A method, program, device, and storage medium for identifying parameters of a ship motion model.

Abstract

This invention discloses a method, program, device, and storage medium for identifying ship motion model parameters, belonging to the field of model parameter identification. The invention employs the AC-PSA algorithm, initializing multiple populations. The position of a single individual in each population represents a ship motion model parameter to be identified, and the overall position vector of the population represents a vector composed of a set of parameters to be identified. The fitness values ​​of each population are calculated, and the population with the highest corresponding fitness value is selected as the optimal population. For each population, the position vector deviation between it and the optimal population in the current iteration is calculated. To accelerate algorithm convergence and improve generalization performance, this invention adds an automatic parameter construction part to the PSA algorithm, enabling simultaneous estimation of the structure and parameters of the ship maneuvering motion regression model. This invention utilizes the parallel optimization advantages and self-organizing mechanism of metaheuristic algorithms to promote rapid algorithm convergence and reduce computational load.

Description

Technical Field

[0001] This invention belongs to the field of model parameter identification, specifically relating to a method, program, device, and storage medium for identifying parameters of a ship motion model. Background Technology

[0002] With the rapid development of autonomous underwater and surface vessel technologies, the demand for accurate ship model parameters is becoming increasingly urgent. Precise ship model parameters are crucial for navigation, control algorithm design, and ship performance optimization. Ship maneuvering motion models are mainly classified into several model structures, including discrete, integrated, responsive, and time-series models. Integrated models offer high accuracy in describing ship maneuvering motion. Current mainstream methods for determining model parameters can be categorized into four types: database or empirical formula methods, constrained model testing methods, theoretical and numerical calculation methods, and system identification methods. With the continuous development of experimental measurement and identification technologies, system identification methods based on free-sailing tests and ship model tests are among the most practical, effective, repeatable, and time-efficient methods. This method only requires the ship's state information and inertia, without needing to measure forces and torques. Furthermore, it can be applied to actual ships, effectively overcoming the scaling effect caused by the difference in Reynolds coefficients between ship models and actual ships. The identified parameter results exhibit good transferability and generalization. Summary of the Invention

[0003] The purpose of this invention is to improve the identification accuracy of ship motion model parameters, and to achieve faster parameter convergence speed and excellent ability to escape local optima.

[0004] This invention provides a method for identifying parameters of a ship motion model, characterized by the following steps:

[0005] Step 1: Obtain the parameters to be identified for the ship motion model, initialize multiple populations, where the position of a single individual in each population represents a parameter of the ship motion model to be identified, and the position vector of the entire population represents a vector composed of a set of parameters to be identified.

[0006] Step 2: Calculate the fitness value of each population and select the population with the highest corresponding fitness value as the optimal population;

[0007] Step 3: For each population, calculate its position vector deviation from the optimal population in the current iteration number;

[0008] Step 4: Based on the position vector deviation between the current iteration number and the previous two iteration numbers, calculate the output value of the PID control and the adjustment factor for zero output;

[0009] Step 5: Update the position vector of each population based on the output value of the PID control and the zero-output control factor;

[0010] Step 6: Calculate the correlation measure of the optimal population, and correct the updated position vector of each population to improve the convergence speed;

[0011] Step 7: If the preset maximum number of iterations is reached or the position vector of the optimal population converges, the calculation ends, and the combination of ship motion model parameters to be identified corresponding to the position vector of the optimal population is output, thus completing the parameter identification of the ship motion model; otherwise, the iteration count is incremented by one, and the process returns to step 2.

[0012] Furthermore, in step 1, the initial population x ij :

[0013] x ij =(u j -l j )·r1+l j ,i=1,2,...,n; j=1,2,...,d

[0014] Where i is the population size, j is the number of parameters to be identified in the ship model; u j and l j These are the upper and lower bounds of the j-th variable, respectively, where r1∈[0,1].

[0015] Furthermore, in step 2, the fitness function f i :

[0016] f i =Fit(x) i )

[0017]

[0018] Among them, J e To estimate the error; The vector of ship model parameters to be identified. This refers to the data vectors collected during ship trials. for The estimated value; for The estimated value;

[0019] When the fitness value Fit approaches 1, the estimation error J e Approaching 0; estimated value Approximately equal to the true value

[0020] Furthermore, in step 3, when the number of iterations is t, the deviation of the population in the previous iteration is e. k-1 (t), the deviation of the population in the first two iterations is e k-2 (t).

[0021] When the iteration number t = 1, let e k-2 (t)=e k-1 (t)=e k (t), the population deviation e from the iteration number t k (t):

[0022] e k (t)=x * (t-1)-x(t-1)

[0023] Where, x * x(t) represents the optimal individual at iteration number t; x(t-1) represents the initial random individual.

[0024] When the number of iterations is t > 1, e k-2 (t)=e k-1 (t-1), the deviation e of the population in the previous iteration. k-1 (t):

[0025] e k-1 (t)=e k (t-1)+x * (t)-x * (t-1)

[0026] Furthermore, in step 5, when the number of iterations is t, the output value Δu(t) of the PID control is:

[0027] Δu(t)=K p ·r2·[e k (t)-e k-1 (t)]+K i ·r3·e k (t)+K d ·r4·[e k (t)-2e k-1 (t)+e k-2 (t)]

[0028] Where r2, r3, and r4 are n-row, 1-column random number vectors ranging from 0 to 1; K p K i K d These are the adjustment coefficients for Proportion, Integral, and Differential, respectively.

[0029] The zero-output adjustment factor o(t);

[0030] o(t)=(cos(1-t / T)+λr5·L)·e k (t)

[0031] λ = [ln(T-t+2) / ln(T)] 2

[0032]

[0033] Where λ is the adjustment coefficient; r5 is an n-row, d-column random number vector from 0 to 1; L is an L′evy flight function; u and v are n-row and d-column random number matrices that follow a standard normal distribution, respectively;

[0034] The population undergoes an initial update x(t+1):

[0035] x(t+1)=x(t)+η·Δu(t)+(1-η)·o(t)

[0036] Where η = r6cos(t / T), T is the maximum number of iterations set, and r6 is an n-row, 1-column random number vector of 0 to 1.

[0037] Furthermore, the correlation measure C of the optimal population r (Φ(l,t),E(1,t)), choose C r The largest k 0max th Item, k 0max =1

[0038]

[0039] If the k-th individual satisfies the normalized correlation coefficient NC r (Φ(l,t),Z(l,t))<ρ, where ρ∈[0,1] is the set parameter, and it has been selected consecutively n times. s Next, the weight variable w k =0, and we get the weight variable matrix w(t). All other weight variables in this matrix are 1. If the value is 1, the individual is retained; otherwise, it is cleared (set to 0).

[0040]

[0041] Updated values ​​x(t+1) for all populations:

[0042] x(t+1) = x(t+1).*w(t)

[0043] This invention also provides a ship motion model parameter identification system, comprising: a data input module for selecting appropriate data and converting it into a unified format based on measured ship navigation data; a parameter setting module for setting the ship motion model type for identification, selecting an appropriate model structure according to actual needs, and setting the necessary initial parameters for the algorithm; a model identification module for iteratively identifying the required model parameters; and a result verification module for displaying the process curve of iterative optimization, the identification results, and the comparison effect with the original data, facilitating timely adjustment of parameters to obtain better identification results.

[0044] The present invention also provides a computer device / equipment / system, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the ship motion model parameter identification method described above.

[0045] The present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps of the ship motion model parameter identification method described in any of the above claims.

[0046] The present invention also provides a computer program product, including a computer program / instruction that, when executed by a processor, implements the steps of the ship motion model parameter identification method described above.

[0047] The beneficial effects of this invention are as follows:

[0048] (1) This invention utilizes the parallel optimization advantage and self-organizing mechanism of metaheuristic algorithms to promote the rapid convergence of the algorithm and reduce the amount of computation.

[0049] (2) This invention utilizes the population diversity of metaheuristic algorithms. Randomization can not only improve the probability of the algorithm escaping local optima, but also its diversity, thereby enhancing the usability of the identification results.

[0050] (3) This invention solves the problem of the black box structure of ships, without requiring precise knowledge of the internal structural parameters and their corresponding relationships, and has good adaptability to solving the problem of ship parameter identification; the algorithm shows high robustness to the interference of currents and waves as well as the static error of the rudder angle. Compared with other metaheuristic algorithms applied to the field of ship parameter identification, this application requires fewer parameters to be optimized and requires less computation time, thus exhibiting high real-time performance.

[0051] (4) The algorithm of this invention has self-organizing structure: for complex multi-parameter models, the model dimension can be automatically simplified according to the importance of the values. Attached Figure Description

[0052] Figure 1This is a schematic diagram of the PID control algorithm of the present invention;

[0053] Figure 2 A schematic diagram illustrating the search principle of the search algorithm provided in this embodiment of the invention;

[0054] Figure 3 This is a schematic diagram illustrating the change of the adjustment coefficient of the zero-output adjustment factor over time in the algorithm provided in this embodiment of the invention.

[0055] Figure 4 A flowchart of the identification algorithm provided in an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the structure of a device according to an embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0058] The present invention will now be further described with reference to the accompanying drawings.

[0059] This invention discloses a system for identifying parameters of a ship motion model, such as... Figure 5 As shown, it includes:

[0060] S1, through the data input module, is used to select appropriate data for unified format conversion based on the measured ship navigation data. Taking a three-degree-of-freedom ship model as an example, it can input the ship's total speed, longitudinal speed, lateral speed, turning angular velocity, real-time rudder angle, sampling time, heading angle, GPS position information, etc., obtained from the ship's full turn or Z-shaped test.

[0061] S2, through the model and initial parameter setting module, is used to set the type of ship operation model for identification. The appropriate model structure can be selected according to actual needs, such as the Nomoto model, the integral model, the split model, etc., and the necessary initial parameters of the algorithm can be set, including basic ship parameters such as ship length, total mass, number of algorithm iterations, identification tolerance, etc.

[0062] S3, through the algorithm iteration identification module, uses the parameters input from S1 and S2 for iterative looping. The AC-PSA algorithm gradually converges to the set conditions and identifies the required model parameters.

[0063] S4 displays the identified model parameters and process data through the result output and effect verification comparison module, such as the process curve of iterative optimization, parameter change curve, error convergence curve, and compares the identification results with the original data, so as to adjust the parameters in a timely manner to obtain better identification results.

[0064] The AC-PSA algorithm is inspired by PID control, a classic control algorithm in the field of automation. This algorithm employs a discretized incremental PID algorithm. Incremental PID control is a proportional-fair integral-derivative control, and its adjustment process is as follows: Figure 1 As shown in the diagram, after the user sets a target value, the incremental PID controller outputs a value to the actuator, which then adjusts the controlled object based on the output value. The actual value of the controlled object after adjustment is collected by the sensor and transmitted back to the incremental PID controller. PID control can effectively correct the deviation of the controlled object, bringing the system towards a stable state. The AC-PSA algorithm is inspired by this process, abstracting the historical best individual in the ship parameters as the target value and each individual as the actual value. By simulating the PID adjustment process and performing self-organizing parameter selection, it corrects the deviation of each individual relative to the best individual, gradually bringing the entire population closer to the optimal state. This algorithm is a high-performance novel swarm intelligence optimization algorithm with a simple structure, requiring only the adjustment of four parameters. As a metaheuristic algorithm based on mathematical formulas, it is very suitable for the field of system identification.

[0065] This invention provides a method for combining this algorithm (AC-PSA) with... Figure 1 and Figure 4 The procedure for identifying ship model parameters is as follows:

[0066] S1. Define the ship model. Currently, this invention supports widely popular ship motion models such as the responsive model (KT), the dissociative model (MMG), and the integral model (Abkowitz). Taking the more complex MMG model as an example: This model decomposes the hydrodynamic forces acting on the ship into hydrodynamic forces acting on the bare hull, the open-water propeller, and the open-water rudder, and then considers the interference between these three. The model formula is as follows:

[0067]

[0068] Among them, I x ,I z ,m x ,m y J x J z These represent the moments of inertia relative to the x-axis and z-axis, respectively, along with the corresponding additional mass and additional moments of inertia. Furthermore, all of the above can be estimated using traditional empirical formulas. The notes will focus on identifying the right-hand side of the equation, namely the hydrodynamic forces and torques acting on the hull, propeller, and rudder: X H ,Y H ,K H N H ,X p ,Y p ,K PN P ,X R ,Y R ,K R N R Note: Footnotes H, P, and R correspond to the English abbreviations for 'hull', 'propeller', and 'rudder', respectively.

[0069] In this model, the forces or moments of the hull are described using a damping force expression. Therefore, the identification scheme proposed in this invention can automatically construct detailed components for different types of ships. Thus, the damping model should be derived from a combination of existing results. The following equation presents the damping model, which can be represented as X. H =X uu u 2 +X uuu u 3 +...+X uvφφ uvφ 2 and Y H ,K H N H These coefficients X uu ,X uuu ,...,N rr These are the hydrodynamic derivatives (ship model parameters) that need to be identified by the proposed identification scheme, where [a1,a2,...,a9], [b1,b2,...,b9] are used. 16 ],[c1,c2,...,c 16 ] and [d1,d2,...,d 16 ]express.

[0070] X H =[a1,a2,...,a9]·[u 2 ,u 3 ,v 2 ,r 2 ,vr,uv 2 ,uvr,ur 2 ,uvφ 2 ] T

[0071] Y H =[b1,b2,...,b 16 ]·[uv,ur,up,uφ,u 2 r,u 2 v,v 3 ,r 3 VR 2 ,v 2 r,uvp 2 ,urp 2 ,|v|v,|r|v,|v|r,|r|r] T

[0072] K H =[c1,c2,...,c 16 ]·[p,p 3 ,uv,ur,u 2 v,u 2 r,v 3 ,r 3 VR 2 ,v 2 r,uvφ 2 ,urφ 2 ,|v|v,|r|v,|v|r,|r|r] T

[0073] N H =[d1,d2,...,d 16 ]·[uv,ur,up,uφ,u 2 r,u 2 v,v 3 ,r 3 VR 2 ,v 2 r,uvp 2 ,urp 2 ,|v|v,|r|v,|v|r,|r|r] T

[0074] Propeller X p ,Y p ,K P N P And Rudder X R ,Y R ,K R N R The force and torque are expressed as follows:

[0075] X p =(1-t) p )T(J p )

[0076] Y p =K p =N p =0

[0077] X R =-(1-t) R )F N sinδ

[0078] Y R =-(1+a) H )F N cosδ

[0079] K R=(1+a) H )Z HR F N cosδ

[0080] N R =-(x R +a H x H )F N cosδ

[0081]

[0082]

[0083] Among them, t P ,n,D P ,K T J p These represent the thrust depreciation fraction, propeller speed per second, propeller diameter, thrust coefficient, and propeller advance ratio, respectively. J0, J1, and J2 are the hydrodynamic coefficients of the open-water propulsion system. For the rudder force / torque, δ, x R Z HR This represents the rudder angle, indicating the center of the normal force acting on the rudder in the x and z directions. Additionally, t... R ,a H ,x H Used to describe the interaction between the hull and the rudder. In the identification scheme, the parameter formula (1-t) R ),(1+α H ),(1+α H )Z HR ,(xR+α H x H This is also considered a coefficient to be estimated. Other variables are obtained through physical measurements, experiments, or empirical formulas, such as: F N ,T(J p ).

[0084] S2. Conduct standard ship maneuvering experiments and collect test data. Requirements: The experiment should incorporate weather forecast information to obtain the wind, wave, and tidal conditions of the test sea area. Select relatively calm waters, with sea state level not exceeding 3, for the actual ship maneuvering tests, such as full-turn experiments at various rudder angles, to obtain a full-scale ship maneuvering dataset. The ship maneuvering dataset mainly consists of ship sensor data, including global navigation satellite system data, inertial navigation data, and radar data. Extract data affecting ship motion characteristics, including eight items: ship roll angle, bow angle, pitch speed, sway speed, position latitude and longitude, control command rudder angle, and speed.

[0085] S3, Data Processing. This is to filter out the influence of currents. In real ocean environments, completely calm conditions are almost impossible; there is usually some environmental disturbance. During turning tests, the trajectories cannot coincide, and the combined environmental forces will cause the ship to deviate in a certain direction. We generally assume that environmental forces do not change significantly over a period of time, and can treat the overall environmental disturbance as an "environmental current" applied to the ship. We then make corrections based on this "environmental current" to obtain ship turning data similar to a still water experiment.

[0086] First, the tangent angle of the actual flight path needs to be calculated based on the position data (X, Y) collected from GPS data:

[0087] χ t =arctan(X) t+1 -X t ,Y t+1 -Y t )

[0088] Find the time when the next tangent angle is the same, and record the time for one revolution under this condition as Δt. Then, the positional offset within each revolution circle under the influence of the flow can be calculated, and the average flow velocity at the current position can be calculated as follows:

[0089]

[0090] After removing the influence of the stream from the collected data, we consider it to be a relatively clean dataset, and then we use the AC-PSA algorithm for computation.

[0091] S4, Initialize algorithm parameters. The problem consists of a set of decision variables, constraints, and an objective function. Let's assume the number of decision variables in the population is d, and the upper and lower bounds of the variables are u and l, respectively. The control parameters of AC-PSA include the maximum number of iterations T and the population size n. Then the initial population can be represented as:

[0092] x ij =(u j -l j )·r1+l j ,i=1,2,...,n; j=1,2,...,d

[0093] Where, x ij Represents the j-th dimension of the i-th individual; u j and l j These are the upper and lower bounds of the j-th variable (dimension), respectively; r1 is a random number from 0 to 1.

[0094] S5. To identify unknown parameters in the ship model, a suitable objective function, i.e., a fitness evaluation function, needs to be constructed to evaluate the output of the AC-PSA algorithm at each update. The input and output can be uniformly written in the following form:

[0095]

[0096] in, The vectors representing the ship model parameters to be identified are [a1,a2,...,a9], [b1,b2,...,b9], and [b1,b2,...,b9]. 16 ],[c1,c2,...,c 16 ] and [d1,d2,...,d 16 ], and This represents a vector of data collected during actual ship trials.

[0097] Let vector The estimated value is Then the estimation error J e It can be calculated using the formula:

[0098]

[0099] The fitness function is:

[0100]

[0101] When the fitness evaluation function value approaches 1, the estimation error approaches 0; that is, the estimated value... Approximately equal to the true value Assuming the objective function is Fir, then the objective function value f for the i-th individual is... i It can be calculated using the following formula:

[0102] f i =Fit(x) i )

[0103] S6, Calculate the system bias. For the minimization problem, calculate the optimal individual x at iteration number t. * (t) represents the individual corresponding to the historical minimum value of the population. The population deviation e from the iteration number t is... k (t) is:

[0104] e k (t)=x * (t-1)-x(t-1)

[0105] For ease of calculation and iterative updates, when the number of iterations is t, the deviation of the population in the previous iteration is denoted as e. k-1 (t), where the deviations of the population in the first two iterations are denoted as e. k-2 (t). Then when t=1, let ek-2 (t)=e k-1 (t)=e k (t) is worthwhile. When t>1, e k-2 (t)=e k-1 (t-1). In the PID algorithm, when t>1, e k-1 It should equal the deviation between the user-defined value and the actual value in the system at the previous moment. In AC-PSA, the user-defined value is the optimal value, but the optimal value after each iteration will not remain unchanged; therefore, in the proposed AC-PSA algorithm, e k-1 (t) cannot be simply equal to e k (t-1), such as Figure 2 As shown. To minimize the space complexity of the algorithm, e is... k (t-1)+x * (t)-x * (t-1) as the final e k-1 (t). Therefore, e k-1 (t) can be represented as

[0106] e k-1 (t)=e k (t-1)+x * (t)-x * (t-1)

[0107] S7, PID control. For example... Figure 3 As shown, in practical problems, the proportional, integral, and derivative factors are adjusted according to different situations and problems. Therefore, the output value Δu(t) of the PID control when the number of iterations is t is:

[0108] Δu(t)=k p ·r2·[e k (t)-e k-1 (t)]+k i ·r3·e k (t)+k d ·r4·[e k (t)-2e k-1 (t)+e k-2 (t)]

[0109] In the formula: r2, r3, r4 are n rows and 1 column random number vectors ranging from 0 to 1; k p k i and k d These are the adjustment coefficients for proportional, integral, and derivative equations, respectively.

[0110] In conventional PID algorithms, an output adjustment value of 0 signifies that the actual value has reached the user's setpoint at a certain moment. However, as time increases, if the controlled object is not adjusted, the actual value will quickly deviate from the setpoint. Therefore, to prevent this, a constant is usually added to the original adjustment value. This is disadvantageous for AC-PSA, as all solutions are extremely close to the optimal solution, especially in the early stages of iteration, which can lead the algorithm into local optima. Therefore, based on the original Δu(t), this algorithm also adds an adjustment factor called zero output to prevent the algorithm from getting trapped in local optima. Zero output is defined in the equation,

[0111] o(t)=(cos(1-t / T)+λr5·L)·e k (t)

[0112] Where r5 is an n-row, d-column random number vector from 0 to 1; λ is an adjustment coefficient, calculated in the formula;

[0113] λ = [ln(T-t+2) / ln(T)] 2

[0114] The variation of λ with t is as follows Figure 3 As shown, λ decreases slowly as t increases, which helps the algorithm to explore fully. In the later stages, λ decreases rapidly, which helps the algorithm shift from exploration to exploitation.

[0115] In the zero-output definition equation: L is an L′evy flight function, defined in the following equation:

[0116]

[0117] Where ε and μ are n-row and d-column random number matrices following a standard normal distribution, respectively; β is a factor set to 1.5; Γ(1+β) is the value of the gamma function at 1+β, which is a generalization of the factorial and is defined as:

[0118]

[0119] The updates of all individuals are related to Δu(t) and o(t). The population update formula is defined as:

[0120] x(t+1)=x(t)+η·Δu(t)+(1-η)·o(t)

[0121] Where η is an n x 1 matrix, represented as:

[0122] η = r6cos(t / T)

[0123] Here, r6 is an n-row, 1-column random number matrix between 0 and 1.

[0124] The ship parameters obtained in this calculation are:

[0125] S8, to accelerate algorithm convergence and improve generalization performance, an automatic parameter construction (AC) part is added to the PSA algorithm. This part can simultaneously estimate the structure and parameters of the ship maneuvering motion regression model. For the generalized linear regression model... Construct a weight vector w(t) = [w1, w2, ..., w n ] T The introduction is of course for w i (t), i = u, v, p, r, with the choice of structure represented by its element '0' or '1'.

[0126] Correlation coefficient C r (α,β) is an important indicator in automatic parameter construction structure estimation. It establishes the correlation measure α(i),β(i), i=1,2,…,N between two time series, with the following formula:

[0127]

[0128] In the formula: C r (Φ k (l,t),E(l,t)) reflects the k-th regression term (Φ k (l,t) represents the k-th row vector of Φ(l,t). This relates to the parameter information E(l,t), C. r (Φ k The correlation magnitude of Z(l,t) reflects the contribution of the k-th regressor to the system output Z(l,t). For units k = 1, 2, ..., n, a normalized correlation coefficient NC is defined. r (Φ k (l,t),Z(l,t)) is:

[0129]

[0130] In the AC-PSA algorithm, C r (Φ k (l,t),E(l,t)) calculate the alternate regression (where ω) at each search step. k =0), select the kth largest Cr. 0max The term is used as a newly added regressor (i.e., k). 0max =1). Then calculate NC. r (Φ k (l,t),Z(l,t)), with the selected regression factor being w. k =1. At the current step size, NC r (Φk If the calculated value of the function (l,t),Z(l,t) is worse than the preset accuracy threshold ρ, it is recorded as one instance. When n consecutive records are made... s The subsequent regression factor is w k =0. After structural selection at each sampling time, the decision vector is... Replace with (.* represents array operations defined in MATLAB). Here The PSA algorithm will be used for updating in this embodiment. Regarding the setting of the accuracy threshold ρ, the basic principle is to retain the propeller-related regressor, i.e., the corresponding w. i Since ρ is not equal to 0, it can be appropriately reduced to improve the model accuracy.

[0131] The final ship parameters output in this algorithm calculation process are as follows:

[0132]

[0133] Recalculate S5. If the fitness does not meet the set conditions, repeat S6-S8 until the set number of iterations is reached or the error converges to the set value. At the end of the algorithm iteration, select the current optimal output value as the optimal parameter for the MMG ship model based on the fitness function result.

[0134] In particular, in some preferred embodiments of the present invention, a computer device is also provided, including a memory and a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the ship motion model parameter identification method described in any of the above embodiments.

[0135] In some other preferred embodiments of the present invention, a computer-readable storage medium is also provided, on which a computer program / instruction is stored, wherein when the computer program is executed by a processor, the steps of the ship motion model parameter identification method described in any of the above embodiments are implemented.

[0136] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above embodiments of the ship motion model parameter identification method, which will not be repeated here.

[0137] It should be noted that the terms "front" and "back" used in this invention are only relative concepts or are based on the normal use of the product, and should not be considered restrictive.

[0138] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various equivalent substitutions or obvious modifications can be made without departing from the concept of the present invention, and all such modifications, achieving the same performance or application, should be considered within the scope of protection of the present invention.

Claims

1. A method for identifying parameters of a ship motion model, characterized in that: Includes the following steps: Step 1: Obtain the parameters to be identified for the ship motion model, initialize multiple populations, where the position of a single individual in each population represents a parameter of the ship motion model to be identified, and the position vector of the entire population represents a vector composed of a set of parameters to be identified. Step 2: Calculate the fitness value of each population and select the population with the highest corresponding fitness value as the optimal population; Step 3: For each population, calculate its position vector deviation from the optimal population in the current iteration number; When the number of iterations is At that time, the deviation of the population in the previous iteration was The deviation of the population in the first two iterations is ; When the number of iterations season Population iteration count deviation : ； in, When the number of iterations The optimal individual at that time; These are initial random individuals; When the number of iterations is hour, The bias of the population in the previous iteration : ； Step 4: Based on the position vector deviation between the current iteration number and the previous two iteration numbers, calculate the output value of the PID control and the adjustment factor for zero output; When the number of iterations is At that time, the output value of PID control : ； in, , , for A vector of random numbers, row 1, column 0 to 1; , , These are the adjustment coefficients for Proportion, Integral, and Differential, respectively. The zero-output adjustment factor ; ； ； ； in, For adjustment coefficients; for OK A vector of random numbers ranging from 0 to 1; For an L'evy flight function; and They are respectively following a standard normal distribution. lines and Columnar random number matrix; , ; Step 5: Update the position vector of each population based on the output value of the PID control and the zero-output control factor; Step 6: Calculate the correlation measure of the optimal population, and correct the updated position vector of each population to improve the convergence speed; Step 7: If the preset maximum number of iterations is reached or the position vector of the optimal population converges, the calculation ends, and the combination of ship motion model parameters to be identified corresponding to the position vector of the optimal population is output, thus completing the parameter identification of the ship motion model; otherwise, the iteration count is incremented by one, and the process returns to step 2.

2. The method for identifying ship motion model parameters according to claim 1, characterized in that: In step 1, the initial population : ； in, Population size This represents the number of parameters that need to be identified in the ship model. and They are the first Upper and lower bounds of each variable. .

3. The method for identifying ship motion model parameters according to claim 1, characterized in that: In step 2, the fitness function : ； ； ； in, To estimate the error; , The vector of ship model parameters to be identified. , This refers to the data vectors collected during ship trials. for The estimated value; for The estimated value; When fitness value When the value approaches 1, the estimation error Approaching 0; estimated value Approximately equal to the true value ; 4. The method for identifying ship motion model parameters according to claim 1, characterized in that: In step 5 The population underwent a preliminary update. : ； in, , The maximum number of iterations is set. for A vector of random numbers in row 1 and column 0 to 1.

5. The method for identifying ship motion model parameters according to claim 4, characterized in that: The correlation measure of the optimal population ,choose The largest item, , ; ； If the Each individual satisfies the normalized correlation coefficient. , A preset precision threshold has been selected consecutively. Next, weight variables The weight variable matrix is ​​obtained. In this matrix, all other weight variables are set to 1. If the value is 1, the individual is retained; otherwise, it is removed (set to 0). ； Update values ​​for all populations : 。 6. A system for identifying parameters of a ship motion model, characterized in that: For performing the parameter identification method of claim 1, comprising: The data input module is used to select appropriate data and convert it into a unified format based on the measured ship navigation data. The parameter setting module is used to set the type of ship operation model used for identification, select the appropriate model structure according to actual needs, and set the necessary initial parameters for the algorithm. The model identification module is used to iteratively identify the required model parameters. The results verification module is used to display the process curves of iterative optimization, identify the results and compare them with the original data, so as to make timely adjustments to the parameters to obtain better identification results.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that: When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 5.

Images