic correction information, triggering the outer optimization strategy and the inner optimization strategy to update the action strategy again through an interruption mechanism, and ensuring that the fin root tracks the fin root expected angular displacement through the trajectory tracking controller.

[0045] The application also provides a computer readable storage medium containing motion control program instructions, characterized in that the motion control program instructions are stored in the computer readable storage medium composed of a Raspberry Pi as a host and a field programmable logic gate array processing unit as a slave, and the computer readable storage medium is used to realize any one of the above fin strip driving type bionic pectoral fin spanwise and chordwise cooperative motion control methods when the motion control program instructions are executed.

[0046] Additional aspects and advantages of the present application will be partially given in the following description, partially will become apparent from the following description, or will be learned by practice of the present application.

[0047] The beneficial effects of the present application are:

[0048] (1) The optimized fin drive type bionic pectoral fin spanwise and chordwise cooperative motion control method is suitable for a bionic pectoral fin comprising at least three independently drivable fin strips, a single spanwise fluctuation control is implemented on the first fin strip; a spanwise fluctuation and adduction and abduction motion control is implemented on the second fin strip, the spanwise fluctuation has a set phase difference relative to the first fin strip; a similar control strategy as the second fin strip is implemented on the third fin strip, the spanwise fluctuation has different set phase differences relative to the first and second fin strips; the chordwise fluctuation of the fin surface as a whole is formed according to the phase differences of the spanwise fluctuations of the three fin strips.

[0049] (2) In order to achieve a balance between fluctuation performance and structural stability, the three fin strips are preferably made of high-elastic and light-weight materials, such as thermoplastic polyurethane elastomer or super-elastic alloy material, to ensure sufficient frequency response characteristics and deformation recovery capability.

[0050] (3) According to the gait characteristic data of the pectoral fin, a kinematics model of the fin root is constructed, so that the pectoral fin has good bionic characteristics in terms of motion function. The introduction of the rhythm signal generator not only enables effective approximation of the original control signal, but also maintains the rhythm, smoothness and derivability of the output trajectory on this basis, providing more stable and natural motion driving signals for bionic control.

[0051] (4) The outer layer optimization strategy adopts a weighted combination of maximum propulsion performance and maximum attitude stability as the optimization objective, and optimizes the optimal fin root motion control parameters to achieve a balance between propulsion performance and attitude stability, and enhance the adaptability of the system to environmental changes.

[0052] (5) The inner layer optimization strategy establishes an equivalent stiffness adjustment model according to the structural response characteristics of each fin strip, and dynamically adjusts the equivalent stiffness through a nested optimization strategy, so that the structural response of the fin strip is closer to the optimal dynamic deformation state, thereby improving the structural compliance.

[0053] (6) The present application introduces a trajectory tracking controller established based on the model predictive control method, so that the driving mechanism realizes fast, smooth and undamped precise adjustment and tracking response, thereby constructing a complete perception-decision-control closed loop path.

[0054] (7) According to the method of the present application, the spanwise fluctuation, chordwise phase and adduction and abduction motion of the fin root are precisely controlled, the efficient cooperative fluctuation of the fin surface in space and time is realized, thereby significantly improving the propulsion performance and attitude stability of the multi-fin strip driven pectoral fin, and enabling it to have stable operation capability. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 Flow chart of the spanwise and chordwise cooperative fluctuation control method of the fin strip driven bionic pectoral fin of the embodiment of the present application.

[0056] Figure 2 Stereoscopic structure diagram of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0057] Figure 3 The figure is a fin root fluctuation angle displacement control signal diagram of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0058] Figure 4 The figure is a fin root deflection angle displacement control signal diagram of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0059] Figure 5 The figure is a fin root fluctuation smoothing control schematic diagram of the bionic pectoral fin rhythm signal generator in the embodiment of the present application.

[0060] Figure 6 The figure is a fin root deflection smoothing control schematic diagram of the bionic pectoral fin rhythm signal generator in the embodiment of the present application.

[0061] Figure 7 The figure is a sensor arrangement and collection schematic diagram in the bionic pectoral fin sensing system in the embodiment of the present application.

[0062] Figure 8 The figure is an outer layer optimization strategy flow chart in the spanwise and chordwise cooperative fluctuation process of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0063] Figure 9 The figure is an inner layer optimization strategy flow chart in the spanwise and chordwise cooperative fluctuation process of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0064] Figure 10 The figure is a DDPG network training flow chart in the spanwise and chordwise cooperative fluctuation process of the fin strip driven bionic pectoral fin in the embodiment of the present application.

[0065] Figure 11 The figure is an inner-outer layer optimization strategy relationship diagram of the spanwise and chordwise cooperative fluctuation of the bionic pectoral fin in the embodiment of the present application.

[0066] Figure 12 The figure is a control block diagram of the bionic pectoral fin trajectory tracking controller in the embodiment of the present application.

[0067] Figure 13 The figure is a structural schematic diagram of the pectoral fin spanwise and chordwise cooperative fluctuation control device in the embodiment of the present application. DETAILED DESCRIPTION

[0068] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be clearly and completely described below with reference to the drawings in the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments; based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the protection scope of the present application.

[0069] Please refer to the accompanying Figures 1 to 13 The embodiment of the present application provides a fin strip driving type bionic pectoral fin spanwise and chordwise cooperative motion control method, device and medium, which comprises the following steps: (1) based on the skeletal structure of fish pectoral fin, a fin strip driving type bionic pectoral fin composed of independently driven fin roots, elastic fin strips and fin surfaces is designed, the spanwise undulation of the fin surface is formed by the active undulation of the fin roots and the passive undulation of the fin strips, the undulation phase difference between the fin roots forms the chordwise undulation of the fin surface, and the deflection phase difference between the fin roots forms the adduction and abduction motion of the fin surface; at the same time, according to the gait characteristic data of the pectoral fin, a kinematic model of the fin root is constructed, and c(t) is defined as the fin root motion control parameter; (2) in order to make the kinematic model of the fin root continuously output smooth, continuous and rhythmic control signals, a central pattern generator based on a Hopf oscillator is used as a rhythm signal generator; the rhythm signal generator is matched with the input parameters of the kinematic model of the fin root, the output of the kinematic model of the fin root is processed by the rhythm signal generator after the fin root motion control parameter c(t) is changed, and then the smooth fin root angular displacement is output, so that the three-dimensional cooperative motion of the spanwise, chordwise, adduction and abduction of the fin strip driving type bionic pectoral fin is realized, and the smooth fin root angular displacement is defined as (3) during the motion of the pectoral fin, the pectoral fin is installed on a six-dimensional force sensor to collect the thrust and moment of the pectoral fin in real time, a strain gauge array is used to obtain the local fluid load P i rays an angle encoder is used to obtain the actual deflection angular displacement of the fin root during motion a bending sensor is used to collect the actual undulation angular displacement of the fin root during motion and the passive undulation angular displacement of the fin strip the fin strip motion line displacement and the equivalent stiffness K i are calculated by using the actual deflection angular displacement of the fin root the actual undulation angular displacement of the fin root and the passive undulation angular displacement of the fin strip , and the actual angular displacement of the fin root is defined as (4) the inner layer optimization strategy and the outer layer optimization strategy of reinforcement learning are constructed based on a deep deterministic policy gradient algorithm; the outer layer optimization strategy adopts a weighted combination of maximum propulsion performance and maximum attitude stability as the optimization objective, and the optimal fin root motion control parameter is obtained by optimization achieve the balance between propulsion performance and attitude stability, and enhance the adaptability of the system to environmental changes; the inner layer optimization strategy establishes an equivalent stiffness adjustment model for the structural response characteristics of each fin strip, and dynamically adjusts the equivalent stiffness K through the nested optimization strategy i , so that the structural response of the fin strip is closer to the optimal dynamic deformation state, thereby improving the structural compliance; (5) To ensure accurate tracking of the fin root expected angular displacement generated by the outer layer optimization strategy , a trajectory tracking controller based on the model predictive control method is introduced, which is constructed on the basis of the outer layer control trajectory and the inner layer equivalent stiffness adjustment, based on the trajectory prediction and rolling optimization mechanism, receives the error between the current fin root actual angular displacement and the fin root expected angular displacement , predicts the future state evolution, and solves the optimal control input by rolling , so as to achieve high-precision, low-overshoot, and strong robustness trajectory tracking execution.

[0070] The specific embodiments of the present application are further described below.

[0071] Referring to Figure 1 , the method and device are based on the structural characteristics of the fish pectoral fin, and a fin strip driven bionic pectoral fin mechanism is designed, and a three-dimensional cooperative control system integrating structure driving, rhythm generation, state sensing, control optimization and execution feedback is constructed on this basis. The system consists of three independently driven fin strips, a rhythm signal generator, a state sensing module, a double-layer DDPG reinforcement learning optimization controller and an MPC execution controller, and forms a complete sensing-decision-execution closed loop path between the functional modules, realizing the coordinated control and efficient driving of the bionic pectoral fin in complex hydrodynamic environment.

[0072] S1、Referring to Figure 2 , the fin strip driven bionic pectoral fin structure provided in the embodiment of the present application comprises a rigid fin bone 1, a rigid fin root 2, an elastic fin strip 3 and a flexible fin membrane 4; the rigid fin bone 1 and the rigid fin root 2 are made of polylactic acid (PLA) material and are integrally formed by three-dimensional printing process, which plays the role of supporting and connecting the movement of the pectoral fin structure; the fin strip 3 is made of super-elastic shape memory alloy sheet material, which has a large deformation capacity and excellent structural compliance, and is used to drive the pectoral fin to perform multi-degree-of-freedom undulating motion; the fin membrane 4 is made of hydrogel material with good flexibility, and is arranged between the fin strips, which is used to maintain the continuity of the fin surface and enhance the hydrodynamic coupling effect.

[0073] Specifically, the first fin root 201 can only perform a spanwise undulatory motion, the second fin root 202 and the third fin root 203 further have the ability of adduction and abduction motion on the basis of the spanwise undulation; the undulation of the fin roots drives the passive undulation of the fin strips to form the spanwise undulation of the flexible fin membrane 4, the phase difference of the undulation of each fin root 3 constitutes the chordwise undulation on the flexible fin membrane 4, and the phase difference of the deflection of each fin root 3 realizes the adduction and abduction of the fin surface, so as to realize the three-dimensional dynamic propulsion of the pectoral fin in space;

[0074] Specifically, since the fin strip 3 has the thin plate structure characteristic, when the fin root 2 generates the undulatory motion, the fin strip 3 can obviously bend in the normal direction; and when the fin root 2 performs the deflection motion, the fin strip 3 mainly rotates in the tangential direction, the structural rigidity is large, and almost no bending deformation occurs;

[0075] Specifically, in order to realize the spanwise, chordwise, adduction and abduction motion control of the fin strip driving type bionic pectoral fin, the fin root kinematics model is constructed in combination with the gait characteristics and bionic law of the pectoral fin of fish; first, the fin root undulation angular displacement α i (t) is modeled by using the following piecewise cosine function:

[0076]

[0077] In the formula, α i (t) is the undulation angular displacement of the i-th fin root at t, is the amplitude of the undulation angular displacement of the i-th fin root, is the bias of the undulation angular displacement of the i-th fin root, r i 1 ,r i 2 ,r i 3 ,r i 4 is the undulation motion period ratio of the i-th fin root, the motion period is divided into four parts, r i 1 +r i 2 +r i 3 +r i 4 =1, T α is the undulation motion period, and t is time;

[0078] Specifically, referring to Figure 3 , when T α =2, r1 1 =0.15, r1 2 =0.35, r1 3 =0.35, r1 4 =0.15, the 1st, 2nd and 3rd fin root fluctuation angular displacement.

[0079] Secondly, the fin root deflection angular displacement β i (t) is modeled as follows:

[0080]

[0081] where β i (t) is the i-th fin root deflection angular displacement at time t, is the amplitude of the i-th fin root deflection angular displacement, is the bias of the i-th fin root deflection angular displacement, r i is the i-th fin root deflection motion period ratio, dividing the period into two parts, 0 < r i < 1, T β is the deflection motion period, and t is time;

[0082] Specifically, refer to Figure 4 when T β = 2, r1 = 0.3, r2 = 0.5, r3 = 0.8, the 1st, 2nd and 3rd fin root deflection angular displacement.

[0083] Further, the fin root motion control parameters i is 1, 2, or 3; the elastic fin strip relies on the elastic properties of its own material to generate fluctuations under the active drive of the fin root, and drives the flexible fin surface to achieve passive spanwise fluctuations.

[0084] S2, in order to make the fin root kinematics model continuously output smooth, continuous, rhythmic and high biological simulation control signals during execution, an improved central pattern generator based on Hopf oscillator is further introduced as a rhythm signal generator to dynamically modulate and real-time smooth the control trajectory.

[0085] Specifically, the rhythm signal generator is based on a Hopf-type oscillation equation, and a double-channel oscillator structure composed of fin root fluctuation angular displacement and fin root deflection motion angular displacement is constructed; the core goal of this structure is to dynamically respond to and second-order filter the fin root control signal, avoid the discontinuous execution phenomenon caused by angle mutation, and enhance the controllability, smoothness and actuator response stability of the control command;

[0086] Specifically, the central pattern generator is embedded into the fin root kinematics model control path, so that the input of the rhythm signal generator is the fin root angular displacement α i (t) and β i (t), which are defined as follows:

[0087]

[0088] wherein, is the α i (t) output after smoothing by the metronome signal generator, is the output of the metronome signal generator with respect to α i (t), is the β i (t) output after smoothing by the metronome signal generator, is the output of the metronome signal generator with respect to β i (t), k α ,k β is the response frequency (the greater the response, the faster), is the damping ratio (controls the degree of smoothing), t is time, and the smoothed fin root angular displacement is i takes the values 1, 2, 3;

[0089] Specifically, by changing the fin strip motion control parameter c(t), the control signal output by the metronome signal generator is the same as the control signals α i (t), β i (t) of the fin strip kinematic model, achieving three-dimensional coordinated motion of the fin strip driving type bionic pectoral fin in the spanwise, chordwise, adduction and abduction directions;

[0090] Specifically, in order to evaluate the effect of the metronome signal generator as a smoother in the generation of periodic signals, reference is made to Figure 5 When the fin root fluctuation parameter is set to T α = 2 in the first cycle, r1 1 = 0.15, r1 2 = 0.35, r1 3 = 0.35, r1 4 = 0.15, and when the fin root fluctuation parameter is changed to T α = 2 in the second cycle, r1 1 = 0.25, r1 2 = 0.25, r1 3 = 0.15, r1 4 = 0.35, As can be seen from the local enlarged view, the control signal after processing by the metronome signal generator is smoother than the original control signal αi The output of (t) is smoother; see Figure 6 When the fin root deflection parameter is set as T β = 2, When the fin root deflection parameter is changed as T β = 2, When r1=0.8, r2=0.4, r3=0.6, it can be seen from the local enlarged view that the control signal processed by the rhythm signal generator is smoother than the original control signal β i (t).

[0091] Further, the module is located between the trajectory generation layer and the execution layer in the actual control path, can be seamlessly coupled with the reinforcement learning strategy and the state perception module, and outputs a three-channel smooth signal with rhythm characteristics and are used to drive the synchronous control of the three fins in the spanwise, chordwise, adduction and abduction directions, respectively, to build a multi-fin strip collaborative propulsion mechanism with physical continuity and neural rhythm characteristics.

[0092] Further, the above rhythm signal generator not only has good approximation ability to the input signal of the fin strip kinematics model, but also can significantly improve the rhythm, smoothness and dynamic derivability of the output angle trajectory, and is particularly suitable for processing segmented definition functions, trapezoidal waves, cosine superposition and other non-continuous or edge mutation control signals.

[0093] S3, in the process of pectoral fin movement, in order to realize the comprehensive perception and accurate observation of the propulsion state of the system, an integrated multi-source sensor perception system is constructed; the system acquires multi-dimensional physical quantity data of the pectoral fin in the movement process by reasonably arranging sensors on key structure parts, and then provides high-quality input required for state estimation for the controller; among them, the pectoral fin combined thrust and moment are collected in real time by a six-dimensional force sensor, the local fluid load P i rays is obtained by using a strain gauge array, the actual deflection angle displacement of the fin root is obtained by using a bending sensor, the actual fluctuation angle displacement of the fin root and the passive fluctuation angle displacement of the fin strip are obtained, the fin strip movement line displacement and equivalent stiffness K i are calculated by the actual deflection angle displacement of the fin root the actual fluctuation angle displacement of the fin root and the passive fluctuation angle displacement of the fin strip ;

[0094] For details, please refer to Figure 7 The sensing system consists of a six-dimensional force sensor, a strain gauge array, an angle encoder, and a bending sensor.

[0095] To elaborate further, the data acquisition and processing flow of the sensing system is as follows:

[0096] Furthermore, by mounting the bionic pectoral fin below a six-dimensional force sensor, the resultant thrust and resultant torque acting on the pectoral fin surface in the three-axis directions are collected in real time during movement. The measured resultant thrust and torque vectors are:

[0097] τ global =[F x ,F y ,F z M x M y M z ], where F x ,F y ,F z M represents the resultant thrust of the pectoral fin surface in the x, y, and z directions, respectively. x M y M z These are the resultant moments of the pectoral fin surfaces about the x, y, and z axes, respectively.

[0098] Furthermore, strain gauge arrays were arranged at the fin root, the fin root, the middle section of the fin, and the fin tip to measure the distribution of the local fluid load as P. i rays =[P i root ,P i mid ,P i end ], P i root ,P i mid ,P i end Let be the fluid loads at the root, middle, and end of the i-th fin, respectively. Based on these, a quadratic interpolation is used to construct the fin fluid pressure function along the fin length direction as follows:

[0099] f i (l)=al 2 +bl+c,l∈[0,L i ];

[0100] In the formula, L i Let be the length of the fin installed on the i-th fin root, where a, b, and c are constants, obtained from the fluid load data of the fin root, the middle section of the fin, and the end of the fin.

[0101] Specifically, the fluid pressure on the end of the i-th fin strip is The fluid torque on the end of the i-th fin strip is

[0102] Further, the actual deflection angle displacement of the fin root is collected by using an angle encoder And the actual fluctuation angle displacement of the fin root is obtained by using a flexible bending sensor And the passive fluctuation angle displacement of the fin strip is The motion line displacement of the end of the fin strip is calculated by combining the geometric model and the material response relationship:

[0103]

[0104] In the formula, is the normal end displacement of the i-th fin strip at time t, is the tangential end displacement of the i-th fin strip at time t, L i is the length of the fin strip installed on the i-th fin root;

[0105] Specifically, the equivalent stiffness K i is calculated by the following formula:

[0106]

[0107] In the formula, K i is the equivalent stiffness of the i-th fin strip (the greater the more difficult to bend), is the passive fluctuation angle displacement of the i-th fin strip after being subjected to the fluid torque;

[0108] Specifically, in order to verify the accuracy of the fluid load inversion result based on strain measurement, the equivalent stiffness of the i-th fin strip can be calculated by material properties And it is compared with the equivalent stiffness K i inverted by the fluid torque; if they are close, it means that the estimated fluid load distribution has good physical consistency, wherein E i is the elastic modulus of the i-th fin strip, and I i is the sectional moment of inertia of the i-th fin strip.

[0109] Further, all the sensor data are uniformly input into a state perception module, and after filtering, normalization and time sequence alignment, a structured state vector required by a reinforcement learning controller is formed:

[0110]

[0111] In the formula,

[0112] Specifically, the state vector is input as an environment observation of the outer optimization strategy, for evaluating the propulsion performance and the attitude stability; meanwhile, wherein P i rays The data is also synchronously transmitted to the inner optimization strategy, for dynamically correcting the fin equivalent stiffness.

[0113] S4, in order to balance the relationship among propulsion performance, attitude stability and fin motion flexibility, an optimization strategy is constructed based on deep deterministic policy gradient algorithm; the optimization strategy design adopts a double-layer reinforcement learning structure of inner-outer nesting, the inner optimization strategy is used for local structure stiffness adjustment, and the outer optimization strategy is used for global motion parameter optimization, and the two work together to form a dynamic adaptive control closed loop.

[0114] Further, referring to Figure 8 , the outer optimization strategy adopts a weighted combination of maximum propulsion performance and maximum attitude stability as the outer optimization target, and the optimal fin motion control parameters (spanwise amplitude, spanwise frequency, spanwise phase, spanwise motion period ratio, inboard and outboard amplitude, inboard and outboard frequency, inboard and outboard phase, inboard and outboard motion period ratio) are obtained by DDPG algorithm, to realize the balance between propulsion performance and stability, and enhance the adaptability of the system to environmental changes;

[0115] Specifically, the equation of the outer optimization target is:

[0116] J outer =w1(1-η thrust )+w2ξ yaw +w3ξ pitch ;

[0117] In the formula, w1, w2, w3 are weight coefficients, when w1>w2, w1>w3, the propulsion performance is given priority, and when w1<w2, w1<w3, the attitude stability is given priority;

[0118] Specifically, the propulsion performance calculation formula is:

[0119]

[0120] In the formula, η thrust is the effective thrust generated by unit power consumption, which represents the propulsion performance of the pectoral fin; P is the power consumption, and F x is the thrust in the x direction.

[0121] Specifically, the attitude stability calculation formula is:

[0122]

[0123] In the formula, ξyaw represents the attitude disturbance of the pectoral fin in the yaw degree of freedom, ξ pitch represents the attitude disturbance of the pectoral fin in the pitch degree of freedom, Var(·) represents a variance function, and the variance represents the fluctuation degree of force and moment, when ξ yaw , ξ pitch is smaller, the attitude stability is better.

[0124] Further, referring to Figure 9 , the inner-layer optimization strategy takes the maximization of structural compliance in the fin strip movement as an inner-layer optimization objective function under the premise that the optimal fin strip movement control parameter is kept unchanged, and adjusts the equivalent stiffness parameter K i of each fin strip to make the actual deformation trajectory of the fin strip under the fluid load have good structural compliance.

[0125] Specifically, the equation of the inner-layer optimization objective is as follows:

[0126]

[0127] In the formula, δ i (t, K i ) is the passive angular displacement of the i-th fin strip when the equivalent stiffness is K i , is the normal end displacement of the i-th fin strip when the equivalent stiffness is K i , w4 and w5 are weighting coefficients of the two parts, T is the pectoral fin movement period, and t is time.

[0128] Specifically, the first term in the above formula measures the passive bending response of the fin strip under the action of the fluid, and controls the compliant deformation degree, and the second term measures the smoothness of the fin strip movement curve (limits the adjustment of the equivalent stiffness to be too frequent or too intense, so as to avoid material fatigue or uncontrollable deformation).

[0129] Further, referring to Figure 10 , the DDPG training process includes a policy network and an evaluation network, and combines an experience replay and a target network soft update mechanism to ensure the training stability and convergence effect.

[0130] Further, referring to Figure 11 , in the double-layer reinforcement learning optimization strategy of inner-outer nesting, the outer-layer optimization strategy realizes high-level decision control of the fin root movement behavior target, the inner-layer optimization strategy adjusts the stiffness to ensure the compliance of the fin strip dynamic response, and the two-layer strategies work together to effectively improve the environmental adaptability, attitude stability and propulsion performance of the bionic pectoral fin control system under dynamic working conditions.

[0131] S5, in order to ensure accurate tracking of the fin root expected angular displacement generated by the outer-layer optimization strategy The system introduces a trajectory tracking controller based on model predictive control method in the execution layer; refer to Figure 12 The trajectory tracking controller is constructed based on outer layer control trajectory and inner layer equivalent stiffness adjustment, based on trajectory prediction and rolling optimization mechanism, displacement error is calculated by using the fin root desired angular displacement and the fin root actual angular displacement Dynamic evolution model within the prediction window is constructed, the problem is solved in real time by minimizing the square sum of displacement error and minimizing the weighted square sum of control input, a series of control input in the future is output, and the optimal control input is obtained in each control period, so as to realize high-precision, low overshoot and strong robustness trajectory tracking execution.

[0132] Specifically, the objective function of the trajectory tracking controller is:

[0133]

[0134] In the formula, is the desired angular displacement of the i-th fin root at the prediction step k at time t, is the actual angular displacement of the i-th fin root at the prediction step k, is the optimal control input of the i-th fin root at the prediction step k, and the optimal control input The superscript “^” only represents the optimal value, and does not change the meaning represented by the letter; Q is an error weight matrix, R is a control cost matrix, N p is the prediction step.

[0135] Specifically, the objective function of the trajectory tracking controller of the first fin root is solved as:

[0136]

[0137] In the formula, is the desired angular displacement of the first fin root at the prediction step k at time t, is the actual angular displacement of the first fin root at the prediction step k, is the optimal control input of the first fin root at the prediction step k, and the optimal control input Q is an error weight matrix, R is a control cost matrix, N p is the prediction step, A and B represent linear system state space matrices, the superscript “min” represents the minimum value, and the superscript “max” represents the maximum value;

[0138] Specifically, the objective function of the trajectory tracking controller of the second fin root is solved as follows:

[0139]

[0140] wherein, is the expected angular displacement of the second fin root at the prediction step k at time t, is the actual angular displacement of the second fin root at the prediction step k, is the optimal control input of the second fin root at the prediction step k, the optimal control input being obtained in each control period Q is an error weight matrix, R is a control cost matrix, N p is a prediction step, A and B represent linear system state space matrices, the superscript "min" represents a minimum value, and the superscript "max" represents a maximum value;

[0141] Specifically, the objective function of the trajectory tracking controller of the third fin root is solved as follows:

[0142]

[0143] wherein, is the expected angular displacement of the third fin root at the prediction step k at time t, is the actual angular displacement of the third fin root at the prediction step k, is the optimal control input of the third fin root at the prediction step k, the optimal control input being obtained in each control period Q is an error weight matrix, R is a control cost matrix, N p is a prediction step, A and B represent linear system state space matrices, the superscript "min" represents a minimum value, and the superscript "max" represents a maximum value.

[0144] Further, the trajectory tracking controller is used to enable the driving mechanism to achieve a fast, smooth, and undamped target angular tracking response, thereby constructing a complete perception-decision-control closed loop path.

[0145] A fin strip driving type bionic pectoral fin spanwise and chordwise cooperative motion control device provided by the application is described below. The active stability control device of the underwater vehicle described below can be correspondingly referred to the active stability control method of the underwater vehicle described above.

[0146] Specifically, referring to Figure 13 The fin strip driving type bionic pectoral fin spanwise and chordwise cooperative motion control device comprises a driving control module, a state sensing module, a signal processing module, and a feedback adjustment module.

[0147] Specifically, the drive control module is configured to output a control signal of the rhythm signal generator and drive an actuator at the fin root to perform corresponding movement;

[0148] Specifically, the state sensing module is configured to collect a state vector at a first moment in the movement of the pectoral fin through a sensor, the state vector including a thrust and torque of the pectoral fin, a local fluid load, an actual fluctuation angular displacement of the fin root, a passive fluctuation angular displacement of the fin strip, and an actual deflection angular displacement of the fin root, for real-time calling by the feedback regulation module.

[0149] Specifically, the signal processing module is configured to generate the fin root movement control parameter through calculation of an outer optimization strategy, an inner optimization strategy, and a trajectory tracking controller.

[0150] Specifically, the feedback regulation module is configured to compare data results of the state sensing module with a current control parameter, generate dynamic correction information, trigger the outer optimization strategy and the inner optimization strategy to reupdate an action strategy through an interruption mechanism, and ensure that the fin root tracks the fin root expected angular displacement through the trajectory tracking controller.

[0151] In another aspect, the application also provides a motion control program instruction stored in the computer readable storage medium composed of a Raspberry Pi as a host and a field programmable logic gate array processing unit as a slave, the computer readable storage medium being used to execute the motion control program instruction to implement the fin strip driving type bionic pectoral fin spanwise and chordwise coordinated motion control method provided by the above method.

[0152] Specifically, the above motion control program instruction can exist in the form of source code, bytecode, intermediate representation, or platform executable file.

[0153] Specifically, during the running of the motion control program instruction, the control logic and data flow are as follows: (1) the current propulsion task target and flow field information are input by an external task management system; (2) the outer DDPG strategy network generates movement parameters according to the current state vector; (3) the inner stiffness optimization module optimizes the fin strip equivalent stiffness K based on the reference trajectory and the feedback displacement; (3) the MPC module predicts future control input, and the instruction drive module outputs an execution signal; (4) the data acquisition module synchronously collects and returns sensing information for next period control iteration.

[0154] In yet another aspect, the present application also provides a computer readable storage medium composed of a master-slave control architecture, a control program of which is deployed in an integrated computing platform, a host adopts a Raspberry Pi microcomputer, and a slave adopts a field programmable logic gate array processing unit; and the computer readable storage medium has stored thereon motion control program instructions, which are executed by a controller to implement the fin driving type bionic pectoral fin spanwise and chordwise coordinated motion control method provided by the above method.

[0155] It should be noted that the device structure and the module division thereof described in the above description are only illustrative descriptions of the embodiments of the present application, and do not constitute a limitation on the protection scope of the present application. In the specific implementation process, the function modules or units described can be physically integrated according to actual needs, or can be logically independent of each other. For example, as an independently described module, it can be a physically separated independent unit in actual deployment, or it can be integrated in the same controller or computing platform; the communication connection between the modules can be realized in various forms such as signal line, data bus, wireless transmission, etc. Those skilled in the art can make reasonable adjustments and implementation according to the technical solutions and descriptions of the present application without creative labor, combined with specific application scenarios.

[0156] The background section of the present application can contain background information about the problems or environment of the present application, and does not necessarily describe the prior art. Therefore, the content contained in the background section is not an admission by the applicant of prior art.

[0157] The above is a further detailed description of the present application in combination with specific embodiments, and the specific implementation of the present application cannot be limited to these descriptions. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and changes.

Claims

1. A fin-ray driven biomimetic pectoral fin spanwise and chordwise coordinated motion control method, characterized in that, Comprise the following steps: S1, based on the skeletal structure of fish pectoral fin, design the fin strip driving type bionic pectoral fin composed of independently driven fin root, elastic fin strip and fin surface, the spanwise fluctuation of the fin surface is formed by the active fluctuation of the fin root and the passive fluctuation of the fin strip, the chordwise fluctuation of the fin surface is formed by the fluctuation phase difference between the fin roots, and the adduction and abduction movement of the fin surface is formed by the deflection phase difference between the fin roots; at the same time, according to the gait characteristic data of pectoral fin, the kinematic model of fin root is constructed, and c(t) is defined as the fin root motion control parameter; S2, in order to make the fin root kinematics model output continuously smooth and continuous and rhythmic control signal, using the improved central pattern generator of Hopf oscillator as the rhythm signal generator, the improved central pattern generator has the double channel Hopf oscillator structure of fin root fluctuation angle displacement and fin root deflection motion angle displacement; the rhythm signal generator is matched with the input parameter of the fin root kinematics model, the output of the fin root kinematics model is processed by changing the fin root motion control parameter c(t) and using the rhythm signal generator, and smooth fin root angle displacement is output, the three-dimensional collaborative motion of the fin strip driving type bionic chest fin in spanwise, chordwise, internal collection and external development is realized, and the smooth fin root angle displacement is defined as S3, during the chest fin movement, the chest fin is installed on a six-dimensional force sensor to collect the chest fin combined thrust and torque in real time, and a strain gauge array is used to acquire local fluid load The actual deflection angle displacement of the fin root during movement is obtained by using an angle encoder The actual fluctuation angle displacement of the fin root during movement is obtained by using a bending sensor And the passive fluctuation angle displacement of the fin strip Through the actual deflection angle displacement of the fin root The actual fluctuation angle displacement of the fin root And the passive fluctuation angle displacement of the fin strip The fin strip movement line displacement and equivalent stiffness K are calculated i The actual angle displacement of the fin root is defined as S4, constructing an inner-layer optimization strategy and an outer-layer optimization strategy of reinforcement learning based on a deep deterministic policy gradient algorithm; the outer-layer optimization strategy adopts a weighted combination of maximum propulsion performance and maximum attitude stability as an optimization target, and optimizes to obtain optimal fin root motion control parameters The inner-layer optimization strategy establishes an equivalent stiffness adjustment model for the structural response characteristics of each fin strip, and dynamically adjusts the equivalent stiffness K through a nested optimization strategy i ​ S5, to ensure accurate tracking of the fin root desired angular displacement generated by the outer layer optimization strategy A trajectory tracking controller based on model predictive control method is introduced, which is constructed on the basis of outer layer control trajectory and inner layer equivalent stiffness adjustment, based on trajectory prediction and rolling optimization mechanism, receives the current actual angular displacement of the fin root The error between the desired angular displacement of the fin root Predicts the future state evolution and solves the optimal control input by rolling Thus realizing high-precision, low overshoot and strong robustness trajectory tracking execution.

2. The method of claim 1, wherein, The fin strip driving type bionic pectoral fin is composed of independently driven fin root, passively deformed elastic fin strip and passively deformed flexible fin surface, wherein the first fin root realizes spanwise fluctuation, the second and third fin roots further have the coupling movement ability of adduction and abduction on the basis of spanwise fluctuation, the spanwise fluctuation of the whole fin surface is realized by the active fluctuation of the fin root and the passive fluctuation of the fin strip, the chordwise fluctuation of the whole fin surface is realized by the fluctuation phase difference, and the adduction and abduction movement of the whole fin surface is realized by the deflection phase difference; the equation of the kinematic model of the fin root is: wherein α i (t) is the i-th fin root fluctuation angular displacement at time t, is the amplitude of the i-th fin root fluctuation angular displacement, is the bias of the i-th fin root fluctuation angular displacement, is the i-th fin root fluctuation motion period ratio, dividing the period into four parts, T α is the fluctuation motion period; β i (t) is the i-th fin root deflection angular displacement at time t, is the amplitude of the i-th fin root deflection angular displacement, is the bias of the i-th fin root deflection angular displacement, r i is the i-th fin root deflection motion period ratio, dividing the period into two parts, 0 < r i < 1, T β is the deflection motion period, t is time, and the fin root motion control parameters i takes values of 1, 2, 3, the elastic fin strip generates fluctuation by virtue of the elastic characteristics of its own material under the active drive of the fin root, and drives the flexible fin surface to achieve passive spanwise fluctuation.

3. The method of claim 1, wherein, The rhythm signal generator is constructed by the improved central pattern generator and the kinematic model of the fin root, and the equation of the rhythm signal generator is as follows: wherein α i (t) is the i-th fin root fluctuation angle displacement at i-th time, is the output of the rhythm signal generator after smoothing processing of α i (t), θ i α (t) is the output of the rhythm signal generator about α i (t), β i (t) is the i-th fin root deflection angle displacement at t-th time, is the output of the rhythm signal generator after smoothing processing of β i (t), is the output of the rhythm signal generator about β i (t), k α , k β is the response frequency, the greater the value, the faster the response, is the damping ratio, controls the smoothing degree, t is the time, and the smoothed fin root angle displacement is The rhythm signal generator not only has good approximation ability to the original control signal, but also has the ability to maintain the rhythm, smoothness and differentiability of the trajectory.

4. The method of claim 1, wherein, The acquisition and processing procedure of the total thrust and moment of the pectoral fin, the local fluid load, the actual fluctuation angular displacement of the fin root, the passive fluctuation angular displacement of the fin strip, the actual deflection angular displacement of the fin root and the motion line displacement of the fin strip is as follows: first, the six-dimensional force sensor installed above the base of the pectoral fin is used to acquire the total thrust and moment of the pectoral fin in three-axis directions in real time, and the total thrust and moment of the pectoral fin is denoted as τ global =[F x ,F y ,F z ,M x ,M y ,M z ],F x ,F y ,F z are the thrusts of the pectoral fin in x, y and z directions respectively, M x , M y , M z are the moments of the pectoral fin around x, y and z axes respectively; meanwhile, strain gauge arrays are arranged in the fin root, the middle section of the fin strip and the end of the fin strip to measure the distribution of the local fluid load as P i rays =[P i root ,P i mid ,P i end ],P i root ,P i mid ,P i end are the fluid loads of the i-th fin strip root, the i-th fin strip middle section and the i-th fin strip end respectively, and a quadratic interpolation is used to construct the fluid pressure function of the fin strip along the length direction of the fin strip as f i (l) = al 2 + bl + c, l e [0, L i ] ; where L i is the length of the i-th fin strip attached to the root, a, b, c are constants determined from the fluid load data at the fin root, the mid-section of the fin strip and the tip of the fin strip, and the fluid pressure at the tip of the i-th fin strip is the fluid moment at the tip of the i-th fin strip is Secondly, the actual deflection angle displacement of the fin root is collected by using an angle encoder and the actual fluctuation angle displacement of the fin root is obtained by using a flexible bending sensor and the passive fluctuation angle displacement of the fin strip The motion line displacement of the fin strip end is calculated by combining the geometric model and the material response relationship: wherein, is the end displacement of the i-th fin strip normal to the water surface at time t, is the end displacement of the i-th fin strip tangential to the water surface at time t, L i is the length of the i-th fin strip; the equivalent stiffness K i is calculated by the formula: In the formula, K i is the equivalent stiffness of the i-th fin, the larger the value, the more difficult to bend, is the passive wave angle displacement of the i-th fin after the fluid torque, i is 1, 2, 3; Finally, after unified sampling, filtering, normalization and feature fusion processing by the state perception module, the state vector is formed as: In the formula, The state vector is input into the outer optimization strategy as an environmental observation for evaluating the system propulsion performance and attitude stability; meanwhile, the state vector is used to calculate the control input of the system. The data is also transmitted to the inner optimization strategy for dynamically modifying the equivalent stiffness of the fin.

5. The method of claim 1, wherein, The outer layer optimization strategy takes maximum propulsion performance and maximum attitude stability as the outer layer optimization objectives, and the optimal fin root motion control parameters are learned by a deep deterministic policy gradient algorithm is an amplitude of the i-th fin root fluctuation angular displacement, is a bias of the i-th fin root fluctuation angular displacement, is a fluctuation motion period ratio of the i-th fin root, and the period is divided into four parts, T α is a fluctuation motion period, is an amplitude of the i-th fin root deflection angular displacement, is a bias of the i-th fin root deflection angular displacement, T β is a deflection motion period, r i is a deflection motion period ratio of the i-th fin root, and the superscript " " only represents the optimal value of the motion control parameter, without changing the meaning represented by the letter; and an equation of the outer layer optimization objective is: J outer = w1(1 - η thrust ) + w2ξ yaw + w3ξ pitch ; In the formula, w1, w2, and w3 are weight coefficients, η thrust , ξ yaw , ξ pitch The calculation equation is: ξ yaw = Var(F y ) + Var(M z ) + Var(M x ); ξ pitch = Var(F z ) + Var(M y ) + Var(M x ) In the formula, η thrust is the effective thrust generated by unit power consumption, P is power consumption, F x , F y , F z are the thrusts of the pectoral fins in x, y, z directions respectively, M x , M y , M z are the moments of the pectoral fins around x, y, z axes respectively, ξ yaw represents the attitude disturbance of the pectoral fins in the yaw degree of freedom, ξ pitch represents the attitude disturbance of the pectoral fins in the pitch degree of freedom, Var(·) represents a variance function, when ξ yaw , ξ pitch is smaller, the attitude stability is better; the inner layer optimization strategy takes the maximization of structural compliance in the process of fin root motion as an inner layer optimization target under the premise of keeping the optimal fin root motion control parameters unchanged, adjusts the equivalent stiffness K i of each fin strip, and the actual deformation trajectory of each fin strip under fluid load has good structural compliance; the equation of the inner layer optimization target is: where the first term measures the passive bending response of the fin strip under fluid action, controlling the degree of compliant deformation, and the second term measures the smoothness of the fin strip motion curve, limiting excessive frequency or intensity of equivalent stiffness adjustment to avoid material fatigue or uncontrollable deformation; δ i (t, K i ) is the passive wave displacement angle of the i-th fin strip when the equivalent stiffness is K i is the normal end displacement of the i-th fin strip when the equivalent stiffness is K i w4, w5 are the weighting coefficients of the two parts, T is the pectoral fin motion period, and t is time; the outer optimization strategy realizes high-level decision control of the fin root motion behavior target, and the inner optimization strategy ensures the compliance of the fin strip dynamic response by adjusting the equivalent stiffness, and the two strategies work together to effectively improve the environmental adaptability, posture stability, and propulsion performance of the bionic pectoral fin control system under dynamic working conditions.​ 6. The method of claim 1, wherein, The trajectory tracking controller ensures accurate tracking of the fin root desired angular displacement generated by the outer optimization strategy using a model predictive control method Based on the trajectory prediction and rolling optimization mechanism, the fin root desired angular displacement And the fin root actual angular displacement The displacement error is calculated, a dynamic evolution model within the prediction window is constructed, the problem is solved in real time by minimizing the weighted sum of the square sum of the displacement error and the square sum of the control input, a series of control inputs in the future is output, and the optimal control input is obtained in each control period Thus, accurate tracking of the target trajectory is achieved; the equation of the trajectory tracking controller is: wherein, is the expected angular displacement of the i-th fin root at the prediction step k at time t, is the actual angular displacement of the i-th fin root at the prediction step k, is the optimal control input of the i-th fin root at the prediction step k, which is obtained in each control period The superscript " " only represents the optimal value of the control input, and does not change the meaning represented by the letter; Q is an error weight matrix, R is a control cost matrix, N p is the prediction step; the trajectory tracking controller is used to make the driving mechanism achieve a fast, smooth and no overshoot target angle tracking response, thereby constructing a complete perception-decision-control closed loop path.

7. The method of claim 1, wherein, The fin strip driving type bionic pectoral fin is made of rigid fin bone, rigid fin root, elastic fin strip and flexible fin membrane; the fin bone and the fin root are made of materials with good structural strength and forming precision, the fin strip is selected from materials with high elasticity and light weight characteristics, and the fin membrane is made of flexible and deformable materials.

8. A fin drive type bionic chest fin spanwise and chordwise cooperative motion control device, characterized in that, The device is used for executing the spanwise and chordwise cooperative movement control method of the fin strip driving type bionic pectoral fin according to claim 1, and the device comprises: a driving control module for outputting the control signal of the rhythm signal generator and driving the actuator at the fin root to execute corresponding movement; A state perception module acquires the state vector at the first moment in the pectoral fin movement through a sensor, and the state vector includes the pectoral fin combined thrust and torque, the local fluid load, the actual fluctuation angular displacement of the fin root, the passive fluctuation angular displacement of the fin strip and the actual deflection angular displacement of the fin root, which are called by a feedback adjustment module in real time; A signal processing module generates the fin root motion control parameter through the calculation of an outer optimization strategy, an inner optimization strategy and a trajectory tracking controller; A feedback adjustment module compares the data results of the state perception module with the current control parameter, generates dynamic correction information, triggers the outer optimization strategy and the inner optimization strategy to update the action strategy again through an interrupt mechanism, and ensures that the fin root tracks the fin root expected angular displacement through a trajectory tracking controller.

9. A computer readable storage medium containing motion control program instructions, characterized in that, The motion control program instructions are stored in the computer readable storage medium composed of a Raspberry Pi as a host and a field programmable logic gate array processing unit as a slave, and the computer readable storage medium is used to execute the motion control program instructions to implement the spanwise and chordwise cooperative movement control method of the fin strip driving type bionic pectoral fin according to any one of claims 1 to 7.

Images