CN116880184A - Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium - Google Patents
Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium Download PDFInfo
- Publication number
- CN116880184A CN116880184A CN202310856720.7A CN202310856720A CN116880184A CN 116880184 A CN116880184 A CN 116880184A CN 202310856720 A CN202310856720 A CN 202310856720A CN 116880184 A CN116880184 A CN 116880184A
- Authority
- CN
- China
- Prior art keywords
- model
- unmanned ship
- hybrid physical
- track
- error dynamics
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 75
- 238000013528 artificial neural network Methods 0.000 claims abstract description 47
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 238000004364 calculation method Methods 0.000 claims abstract description 25
- 238000004088 simulation Methods 0.000 claims abstract description 19
- 238000006243 chemical reaction Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 56
- 230000006870 function Effects 0.000 claims description 34
- 238000005457 optimization Methods 0.000 claims description 17
- 230000009466 transformation Effects 0.000 claims description 6
- 230000007613 environmental effect Effects 0.000 abstract description 8
- 239000012530 fluid Substances 0.000 abstract description 8
- 238000013461 design Methods 0.000 description 5
- 238000013459 approach Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005312 nonlinear dynamic Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 230000007723 transport mechanism Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a unmanned ship track tracking prediction control method, a unmanned ship track tracking prediction control system and a storage medium, which are applied to the technical field of unmanned ship intelligent control, can relieve modeling errors and model prediction control algorithm calculation efficiency problems caused by uncertain fluid dynamics or environmental interference of an unmanned ship, and improve control stability. The method comprises the following steps: acquiring a training data set; wherein the training data set is obtained through analog simulation; training a preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model; constructing a hybrid physical learning model according to the learning model and the unmanned ship physical model; the hybrid physical learning error dynamics model is constructed and obtained by carrying out coordinate system conversion on the hybrid physical learning model; and carrying out unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
Description
Technical Field
The invention relates to the technical field of unmanned ship intelligent control, in particular to an unmanned ship track tracking prediction control method, an unmanned ship track tracking prediction control system and a storage medium.
Background
Model Predictive Control (MPC) has been widely used in industrial processes and robotics related fields, including autopilot and unmanned aerial vehicles. The practical use of MPC methods on unmanned vessels (USV, unmanned Surface Vehicle) is still rare. In the related art, on one hand, as a typical model-based control method, the performance of MPC is severely dependent on the accuracy of the system model. On the other hand, there is typically a tradeoff between model accuracy and computational burden or controller design complexity. Furthermore, it is often difficult to obtain accurate USV models due to general marine environmental disturbances and uncertain USV fluid dynamics. Even though the high-precision USV hydrodynamic model is ready, the heavy computational burden of MPC online computing is not satisfactory for real-time applications. Therefore, the above problems need to be solved.
Disclosure of Invention
In order to solve at least one of the technical problems, the invention provides an unmanned ship track tracking prediction control method, an unmanned ship track tracking prediction control system and a storage medium, which can relieve modeling errors and model prediction control algorithm calculation efficiency problems caused by uncertain fluid dynamics or environmental interference of an unmanned ship and improve control stability.
On the one hand, the embodiment of the invention provides an unmanned ship track tracking prediction control method, which comprises the following steps:
acquiring a training data set; wherein the training data set is obtained through analog simulation;
training a preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model;
constructing a hybrid physical learning model according to the learning model and the unmanned ship physical model;
the hybrid physical learning error dynamics model is constructed and obtained by carrying out coordinate system conversion on the hybrid physical learning model;
and carrying out unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
According to some embodiments of the invention, the acquiring a training data set includes:
performing track tracking simulation through an unmanned ship model and the unmanned ship physical model to obtain a first data set;
and normalizing the first data set to obtain the training data set.
According to some embodiments of the present invention, the constructing the hybrid physical learning error dynamics model by performing coordinate system transformation on the hybrid physical learning model includes:
Converting a preset reference track into an unmanned ship body coordinate system, and calculating a difference value between an expected track and a current track in the unmanned ship body coordinate system through the hybrid physical learning model to obtain an error kinematic model;
and obtaining the hybrid physical learning error dynamics model by carrying out preset differential operation on the error dynamics model.
According to some embodiments of the invention, after performing the step of constructing the hybrid physical learning error dynamics model by performing coordinate system transformation on the hybrid physical learning model, the method further includes:
calculating a preset maximum invariant set according to the hybrid physical learning error dynamics model;
and constructing a terminal area through the preset maximum unchanged set.
According to some embodiments of the invention, the calculating the preset maximum invariant set according to the hybrid physical learning error dynamics model includes:
linearizing the hybrid physical learning error dynamics model at an origin to obtain a first linearization function;
solving a Lyapunov equation according to the first linearization function to obtain a terminal weighting matrix;
and calculating according to the terminal weighting matrix to obtain the preset maximum unchanged set.
According to some embodiments of the invention, in the step of performing the unmanned ship trajectory tracking prediction control by the hybrid physical learning error dynamics model, the method further includes:
acquiring a seed control track according to the current motion state and the optimal control sequence at the previous moment;
constructing an incremental model according to the seed control track and the hybrid physical learning error dynamics model;
carrying out Jacobian matrix calculation according to the incremental model to obtain a continuous Jacobian matrix; wherein the continuous jacobian matrix comprises a continuous jacobian state matrix and a continuous jacobian control matrix;
discretizing the continuous jacobian matrix, and carrying out system state prediction according to the incremental model to construct and obtain a quadratic programming optimization function;
and continuously linearizing the hybrid physical learning error dynamics model through the quadratic programming optimization function.
According to some embodiments of the invention, the constructing an incremental model according to the seed control track and the hybrid physical learning error dynamics model includes:
carrying out Taylor expansion on the hybrid physical learning error dynamics model at the seed control track to obtain a model system track; wherein the model system trajectory is the sum of the seed control trajectory and the increment of the seed control trajectory;
Constructing and obtaining a deviation model according to the model system track and the seed control track;
and constructing and obtaining the incremental model according to the deviation model and the seed state updating function.
On the other hand, the embodiment of the invention also provides an unmanned ship track tracking prediction control system, which comprises the following steps:
a first module for acquiring a training data set; wherein the training data set is obtained through analog simulation;
the second module is used for training a preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model;
the third module is used for constructing and obtaining a hybrid physical learning model according to the learning model and the unmanned ship physical model;
the fourth module is used for constructing and obtaining a hybrid physical learning error dynamics model by carrying out coordinate system conversion on the hybrid physical learning model;
and the fifth module is used for carrying out unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
On the other hand, the embodiment of the invention also provides an unmanned ship track tracking prediction control system, which comprises the following steps:
at least one processor;
at least one memory for storing at least one program;
The at least one program, when executed by the at least one processor, causes the at least one processor to implement the unmanned ship trajectory tracking predictive control method as described in the above embodiment.
In another aspect, an embodiment of the present invention further provides a computer storage medium, in which a program executable by a processor is stored, where the program executable by the processor is used to implement the unmanned ship trajectory tracking prediction control method according to the above embodiment.
The unmanned ship track tracking prediction control method provided by the embodiment of the invention has at least the following beneficial effects: according to the embodiment of the invention, the training data set is firstly obtained, so that the preset neural network is trained through a preset neural network training algorithm according to the training data set, and a learning model is obtained. The training data set is obtained through simulation. Then, the embodiment of the invention constructs the hybrid physical learning model according to the learning model and the unmanned ship physical model, and converts the coordinate system of the hybrid physical learning model to construct the hybrid physical learning error dynamics model, so that the unmanned ship track tracking prediction control is performed according to the hybrid physical learning error dynamics model, and the accuracy and the efficiency of the unmanned ship track tracking prediction control are improved. Meanwhile, the method for training the preset neural network through the preset neural network training algorithm effectively relieves the modeling error problem caused by uncertain fluid dynamics or environmental interference of the unmanned ship. In addition, the embodiment of the invention can effectively relieve the problem of the calculation efficiency of the model predictive control algorithm and improve the control stability by constructing the obtained hybrid physical learning error dynamics model.
Drawings
FIG. 1 is a flowchart of an unmanned ship track tracking prediction control method provided by an embodiment of the application;
fig. 2 is a schematic diagram of a method for tracking, predicting and controlling a trajectory of an unmanned ship according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a unmanned ship track tracking predictive control system provided by an embodiment of the application;
fig. 4 is a schematic structural diagram of an unmanned ship track tracking prediction control system according to an embodiment of the present application.
Detailed Description
The embodiments described herein should not be construed as limiting the application, and all other embodiments, which may be made by those of ordinary skill in the art without the benefit of the present disclosure, are intended to be within the scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the application only and is not intended to be limiting of the application.
Before describing embodiments of the present application, related terms referred to in the present application will be first described.
Model predictive control (Model Predictive Control, MPC): is a control method for solving the optimization problem of a multivariable, nonlinear and constrained dynamic system. The model predictive control predicts based on the system model and finds the current optimal control strategy through optimization.
Deep Neural Network (DNN): is a feed-forward neural network that extracts and learns advanced features in data through multiple nonlinear transformation layers. Deep neural networks find application in many fields, such as computer vision, natural language processing, and speech recognition.
Jacobian matrix (Jacobian matrix) is a matrix that describes the partial derivatives of a multivariate function. Wherein if a function has n independent variables and m dependent variables, its jacobian matrix is an mxn matrix. Jacobian, which provides a means of linear approximation, can be used to describe the local behavior of nonlinear functions. By means of the jacobian matrix, a nonlinear function can be approximated as a linear function around a certain point.
Model Predictive Control (MPC) has been widely used in industrial processes and related fields of robotics, such as automatic driving automobiles and unmanned aerial vehicles. However, the model predictive control method is still rarely practiced on unmanned vessels (Unmanned Surface Vehicle, USV). On the one hand, as a typical model-based control method, the performance of model predictive control depends heavily on the accuracy of the system model. On the other hand, there is typically a tradeoff between model accuracy and computational burden or calculator design complexity. In the related art, due to general ocean environment disturbance and uncertain unmanned ship fluid dynamics, an accurate unmanned ship model is generally difficult to obtain, and the heavy calculation burden of model predictive control on-line calculation cannot meet the requirement of real-time application. In addition, most of the existing efforts of unmanned ship control based on model predictive control assume that unmanned ship models are precisely known to design special methods to address the effects of environmental disturbances. For example, for modeling error problems in ship dynamic positioning control, modeling error is estimated by involving a nonlinear disturbance observer, and the estimation result is used in a robust nonlinear model predictive control scheme. But this approach typically requires strict assumptions and complex mathematical derivations to ensure that the estimation errors converge. Meanwhile, the use of a high-precision system model in the controller design means a heavy calculation load, and thus, the efficiency of model predictive control on-line calculation needs to be studied. For example, dynamics are decoupled and split into three questions that are then resolved by distributed computing. However, this type of decoupling can only be achieved under zero yaw rate conditions, which is difficult to achieve in most maneuvering situations. Other methods focus on efficient optimization algorithm design, such as gaussian-newton optimization strategies, but the convergence of the algorithm cannot be guaranteed.
Based on the above, an embodiment of the invention provides a unmanned ship track tracking prediction control method, a system and a storage medium, which can relieve modeling errors and model prediction control algorithm calculation efficiency problems caused by uncertain fluid dynamics or environmental interference of an unmanned ship and improve control stability. Referring to fig. 1, the method of the embodiment of the present invention includes, but is not limited to, step S110, step S120, step S130, step S140, and step S150.
Specifically, the method application process of the embodiment of the invention includes, but is not limited to, the following steps:
s110: a training dataset is acquired. The training data set is obtained through analog simulation.
S120: training the preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model.
S130: and constructing and obtaining a hybrid physical learning model according to the learning model and the unmanned ship physical model.
S140: and carrying out coordinate system conversion on the hybrid physical learning model, and constructing to obtain the hybrid physical learning error dynamics model.
S150: and carrying out unmanned ship track tracking prediction control through a hybrid physical learning error dynamics model.
In this embodiment, the embodiment of the present invention first acquires a training data set. In particular, the invention is true The training data set in the embodiment refers to a data set for model training for performing a learning model. In the embodiment of the invention, the training data set of the learning model is obtained through simulation. For example, the embodiment of the invention can establish the unmanned ship motion model through a micro nonlinear equation so as to perform simulation through the unmanned ship motion model, thereby obtaining a corresponding training data set. Then, according to the training data set, the preset neural network is trained through a preset neural network training algorithm, and a learning model is obtained. Specifically, the preset neural network training algorithm in the embodiment of the invention refers to an algorithm, such as a back propagation algorithm, a random gradient descent algorithm, a regularization algorithm, and the like, which can make the neural network fit training data better by adjusting the weight and the bias value of the neural network, so as to predict the input data accurately. According to the embodiment of the invention, the Deep Neural Network (DNN), namely the preset neural network, is subjected to model training through the obtained training data set, so that a learning model is obtained. It is easy to understand that the embodiment of the invention learns unmanned ship modeling errors by approximating the nonlinear dynamics model through a Deep Neural Network (DNN). Further, the embodiment of the invention constructs a hybrid physical learning model according to the learning model and the unmanned ship physical model. Specifically, in the embodiment of the invention, the unmanned ship physical model refers to a physical model obtained by identifying the unmanned ship complete model. Correspondingly, the embodiment of the invention obtains the learning model f through training l With unmanned ship physical model f p And combining to obtain the hybrid physical learning model h. The hybrid physical model constructed in the embodiment of the invention is shown in the following formula (1):
h＝f l +f p (1)
further, the embodiment of the invention constructs the hybrid physical learning error dynamics model by carrying out coordinate system conversion on the hybrid physical learning model. In order to facilitate stability analysis and calculation, the embodiment of the invention converts the hybrid physical learning model into a coordinate system, thereby converting the hybrid physical learning model into a physical learning error dynamics model. Then, the embodiment of the invention carries out unmanned ship track tracking prediction control through the hybrid physical learning error dynamic model so as to realize the track tracking prediction control of the unmanned ship and improve the accuracy and efficiency of the track tracking prediction control of the unmanned ship. It should be noted that, in the embodiment of the invention, the mode of training the preset neural network by the preset neural network training algorithm effectively relieves the modeling error problem caused by uncertain fluid dynamics or environmental interference of the unmanned ship. In addition, the embodiment of the invention can effectively relieve the problem of the calculation efficiency of the model predictive control algorithm by mixing the physical learning error dynamic model, and improve the control stability.
In some embodiments of the present invention, a training data set is acquired, including, but not limited to, the steps of:
and performing track tracking simulation through the unmanned ship model and the unmanned ship physical model to obtain a first data set.
The first data set is normalized to obtain a training data set.
In this embodiment, the first data set is obtained by performing a trajectory tracking simulation through the unmanned ship model and the unmanned ship physical model. Specifically, in order to obtain data for training the deep neural network, i.e., a training data set, embodiments of the present invention use a complete unmanned ship model f and an identified unmanned ship physical model f p And performing simulation for a plurality of times, thereby obtaining a corresponding first data set. Wherein, aiming at the unmanned ship track tracking problem, in the jth test (namely a complete track tracking simulation), the embodiment of the invention records the obtained data set as D j ＝{a j (k) Where k refers to discrete time steps, and k=1, 2. Correspondingly, the data obtained in each time step in the embodiment of the invention is a j (k)＝([v j (k)，τ j (k)]，d j (k) A kind of electronic device. Wherein v= [ uvr ]]Longitudinal speed, transverse speed and bow-turning angular speed of unmanned ship respectively, τ= [ τ ] 1 τ 2 τ 3 ]Representing unmanned ship control vectors. In addition, d=v in the embodiment of the present invention r -v p Representing modeling errorsThe difference being derived from the actual velocity vector v r And unmanned ship physical model f p Is the predicted velocity vector v of (2) p The difference between them is obtained. Further, the embodiment of the invention normalizes the first data set obtained by simulation to obtain the training data set. Illustratively, at the kth step of the jth trial, embodiments of the present invention normalize the first data set to obtain a training data set as shown in equation (2) below:
wherein, in the formulaIs the average value of training data, sigma a Is the standard value of training data, < >>By->Composition, sigma a By->Composition is prepared.
It is easy to understand that after the training data set is obtained, the embodiment of the invention carries out deep neural network training through the training data set, selects the network layer number m and the activation function of the neural network, and carries out the steps of [ v, tau ]]As an input to the neural network, d as an output of the neural network. Then, the embodiment of the invention trains the preset neural network through a preset neural network training algorithm, thereby obtaining a learning model f l 。
In some embodiments of the present invention, the hybrid physical learning error dynamics model is constructed by transforming the hybrid physical learning model in a coordinate system, including but not limited to the following steps:
And converting the preset reference track into the unmanned ship body coordinate system, and calculating the difference value between the expected track and the current track in the unmanned ship body coordinate system through the hybrid physical learning model to obtain the error kinematic model.
And obtaining the hybrid physical learning error dynamics model by carrying out preset differential operation on the error dynamics model.
In the embodiment of the invention, a preset reference track is firstly converted into an unmanned ship body coordinate system, and then a difference value between an expected track and a current track in the unmanned ship body coordinate system is calculated through a hybrid physical learning model, so that an error kinematic model is obtained. It is easy to understand that the embodiment of the invention facilitates stability analysis and calculation by converting the hybrid physical learning model h into the hybrid physical learning error dynamics model under the unmanned ship body coordinate system. Correspondingly, the embodiment of the invention firstly converts the preset reference track into the unmanned ship body coordinate system so as to calculate the difference value between the expected track and the current track in the coordinate system through the mixed physical learning model, thereby obtaining an error kinematic model, wherein the error kinematic model is shown in the following formula (3):
Wherein x is e Representing the difference between the desired position and the current position in the north-positive direction, y e Representing the difference between the desired position in the forward direction and the current position, u e Representing the difference between the desired longitudinal speed and the actual longitudinal speed, v e Representing the difference between the desired and actual lateral velocities, r e Representing the difference between the desired and actual yaw rates, ψ represents the ship heading, ψ e Representing the difference in desired and actual heading, ψ d Indicating the desired heading, x d Representing the position in the desired forward direction, v d Indicating the desired transverse velocity, r d Indicating a desired angular velocity, r indicating an angular velocity, x indicating a position in the forward direction, and y indicating a position in the forward direction.
Further, the embodiment of the invention obtains the hybrid physical learning error dynamics model by carrying out preset differential operation on the error dynamics model. According to the embodiment of the invention, the two sides of the error kinematic model equation are differentiated with respect to time, so that a hybrid physical learning error kinematic model is obtained, and the hybrid physical learning error kinematic model is shown in the following formula (4):
wherein x is d ＝[η d ，v d ]Representing the desired trajectory. η (eta) d ＝[x d y d ψ d ]Indicating the desired position. v d ＝[u d v d r d ]Indicating the desired speed. Wherein x is d And y d Indicating the desired position in the northbound and eastern directions, ψ d To expect the ship heading, u d Representing the desired longitudinal speed of the unmanned ship, v d Representing the desired lateral velocity of the unmanned ship, r d Indicating the desired yaw rate of the unmanned ship. Accordingly, h in the embodiment of the invention e (x e ，x d τ) is represented by the following formulas (5) and (6):
h e (x e ，x d ，τ)＝f p (x e +x d ，τ)+f l (v e +v d ，τ)-f d (x e ，x d ) (5)
it is easy to understand that the unmanned ship track tracking problem is converted into the stability problem by constructing the hybrid physical learning error dynamics model. Wherein, the control target in the embodiment of the invention is to control the error state x e To the origin, i.e. the unmanned ship trajectory approaches the desired trajectory. Accordingly, if the prediction range is set to N, the formulas of the corresponding mixed-case learning model prediction control problem are shown in the following formulas (7) and (8):
where i=0, 1,..n-1 and (i|k) represent the i-th prediction step at time step k. In addition, in the embodiment of the inventionAnd ∈10 in the embodiment of the present invention>Wherein (1)>And->Representing a matrix of positive weights to be determined,refers to a terminal weight matrix. In addition, omega in the embodiment of the invention x Representing a set of error state constraints, Ω τ Representing a set of control constraints, Ω is a designed termination region containing the origin, where the predicted termination state x in the termination region e (n|k) will end, i.e. the system state will stabilize in this region, guaranteeing closed loop stability.
In some embodiments of the present invention, after performing the step of obtaining a hybrid physical learning error dynamics model by performing coordinate system transformation on the hybrid physical learning model, the unmanned ship trajectory tracking prediction control method provided by the embodiment of the present invention further includes, but is not limited to, the following steps:
and calculating a preset maximum invariant set according to the hybrid physical learning error dynamics model.
And constructing a terminal area through a preset maximum unchanged set.
In this embodiment, the embodiment of the present invention first calculates a preset maximum invariant set according to the hybrid physical learning error dynamics model, so as to construct a terminal area through the preset maximum invariant set. Specifically, the infinite prediction range n= infinity in the embodiment of the invention ensures the closed-loop stability of the hybrid physical learning error dynamics model. Accordingly, for linear model predictive control models with limited prediction ranges, terminal cost and constraint approaches are typically employed to ensure stability of the closed loop system. In the classical approach, the stability of the system is ensured by designing a conserved elliptical termination region. Wherein the feedback controller of the linearization system can replace the nonlinear system. Further, when the linearity error is small, it can be assumed that the linearization controller approximates the nonlinear controller throughout the experimental region. Therefore, the embodiment of the invention constructs the terminal area by calculating the maximum invariant set, namely the preset maximum invariant set, and constructs the terminal area by the preset maximum invariant set, so that a large prediction range is not needed to avoid heavy calculation burden.
In some embodiments of the present invention, the preset maximum invariant set is calculated from a hybrid physics learning error dynamics model, including, but not limited to, the steps of:
and linearizing the hybrid physical learning error dynamics model at the origin to obtain a first linearization function.
And solving the Lyapunov equation according to the first linearization function to obtain a terminal weighting matrix.
And calculating according to the terminal weighting matrix to obtain a preset maximum unchanged set.
In this embodiment, the embodiment of the present invention first linearizes the hybrid physical learning error dynamics model at the origin to obtain a first linearization function. Illustratively, in the embodiment of the present invention, the hybrid physical learning error dynamics model fe is first linearized at the origin to obtain a first linearization function as shown in the following formula (9):
wherein, in the formulaAnd->It will be readily appreciated that when equation (9) is stable, then the resulting state feedback control rate τ=kx can be derived e 。
Then, according to the embodiment of the invention, a Lyapunov equation is solved according to a first linearization function, and a terminal weighting matrix is obtained. Specifically, the embodiment of the invention selects one of [0, ] to satisfy k < -lambda ] max (A K ) Is set to k of (c). Wherein in the embodiment of the invention, A K =a+bk. Then, embodiments of the present invention solve the Lyapunov equation (A K +kI) T P+P(A K +kl) = -Q), resulting in a terminal weighting matrix P of positive symmetry. Wherein, in the embodiment of the inventionA weight matrix of the cost function with respect to the states is controlled for model prediction. Further, according to the embodiment of the invention, the preset maximum invariant set is obtained through calculation according to the terminal weighting matrix. Specifically, in calculating the maximum invariant set of the linearization system, i.e. the preset maximum invariant set Ω, the embodiment of the invention first makes k=0,/and/or->The following calculation is then repeated, as shown in the following formula (10):
it will be readily appreciated that embodiments of the present invention provide for the determination of omega by repeating the above-described calculation process until omega k+1 ＝Ω k Obtaining a preset maximum unchanged set omega=omega k+1 。
In some embodiments of the present invention, in performing the step of performing unmanned ship trajectory tracking prediction control through the hybrid physical learning error dynamics model, the unmanned ship trajectory tracking prediction control method provided by the embodiment of the present invention further includes, but is not limited to:
and obtaining a seed control track according to the current motion state and the optimal control sequence at the last moment.
And constructing an incremental model according to the seed control track and the hybrid physical learning error dynamics model.
And carrying out Jacobian matrix calculation according to the incremental model to obtain a continuous Jacobian matrix. Wherein the continuous jacobian matrix comprises a continuous jacobian state matrix and a continuous jacobian control matrix.
Discretizing the continuous jacobian matrix, and carrying out system state prediction according to the incremental model to construct and obtain a quadratic programming optimization function.
And continuously linearizing the hybrid physical learning error dynamics model through a quadratic programming optimization function.
In this embodiment, in order to improve the calculation efficiency of the hybrid physical learning error dynamics model, from the perspective of reducing the complexity of the prediction model and the optimization problem, the embodiment of the invention continuously linearizes the hybrid physical learning error dynamics model by taking the calculation efficiency and the calculation accuracy as the measurement. Specifically, according to the embodiment of the invention, the seed control track is obtained according to the current motion state and the optimal control sequence of the commercial moment. Illustratively, embodiments of the present invention obtain the seed input τ at the kth step 0 (i|k). Meanwhile, the embodiment of the invention takes the optimal control sequence in the last step as seed input 0 Represents the seed trajectory (seed control trajectory). Correspondingly, according to the model predictive control principle, the first element tau (0|k-1) of the optimal control sequence is applied to the system, and the rest is ignored. In addition, in the embodiment of the present invention, for linearization of model prediction control in the kth step, the following formula (11) is defined:
τ 0 (i|k)＝τ(i+1|k-1) (11)
Wherein i=0, 1, N-2, and τ 0 (N-1|k) =τ (N-1|k-1), N being the prediction and control step size.
Accordingly, at step k, the embodiment of the present invention inputs τ through the seed 0 (i|k) current stateObtaining seed status by mixed physical learning error dynamics model>Further, the embodiment of the invention generates a seed state track +.>In the embodiment of the invention, the seed state update formula is shown in the following formula (12):
further, according to the embodiment of the invention, an incremental model is constructed according to the seed control track and the hybrid physical learning error dynamics model. According to the embodiment of the invention, the hybrid physical learning error power model is subjected to Qin Le expansion, so that an incremental model between the system track and the seed track is constructed. Then, the embodiment of the invention calculates the jacobian matrix according to the incremental model to obtain the continuous jacobian matrix. Specifically, the continuous jacobian matrix in the embodiment of the present invention includes a continuous jacobian state matrix and a continuous jacobian control matrix. The incremental model of the embodiment of the invention comprises a continuous jacobian state matrixA continuous Jacobian control matrix +. >Wherein, the jacobian state matrix in the embodiment of the invention/>A continuous Jacobian control matrix +.>The following formulas (13) and (14) are respectively shown:
accordingly, in the embodiment of the present invention, the jacobian state matrix and the jacobian control matrix of the deep neural network model are respectively shown in the following formulas (15) and (16):
wherein, for continuous linearization, the activation function in the embodiment of the invention is a derivative function, such as tanh, so that the following formulas (17) and (18) can be obtained:
wherein in the formula I m-1 For every m-layer input of the neural network,for the weight of the neural network per m layers, m=1, 2,..m, m represents the number of layers of the neural network.
Further, the embodiment of the invention discretizes the continuous jacobian matrix, predicts the system state according to the incremental model, and constructs a quadratic programming optimization function. In particular, forThe embodiment of the invention predicts the system state +.>Accordingly, at the kth step, it can be expressed as the following formulas (19) to (21):
then, the nonlinear optimization problem of the hybrid physical learning error dynamics model is converted into a quadratic programming problem through the following formulas (22) and (23):
Accordingly, X in the embodiment of the invention e (k) Can be compactly represented by the following formula (24):
thus, the nonlinear problem of the hybrid physics learning error dynamics model in the embodiments of the present invention is converted into a quadratic programming problem as shown in the following formulas (25) and (26):
wherein, in the formulaAs shown in the following formula (26)>As shown in the following formula (27)>As shown in the following formula (28)>The following formula (29):
correspondingly, the embodiment of the invention carries out continuous linearization on the hybrid physical learning error dynamics model through the quadratic programming optimization function obtained by construction, thereby being capable of relieving the complexity problem of the prediction model caused by introducing a Deep Neural Network (DNN) and effectively relieving the calculation efficiency problem of the model.
In some embodiments of the present invention, an incremental model is constructed from a seed control trajectory and a hybrid physical learning error dynamics model, including, but not limited to, the steps of:
and developing the hybrid physical learning error dynamics model at the seed control track by Qin Le to obtain a model system track. Wherein the model system trajectory is the sum of the seed control trajectory and the increment of the seed control trajectory.
And constructing and obtaining a deviation model according to the model system track and the seed control track.
And constructing and obtaining an incremental model according to the deviation model and the seed state updating function.
In this embodiment, the embodiment of the present application firstly expands Qin Le the hybrid physical learning error dynamics model at the seed control track to obtain a model system track, and constructs a deviation model according to the model system track and the seed control track, so as to construct an incremental model according to the deviation model and the seed state update function. Specifically, in the embodiment of the application, the model system track is the sum of the seed control track and the increment of the seed control track. In the embodiment of the application, the deviation between the model system track and the seed control track is defined asThe corresponding bias model is shown in the following equation (30):
further, according to the deviation model, the embodiment of the application applies the state update formula to the seed control trackThe expansion at Qin Le and ignoring the corresponding higher order term yields the following equation (31):
further, the embodiment of the application updates the formula according to the above formula and the seed state, namely the formula (12), to the leftAnd right->And (3) performing cancellation to construct a final incremental model, wherein the final incremental model is shown in the following formula (32):
with reference to fig. 2, the complete implementation process of the unmanned ship track tracking prediction control method in the technical scheme of the application is as follows:
According to the embodiment of the invention, the training data set is obtained through analog simulation. Specifically, in the embodiment of the invention, the first data set is obtained by performing track tracking simulation through the unmanned ship model and the unmanned ship physical model, and the training data set is obtained by normalizing the first data set. Then, according to the embodiment of the invention, the preset neural network is trained through the preset neural network training algorithm according to the training data set so as to obtain a learning model, and a hybrid physical learning model is constructed according to the learning model and the unmanned ship physical model, so that the modeling error problem caused by uncertain fluid dynamics or environmental interference of the unmanned ship is effectively relieved in a mode that the preset neural network is trained through the preset neural network training algorithm. Further, in the embodiment of the invention, the hybrid physical learning error dynamics model is constructed and obtained by converting the coordinate system of the hybrid physical learning model, so that the unmanned ship track tracking prediction control is performed according to the hybrid physical learning error dynamics model, and the accuracy and the efficiency of the unmanned ship track tracking prediction control are improved. Specifically, the embodiment of the invention converts the preset reference track into the unmanned ship equipment body coordinate system, calculates the difference value between the expected track and the current track in the unmanned ship equipment body coordinate system through the mixed physical learning model to obtain the error kinematic model, and carries out preset differential operation on the error kinematic model to obtain the mixed physical learning error dynamic model, and carries out unmanned ship track tracking prediction control through the mixed physical learning error dynamic model, thereby effectively relieving the problem of the calculation efficiency of a model prediction control algorithm and improving the control stability. In addition, in order to perform relevant optimization on the hybrid physical learning error dynamics model, the embodiment of the invention can be used for relieving the stability problem of the system by expanding the terminal area by means of the maximum invariant set. Specifically, the embodiment of the invention firstly calculates a preset maximum unchanged set according to a hybrid physical learning error dynamics model. Then, the embodiment of the invention constructs the terminal area by presetting the maximum invariant set, thereby realizing the expansion of the terminal area. In the embodiment of the invention, in the process of calculating the preset maximum invariant set, firstly, the hybrid physical learning error dynamics model is linearized at the original point to obtain a first linearization function, and the Lyapunov equation is solved according to the first linearization function to obtain a terminal weighting matrix, and then the maximum invariant set is calculated according to the terminal weighting matrix, namely the preset maximum invariant set, so that the calculation of the preset maximum invariant set is realized.
Furthermore, the embodiment of the invention relieves the complexity problem of the prediction model and the calculation efficiency problem caused by the introduction of the deep neural network by introducing a continuous linearization mode. Specifically, according to the embodiment of the invention, firstly, a seed control track is obtained according to the current motion state and the optimal control sequence at the last moment, and an incremental model is constructed according to the seed control track and the hybrid physical learning error dynamics model. According to the embodiment of the invention, the mixed physical learning error dynamics model is expanded Qin Le at the seed control track to obtain a model system track, and a deviation model is constructed according to the model system track and the seed control track, so that an incremental model is constructed according to the deviation model and a seed state updating function. Accordingly, in the embodiment of the present invention, the model system track is the sum of the seed control track and the increment of the seed control track. Further, according to the embodiment of the invention, the jacobian matrix is calculated according to the incremental model to obtain the continuous jacobian matrix. Accordingly, the continuous jacobian matrix in the embodiment of the present invention includes a continuous jacobian state matrix and a continuous jacobian control matrix. Then, the embodiment of the invention discretizes the continuous jacobian matrix, predicts the system state according to the incremental model, and constructs a quadratic programming optimization function, so that the hybrid physical learning error dynamics model is continuously linearized through the quadratic programming optimization function.
Referring to fig. 3, an embodiment of the present invention further provides an unmanned ship trajectory tracking prediction control system, including:
a first module 210 for acquiring a training data set. The training data set is obtained through analog simulation.
The second module 220 is configured to train the preset neural network through a preset neural network training algorithm according to the training data set, so as to obtain a learning model.
And a third module 230, configured to construct a hybrid physical learning model according to the learning model and the unmanned ship physical model.
And a fourth module 240, configured to construct a hybrid physical learning error dynamics model by performing coordinate system transformation on the hybrid physical learning model.
And a fifth module 250, configured to perform unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
Referring to fig. 4, an embodiment of the present invention further provides an unmanned ship trajectory tracking prediction control system, including:
at least one processor 310.
At least one memory 320 for storing at least one program.
The at least one program, when executed by the at least one processor 310, causes the at least one processor 310 to implement the unmanned ship trajectory tracking predictive control method as described in the above embodiments.
An embodiment of the present invention also provides a computer-readable storage medium storing computer-executable instructions for execution by one or more control processors, e.g., to perform the steps described in the above embodiments.
Those of ordinary skill in the art will appreciate that all or some of the steps, systems, and methods disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.
While the preferred embodiment of the present invention has been described in detail, the present invention is not limited to the above embodiment, and various equivalent modifications and substitutions can be made by those skilled in the art without departing from the spirit of the present invention, and these equivalent modifications and substitutions are intended to be included in the scope of the present invention as defined in the appended claims.
Claims (10)
1. The unmanned ship track tracking prediction control method is characterized by comprising the following steps of:
acquiring a training data set; wherein the training data set is obtained through analog simulation;
training a preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model;
constructing a hybrid physical learning model according to the learning model and the unmanned ship physical model;
the hybrid physical learning error dynamics model is constructed and obtained by carrying out coordinate system conversion on the hybrid physical learning model;
and carrying out unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
2. The unmanned ship trajectory tracking predictive control method of claim 1, wherein the acquiring a training dataset comprises:
Performing track tracking simulation through an unmanned ship model and the unmanned ship physical model to obtain a first data set;
and normalizing the first data set to obtain the training data set.
3. The unmanned ship track following prediction control method according to claim 1, wherein the constructing the hybrid physical learning error dynamics model by performing coordinate system transformation on the hybrid physical learning model comprises:
converting a preset reference track into an unmanned ship body coordinate system, and calculating a difference value between an expected track and a current track in the unmanned ship body coordinate system through the hybrid physical learning model to obtain an error kinematic model;
and obtaining the hybrid physical learning error dynamics model by carrying out preset differential operation on the error dynamics model.
4. The unmanned ship trajectory tracking prediction control method according to claim 1, wherein after the step of constructing a hybrid physical learning error dynamics model by performing coordinate system conversion on the hybrid physical learning model is performed, the method further comprises:
calculating a preset maximum invariant set according to the hybrid physical learning error dynamics model;
And constructing a terminal area through the preset maximum unchanged set.
5. The unmanned ship trajectory tracking predictive control method according to claim 4, wherein the calculating a preset maximum invariant set according to the hybrid physical learning error dynamics model comprises:
linearizing the hybrid physical learning error dynamics model at an origin to obtain a first linearization function;
solving a Lyapunov equation according to the first linearization function to obtain a terminal weighting matrix;
and calculating according to the terminal weighting matrix to obtain the preset maximum unchanged set.
6. The unmanned ship trajectory tracking predictive control method according to claim 1, wherein, during the step of performing the unmanned ship trajectory tracking predictive control by the hybrid physical learning error dynamics model, the method further comprises:
acquiring a seed control track according to the current motion state and the optimal control sequence at the previous moment;
constructing an incremental model according to the seed control track and the hybrid physical learning error dynamics model;
carrying out Jacobian matrix calculation according to the incremental model to obtain a continuous Jacobian matrix; wherein the continuous jacobian matrix comprises a continuous jacobian state matrix and a continuous jacobian control matrix;
Discretizing the continuous jacobian matrix, and carrying out system state prediction according to the incremental model to construct and obtain a quadratic programming optimization function;
and continuously linearizing the hybrid physical learning error dynamics model through the quadratic programming optimization function.
7. The unmanned ship trajectory tracking predictive control method according to claim 6, wherein the constructing an incremental model from the seed control trajectory and the hybrid physical learning error dynamics model comprises:
carrying out Taylor expansion on the hybrid physical learning error dynamics model at the seed control track to obtain a model system track; wherein the model system trajectory is the sum of the seed control trajectory and the increment of the seed control trajectory;
constructing and obtaining a deviation model according to the model system track and the seed control track;
and constructing and obtaining the incremental model according to the deviation model and the seed state updating function.
8. An unmanned ship track tracking predictive control system, comprising:
a first module for acquiring a training data set; wherein the training data set is obtained through analog simulation;
The second module is used for training a preset neural network through a preset neural network training algorithm according to the training data set to obtain a learning model;
the third module is used for constructing and obtaining a hybrid physical learning model according to the learning model and the unmanned ship physical model;
the fourth module is used for constructing and obtaining a hybrid physical learning error dynamics model by carrying out coordinate system conversion on the hybrid physical learning model;
and the fifth module is used for carrying out unmanned ship track tracking prediction control through the hybrid physical learning error dynamics model.
9. An unmanned ship track tracking predictive control system, comprising:
at least one processor;
at least one memory for storing at least one program;
when the at least one program is executed by the at least one processor, the at least one processor is caused to implement the unmanned ship trajectory tracking predictive control method as claimed in any one of claims 1 to 7.
10. A computer storage medium in which a processor-executable program is stored, characterized in that the processor-executable program, when executed by the processor, is for implementing the unmanned ship trajectory tracking predictive control method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310856720.7A CN116880184A (en) | 2023-07-12 | 2023-07-12 | Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310856720.7A CN116880184A (en) | 2023-07-12 | 2023-07-12 | Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116880184A true CN116880184A (en) | 2023-10-13 |
Family
ID=88256317
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310856720.7A Pending CN116880184A (en) | 2023-07-12 | 2023-07-12 | Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116880184A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117252905A (en) * | 2023-11-20 | 2023-12-19 | 暗物智能科技(广州)有限公司 | Pedestrian track prediction method and system based on neural differential equation |
-
2023
- 2023-07-12 CN CN202310856720.7A patent/CN116880184A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117252905A (en) * | 2023-11-20 | 2023-12-19 | 暗物智能科技(广州)有限公司 | Pedestrian track prediction method and system based on neural differential equation |
| CN117252905B (en) * | 2023-11-20 | 2024-03-19 | 暗物智能科技(广州)有限公司 | Pedestrian track prediction method and system based on neural differential equation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110647042B (en) | Robot robust learning prediction control method based on data driving | |
| CN108153153B (en) | Learning variable impedance control system and control method | |
| Sokolov et al. | Complete stability analysis of a heuristic approximate dynamic programming control design | |
| CN111459051B (en) | Discrete terminal sliding mode model-free control method with disturbance observer | |
| US20230078812A1 (en) | Iterative learning control method for multi-particle vehicle platoon driving system | |
| CN111258218B (en) | Intelligent vehicle path tracking method based on maximum correlation entropy criterion | |
| Zhang et al. | A novel event-triggered robust neural formation control for USVs with the optimized leader–follower structure | |
| Qin et al. | Fast fixed-time nonsingular terminal sliding-mode formation control for autonomous underwater vehicles based on a disturbance observer | |
| Si et al. | Decentralized adaptive neural control for high-order stochastic nonlinear strongly interconnected systems with unknown system dynamics | |
| CN116880184A (en) | Unmanned ship track tracking prediction control method, unmanned ship track tracking prediction control system and storage medium | |
| Huang et al. | Prescribed performance fuzzy back-stepping control of a flexible air-breathing hypersonic vehicle subject to input constraints | |
| Çetin et al. | Adaptive uncertainty compensation-based nonlinear model predictive control with real-time applications | |
| Gao et al. | Online optimal control for dynamic positioning of vessels via time-based adaptive dynamic programming | |
| CN113110511A (en) | Intelligent ship course control method based on generalized fuzzy hyperbolic model | |
| Xu et al. | DBN based SD-ARX model for nonlinear time series prediction and analysis | |
| Tabuada et al. | Data-driven control for SISO feedback linearizable systems with unknown control gain | |
| Ghazali et al. | A multiple-node hormone regulation of neuroendocrine-PID (MnHR-NEPID) control for nonlinear MIMO systems | |
| Kim et al. | TOAST: trajectory optimization and simultaneous tracking using shared neural network dynamics | |
| Hager et al. | Adaptive Neural network control of a helicopter system with optimal observer and actor-critic design | |
| Kalaria et al. | Delay-aware robust control for safe autonomous driving | |
| Song et al. | Adaptive dynamic programming: single and multiple controllers | |
| Kumar et al. | Hybrid neural network control for uncertain nonlinear discrete-time systems with bounded disturbance | |
| Sun et al. | Extreme learning control of surface vehicles with unknown dynamics and disturbances | |
| Liu et al. | Robust Adaptive Self‐Structuring Neural Network Bounded Target Tracking Control of Underactuated Surface Vessels | |
| CN111352344B (en) | Data-driven self-adaptive predictor of dynamic completely unknown system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |