Fishery holder equipment self-adjustment control optimization method

Abstract

The invention provides a fishery pan-tilt equipment self-adjustment control optimization method, and belongs to the field of pan-tilt control, and the method specifically comprises the steps: S1, constructing a fishery pan-tilt equipment control model which comprises a corner control model, an improved artificial travel mouse algorithm and a position type PID controller model; s2, improving a standard artificial squirrel algorithm, and setting a proportionality coefficient, an integral coefficient and a differential coefficient of a position type PID of the holder equipment controller by using the improved artificial squirrel algorithm to realize optimization of the holder equipment controller; and S3, inputting a difference value between the target rotation angle and the real-time rotation angle of the holder into a holder equipment controller, outputting a control signal through a position type PID, inputting the control signal into a rotation angle control model, and outputting a real-time rotation angle value of the holder. Experimental simulation proves that the method disclosed by the invention can solve the problems of easy oscillation and overshoot under the conditions of large rotation angle range and frequent speed abrupt change of the holder.

Description

Technical Field

[0001] This invention belongs to the field of gimbal control, specifically relating to a self-adjusting control optimization method for fishery gimbal equipment. Background Technology

[0002] As modern fisheries develop towards intelligence and precision, and with the increasing prevalence of technologies such as image recognition, water quality monitoring, and behavior analysis, gimbals, as key actuators in these systems, are responsible for the real-time positioning and attitude adjustment of sensing and operational equipment such as cameras, water samplers, and feeders. The response performance of the gimbal directly affects the system's observation accuracy and control effectiveness. Traditional gimbal control systems mostly use proportional-integral-derivative (PID) controllers. In actual fisheries applications, gimbal control faces the following challenges: factors such as water surface reflection, fish movement, and water quality disturbances cause rapid changes in camera targets, requiring the gimbal to have excellent tracking capabilities; under conditions of large rotation angle range and frequent speed changes, oscillation and overshoot problems are prone to occur; wind, waves, and water flow interference cause deviations in the gimbal's motion trajectory, which traditional controllers struggle to adapt to.

[0003] The Artificial Lemming Algorithm (ALA) is a newly proposed biomimetic heuristic optimization algorithm inspired by four typical behaviors of lemmings in nature: long-distance migration, burrowing, foraging, and evading predators. This algorithm is primarily used to solve complex engineering optimization problems. However, as a global optimization algorithm, ALA is suitable for offline parameter tuning due to its insufficient convergence speed, but not for real-time dynamic decision-making in the self-adjustment of fishery gimbal equipment. Furthermore, while ALA has a stronger global scope, in the fine-grained local adjustment problem of gimbal control, insufficient convergence accuracy during the development phase leads to inadequate optimization accuracy of the PID control parameters for the self-adjustment of the fishery gimbal equipment, thus affecting the self-adjustment control of the fishery gimbal equipment. Summary of the Invention

[0004] To address the aforementioned shortcomings in existing technologies, this invention provides a self-adjusting control optimization method for fishery gimbal equipment. By introducing an improved artificial lemming algorithm to optimize the PID controller of the fishery gimbal equipment, the improved artificial lemming algorithm is suitable for real-time dynamic decision-making for the self-adjustment of fishery gimbal equipment and has higher accuracy in optimizing the control parameters of the controller's position PID control algorithm. This solves the problems of oscillation and overshoot that easily occur when the gimbal has a large rotation angle range and frequent speed changes.

[0005] To achieve the aforementioned objectives, the present invention employs the following technical solution: a self-adjusting control optimization method for a fishery gimbal device, comprising the following specific steps: S1. Construct a control model for fishery gimbal equipment. The control model includes a rotation control model, an improved artificial lemming algorithm, and a positional PID controller model. S2. Improvements to the standard artificial lemming algorithm: The improved algorithm is used to tune the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device, thereby optimizing the gimbal device controller. The improved artificial lemming algorithm includes: introducing an ecology-driven migration mechanism, constructing an ecological environment quality field, and using a local gradient guidance strategy to drive the search individuals to migrate to areas with higher ecological suitability, achieving directional exploration and updating the search individual's position; and introducing a local competitive pressure index. An improved escape factor was constructed to improve the mathematical model for updating the search individual's position during the predator avoidance phase; S3. Input the difference between the target rotation angle and the real-time rotation angle of the gimbal into the gimbal device controller, and output a control signal through a position-type PID controller. ,Will Input the angle control model and output the real-time rotation angle value of the gimbal.

[0006] Furthermore, the platform for gimbal rotation is a BLDC motor, and the position-based PID control algorithm outputs control signals. The control signal controls the rotation angle of the BLDC motor to achieve the gimbal's rotation angle; that is, the rotation angle of the BLDC motor equals the rotation angle of the gimbal. A Hall sensor feeds back the real-time rotation angle of the BLDC motor to the input unit, thereby calculating the difference between the target rotation angle and the real-time rotation angle of the gimbal each time. This difference is input to the gimbal controller to control the rotation of the BLDC motor, thus controlling the rotation angle of the gimbal, until the target rotation angle is achieved. The input to the angle control model is the real-time control signal. The output is the real-time rotation angle value; the values ​​of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the positional PID control algorithm are obtained by optimization using an improved artificial lemming algorithm.

[0007] Furthermore, the positional PID control algorithm is implemented in discrete time, satisfying the condition at all times. The proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the discrete position PID control algorithm are correlated with the search individual positions of the improved artificial lemming algorithm. The improved artificial lemming algorithm updates the search individual's position using migration, burrowing, foraging, and predator avoidance strategies, ultimately obtaining the optimal search individual position. This is analyzed using the optimal proportional coefficient, integral coefficient, and differential coefficient. The search individual position update is driven by a fitness function. The fitness function is designed based on the time required to first reach and stabilize within the target value ±2%, taking into account control accuracy. The index is the integral of the absolute value of the control signal divided by the total simulation time. The mathematical model is as follows: ; In the formula, The total simulation duration is... For the first The real-time rotation angle value at any given moment. Rotate by the target angle.

[0008] Furthermore, an ecologically driven migration mechanism is introduced to simulate the perception of environmental quality fields by search individuals during migration within an ecosystem and their decision-making accordingly. Starting from ecological modeling and non-uniform field-driven mechanisms, unlike the standard artificial lemming optimization algorithm which uses random directions and Brownian motion noise to drive long-distance migration, the ecologically driven migration mechanism constructs an ecological environment quality field and uses a local gradient guidance strategy to drive search individuals (lemming individuals) to migrate to areas with higher ecological suitability, achieving more directional exploration and updating. Areas with higher ecological suitability are denoted as locations where the search individual's fitness value is lower. This strategy simulates the behavior of animals in natural ecosystems making choices based on habitat quality to construct a mathematical model for updating the search individual's location.

[0009] Furthermore, the location of searched individuals is updated through an ecologically driven migration mechanism, specifically as follows: S301. Construct an ecological environment quality field, and define the search space as... The search space dimension is determined based on the question dimension Dim, defining the location (search point) of each search entity. ecological quality The mathematical model is: ; In the formula, Let i be the position of the i-th search entity in the t-th iteration. The fitness value for the location of the searched individual. As an ecological suppressor, it adjusts the weight of ecological quality and crowding. The mathematical model for the crowding around the current searched individual is: ; in, Let be the position of the j-th search individual near the i-th search individual in the t-th iteration, n be the number of nearby search individuals, and h be the neighborhood perception factor; Among them, ecological suppression factors The mathematical model is as follows: ; In the formula, For the ecological quality of the largest neighboring point, For the minimum neighboring point ecological quality, Let be the ecological quality value of the i-th search individual; S302. The virtual gradient of ecological quality is used to guide the search direction for individuals. The mathematical model is as follows: ; In the formula, Let be the virtual gradient value for the i-th search individual in the t-th iteration, where the gradient points in the direction of the fastest improvement in ecological quality. To select k higher-quality neighboring points from the searched individual, The mathematical model for poor ecological quality is as follows: μ is used to prevent the value from returning to zero. S303. An ecologically driven migration strategy is constructed to update the location of search individuals through an ecologically driven migration mechanism. The mathematical model is as follows: ; In the formula, The maximum number of iterations, This is the current best location for the individual being searched.

[0010] Furthermore, this mechanism constructs an eco-inspired strongly guided exploration method through ecological quality modeling, local collaborative gradient perception, and time convergence control. This method retains the self-organizing nature of the artificial lemming algorithm while achieving precise control over its behavior. Specifically, the ecological environment quality field guides search individuals towards optimal solutions while avoiding premature clustering that could affect solution accuracy. A virtual gradient of ecological quality guides the direction of search individuals, not based on a single optimal individual, but on the weighted gradient directions of multiple excellent neighbors, preserving local solution diversity and enhancing robustness. An ecologically driven migration strategy is constructed to update the location of search individuals through an exponential decay factor. By controlling the search intensity and incorporating an optimal individual guiding term, convergence can be considered. The improved artificial lemming algorithm utilizes an ecologically driven migration strategy to optimize the solution, which can prevent getting trapped in local optima, improve the stability of the PID control parameter values ​​of the adaptive controller of the fishery gimbal equipment, guide the PID parameters to evolve towards the optimal direction, achieve stable convergence, avoid oscillations, and make the parameter search process smooth.

[0011] Furthermore, the linear decay of the escape coefficient in the standard ALA algorithm, considering only the time factor, while showing a reasonable trend, lacks adaptive intelligence. For the algorithm itself, each search entity is subject to uniform control, lacking personalized responses. For the self-adjusting control of fishery gimbal equipment, this leads to overly limited control parameter solutions for the core controller's position PID control algorithm, preventing it from reaching the optimal solution. Introducing a local competition pressure index... The decision to escape is based on comparisons with surrounding individuals and the crowding level near the current individual. The mathematical model is as follows: ; In the formula, Let i be the survival benefit value of the i-th search individual. Let be the local neighbor survival benefit value of the j-th search individual. , Let r be the number of local neighbors of the i-th search individual, and r be a random number between 1 and 2. The mathematical model is as follows: ; In the formula, Let i be the crowding density near the i-th search individual in the t-th iteration; Based on local competitive pressure index Constructing an improved escape factor: ; In the formula, Let be the escape factor value of the i-th search individual in the t-th iteration; Then, the mathematical model for updating the search individual's position during the predator avoidance phase is improved using an improved escape factor: ; In the formula, rand is a normally distributed random number between 0 and 1, Dim is the question dimension, set to 3; Levy is the Levy distribution formula. This is the current best location for the individual being searched.

[0012] Furthermore, a sine function and local distance tensor modulation are used to simulate the local disturbance resonance of the improved ALA algorithm in the control of the fishery gimbal equipment. If a neighbor search individual triggers high resonance, i.e., the fishery gimbal equipment rotates too much, the position PID control parameter optimization process is prone to getting trapped in local optima during the optimization control process. Therefore, in the improved AFA algorithm and PID controller parameter optimization... The nonlinear increase leads to a larger local competitive pressure index, which in turn increases the improved escape factor, allowing the algorithm to escape local optima. This enables the AFA algorithm to adapt to changes in the gimbal and continue to adaptively adjust the parameter values ​​of the position PID controller. The introduction of a nonlinear fluctuation mechanism using a sine function avoids judging the danger solely by the distance between the searched individual's position (the difference in solutions), simulates real fluctuations, and prevents the searched individual's position update from getting stuck in local optima.

[0013] Furthermore, the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device are tuned using an improved artificial lemming algorithm. The specific method is as follows: S21. Initialize the maximum number of iterations T, problem dimension D, upper bound ub and lower bound lb of the search space, and maximum population size N of the improved artificial lemming algorithm. S22. Randomly initialize the search individual positions according to the search space boundaries, calculate the fitness value of the search individual positions after initialization, and take the search individual position corresponding to the minimum fitness value as the current best search individual position. S23. Calculate the capability factor value for the current iteration. If the energy factor value is greater than 1, the global search phase position update mathematical model is executed; otherwise, the local development phase position update mathematical model is executed. S24. Calculate the ecological suppression factor value and the crowding value near the current search individual, construct the ecologically driven migration mechanism, improve the mathematical model for position update in the global search stage, and update the position of the search individual. S25. Introduce a local competitive pressure index to calculate an improved escape factor value, and use the improved escape factor value to improve the local development stage position update mathematical model and update the position of the search individual. S26. Calculate the fitness value of the current search individual position, update the current best search individual position, and execute t=t+1 for the current iteration number t; if t≤T is satisfied, return to execute S23, otherwise take the current best search individual position as the global best individual position; S27. Output the global optimal individual position and resolve it into the optimal proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd values.

[0014] Furthermore, the fitness function calculates the fitness value of the search individual's position by resolving the search individual's position into proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd values, inputting these values ​​into the fishery gimbal control model, and indirectly calculating the fitness value through the control signal value output by the control model and the real-time rotation angle value of the gimbal. The smaller the fitness value, the higher the self-adjusting control accuracy of the fishery gimbal. The fitness value directly reflects the better accuracy of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd, thus indicating a better current search individual position. By finding the current optimal search individual position through each algorithm iteration, the improved artificial lemming algorithm is used to tune the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal.

[0015] Furthermore, the ecologically driven migration mechanism and burrowing constitute the global search phase of the improved artificial lemming algorithm; the foraging strategy and the predator avoidance strategy based on the improved escape factor value constitute the local development phase of the improved artificial lemming algorithm.

[0016] Compared with existing technologies, the self-adjusting control method for fishery gimbal equipment proposed in this invention has the following technical effects: By introducing an ecologically driven migration mechanism and a local competitive pressure index through an improved Artificial Lemming Algorithm (ALA), it effectively solves the problems of insufficient real-time response, easy getting trapped in local optima, and low optimization accuracy of traditional algorithms; The innovative design of a comprehensive fitness function that integrates integral absolute error, response time, overshoot, energy consumption, and jitter suppression significantly improves the optimization accuracy and control stability of control parameters; In complex fishery environments with wind, waves, water flow disturbances, and rapid target movement, it can make real-time dynamic decisions and quickly and accurately adjust the gimbal position, greatly reducing overshoot and oscillation phenomena, significantly reducing equipment energy consumption and jitter, and extending the service life of the equipment; It improves the adaptability and robustness of the system in dynamic environments and optimizes the overall control performance of fishery gimbal equipment. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating the implementation steps of the self-adjusting control optimization method for fishery gimbal equipment proposed in this invention; Figure 2 This is a schematic diagram illustrating the tuning process of the proportional, integral, and derivative coefficients of the positional PID controller for the gimbal device using the improved artificial lemming algorithm. Figure 3 This is a comparison chart of the fitness values ​​of the improved ALA algorithm and the standard ALA algorithm for the solutions of PID control parameters of the gimbal. Figure 4 This is a graph showing the effect of the method of the present invention in optimizing the parameter values ​​of Kp, Ki, and Kd; Figure 5 This is a graph showing the effect of the standard method in optimizing the parameter values ​​of Kp, Ki, and Kd; Figure 6 This is a comparison chart of the effects of the fishery gimbal control method of this invention and the standard ALA-PID control method. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention provides a technical solution: a self-adjusting control optimization method for fishery gimbal equipment, according to as follows Figure 1 The steps shown are implemented, specifically steps S1 to S3.

[0020] S1. Construct a control model for fishery gimbal equipment. The control model includes a rotation control model, an improved artificial lemming algorithm, and a positional PID controller model.

[0021] Preferably, the carrier for gimbal rotation is a BLDC motor, and the position-based PID control algorithm outputs control signals. The relationship between the control signal and the rotation angle of the BLDC motor is established through the rotation angle control model. The control signal controls the rotation angle of the BLDC motor to achieve the rotation angle of the gimbal. That is, the rotation angle of the BLDC motor is equal to the rotation angle of the gimbal. The real-time rotation angle of the BLDC motor is fed back to the input unit through the Hall sensor. The difference between the target rotation angle of the gimbal and the real-time rotation angle is calculated each time. The difference is input to the gimbal device controller to control the rotation of the BLDC motor, that is, the rotation angle of the gimbal, until the target rotation angle of the gimbal is reached.

[0022] Preferably, the angle control model, the improved artificial lemming algorithm, and the positional PID controller model are implemented in MATLAB code. Specifically, based on the characteristics of the BLDC motor angle control system of the standard gimbal device, a second-order system time-domain model is used to construct the angle control model. MATLAB code is written based on the angle control model, and then converted into a Laplace transform form as the controlled function of the Siumulink simulation model. The time-domain mathematical model of the angle control model is as follows: ; In the formula, The moment of inertia is set to 0.02. The damping coefficient is set to 0.01. For the first The real-time rotation angle value at any given moment. The motor torque constant is set to 0.5, and the simulation step size is 0.01. Based on the set parameter values, it is converted into a Laplace transform form, and the mathematical model is as follows: .

[0023] S2. Improvements to the standard artificial lemming algorithm: The improved algorithm is used to tune the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device, thereby optimizing the gimbal device controller. The improved artificial lemming algorithm includes: introducing an ecology-driven migration mechanism, constructing an ecological environment quality field, and using a local gradient guidance strategy to drive the search individuals to migrate to areas with higher ecological suitability, achieving directional exploration and updating the search individual's position; and introducing a local competitive pressure index. An improved escape factor is constructed to improve the mathematical model for updating the search individual's position during the predator avoidance phase.

[0024] Preferably, the standard artificial lemming algorithm code is written in Matlab. Based on this, the standard artificial lemming algorithm is improved by introducing an ecology-driven migration mechanism, constructing an ecological environment quality field, and using a local gradient guidance strategy to drive the search individuals to migrate to areas with higher ecological suitability. This achieves directional exploration and updates the search individual's location. The Matlab code is written based on the following improved artificial lemming algorithm mathematical model, and the specific method is as follows: S301. Construct an ecological environment quality field, and define the search space as... The search space dimension is determined based on the question dimension Dim, defining the location (search point) of each search entity. ecological quality The mathematical model is: ; In the formula, Let i be the position of the i-th search entity in the t-th iteration. The fitness value for the location of the searched individual. As an ecological suppressor, it adjusts the weight of ecological quality and crowding. The mathematical model for the crowding around the current searched individual is: ; in, Let be the position of the j-th search individual near the i-th search individual in the t-th iteration, n be the number of nearby search individuals (a fixed value of 5), and h be the neighborhood perception factor (a random value between 0 and 1). Among them, ecological suppression factors The mathematical model is as follows: ; In the formula, For the ecological quality of the largest neighboring point, For the minimum neighboring point ecological quality, Let be the ecological quality value of the i-th search individual; S302. The virtual gradient of ecological quality is used to guide the search direction for individuals. The mathematical model is as follows: ; In the formula, Let be the virtual gradient value for the i-th search individual in the t-th iteration, where the gradient points in the direction of the fastest improvement in ecological quality. To select k higher-quality neighboring points from the searched individual, The mathematical model for poor ecological quality is as follows: μ is used to prevent the value from returning to zero. S303. An ecologically driven migration strategy is constructed to update the location of search individuals through an ecologically driven migration mechanism. The mathematical model is as follows: ; In the formula, The maximum number of iterations is set to 50. This is the current best location for the individual being searched; The MATLAB code is as follows: %S301: Calculate the ecological quality Q(Z_i^t) Q = zeros(N,1); alpha = zeros(N,1); y = zeros(N,1); for i = 1:N % Calculate the nearest neighbor congestion level; Z_i = Z(i,:); dist = sqrt(sum((Z - Z_i).^2, 2)); % Euclidean distance y(i) = sum(exp(-dist.^2 / (2*h^2))); %Calculate ecological target factors; neighbors = find(dist<= h&(1:N)' ~= i); % Find neighbors within radius h if ~isempty(neighbors) Q_neighbors = -f_values(neighbors) + alpha(neighbors).*y(neighbors); alpha(i) = (1 / sqrt(Dim))*(ub(1)-lb(1)) * (max(Q_neighbors) - (-f_values(i)+alpha(i)*y(i))) / (max(Q_neighbors) - min(Q_neighbors) + eps); else alpha(i) = 0; end % Calculate ecological quality; Q(i) = -f_values(i) + alpha(i)*y(i); % formula (0008-S301) end %S302: Calculate the virtual gradient; B = zeros(N,Dim); for i = 1:N Select k higher-quality neighbors (Q_j > Q_i) better_idx = find(Q>Q(i)); [~, dist_sort] = sort(vecnorm(Z(better_idx,:) - Z(i,:), 2, 2)); selected = better_idx(dist_sort(1:min(k, length(better_idx)))); %Calculate the gradient direction; sum_grad = zeros(1,Dim); for j = selected' w_j = Q(j) - Q(i); diff = Z(j,:) - Z(i,:); norm_diff = norm(diff) + mu; sum_grad = sum_grad + w_j*(diff / norm_diff); % formula (0009-S302) end B(i,:) = sum_grad; end S303: Location update; Z_new = Z + exp(-0.5*t / T).*B + rand(N,Dim).*(Z_best - Z); % formula (0009-S303) end.

[0025] Preferably, a local competitive pressure index is introduced. An improved escape factor was constructed to improve the mathematical model for updating the search individual's position during the predator avoidance phase. Matlab code for the improved artificial lemming algorithm was then implemented based on the following mathematical model; a local competitive pressure index was introduced. The decision to escape is based on comparisons with surrounding individuals and the crowding level near the current individual. The mathematical model is as follows: ; In the formula, Let i be the survival benefit value of the i-th search individual. Let be the local neighbor survival benefit value of the j-th search individual. , Let r be the number of local neighbors of the i-th search individual, and r be a random number between 1 and 2. The mathematical model is as follows: ; In the formula, Let i be the crowding density near the i-th search individual in the t-th iteration; Furthermore, based on the local competitive pressure index Constructing an improved escape factor: ; In the formula, Let be the escape factor value of the i-th search individual in the t-th iteration; Furthermore, an improved escape factor is used to refine the mathematical model for updating the search individual's position during the predator avoidance phase: ; In the formula, rand is a normally distributed random number between 0 and 1, Dim is the question dimension, set to 3; Levy is the Levy distribution formula. This is the current best location for the individual being searched; The MATLAB code is as follows: Step 1: Calculate the local competitive pressure index; lambda = zeros(N,1); pi_values ​​= zeros(N,1); % Calculate the survival benefit value for each individual; for i = 1:N Finding Neighbors (Euclidean Distance) <h) dist = vecnorm(Z - Z(i,:), 2, 2); neighbors = find(dist <h&(1:N)' ~= i); %Calculate π_i; sum_term = 0; for j = neighbors' diff = Z(i,:) - Z(j,:); sum_term = sum_term + abs(sin(gamma(i)) * norm(diff); end pi_values(i) = sum_term; end % Calculate λ_i^t; for i = 1:N dist = vecnorm(Z - Z(i,:), 2, 2); neighbors = find(dist <h&(1:N)' ~= i); N_i = length(neighbors); sum_diff = 0; for j = neighbors' sum_diff = sum_diff + max(0, pi_values(j) - pi_values(i)); end lambda(i) = sum_diff / sqrt(N_i + r + eps); % Zero-prevention processing end Step 2: Construct an improved decreasing factor; G = 1 + log(1 + lambda + eps); % Formula 0025 Step 3: Position Update (Formula 0025) for i = 1:N %Generate Levi's flight stride beta = 1.5; % Typical Lévy parameter sigma = (gamma(1) / (gamma(1+beta)*sin(pi*beta / 2)))^(1 / beta); u = normrnd(0, sigma, [1,Dim]); v = normrnd(0, 1, [1,Dim]); levy = u . / (abs(v).^(1 / beta)); % Calculate location update rand_term = rand(1,Dim); delta = Z_best - Z(i,:); Z_new(i,:) = Z_best + rand_term .* G(i) .* levy .* delta; end end.

[0026] Preferably, based on the improved part, an improved artificial lemming algorithm is established to tune the proportional coefficient, integral coefficient, and derivative coefficient of the positional PID controller of the gimbal device.

[0027] Preferably, firstly, the positional PID control algorithm is implemented in discrete time, satisfying the following conditions at all times. The proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the discrete position PID control algorithm are correlated with the search individual positions of the improved artificial lemming algorithm. The improved artificial lemming algorithm updates the search individual's position using migration, burrowing, foraging, and predator avoidance strategies, ultimately obtaining the optimal search individual position. This is analyzed using the optimal proportional coefficient, integral coefficient, and differential coefficient. The search individual position update is driven by a fitness function. The fitness function is designed based on the time required to first reach and stabilize within the target value ±2%, taking into account control accuracy. The index is the integral of the absolute value of the control signal divided by the total simulation time. The mathematical model is as follows: ; In the formula, The total simulation duration is... For the first The real-time rotation angle value at any given moment. Rotate by the target angle; The improved artificial lemming algorithm includes the following code for associating the search individual's location: [Male_ALA_score,Male_ALA_pos,ALA_cg_curve,Position_curve]=ALA(SearchAgents_no,Max_iteration,lb,ub,dim,fobj); [IMale_IALA_score,IMale_IALA_pos,IALA_cg_curve,Position_curve1]=IALA(SearchAgents_no,Max_iteration,lb,ub,dim,fobj); Wherein, IALA_cg_curve is the fitness value of the improved artificial lemming algorithm (IALA), IMale_IALA_pos is the optimal proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd value of the improved artificial lemming algorithm (IALA), Position_curve1 is the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd value in the optimization process of the improved artificial lemming algorithm (IALA), and IMale_IALA_score is the control signal value.

[0028] Preferably, secondly, the improved artificial lemming algorithm tunes the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device, specifically as follows: S21. Initialize the maximum number of iterations T, problem dimension Dim, upper bound ub and lower bound lb of the search space, and maximum population size N of the improved artificial lemming algorithm. S22. Randomly initialize the search individual positions according to the search space boundaries, calculate the fitness value of the search individual positions after initialization, and take the search individual position corresponding to the minimum fitness value as the current best search individual position. S23. Calculate the capability factor value for the current iteration. If the energy factor value is greater than 1, the global search phase position update mathematical model is executed; otherwise, the local development phase position update mathematical model is executed. S24. Calculate the ecological suppression factor value and the crowding value near the current search individual, construct the ecologically driven migration mechanism, improve the mathematical model for position update in the global search stage, and update the position of the search individual. S25. Introduce a local competitive pressure index to calculate an improved escape factor value, and use the improved escape factor value to improve the local development stage position update mathematical model and update the position of the search individual. S26. Calculate the fitness value of the current search individual position, update the current best search individual position, and execute t=t+1 for the current iteration number t; if t≤T is satisfied, return to execute S23, otherwise take the current best search individual position as the global best individual position; S27. Output the global optimal individual position and resolve it into the optimal proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd values.

[0029] Preferably, the search individual positions are randomly initialized according to the search space boundaries, and the code is implemented according to the following mathematical model: ; In the formula, Let j be the lower bound of the search space in dimension j. Let be the upper bound of the j-th dimension of the search space, rand be a normally distributed random number between 0 and 1, N be the maximum population size, set to 30, and Dim be the problem dimension, set to 3.

[0030] Preferably, the capability factor value Implement the code according to the following mathematical model: ; In the formula, The maximum number of iterations is 50, t is the current iteration number, and rand is a normally distributed random number between 0 and 1.

[0031] Preferably, the improved global search phase position update is divided into an ecosystem-driven migration strategy and a standard mining strategy. The improved global search phase position update mathematical model is implemented according to the following code: .

[0032] Preferably, the improved escape factor improves the predator avoidance phase, and the improved local development phase position update mathematical model is implemented according to the following code: ; Where *spiral* is the spiral term, and the mathematical model is: ; in, The j-th dimension value represents the current best search individual position. This represents the j-th dimension value of the current searched individual's location.

[0033] S3. Input the difference between the target rotation angle and the real-time rotation angle of the gimbal into the gimbal device controller, and output a control signal through a position-type PID controller. ,Will Input the angle control model and output the real-time rotation angle value of the gimbal.

[0034] Preferably, according to the mathematical model Calculate the difference in rotation angles ,Will The positional PID control algorithm takes the input value as input, and outputs the control signal after passing through the positional PID mathematical model. Implement the code according to the following mathematical model: ; In this formula, Kp, Ki, and Kd are the optimal values ​​of the proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd. The total simulation duration was set to 60 seconds, with the gimbal target rotation angles set to a small rotation angle of 10 degrees and a large rotation angle of 50 degrees, respectively, to test the control performance of the control method of the present invention on the fishery gimbal during large and small angle rotations.

[0035] Preferably, the sub-file code for the standard artificial lemming algorithm, the improved artificial lemming algorithm, the corner control model, and the positional PID control model is completed according to the above code. Then, the Main function code is designed, setting the maximum number of iterations T=50, the problem dimension Dim=3, the upper bound of the search space ub=[80,80,80] and the lower bound lb=[0,0.001,0], and the maximum population size N=30. A Simulink simulation model is completed according to the above code. The Main program is run to complete the self-adjusting control optimization process of the fishery gimbal equipment according to the method of this invention, and the output is as follows: Figures 3 to 6The control optimization results are shown in the figure.

[0036] Preferably, such as Figure 3 As shown, under the design conditions of this invention, the standard ALA algorithm converges slowly in complex scenarios and enters a false optimum after 9 iterations, getting trapped in a local optimum. Under the same calibration conditions, the fitness value of the improved ALA algorithm is significantly reduced to 0.5, indicating that after optimization through the ecological driving mechanism and the local competitive pressure index, the global search capability and convergence accuracy are greatly improved. The final fitness value is smaller than that of the standard ALA algorithm under the design conditions of this invention, indicating that the proportional coefficient, integral coefficient and derivative coefficient of the positional PID control algorithm are more accurate.

[0037] Preferably, the standard artificial lemming algorithm is used to tune the proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device as follows: Figure 4 As shown in the corresponding fitness value variation graph, stability was achieved after 9 iterations, with Kp=3.27535; Ki=3.77174; Kd=2.00316. The tuning results of the proportional coefficient, integral coefficient, and derivative coefficient of the positional PID controller of the gimbal device using the improved artificial lemming algorithm of this invention are as follows. Figure 5 As shown, it reaches stability after 34 iterations, with Kp=21.0456; Ki=2.08650; Kd=1.38812.

[0038] Preferably, such as Figure 6 As shown, the control method of the present invention is superior to the ALA-PID control method in terms of fast response, stability, and control of large and small angle rotation of the fishery gimbal. Specifically, the control method of the present invention exhibits a faster response speed to control the fishery gimbal, especially when rotating at a large angle (50 degrees), it can quickly stabilize to the target angle without obvious overshoot. When rotating at a small angle, the control effect is more stable, the system adjusts quickly, and eventually stabilizes at the target angle. In contrast, the ALA-PID control method is slightly inferior in response to both angles, especially when rotating at a large angle, the overshoot is more obvious, and there is a large oscillation when approaching the target angle. When rotating at a small angle, it is also slightly slower than the control method of the present invention, with some overshoot and oscillation.

Claims

1. A self-adjusting control optimization method for a fishery gimbal device, characterized in that, The specific steps are as follows: S1. Construct a control model for fishery gimbal equipment, which includes a rotation control model, an improved artificial lemming algorithm, and a positional PID controller model; S2. Improve the standard artificial lemming algorithm by using the improved artificial lemming algorithm to tune the proportional coefficient, integral coefficient and derivative coefficient of the positional PID controller of the gimbal device, thereby optimizing the gimbal device controller. The improved artificial lemming algorithm includes: introducing an ecology-driven migration mechanism, constructing an ecological environment quality field, and using a local gradient guidance strategy to drive search individuals to migrate to areas with higher ecological suitability, thereby updating the search individual's location through directional exploration; and introducing a local competitive pressure index. An improved escape factor was constructed to improve the mathematical model for updating the search individual's position during the predator avoidance phase; S3. Input the difference between the target rotation angle and the real-time rotation angle of the gimbal into the gimbal device controller, and output a control signal through a position-type PID controller. , will the Input the angle control model and output the real-time rotation angle value of the gimbal.

2. The self-adjusting control optimization method for a fishery gimbal device according to claim 1, characterized in that, The positional PID is implemented in discrete time, satisfying the condition at all times. The proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the discrete position PID control algorithm are correlated with the search individual positions of the improved artificial lemming algorithm. The improved artificial lemming algorithm updates the search individual's position using migration, burrowing, foraging, and predator avoidance strategies, ultimately obtaining the optimal search individual position. This is analyzed using the optimal proportional coefficient, integral coefficient, and differential coefficient. The search individual position update is driven by a fitness function. The fitness function is designed based on the time required to first reach and stabilize within the target value ±2%, taking into account control accuracy. The index is the integral of the absolute value of the control signal divided by the total simulation time. The mathematical model is as follows: ； In the formula, The total simulation duration is... For the first The real-time rotation angle value at any given moment. Rotate by the target angle.

3. The self-adjusting control optimization method for a fishery gimbal device according to claim 2, characterized in that, The ecologically driven migration mechanism simulates the model of how individuals perceive the environmental quality field and make decisions based on it when migrating in an ecosystem, starting from ecological modeling and non-uniform field driving mechanisms. The regions with higher fitness are denoted as locations with lower fitness values ​​for the search individuals. This strategy simulates the behavior of animals in natural ecosystems making choices based on habitat quality to construct a mathematical model for updating the search individual's location. The specific method is as follows: S301. Construct an ecological environment quality field, and define the search space as... The search space dimension is determined based on the question dimension Dim, defining the location (search point) of each search entity. ecological quality The mathematical model is: ； In the formula, Let i be the position of the i-th search entity in the t-th iteration. The fitness value for the location of the searched individual. As an ecological suppressor, it adjusts the weight of ecological quality and crowding. The mathematical model for the crowding around the current searched individual is as follows: ； in, Let be the position of the j-th search individual near the i-th search individual in the t-th iteration, n be the number of nearby search individuals, and h be the neighborhood perception factor; Among them, ecological suppression factors The mathematical model is as follows: ； In the formula, For the ecological quality of the largest neighboring point, For the minimum neighboring point ecological quality, Let be the ecological quality value of the i-th search individual; S302. The virtual gradient of ecological quality is used to guide the search direction for individuals. The mathematical model is as follows: ； In the formula, Let be the virtual gradient value for the i-th search individual in the t-th iteration, where the gradient points in the direction of the fastest improvement in ecological quality. To select k higher-quality neighboring points from the searched individual, The mathematical model for poor ecological quality is as follows: μ is used to prevent the value from returning to zero. S303. An ecologically driven migration strategy is constructed to update the location of search individuals through an ecologically driven migration mechanism. The mathematical model is as follows: ； In the formula, The maximum number of iterations, This is the current best location for the individual being searched.

4. The self-adjusting control optimization method for a fishery gimbal device according to claim 3, characterized in that, The introduction of a local competition pressure index The improved escape factor is constructed using the following mathematical model: ； In the formula, The survival benefit value for the i-th search individual: Let be the local neighbor survival benefit value of the j-th search individual. , Let r be the number of local neighbors of the i-th search individual, and r be a random number between 1 and 2. The mathematical model is as follows: ； In the formula, Let i be the crowding density near the i-th search individual in the t-th iteration; Based on local competitive pressure index Constructing an improved escape factor: ； In the formula, Let be the escape factor value of the i-th search individual in the t-th iteration; Then, the mathematical model for updating the search individual's position during the predator avoidance phase is improved using an improved escape factor: ； In the formula, rand is a normally distributed random number between 0 and 1, Dim is the question dimension, set to 3; Levy is the Levy distribution formula. This is the current best location for the individual being searched.

5. The self-adjusting control optimization method for a fishery gimbal device according to claim 4, characterized in that, The proportional, integral, and derivative coefficients of the positional PID controller of the gimbal device are tuned using an improved artificial lemming algorithm. The specific method is as follows: S21. Initialize the maximum number of iterations T, problem dimension D, upper bound ub and lower bound lb of the search space, and maximum population size N of the improved artificial lemming algorithm. S22. Randomly initialize the search individual positions according to the search space boundaries, calculate the fitness value of the search individual positions after initialization, and take the search individual position corresponding to the minimum fitness value as the current best search individual position. S23. Calculate the capability factor value for the current iteration. If the energy factor value is greater than 1, the global search phase position update mathematical model is executed; otherwise, the local development phase position update mathematical model is executed. S24. Calculate the ecological suppression factor value and the crowding value near the current search individual, construct the ecologically driven migration mechanism, improve the mathematical model for position update in the global search stage, and update the position of the search individual. S25. Introduce a local competitive pressure index to calculate an improved escape factor value, and use the improved escape factor value to improve the local development stage position update mathematical model and update the position of the search individual. S26. Calculate the fitness value of the current search individual position, update the current best search individual position, and execute t=t+1 for the current iteration number t; if t≤T is satisfied, return to execute S23, otherwise take the current best search individual position as the global best individual position; S27. Output the global optimal individual position and resolve it into the optimal proportional coefficient Kp, integral coefficient Ki, and differential coefficient Kd values.

Images