Navigation signal lamp remote control method, device and system

Abstract

The invention relates to the technical field of marine navigation control and safety monitoring, and discloses a navigation signal lamp remote control method, device and system, and the method comprises the following steps: collecting ship dynamic and environment information from a plurality of sensors, encrypting the information, and outputting the information; receiving encrypted information to construct a space-time constraint matrix, and integrating a ship motion prediction item, a position and signal intensity item and a flow field compensation item; and performing hybrid prediction on the future trajectory of the ship according to the space-time constraint matrix, and generating a hybrid prediction result in combination with correction of the data driving model and the physical model. According to the method, a hybrid prediction technical scheme of fusing data driving and a physical model is adopted, the technical effect of improving the prediction accuracy and reliability of the future trajectory of the ship in the changeable marine environment is achieved, and the defects of low prediction precision and poor adaptability in the complex or unseen marine environment are overcome.

Description

Technical Field

[0001] This invention relates to the field of maritime navigation control and safety monitoring technology, and in particular to a method, device and system for remote control of maritime signal lights. Background Technology

[0002] Typically, ships are equipped with navigation lights, and the flashing sequence or combination of these lights is used to signal at night and when ships are entering or leaving port, thereby preventing collisions, etc.

[0003] However, in the field of ship trajectory prediction, mainstream methods mostly rely on statistical patterns of historical track data or simply apply physical motion models. The former is prone to "going astray" under sudden weather or complex flow fields, while the latter ignores the nonlinear changes in the marine environment. As a result, the prediction effect is inconsistent and it is difficult to meet the high-precision requirements under congested or harsh conditions. Regarding the security verification of control commands, most systems currently use traditional cryptographic, token, or fixed key mechanisms. Although simple to implement, they are extremely vulnerable to external attacks or internal abuse. Once the credentials are leaked, the commands can be misused. At the command generation level, most systems focus only on optimizing a single objective, such as maximizing speed or minimizing energy consumption. This either sacrifices fuel efficiency or reduces navigation safety. Even worse, the decision-making logic under fixed rules becomes rigid when encountering new scenarios, unable to flexibly balance the contradictions between different objectives. Furthermore, when it is necessary to send signals to the ship, staff have to manually turn on multiple corresponding switches, which is cumbersome. There is a lack of intelligent diagnosis and repair methods for complex faults. Once sensor data is lost or communication interference occurs, it is difficult to restore normal operation in a timely manner. Summary of the Invention

[0004] The purpose of this invention is to provide a remote control method, device and system for marine signal lights, which solves the problems of low prediction accuracy, poor adaptability, rigid decision-making, inability to take into account multiple constraints and objectives, single verification mechanism, easy impersonation and lack of non-repudiation of instructions, slow fault response, single recovery method and cumbersome manual operation of multiple switches by staff in complex or unseen marine environments.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a remote control method for navigational signal lights, comprising the following steps:

[0006] S1. Collect ship dynamics and environmental information from multiple sensors, encrypt the information, and output it.

[0007] S2. Receive the encrypted information to construct a spatiotemporal constraint matrix, and integrate the ship motion prediction term, position and signal strength term, and flow field compensation term.

[0008] S3. Based on the spatiotemporal constraint matrix, perform hybrid prediction of the ship's future trajectory, combine the data-driven model and the physical model for correction, and generate hybrid prediction results;

[0009] S4. Based on the hybrid prediction results and the current environmental information, a multi-objective optimization problem is formed. The optimal weights of the control commands are obtained through the hybrid solver, and the target control commands are generated according to the optimal weights.

[0010] S5. Encode the target control instructions and verify their validity through biometric identification and blockchain evidence storage before distributing the execution of the control instructions.

[0011] S6. After the control commands are executed in a distributed manner, the operating status of the navigation signal lights is monitored regularly. The health of the signal lights is evaluated by collecting status data in real time, and a fault-tolerant recovery mechanism is triggered when an anomaly occurs.

[0012] Preferably, the information from multiple sensors in step S1 includes an AIS receiver, a radar sensor, and a weather sensor. This information is encrypted using the SM4 algorithm, and the SM4 encryption formula is as follows:

[0013] ;

[0014] In its formula, This represents the original collected data. This represents a symmetric encryption algorithm. Represents the key. This indicates the encrypted data.

[0015] Preferably, the step of constructing the spatiotemporal constraint matrix in step S2 includes: calculating the ship's motion terms, position and signal strength matrices, and solving the fluid dynamics equations to obtain flow field compensation. The fluid dynamics equations are as follows:

[0016] ;

[0017] In its formula, This indicates the movement of the fluid in this direction. This indicates the movement of the fluid in this direction. This indicates how fluid motion changes over time. This indicates the position of the fluid in the horizontal displacement direction. This indicates the position of the fluid in the vertical displacement direction. This represents the pressure value at a specific point in a fluid. This represents the mass per unit volume of a fluid. express The second derivative in space.

[0018] Preferably, the hybrid prediction step for the future trajectory in step S3 includes data-driven prediction using an LSTM network and correction of the physical model by combining fluid dynamics to generate a hybrid prediction result. The hybrid prediction formula is as follows:

[0019] ;

[0020] In its formula, Indicates time The predicted result or trajectory at any given moment. Represents the weight parameters. Indicates time The prediction results are always based on data-driven models. Indicates time The prediction results are based on the physical model at all times.

[0021] Preferably, the multi-objective optimization problem formation step in step S4 includes constructing a quadratic programming model based on the hybrid prediction results and current environmental information, and using a hybrid solution method of quantum annealing algorithm and branch and bound method to generate optimal weights for control commands. Subsequently, target control commands are generated according to the optimal weights. The formula for the quantum annealing algorithm is:

[0022] ;

[0023] In its formula, Indicates the system's energy The probability of a state. Indicates energy state, This represents the energy of the current optimal solution. Represents Boltzmann's constant. Indicates the system temperature;

[0024] The formula for the branch and bound method is:

[0025] ;

[0026] In its formula, Indicates through decision variables sum coefficient vector The calculated results Representation and decision variables The relevant weights or costs, This represents the set of variables that need to be determined in an optimization problem.

[0027] Preferably, the verification step of the target control command in step S5 utilizes a biometric identification method and blockchain notarization, and the formula for the biometric identification method is as follows:

[0028] ;

[0029] In its formula, This indicates the output biometric verification result or score. This represents a pre-trained biometric matching function or model. This represents a feature vector extracted from biometric data collected in real time by sensors.

[0030] The blockchain formula is as follows:

[0031] ;

[0032] In its formula, This indicates a new block that has been created and added to the blockchain. This represents the hash value of the current block header. This represents the data within the block. This represents the hash value of the previous block. Indicates the timestamp of block creation. This represents the random number found during the consensus process.

[0033] Preferably, monitoring the operational status of the navigation signal lights in step S6 includes real-time assessment of the health of the navigation signal lights and using a health assessment model to ensure the safety of their operational status.

[0034] Preferably, the fault-tolerant recovery mechanism in step S6 includes a combined application of manifold learning reconstruction and quantum error-correcting coding, wherein the manifold learning reconstruction formula is:

[0035] ;

[0036] In its formula, This represents the idealized health state vector reconstructed by the model. This represents a pre-trained manifold learning model. This represents a real-time status data vector that has been collected from the sensor and has been identified as abnormal.

[0037] The quantum error correction coding formula is as follows:

[0038] ;

[0039] In its formula, The expression representing a quantum state, Represents the normalization constant. express The ground state of a qubit. This indicates that the state is repeated. Second-rate, express The excited state of a qubit.

[0040] A remote control system for marine signal lights, comprising:

[0041] The information collection module collects dynamic and environmental information of the ship from multiple sensors, encrypts the collected data, and outputs it.

[0042] The spatiotemporal constraint matrix module receives encrypted information, constructs the spatiotemporal constraint matrix, and integrates ship motion prediction, position information, signal strength, and flow field compensation terms.

[0043] The hybrid prediction module performs hybrid prediction of the ship's future trajectory based on the spatiotemporal constraint matrix, and generates accurate hybrid prediction results by combining the correction of the data-driven model and the physical model.

[0044] The control command generation module is optimized by forming a multi-objective optimization problem based on the hybrid prediction results and the current environmental information, and obtaining the optimal weights of the control commands through a hybrid solver to generate the target control commands.

[0045] The verification and execution module encodes the target control commands and verifies their validity through biometric identification and blockchain notarization, and performs distributed execution of the control commands.

[0046] The status monitoring and fault-tolerant recovery module periodically monitors the operational status of the navigation lights after the control commands are executed, collects status data in real time to assess the health of the lights, and triggers the fault-tolerant recovery mechanism when an anomaly occurs.

[0047] An apparatus includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the aforementioned method for remote control of a navigation signal light.

[0048] In summary, the present invention has at least one of the following beneficial technical effects:

[0049] 1. This invention adopts a hybrid prediction technology that integrates data-driven and physical models, which improves the accuracy and reliability of ship trajectory prediction in variable marine environments. Compared with existing technologies that rely solely on historical data or isolated physical motion formulas for prediction, this invention solves the shortcomings of low prediction accuracy and poor adaptability in complex or unseen marine environments.

[0050] 2. This invention adopts a technical solution that introduces a multi-objective hybrid solver to generate optimal control commands. It achieves the technical effect of weighing multiple conflicting objectives such as navigation efficiency, energy consumption and safety, and generating globally optimal control decisions. Compared with the existing control strategies based on fixed rules or pursuing the optimization of a single index, this invention solves the shortcomings of rigid decision-making and inability to take into account multiple constraints and objectives.

[0051] 3. This invention adopts a dual verification technology solution that combines biometric identification and blockchain notarization, which achieves the technical effect of ensuring the security, traceability and non-repudiation of the control command issuance process. Compared with the existing technology that only relies on a single certificate such as a password or authorization code for verification, it solves the shortcomings of the single verification mechanism, easy impersonation and lack of non-repudiation of the command.

[0052] 4. This invention adopts a fault-tolerant recovery technology solution that combines a health assessment model, manifold learning reconstruction, and quantum error correction coding. It achieves the technical effects of proactive prediction, intelligent repair, and reliable recovery of critical equipment failures. Compared with the existing technology that relies on passive alarms or simple redundant switching for fault handling and requires manual opening of multiple switches to restore the signal, this invention solves the problems of slow fault response, single recovery methods, and the need for staff to manually open multiple switches. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0054] Figure 2 This is a schematic diagram of the system flow of the present invention;

[0055] Figure 3 This is a schematic diagram of the device of the present invention. Detailed Implementation

[0056] The following is in conjunction with the appendix Figure 1 The present invention will be further described in detail below.

[0057] This invention provides a method for remote control of marine signal lights, comprising the following steps:

[0058] S1. Collect ship dynamics and environmental information from multiple sensors, encrypt the information, and output it.

[0059] Specifically, the first step is to establish a data collection model, which includes the following sensor components:

[0060] AIS receiver: used to obtain real-time information on a ship's location and navigation status, such as longitude, latitude, speed, and heading;

[0061] Radar sensors: provide target detection and obstacle location, including relative distance and angle information, and can monitor dynamic changes in the surrounding environment;

[0062] Meteorological sensors: acquire information on environmental variables such as wind speed, wind direction, temperature, and wave height to assess navigation conditions.

[0063] The collected information can be represented as the following dataset:

[0064] ;

[0065] In its formula, This indicates the data returned by the AIS receiver, including location data and flight speed. and heading , This indicates the data returned by the radar sensor. This indicates the data returned by the weather sensors, including wind speed. ,wind direction Temperature Wave height .

[0066] This information is then integrated to generate a comprehensive information set. , can be represented as:

[0067] ;

[0068] In its formula, Represents a comprehensive information set. It is a weighted fusion function. This indicates the data returned by the AIS receiver, including location data and flight speed. and heading , This indicates the data returned by the radar sensor. This indicates the data returned by the weather sensors, including wind speed. ,wind direction Temperature Wave height .

[0069] Subsequently, the collected ship dynamics and environmental information is encrypted using the SM4 algorithm, employing the following formula:

[0070] ;

[0071] In its formula, This represents the original collected data. This represents a symmetric encryption algorithm. Represents the key. This indicates the encrypted data.

[0072] The specific processing steps of SM4 encryption are as follows:

[0073] Data grouping: grouping plaintext data The data is divided into multiple 128-bit blocks. If the last data block is less than 128 bits, it is padded to ensure integrity.

[0074] Encryption operations: Each data block undergoes an initial transformation and multiple rounds of iterative operations, including byte substitution, row shifting, column mixing, and round key addition.

[0075] Output encrypted data: Generate encrypted ciphertext After encryption, the processed data will be sent to the control center via a secure communication module.

[0076] The communication module used will employ transport layer security protocols (such as TLS) to ensure the integrity and confidentiality of data during transmission and prevent data from being tampered with or stolen.

[0077] S2. Receive the encrypted information to construct a spatiotemporal constraint matrix, and integrate the ship motion prediction term, position and signal strength term, and flow field compensation term.

[0078] Specifically, firstly, the decrypted original information I contains the ship's motion terms, position, and signal strength. This information is used to construct the spatiotemporal constraint matrix T, whose elements are defined as follows:

[0079] ;

[0080] In its formula, Indicates the ship's current x-axis position. Indicates the ship's current y-axis position. This represents a time variable, indicating the current timestamp. Indicates the received signal strength. The flow field compensation term is obtained by solving the fluid dynamics equations. This represents the predicted motion vector of the ship, including its speed and direction.

[0081] Flow field compensation term It is obtained by solving the following fluid dynamics equations:

[0082] ;

[0083] In its formula, This indicates the movement of the fluid in this direction. This indicates the movement of the fluid in this direction. This indicates how fluid motion changes over time. This indicates the position of the fluid in the horizontal displacement direction. This indicates the position of the fluid in the vertical displacement direction. This represents the pressure value at a specific point in a fluid. This represents the mass per unit volume of a fluid. express The second derivative in space, The symbol represents the partial derivative.

[0084] To solve this equation, the following steps are required:

[0085] Prepare initial conditions: Obtain the initial velocity of the flow field. and initial pressure ;

[0086] Discretized equations: The fluid dynamics equations are discretized in space and time, and the discrete equations are obtained using the finite difference method or the finite element method.

[0087] Solving the equations: Numerical methods (such as Jacobi iteration or Gauss-Seidel iteration) are used to solve the discretized equations to obtain the velocity field at each time step;

[0088] Calculate the flow field compensation term: Based on the obtained fluid velocity field, calculate the influence of the flow field and obtain the flow field compensation term. .

[0089] After calculating the flow field compensation term, it is integrated into the complete spatiotemporal constraint matrix. .

[0090] S3. Based on the spatiotemporal constraint matrix, perform hybrid prediction of the ship's future trajectory, combine the data-driven model and the physical model for correction, and generate hybrid prediction results;

[0091] Specifically, first, the complete spatiotemporal constraint matrix is ​​received.

[0092] ;

[0093] In its formula, Represents the complete spatiotemporal constraint matrix , Indicates the ship's current x-axis position. Indicates the ship's current y-axis position. Indicates the timestamp of information collection. Indicates the received signal strength. Represents the flow field compensation term. This represents the ship motion prediction vector, which includes velocity. and .

[0094] The first step in data-driven forecasting is to use Long Short-Term Memory (LSTM) networks to forecast ship dynamics data.

[0095] LSTM networks can learn time series features and generate predictions of future states. The output format of the LSTM network is as follows:

[0096] ;

[0097] In its formula, This represents a future state prediction generated based on LSTM, including position and velocity information. This represents the input feature set of the LSTM network. This represents the weight parameters of the LSTM network, which are continuously updated during training. This represents the bias term, which helps optimize the model output.

[0098] Next, the prediction results are integrated using the following hybrid prediction formula, incorporating corrections to the fluid dynamics physical model:

[0099] ;

[0100] In its formula, Indicates time The predicted result or trajectory at any given moment. Represents the weight parameters. Indicates time The prediction results are always based on data-driven models. Indicates time The prediction results are based on the physical model at all times.

[0101] The hybrid prediction process is achieved by performing the following steps:

[0102] Data preparation and preprocessing: from the spatiotemporal constraint matrix Extract relevant data to construct the input features of the LSTM model, including current state information. .

[0103] LSTM network training: Using historical motion data and environmental information, an LSTM network is trained to obtain future state predictions. The optimized LSTM can better capture the trends of ship motion, improving accuracy.

[0104] Solution of fluid dynamics model: Based on the fluid dynamics equations, the influence of the flow field is calculated, and the prediction results of the physical model are obtained. The formula is:

[0105] ;

[0106] In its formula, This represents the prediction results from the physical model. This represents the initial state of the object at the start of the calculation (or the previous moment). Axis position, This indicates the object's initial y-axis position at the start of the calculation. Indicates that the object is in Velocity components in the axial direction, This represents the velocity component of the object along the y-axis. This represents a discrete time step or time interval. Indicates the time interval Internally, the change in position along the x-axis caused by factors such as acceleration or external force. Indicates the time interval Internal, the change in position along the y-axis caused by factors such as acceleration or external force.

[0107] Hybrid prediction result calculation: In each iteration, the hybrid prediction formula is used to combine the data-driven prediction results with the prediction results of the physical model to generate the final hybrid prediction state. .

[0108] Output prediction results: Combine prediction results Used for subsequent ship path planning and signal control to improve navigation safety and efficiency.

[0109] S4. Based on the hybrid prediction results and the current environmental information, a multi-objective optimization problem is formed. The optimal weights of the control commands are obtained through the hybrid solver, and the target control commands are generated according to the optimal weights.

[0110] Specifically, first, the future trajectory state vector output by the hybrid prediction module is received. and real-time environmental information obtained from sensors Based on these inputs, a multi-objective optimization problem is constructed.

[0111] The problem is constructed as a quadratic programming model, aiming to find the optimal balance between navigation safety, energy consumption, and navigation efficiency. The objective function of the quadratic programming model can be expressed as:

[0112] ;

[0113] In its formula, Represents the minimization operator. Describe the objective function. Represents the decision vector or weight vector. Represents the coefficient matrix of the quadratic term. Represents the vector of coefficients of linear terms. This represents the transpose symbol.

[0114] This optimization process is subject to the following constraints: and Limitations are imposed to ensure the effectiveness of weight allocation.

[0115] In its formula, This represents the total number of weights, that is, the number of items or samples to be weighted. Indicates the index, indicating the first... The number of items and the range of values ​​are: , Indicates the first The weighting coefficients of each project.

[0116] Subsequently, in order to solve the above quadratic programming model, a hybrid solution method combining quantum annealing and branch-and-bound is preferably adopted to efficiently obtain the globally optimal weight vector.

[0117] First, the quantum annealing algorithm is used to initially solve the quadratic programming problem. This algorithm maps the original problem to a Hamiltonian of an Ising model, explores the solution space by simulating the quantum tunneling effect, and searches for the ground state energy of the system, which corresponds to the optimal solution of the objective function.

[0118] During annealing, the system evolves from one quantum state to another, and the probability distribution formula for reaching a certain energy state is as follows:

[0119] ;

[0120] In its formula, Indicates the system's energy The probability of a state. Indicates energy state, This represents the energy of the current optimal solution. Represents Boltzmann's constant. This indicates the system temperature.

[0121] To ensure global optimality of the solution or to handle discrete variables, the candidate solutions obtained by the quantum annealing algorithm are used as upper bounds, and combined with the branch and bound method for exact solution. The branch and bound method systematically partitions the solution space and uses a bound function to remove subsets that do not contain the optimal solution.

[0122] In the bounding step, a linear programming relaxation problem can be solved to obtain the lower bound of the subproblems, with the objective function in the form of:

[0123] ;

[0124] In its formula, Indicates through decision variables sum coefficient vector The calculated results Representation and decision variables The relevant weights or costs, This represents the set of variables that need to be determined in an optimization problem.

[0125] Finally, the hybrid solver outputs the optimal weight vector. Based on this optimal weight, the final target control command is generated. The target control command is a weighted combination of multiple basic control strategies based on optimal weights, and its generating function can be expressed as:

[0126] ;

[0127] In its formula, This indicates the final target control command generated. This represents an instruction generation function. This indicates a mixed prediction result. Current environmental information This represents the optimal weight vector calculated by the hybrid solver.

[0128] S5. Encode the target control instructions and verify their validity through biometric identification and blockchain evidence storage before distributing the execution of the control instructions.

[0129] Specifically, firstly, the target control commands obtained from the multi-objective optimization solver are... Encode and encapsulate into a standardized message packet. This is to facilitate subsequent hash calculations and secure transmission.

[0130] ;

[0131] In its formula, This represents the encoded message packet to be processed. This represents the original target control command vector. Indicates the timestamp of the instruction generation. This represents a unique transaction identifier generated for this instruction.

[0132] Before an instruction is issued, its validity must be confirmed through a dual verification mechanism. The first verification is the operator's biometric identification. This process is implemented through a pre-trained biometric matching model, and its verification process can be represented by the following formula:

[0133] ;

[0134] In its formula, This indicates the output biometric verification result or score. This represents a pre-trained biometric matching function or model. This represents the feature vector extracted from biometric data collected in real time by sensors.

[0135] Output results The value is compared with a preset threshold to determine whether the verification passes, thus obtaining a Boolean verification status. .

[0136] The second layer of verification is blockchain-based evidence verification. The previously calculated instruction hash value... The hash value is submitted to a distributed ledger network, where it is stored by creating a new block containing that hash value, thus forming an immutable record. The structure of the new block is defined as follows:

[0137] ;

[0138] In its formula, This indicates a new block that has been created and added to the blockchain. This represents the hash value of the current block header. This represents the data within the block. This represents the hash value of the previous block. Indicates the timestamp of block creation. This represents the random number found during the consensus process.

[0139] When the block contains the instruction hash Once successfully created and reaches network consensus, it is considered to have passed the evidence verification and is recorded as [status]. True. Only when both biometric identification and blockchain verification are successfully passed is the target control instruction confirmed as valid and authorized for execution.

[0140] ;

[0141] In its formula, It is a Boolean value. For biometric verification results, and This serves as a marker for whether blockchain-based evidence storage is successful.

[0142] Finally, when When true, the original target control command The commands are sent to the distributed execution units. The control system includes a central command distribution module and multiple distributed execution units, such as servo control units and main control units. The distribution module is responsible for sending fully verified commands... The instruction is safely transmitted to the appropriate execution unit, thereby completing the final execution of the instruction.

[0143] S6. After the control commands are executed in a distributed manner, the operating status of the navigation signal lights is monitored regularly, the health of the signal lights is evaluated by collecting status data in real time, and a fault-tolerant recovery mechanism is triggered when an anomaly occurs.

[0144] Specifically, firstly, the system periodically collects key operating parameters from multiple sensors deployed on the navigation lights, forming a real-time status data vector. As part of monitoring the operational status of the navigation lights, the system will utilize a health assessment model to quantify the current operational status of the lights, ensuring their safe operation.

[0145] The model processes the collected real-time status data vectors and calculates a comprehensive health index. .

[0146] ;

[0147] In its formula, This represents the calculated overall health index. This represents the total number of key performance indicators used to assess health. Indicates assignment to the first Preset weights for each performance indicator, Represents the first extracted from the real-time state data vector. One raw sensor reading, This represents a predefined normalization function used to normalize the first... Original sensor readings This is mapped to a standardized performance score.

[0148] Furthermore, the system will calculate the health index in real time. With a preset health threshold Continuous comparisons are performed to automatically detect abnormal states. When the health index falls below a certain threshold, an anomaly is detected, and the system will immediately activate the aforementioned fault-tolerant recovery mechanism.

[0149] The first step of this mechanism is to use manifold learning reconstruction technology to repair and reconstruct the currently collected state data vectors, which may be damaged or incomplete, in order to recover the ideal operating state of the system from the observed fault data.

[0150] ;

[0151] In its formula, This represents the idealized health state vector reconstructed by the model. This represents a pre-trained manifold learning model. This represents a real-time status data vector that has been collected from the sensor and has been identified as abnormal.

[0152] Next, to ensure that the ideal state is reconstructed The generated recovery command is absolutely reliable when transmitted and issued to the redundant backup system, and quantum error correction coding technology is preferably used to encode and protect this critical information.

[0153] This process recovers the classical bit information in the instruction and encodes it into a highly fault-tolerant quantum state. It employs an encoding scheme based on multi-qubit entanglement, and the formula for the encoded logical qubit is as follows:

[0154] ;

[0155] In its formula, The expression representing a quantum state, Represents the normalization constant. express The ground state of a qubit. This indicates that the state is repeated. Second-rate, express The excited state of a qubit.

[0156] Finally, the recovery command, protected by quantum error correction encoding, is securely transmitted to the backup control module or hardware unit. Upon receiving and decoding, this unit executes the corresponding recovery action, such as activating the backup light group, adjusting the power module output, or restarting the control software. After the recovery action is completed, the system will re-enter the health monitoring loop to verify whether the fault has been successfully resolved.

[0157] Please see the appendix Figure 2 A remote control system for marine signal lights, comprising:

[0158] The information collection module collects dynamic and environmental information of the ship from multiple sensors, encrypts the collected data, and outputs it.

[0159] The spatiotemporal constraint matrix module receives encrypted information, constructs the spatiotemporal constraint matrix, and integrates ship motion prediction, position information, signal strength, and flow field compensation terms.

[0160] The hybrid prediction module performs hybrid prediction of the ship's future trajectory based on the spatiotemporal constraint matrix, and generates accurate hybrid prediction results by combining the correction of the data-driven model and the physical model.

[0161] The control command generation module is optimized by forming a multi-objective optimization problem based on the hybrid prediction results and the current environmental information, and obtaining the optimal weights of the control commands through a hybrid solver to generate the target control commands.

[0162] The verification and execution module encodes the target control commands and verifies their validity through biometric identification and blockchain notarization, and performs distributed execution of the control commands.

[0163] The status monitoring and fault-tolerant recovery module periodically monitors the operational status of the navigation lights after the control commands are executed, collects status data in real time to assess the health of the lights, and triggers the fault-tolerant recovery mechanism when an anomaly occurs.

[0164] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0165] Please see the appendix Figure 3 The present invention also provides an apparatus comprising: a processor and a memory, wherein the memory stores a processor-executable computer program, and the computer program, when executed by the processor, performs the method described above.

[0166] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for remote control of marine signal lights, characterized in that, Includes the following steps: S1. Collect ship dynamics and environmental information from multiple sensors, encrypt the information, and output it. S2. Receive the encrypted information to construct a spatiotemporal constraint matrix, and integrate the ship motion prediction term, position and signal strength term, and flow field compensation term. S3. Based on the spatiotemporal constraint matrix, perform hybrid prediction of the ship's future trajectory, combine the data-driven model and the physical model for correction, and generate hybrid prediction results; S4. Based on the hybrid prediction results and the current environmental information, a multi-objective optimization problem is formed. The optimal weights of the control commands are obtained through the hybrid solver, and the target control commands are generated according to the optimal weights. S5. Encode the target control instructions and verify their validity through biometric identification and blockchain evidence storage before distributing the execution of the control instructions. S6. After the control commands are executed in a distributed manner, the operating status of the navigation signal lights is monitored regularly. The health of the signal lights is evaluated by collecting status data in real time, and a fault-tolerant recovery mechanism is triggered when an anomaly occurs.

2. The remote control method for a marine signal light according to claim 1, characterized in that, In step S1, the information from multiple sensors includes an AIS receiver, a radar sensor, and a weather sensor. This information is encrypted using the SM4 algorithm. The SM4 encryption formula is as follows: ； In its formula, This represents the original collected data. This represents a symmetric encryption algorithm. Represents the key. This indicates the encrypted data.

3. The remote control method for a marine signal light according to claim 1, characterized in that, The step of constructing the spatiotemporal constraint matrix in step S2 includes: calculating the ship's motion terms, position and signal strength matrices, and solving the fluid dynamics equations to obtain flow field compensation. The fluid dynamics equations are as follows: ； In its formula, This indicates the movement of the fluid in this direction. This indicates the movement of the fluid in this direction. This indicates how fluid motion changes over time. This indicates the position of the fluid in the horizontal displacement direction. This indicates the position of the fluid in the vertical displacement direction. This represents the pressure value at a specific point in a fluid. This represents the mass per unit volume of a fluid. express The second derivative in space.

4. The remote control method for a navigation signal light according to claim 1, characterized in that, The hybrid prediction step for the future trajectory in step S3 includes data-driven prediction using an LSTM network and correction of the physical model by combining fluid dynamics to generate a hybrid prediction result. The hybrid prediction formula is as follows: ； In its formula, Indicates time The predicted result or trajectory at any given moment. Represents the weight parameters. Indicates time The prediction results are always based on data-driven models. Indicates time The prediction results are based on the physical model at all times.

5. The remote control method for a marine signal light according to claim 1, characterized in that, The multi-objective optimization problem formation step in step S4 includes constructing a quadratic programming model based on the hybrid prediction results and current environmental information, and using a hybrid solution method of quantum annealing and branch-and-bound to generate optimal weights for control commands. Subsequently, target control commands are generated based on these optimal weights. The quantum annealing algorithm formula is as follows: ； In its formula, Indicates the system's energy The probability of a state. Indicates energy state, This represents the energy of the current optimal solution. Represents the Boltzmann constant. Indicates the system temperature; The formula for the branch and bound method is: ； In its formula, Indicates through decision variables sum coefficient vector The calculated results Representation and decision variables The relevant weights or costs, This represents the set of variables that need to be determined in an optimization problem.

6. The remote control method for a marine signal light according to claim 1, characterized in that, The verification step of the target control command in step S5 utilizes biometric identification methods and blockchain notarization. The formula for the biometric identification method is as follows: ； In its formula, This indicates the output biometric verification result or score. This represents a pre-trained biometric matching function or model. This represents a feature vector extracted from biometric data collected in real time by sensors. The blockchain formula is as follows: ； In its formula, This indicates a new block that has been created and added to the blockchain. This represents the hash value of the current block header. This represents the data within the block. This represents the hash value of the previous block. Indicates the timestamp of block creation. This represents the random number found during the consensus process.

7. The remote control method for a navigation signal light according to claim 1, characterized in that, The monitoring of the operational status of the navigation signal lights in step S6 includes real-time assessment of the health of the navigation signal lights and the use of a health assessment model to ensure the safety of their operational status.

8. The remote control method for a marine signal light according to claim 1, characterized in that, The fault-tolerant recovery mechanism in step S6 includes a combined application of manifold learning reconstruction and quantum error-correcting coding. The manifold learning reconstruction formula is: ； In its formula, This represents the idealized health state vector reconstructed by the model. This represents a pre-trained manifold learning model. This represents a real-time status data vector that has been collected from the sensor and has been identified as abnormal. The quantum error correction coding formula is as follows: ； In its formula, The expression representing a quantum state, Represents the normalization constant. express The ground state of a qubit. This indicates that the state is repeated. Second-rate, express The excited state of a qubit.

9. A remote control system for marine signal lights, applied to the remote control method for marine signal lights as described in any one of claims 1-8, characterized in that, include: The information collection module collects dynamic and environmental information of the ship from multiple sensors, encrypts the collected data, and outputs it. The spatiotemporal constraint matrix module receives encrypted information, constructs a spatiotemporal constraint matrix, and integrates ship motion prediction, position information, signal strength, and flow field compensation terms. The hybrid prediction module performs hybrid prediction of the ship's future trajectory based on the spatiotemporal constraint matrix, and generates accurate hybrid prediction results by combining the correction of the data-driven model and the physical model. The control command generation module is optimized by forming a multi-objective optimization problem based on the hybrid prediction results and the current environmental information, and obtaining the optimal weights of the control commands through a hybrid solver to generate the target control commands. The verification and execution module encodes the target control commands and verifies their validity through biometric identification and blockchain notarization, and performs distributed execution of the control commands. The status monitoring and fault-tolerant recovery module periodically monitors the operational status of the navigation lights after the control commands are executed, collects status data in real time to assess the health of the lights, and triggers the fault-tolerant recovery mechanism when an anomaly occurs.

10. An apparatus comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-8.

Images