Integrated control system and method for safe sailing and carbon emission reduction of autonomous ship

Abstract

The present invention relates to an integrated control system and method for safe sailing and carbon emission reduction of an autonomous ship, and, more specifically, to an integrated control system and method for safe sailing and carbon emission reduction of an autonomous ship, which improve receptivity of a site in accordance with introduction of autonomous sailing technology through systematic data management, real-time information sharing between ships and land, a crew-centric interface, and the like.

Description

Integrated control system and method for safe navigation and carbon emission reduction of autonomous vessels

[0001] The present invention relates to an integrated control system and method for safe navigation and carbon emission reduction of autonomous vessels, and more specifically, to an integrated control system and method for safe navigation and carbon emission reduction of autonomous vessels that enhances on-site acceptance of the introduction of autonomous navigation technology through systematic data management, real-time information sharing between the vessel and land, and a crew-centered interface.

[0002] The shipping industry is the artery of the global economy, accounting for more than 90% of world trade. However, despite the increase in the size and speed of ships over the past few decades, fundamental innovation in navigation technology has been slow.

[0003] The reality is that analog navigation methods, which still rely on the experience and intuition of skilled sailors, remain dominant. This traditional navigation paradigm reveals limitations in the following aspects.

[0004] First, the risk of accidents caused by human error is ever-present. According to the International Maritime Organization (IMO), more than 80% of maritime accidents are attributed to human negligence. Fatigue, lack of experience, and mistakes by navigators can sometimes lead to major disasters.

[0005] In fact, the negligence of the night watch navigator was identified as the main cause of the chemical tanker capsizing accident that occurred in waters near Japan in 2019. The introduction of automation and autonomous technologies is urgently needed to overcome the cognitive and physical limitations of crew members and to ensure consistent and accurate navigation.

[0006] Second, inefficiency resulting from suboptimal decision-making is rampant. Current flight operations tend to rely heavily on empirical decision-making based on limited information.

[0007] In particular, decisions determining operational efficiency—such as selecting the optimal route, controlling speed, and managing fuel—are often relied upon the crew's intuition rather than systematic data analysis. The failure to fully utilize real-time information on changing ocean conditions and weather is also a problem. As a result, inefficiencies such as unnecessary fuel consumption and inaccurate arrival time predictions are occurring.

[0008] Third, there is a lack of consideration for the environmental impact of ship operations. Despite the IMO's greenhouse gas reduction strategy, carbon emissions in the shipping sector are actually on the rise. This is due to the prevalence of practices that prioritize profitability over environmental sustainability, such as high-speed operation and the use of aging vessels.

[0009] Problems such as air pollution and marine noise caused by ship operations are also at a serious level. To enhance the sustainability of the shipping industry, it is urgent to establish a decision-making system that considers the eco-friendliness of overall operations, along with technological solutions such as energy efficiency innovation and the introduction of clean fuels.

[0010] Fourth, ship management practices centered on reactive response rather than preventive maintenance are dominant. It is common practice to respond only after a breakdown occurs, rather than detecting and addressing potential signs of mechanical equipment abnormalities in advance. Decisions regarding the timing and methods of maintenance also tend to rely primarily on past experience.

[0011] As a result, inefficiencies arise, such as operational suspensions due to unexpected breakdowns and excessive maintenance costs. There is a need to introduce data-driven intelligent ship management technology that enables breakdown prediction and forecasting of optimal maintenance timing.

[0012] Fifth, there is a lack of technological infrastructure to support maritime digital transformation. A system capable of collecting, processing, and utilizing the vast amount of data generated from various sensors and equipment on board vessels has not been established. The absence of onboard networks, low satellite communication quality, and insufficient data standardization act as obstacles.

[0013] Furthermore, communication infrastructure for real-time data linkage and remote monitoring between land and vessels is insufficient. This acts as a limiting factor for big data analysis and the utilization of artificial intelligence technologies for operational optimization.

[0014] Sixth, the workload and manpower supply issues for seafarers are intensifying. With the increase in the size and speed of vessels, the scope of management and workload per seafarer have increased significantly. Conversely, the supply of seafarers is becoming increasingly difficult due to poor working conditions and a decline in the attractiveness of the profession. The shortage of skilled seafarers directly leads to an increased risk of maritime accidents.

[0015] It is necessary to reduce the workload of seafarers and enhance workforce flexibility by automating and remotely managing routine tasks such as navigation and maintenance, and by expanding technical support from land.

[0016] Seventh, despite technological advancements, limitations in the maritime communication environment persist. Issues such as the high cost and low speed of satellite communication remain, and communication blind spots are widespread. These factors limit the utilization of digital technologies, such as real-time data transmission and remote monitoring.

[0017] In addition, the separation of internal and external ship communication networks is pointed out as a limitation, as it makes data integration and real-time connectivity difficult. Along with the advancement of maritime communication infrastructure, the integration of onboard networks and data standardization are urgently required.

[0018] Eighth, the regulatory framework for autonomous navigation technology is inadequate. Existing maritime regulations and international conventions are designed based on manned vessels.

[0019] The lack of clear regulations regarding the safety assessment of autonomous vessels, liability for accidents, and the legal status of crew members is acting as an obstacle to the development and commercialization of related technologies. It is necessary to resolve the institutional uncertainties arising from the introduction of autonomous navigation technology through proactive regulatory overhauls and the establishment of international standards.

[0020] Ninth, social acceptance of artificial intelligence technology is not yet sufficient. It is true that there is a vague sense of anxiety regarding fully autonomous flight without human intervention.

[0021] In particular, doubts persist regarding AI response capabilities in emergency situations and ethical considerations. Furthermore, social concerns such as job losses and privacy infringement resulting from autonomous flight are also being raised. Efforts are needed to enhance the reliability of the technology while narrowing the gap in perception through communication with the general public.

[0022] Tenth, there is a lack of core technologies and players to drive change across the entire industrial ecosystem. While partial technological development is taking place in individual areas such as flight operations, air traffic control, security, and maintenance, integrated solutions encompassing them are insufficient.

[0023] The absence of leading companies or institutions to unite diverse stakeholders and drive cooperation also acts as a factor slowing down digital transformation. It is urgent to establish an industrial innovation ecosystem based on openness and collaboration to advance autonomous navigation technology and create new business models.

[0024] As examined above, the traditional ship operation paradigm is facing limitations in various aspects, such as safety, efficiency, and environmental friendliness.

[0025] To respond to the rapidly changing shipping environment and lead the new maritime era, a paradigm shift based on innovative technologies such as autonomous navigation and artificial intelligence is inevitable. However, the reality is that the development and adoption of related technologies are slow due to technical, institutional, and social constraints.

[0026] Against this backdrop, the present invention originated from the awareness of the need to accelerate the commercialization of autonomous ships and fundamentally improve the structure of the shipping industry. By proposing an integrated control system for autonomous ships that can dramatically enhance safety, efficiency, and eco-friendliness, the invention aims to overcome the limitations of existing technologies and create new value.

[0027] We aimed to maximize synergy effects by advancing core element technologies of autonomous navigation through the convergence of advanced information and communication technologies (ICT), such as multi-sensor fusion, artificial intelligence, and digital twin, while implementing a platform that organically integrates them.

[0028] The present invention has been devised to improve upon the aforementioned problems, and the present invention aims to solve the following problems.

[0029] First and foremost, the core of autonomous navigation lies in environmental perception, situational judgment, and decision-making capabilities. Advanced artificial intelligence technology is required for ships to independently determine optimal routes and respond rapidly to unexpected situations in complex and dynamic marine environments. It must be able to process large volumes of data collected from various sensors, such as radar, lidar, and cameras, in real time, and further perform semantic reasoning by combining this data with maritime knowledge in the form of natural language.

[0030] Secondly, ensuring safety is crucial. Ship collisions or groundings are catastrophic accidents that can lead not only to loss of life but also to marine pollution. Therefore, autonomous vessels must strictly comply with safety regulations set by the International Maritime Organization (IMO) while prioritizing the safety of passengers and crew in the event of an emergency. To achieve this, multiple safety measures are required, including collision avoidance algorithms and fault detection and response systems.

[0031] The third point is to enhance operational efficiency and economic viability. Since fuel costs account for a significant portion of ship operating expenses, it is necessary to minimize fuel consumption while maximizing transport efficiency. This can be achieved through optimal route planning, the utilization of real-time weather information, and monitoring of hull and engine conditions. Furthermore, efforts to improve overall operational efficiency, such as reducing crew labor costs and optimizing maintenance plans, must be pursued in parallel.

[0032] Finally, responding to environmental regulations is emerging as a major challenge. The IMO has set a goal to reduce greenhouse gas emissions from ships by 50% by 2050 compared to 2008 levels. Accordingly, autonomous ships must devise technical measures to dramatically improve energy efficiency and minimize the emission of air pollutants and CO2. Along with the application of alternative fuels such as LNG, hydrogen, and electricity, the introduction of innovative energy-saving technologies, such as waste heat recovery and aerodynamic design, is required.

[0033] To achieve the above objectives, one embodiment of the present invention relates to an integrated control system for safe navigation and carbon emission reduction of an autonomous vessel, comprising: a multi-sensor unit (100) for collecting environmental information around the vessel; a collision risk analysis unit (200) for analyzing data collected from the multi-sensor unit (100) to evaluate the risk of collision with surrounding vessels and obstacles; a path optimization unit (300) for calculating the optimal path of the vessel considering the current location, destination, and weather information; a vessel control unit (400) for directly controlling the vessel's engine, rudder, and propulsion system; a carbon emission monitoring unit (500) for calculating and monitoring the vessel's real-time carbon emissions; an integrated AI control unit (600) for establishing an optimal control strategy by comprehensively analyzing data from the collision risk analysis unit (200), the path optimization unit (300), the vessel control unit (400), and the carbon emission monitoring unit (500); a central control unit (700) for coordinating and controlling each sub-system according to the command of the integrated AI control unit (600); and a communication unit (800) responsible for real-time data transmission and reception between the vessel and a land-based control center. A digital twin unit (900) that creates a virtual model of an actual ship and performs a simulation; a status monitoring unit (1000) that monitors the status of the ship's main equipment in real time; and a predictive maintenance unit (1100) that analyzes the status data of the main equipment to predict failures and optimize maintenance schedules;The integrated AI control unit (600) establishes a control strategy that simultaneously optimizes the safety, efficiency, and environmental friendliness of the ship by comprehensively considering the collision risk assessment result of the collision risk analysis unit (200), the optimal path calculation result of the path optimization unit (300), carbon emission data of the carbon emission monitoring unit (500), equipment status information of the status monitoring unit (1000), and maintenance schedule information of the predictive maintenance unit (1100); the digital twin unit (900) verifies the safety and efficiency by simulating the control strategy established by the integrated AI control unit (600) in a virtual environment, feeds the verification results back to the integrated AI control unit (600) to continuously improve the control strategy, performs simulations to predict the state changes and failure possibilities of the ship's major equipment, performs simulations to predict the state changes and failure possibilities of the ship's major equipment to optimize the maintenance plan based on the prediction results, and the central control unit (700) [manages] the ship according to the control strategy established by the integrated AI control unit (600). The control unit (400) controls the ship's engine, rudder, and propulsion system, and the communication unit (800) transmits the ship's operation data, carbon emission status, and equipment status information in real time to the land control center, receives updated weather information, traffic information, and emergency control commands from the land control center and transmits them to the integrated AI control unit (600), and the integrated AI control unit (600) stores control strategies including engine speed adjustment, optimal propulsion settings, and route adjustment to minimize the ship's fuel consumption based on the optimal route information provided by the route optimization unit (300).

[0034] The present invention relates to an integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, comprising: a step of collecting data through a multi-sensor unit for collecting environmental information around the vessel; a step of analyzing collision risk through a collision risk analysis unit that evaluates the risk of collision with surrounding vessels and obstacles by analyzing the data collected from the multi-sensor unit; a step of calculating an optimal path through a path optimization unit that calculates an optimal path considering the current location, destination, and weather information; a step of executing control commands through a vessel control unit to control the vessel's engine, rudder, and propulsion system; a step of monitoring carbon emissions through a carbon emission monitoring unit that calculates and monitors the vessel's real-time carbon emissions; a step of establishing a control strategy through an integrated AI control unit that establishes an optimal control strategy by comprehensively analyzing data from the collision risk analysis unit, the path optimization unit, the vessel control unit, and the carbon emission monitoring unit; a step of controlling the entire system through a central control unit that coordinates and controls each sub-system according to the commands of the integrated AI control unit; a step of performing a simulation through a digital twin unit that creates a virtual model of the actual vessel and performs a simulation, and providing feedback on the results; and a step of monitoring the status of equipment through a status monitoring unit that monitors the status of the vessel's major equipment in real time. It includes a step of establishing a maintenance plan through a predictive maintenance unit that analyzes the status data of the equipment to predict failures and optimize maintenance schedules.

[0035] The present invention provides an integrated control system that enables the safe, efficient, and eco-friendly operation of autonomous vessels. The proposed system is expected to provide technical and economic benefits in the following aspects.

[0036] First, the core functions of autonomous navigation, such as situational awareness and decision-making capabilities, have been advanced. By applying the latest artificial intelligence technologies—including the fusion of multi-sensor data, deep learning-based situational awareness, and reinforcement learning-based decision-making—significant improvements in autonomy have been achieved compared to existing systems. Furthermore, by linking this with a knowledge graph, reasoning capabilities have been elevated to a new level. Through this, the vessel is now able to make quick and accurate judgments even in unpredictable and unexpected situations.

[0037] Second, multi-layered safety mechanisms, such as collision avoidance and predictive maintenance, have been established to fundamentally eliminate the risk of accidents. In particular, simulation technology utilizing digital twins is employed to comprehensively verify the safety of the vessel and identify hidden risk factors in advance. Furthermore, by enabling immediate warnings and responses in the event of an emergency, human and material damage resulting from accidents can be minimized.

[0038] Third, operational efficiency and economic viability of vessels have been maximized through data-driven scientific operational management. By analyzing vast amounts of real-time operational data to determine optimal speeds, fuel consumption, and maintenance cycles, unnecessary energy losses can be minimized and utilization rates increased. This is expected to not only directly lead to improved profitability for shipping companies but also ultimately contribute to enhancing the competitiveness of the shipping industry as a whole.

[0039] Fourth, eco-friendliness has been secured through energy efficiency innovation and the utilization of clean energy. Not only has propulsion efficiency been significantly improved through the application of waste heat recovery and aerodynamic hull shapes, but air pollutant emissions have also been greatly reduced by expanding the use of alternative fuels such as LNG and hydrogen. In addition, a real-time carbon emission monitoring system induces continuous improvement by "visualizing" the ship's environmental performance.

[0040] Fifth, a virtuous cycle structure has been established for the advancement of autonomous navigation technology. The accumulation of navigation data, the continuous improvement of algorithms through simulation, and the repeated field application and verification will not only elevate the completeness of the technology but also serve as a catalyst for new innovations. Furthermore, the accumulation of such data and know-how is expected to contribute to enhancing competitiveness in the marine AI sector and revitalizing the related industrial ecosystem.

[0041] In summary, this invention can be described as an innovative technology with the potential to accelerate the commercialization of autonomous vessels and transform the paradigm of the shipping industry. Furthermore, by contributing to the creation of social benefits such as the reduction of maritime accidents and environmental protection, it will establish itself as a marine mobility solution for sustainable development.

[0042] FIGS. 1 and FIGS. 2 are schematic diagrams illustrating the concept of an integrated control system for safe navigation and carbon emission reduction of an autonomous vessel according to an embodiment of the present invention.

[0043] FIGS. 3 and FIGS. 4 are schematic diagrams illustrating the relationships between components implementing an integrated control system for safe navigation and carbon emission reduction of an autonomous vessel according to an embodiment of the present invention.

[0044] FIG. 5 is a sequence diagram implementing an integrated control method for safe navigation and carbon emission reduction of an autonomous vessel according to an embodiment of the present invention.

[0045] FIG. 6 is a diagram schematically showing the relationship between components that implement integrated control for safe navigation and carbon emission reduction of an autonomous vessel, such as an integrated AI control unit according to one embodiment of the present invention.

[0046] The present invention as described above will be explained in detail through the attached FIGS. 1 to 6 and embodiments.

[0047] When a technical term used in this invention is a technical term that similarly expresses the concept of this invention, it should be understood as being replaced with a technical term that can be correctly understood by a person skilled in the art (e.g., ~ module, ~ server, ~ part).

[0048] Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings, wherein identical or similar components and functions, regardless of the drawing symbols, are given the same reference number and function as modules, servers, parts, means, devices, etc. having specific functions.

[0049] Furthermore, in describing the present invention, detailed descriptions of related prior art are omitted if it is determined that such descriptions could obscure the essence of the invention. Additionally, it should be noted that the attached drawings are intended only to facilitate an understanding of the concept of the present invention and should not be interpreted as limiting the concept of the present invention.

[0050] In this case, each functional description divided by a distinguishing number describing each embodiment implies that it includes a function or module according to such description. Furthermore, these functions or modules are organically connected to the present invention via a network.

[0051] Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the attached drawings. Identical or similar components are given the same reference number regardless of the drawing symbols, and redundant descriptions thereof will be omitted.

[0052] In this case, each functional description divided by a distinguishing number describing each embodiment implies that it includes a function or module according to such description. Furthermore, these functions or modules are organically connected to the present invention via a network.

[0053] The integrated control system for safe navigation and carbon emission reduction of the autonomous vessel of the present invention is composed of the following components, and the detailed functions and operating methods of each component are as follows.

[0054] The multi-sensor unit (100) is composed of various sensors for collecting environmental information around the ship. Specifically, it includes the following sensors:

[0055] a) The radar (101) uses electromagnetic waves to detect surrounding vessels and obstacles. The radar can operate 24 hours a day regardless of the weather and is particularly effective for detecting large objects. In this system, X-band and S-band radars are used together to effectively detect both near and far objects.

[0056] b) The lidar (102) collects 3D information of the surrounding environment using a laser. Lidar can measure distances more precisely than radar and can detect small objects. In this system, a 360-degree rotating lidar is used to perform an all-around scan of the ship.

[0057] c) The AIS receiver (103) receives identification information and navigation information of surrounding vessels. The Automatic Identification System (AIS) is a system that automatically transmits and receives information such as the name, size, location, speed, and course of a vessel, and this system utilizes this information to identify the movements of surrounding vessels.

[0058] d) The camera (104) collects visual information and performs image processing. It enables 24-hour monitoring by using a high-resolution optical camera during the day and an infrared camera at night. It also performs object recognition and tracking through image analysis AI.

[0059] e) GPS (105) obtains accurate position information of the vessel. In addition to GPS, multiple satellite navigation systems (GNSS) such as GLONASS and Galileo are used together to increase position accuracy.

[0060] f) The weather sensor (106) collects weather information such as wind direction, wind speed, atmospheric pressure, humidity, and temperature. This information is used as an important factor in route planning and ship control.

[0061] Each sensor of the multi-sensor unit (100) operates on a different principle and has its own advantages and disadvantages. In this system, by fusing and processing the data from these sensors, accurate and comprehensive environmental information that cannot be obtained from a single sensor alone can be obtained.

[0062] The collision risk analysis unit (200) analyzes data collected from the multi-sensor unit (100) to evaluate the risk of collision with surrounding vessels and obstacles. This unit consists of the following modules:

[0063] a) The object detection and tracking module (201) fuses sensor data to estimate the position, velocity, and direction of surrounding objects. This module uses an advanced tracking algorithm, such as a Kalman filter, to accurately predict the movement of objects.

[0064] b) The CPA / TCPA calculation module (202) calculates the Closest Point of Approach (CPA) and the Time to Closest Point of Approach (TCPA). This information is a key factor in determining the risk of collision.

[0065] c) The collision risk assessment module (203) calculates the collision risk by considering CPA, TCPA, and other factors. This module uses fuzzy logic to accurately assess the risk for various situations.

[0066] The collision risk analysis unit (200) continuously monitors the surrounding environment and, when a dangerous situation occurs, immediately sends a notification to the integrated AI control unit (600) so that appropriate avoidance measures can be taken.

[0067] The path optimization unit (300) calculates the optimal path by considering the current location, destination, weather information, etc. To calculate the optimal path, it is composed of the following elements:

[0068] a) The nautical chart and route information database (301) stores and manages electronic navigational chart (ENC) information and existing route information. This database is regularly updated to reflect the latest marine information.

[0069] b) The weather and ocean current information collection module (302) collects real-time weather information and ocean current information. This module comprehensively utilizes information from various sources, such as satellite weather data, ocean observation station data, and numerical weather prediction models.

[0070] c) The multi-objective optimization algorithm (303) derives the optimal path by simultaneously considering several factors such as distance, time, fuel consumption, and carbon emissions.

[0071] The multi-objective optimization algorithm (303) solves complex optimization problems using evolutionary computation techniques such as the genetic algorithm.

[0072] The route optimization unit (300) does not simply find the shortest route, but rather presents an optimal route by comprehensively considering various factors such as weather conditions, ocean currents, fuel efficiency, and carbon emissions. In addition, it recalculates the route in real time according to changes in conditions during navigation to ensure that the optimal route is always maintained.

[0073] The ship control unit (400) plays the role of directly controlling the ship's engine, rudder, propulsion system, etc. The ship control unit (400) is composed of the following modules:

[0074] a) The engine control module (401) controls the output of the ship's main engine and auxiliary engine. The engine control module (401) precisely adjusts the fuel injection amount, ignition timing, etc., to optimize engine efficiency and minimize exhaust gas.

[0075] b) The rudder control module (402) controls the rudder that controls the direction of the ship. The rudder control module (402) calculates the optimal rudder angle based on a hydrodynamic model to enable accurate steering.

[0076] c) The propulsion system control module (403) controls a propulsion system such as a propeller or a water jet. The propulsion system control module (403) controls the propulsion system in a way that optimizes propulsion efficiency and minimizes cavitation.

[0077] The ship control unit (400) plays the role of converting control commands received from the integrated AI control unit (600) into actual ship operations. In this process, safe and efficient control is achieved by taking into account the ship's current state and surrounding environment.

[0078] The carbon emission monitoring unit (500) is responsible for calculating and monitoring the real-time carbon emissions of the ship. This unit consists of the following modules:

[0079] a) The engine performance data collection module (501) collects data such as fuel consumption, output, and rotational speed of the engine in real time.

[0080] The engine performance data collection module (501) collects precise data from various sensors attached to the engine and provides the basis for accurate carbon emission calculation.

[0081] b) The carbon emission calculation module (502) calculates real-time carbon emissions based on collected engine performance data.

[0082] The carbon emission calculation module (502) calculates accurate emissions using a carbon emission calculation algorithm that complies with the guidelines of the International Maritime Organization (IMO).

[0083] c) The emissions visualization module (503) visually represents the calculated carbon emissions data.

[0084] The emission visualization module (503) generates a real-time carbon emission status map based on GIS (Geographic Information System) to enable intuitive monitoring.

[0085] The carbon emission monitoring unit (500) does not merely calculate carbon emissions, but provides this information to the integrated AI control unit (600) to utilize it in establishing a ship operation strategy to minimize carbon emissions.

[0086] The integrated AI control unit (600) is a core part of the system and establishes an optimal control strategy by comprehensively analyzing all collected information. This part consists of the following modules:

[0087] a) The data preprocessing module (601) refines and standardizes the data collected from each part.

[0088] The data preprocessing module (601) performs tasks such as noise removal, missing value processing, and feature extraction to obtain high-quality data.

[0089] b) The deep learning-based situation recognition module (602) accurately recognizes the current situation based on the preprocessed data.

[0090] The deep learning-based situation recognition module (602) uses a complex neural network structure combining a convolutional neural network (CNN) and a recurrent neural network (RNN) to accurately interpret complex marine environments.

[0091] c) The reinforcement learning-based decision module (603) determines the optimal response strategy for the perceived situation.

[0092] The reinforcement learning-based decision module (603) uses an advanced reinforcement learning algorithm such as Deep Q-Learning to make decisions that simultaneously consider safety, efficiency, and environmental friendliness.

[0093] d) The control command generation module (604) converts the determined strategy into an actual control command. This module uses a Model Predictive Control technique to generate an optimal control command that takes into account the dynamic characteristics of the vessel.

[0094] The integrated AI control unit (600) improves performance through continuous learning. It enhances the ability to respond to various situations by utilizing both actual flight data and simulation data.

[0095] The central control unit (700) plays the role of coordinating and controlling each subsystem according to the commands of the integrated AI control unit (600). This part performs the following functions:

[0096] a) The system integration management function regulates data flow and operation timing between each subsystem to ensure that the entire system operates organically.

[0097] b) When multiple control commands occur simultaneously, the priority adjustment function determines the priority by considering the urgency and importance of the situation.

[0098] c) The safety monitoring function continuously monitors the operating status of each subsystem and takes immediate response measures when abnormal signs occur.

[0099] d) The backup system management function manages backup systems in preparation for the event of a problem in the main system and switches quickly when necessary.

[0100] The central control unit (700) operates based on a real-time operating system (RTOS) with high reliability and stability, thereby preventing errors caused by minute time differences and ensuring a fast response speed.

[0101] The communication unit (800) is responsible for real-time data transmission and reception with the ground control center. This unit consists of the following modules:

[0102] a) The LTEM module (801) is responsible for high-speed data communication in coastal areas. LTEM (Long Term Evolution for Machines) has advantages such as low power consumption, wide coverage, and high penetration power, making it suitable for ship communication.

[0103] b) The satellite communication module (802) is responsible for data communication in the open ocean. This module supports various marine satellite communication systems such as Inmarsat and Iridium, enabling stable communication anywhere in the world.

[0104] c) The data encryption and security module (803) is responsible for the security of communication data. This module encrypts data using a strong encryption algorithm such as AES (Advanced Encryption Standard) and establishes a secure communication channel through a VPN (Virtual Private Network).

[0105] The communication unit (800) supports multiple communication paths to enable continuous data exchange even when a communication failure occurs. In addition, it ensures priority transmission of important data through a bandwidth management function.

[0106] The digital twin section (900) is responsible for creating a virtual model of the actual ship and performing simulations. This section consists of the following modules:

[0107] a) The 3D ship modeling module (901) generates a precise 3D model of the ship. Based on CAD (Computer-Aided Design) data, this module models not only the external shape of the ship but also its internal structure and systems in detail.

[0108] b) The real-time data synchronization module (902) reflects the sensor data of the actual vessel into the virtual model in real time. This allows the virtual model to accurately reflect the state of the actual vessel.

[0109] c) The simulation engine (903) performs simulations for various situations. This engine provides realistic simulations by including various physical models such as fluid dynamics, thermodynamics, and structural mechanics.

[0110] d) The result analysis and feedback module (904) analyzes the simulation results and provides feedback to the actual system. This module utilizes machine learning techniques to extract meaningful insights from the simulation results and provides them to the integrated AI control unit (600) to improve the control strategy.

[0111] The digital twin unit (900) plays a key role in simulating various situations that may occur during actual operation in advance to predict potential risks and prepare countermeasures.

[0112] The status monitoring unit (1000) monitors the status of the ship's main equipment in real time. This unit consists of the following modules:

[0113] a) Sensor data collection module (1001): Collects data from various sensors attached to major equipment such as engines, generators, and propulsion systems. This module measures various physical quantities such as vibration, temperature, pressure, and current to accurately determine the condition of the equipment.

[0114] b) Data analysis and diagnosis module (1002): Analyzes collected sensor data to diagnose the condition of the equipment. This module combines statistical techniques and machine learning algorithms to detect deviations from the normal state and identify potential problems early.

[0115] The condition monitoring unit (1000) detects signs of abnormality in the equipment at an early stage to enable preventive maintenance, thereby increasing the reliability of the equipment and reducing operating costs.

[0116] The predictive maintenance unit (1100) analyzes the condition data of the equipment to predict failures and optimize the maintenance schedule. This unit consists of the following modules:

[0117] a) Failure prediction algorithm (1101): Predicts the likelihood of equipment failure based on condition monitoring data. This algorithm uses a deep learning-based time series analysis technique to recognize complex patterns and predict the likelihood of future failure.

[0118] b) Maintenance schedule optimization module (1102): Establishes an optimal maintenance schedule by considering the predicted failure probability, the life cycle of parts, and the operation schedule. This module uses a multi-objective optimization algorithm to present a schedule that simultaneously considers the effectiveness and cost efficiency of maintenance.

[0119] The predictive maintenance unit (1100) enables maintenance based on the condition of the equipment, thereby reducing unnecessary maintenance and ensuring that maintenance at critical times is not missed.

[0120] The process by which the above components are organically combined and operate is as follows:

[0121] 1. The multi-sensor unit (100) continuously collects environmental information around the ship.

[0122] 2. The collision risk analysis unit (200) analyzes the collected data in real time to evaluate the collision risk.

[0123] 3. The path optimization unit (300) calculates the optimal path by considering the current location, destination, weather information, collision risk, etc.

[0124] 4. The carbon emission monitoring unit (500) calculates and monitors real-time carbon emissions based on the current operating status of the ship.

[0125] 5. The integrated AI control unit (600) comprehensively analyzes all information, such as collision risk, optimal path, carbon emissions, and equipment status, to establish an optimal control strategy.

[0126] 6. The central control unit (700) controls the actual ship through the ship control unit (400) according to the command of the integrated AI control unit (600).

[0127] 7. The communication unit (800) continuously exchanges data with the ground control center and receives remote control commands when necessary.

[0128] 8. The digital twin unit (900) performs a virtual simulation based on actual ship operation data to verify the safety and efficiency of the control strategy and feeds the results to the integrated AI control unit (600).

[0129] 9. The condition monitoring unit (1000) and the predictive maintenance unit (1100) continuously monitor and analyze the condition of the ship equipment to establish an optimal maintenance plan, and provide this information to the integrated AI control unit (600) to reflect it in the operation strategy.

[0130] Through such an integrated control system, autonomous vessels can maximize safety while simultaneously improving fuel efficiency and minimizing carbon emissions. Furthermore, virtual simulations utilizing digital twin technology enable preparation for various scenarios, and predictive maintenance can enhance equipment reliability.

[0131] This system is expected to lead the digital transformation of the shipping industry and make a significant contribution to marine environmental protection and safe maritime transportation.

[0132] As an example, the configuration and functions of the integrated AI control unit (600) are described in detail as follows.

[0133] First, the data preprocessing module (601) is responsible for processing raw data collected from other systems into a form suitable for AI models to learn and infer. The main functions of this module include data cleaning, normalization and scaling, feature extraction and selection, and data augmentation.

[0134] Data cleaning is a process that improves data quality by handling missing values, removing outliers, and filtering noise. Normalization and scaling are techniques that enhance the stability and convergence speed of model training by consistently adjusting the units and ranges of data; MinMax normalization and Z-score normalization are representative examples. MinMax normalization scales data to values ​​between 0 and 1, while Z-score normalization transforms data so that the mean is 0 and the standard deviation is 1.

[0135] Feature extraction and selection is the process of extracting and selecting features useful for training AI models by utilizing domain knowledge and statistical techniques. To this end, Principal Component Analysis (PCA) can be used to reduce the dimensionality of high-dimensional data while preserving key information, or correlation analysis can be used to analyze correlations between features to remove redundant or unnecessary features.

[0136] Data augmentation is a technique that generates new training samples by transforming existing data using methods such as rotation, inversion, or resizing when training data is insufficient.

[0137] Next, the deep learning-based situation recognition module (602) performs the role of recognizing and classifying the current situation of the ship based on the preprocessed data. This module uses a hybrid model structure that combines a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) network.

[0138] CNN is a neural network that demonstrates excellent performance in extracting features from grid-structured data such as images, and is used to recognize patterns of surrounding objects in radar or camera images. LSTM is a type of recurrent neural network specialized for processing time-series data, capable of capturing pattern changes over time in continuous data collected from sensors.

[0139] This module uses a trained model to classify which situation class the input data belongs to. For example, it can be classified into situations such as normal operation, collision risk, bad weather, equipment malfunction, and emergency. Additionally, by applying ensemble learning techniques to combine the prediction results of multiple models, the accuracy and reliability of situation recognition can be improved. It is also possible to quantify uncertainty through Bayesian inference.

[0140] Recently, Explainable AI (XAI) technology, which provides the rationale for AI model predictions in a form understandable to humans, has also been gaining attention. By applying XAI techniques such as SHAP (SHapley Additive exPlanations) to a context awareness module, the rationale for the recognized situation can be presented visually.

[0141] When a situation is recognized, a decision module (603) intervenes to formulate a strategy to optimally respond to the situation. This module is implemented using reinforcement learning, particularly deep reinforcement learning algorithms. Representative algorithms include Deep QNetwork (DQN) and Proximal Policy Optimization (PPO).

[0142] In reinforcement learning, an agent takes a specific action from a given state and receives a reward from the environment as a result. The agent's goal is to maximize the sum of long-term rewards. To achieve this, the agent observes the current state and learns a policy that selects the optimal action based on past experiences.

[0143] In the decision-making module, the state consists of variables representing the vessel's current situation, such as position, speed, direction, surrounding environment information, and equipment status. Actions refer to decisions related to vessel control, such as speed adjustment, direction change, and engine output control. Compensation is defined as a combination of indicators that evaluate the degree of achievement of goals, such as safe operation, fuel efficiency, and carbon emission reduction.

[0144] The reinforcement learning process is mostly carried out in a simulation environment. Virtual scenarios are generated in conjunction with the digital twin (900), and agents are trained in this environment to develop the ability to respond to real-world situations. By utilizing the Monte Carlo Tree Search (MCTS) technique, long-term decision-making strategies can be established based on the simulation results.

[0145] It is also possible to model each subsystem of a ship, such as the navigation system, engine system, and electrical system, as individual agents and optimize the entire system through cooperative interactions among them. To achieve this, a Multi-Agent Reinforcement Learning framework can be utilized.

[0146] The strategy derived by the decision module is converted into a specific control signal by the control command generation module (604). This module determines the optimal control input while predicting future state changes using a Model Predictive Control (MPC) technique based on a ship dynamics model.

[0147] MPC is a technique that simulates the operation of a system starting from its current state up to a specific point in the future and derives an optimal control sequence based on the results. In this process, the ship's equations of motion, physical constraints, and environmental disturbances are considered. Control commands are appropriately modulated considering the characteristics of the actuators, and robustness against disturbances is ensured through closed-loop feedback control.

[0148] When a control command is generated, the safety verification module (606) operates to determine whether the command threatens safe navigation. To this end, it examines whether the control command exceeds the physical limits of the vessel, whether there is a risk of collision, etc., through simulation and logical analysis.

[0149] If the control command does not satisfy safety requirements, the safety constraints are strengthened to re-perform the MPC process, or an alternative strategy is sought within the scope where safety is guaranteed. The control command that passes verification is transmitted to each subsystem via the central control unit (700) and executed.

[0150] Meanwhile, the explainability module (607) serves to provide the decision-making process and rationale of the AI ​​system in a form that humans can understand. This is important for increasing the transparency and reliability of the AI ​​system and for enhancing user acceptance.

[0151] The explainability module operates in two main ways. One is to provide ex post explanations for individual decisions, using techniques such as Local Interpretable ModelAgnostic Explanations (LIME) or Shapley Additive Explanations (SHAP). The other is to construct the AI ​​model itself in a form that is human-interpretable, utilizing decision trees or fuzzy rule-based systems.

[0152] The Explainability module visualizes the features that played a major role in the situational awareness and decision-making processes and their logical relationships, or presents the rules serving as the basis for decisions in natural language. Additionally, it can provide analysis results regarding assumptions such as "What would have happened if ~ had not happened?" through counterfactual reasoning.

[0153] Finally, the human AI interaction module (608) effectively conveys the system's situational awareness results and proposed decisions to the operator and acts as an interface that supports the operator's intervention.

[0154] This module facilitates rapid situational awareness by providing contextual information and control strategies in intuitive formats such as graphics, charts, and natural language. It raises awareness of potential hazards by generating warning messages or behavioral recommendations when necessary. Furthermore, it offers a mechanism to seamlessly transfer control to the AI ​​system if the operator determines that intervention is required.

[0155] Operator intervention can contribute to the continuous learning and improvement of the system, going beyond merely supplementing AI judgment. This is because accumulated data on the operator's control behaviors can be utilized to fine-tune the AI ​​model, thereby incorporating expert experience and know-how into the model.

[0156] As described above, the integrated AI control unit (600) is a core module that organically integrates a series of processes ranging from data preprocessing to situation recognition, decision making, generation of control commands, safety verification, securing explainability, and human-AI collaboration.

[0157] As the brain of the entire system, it realizes the intelligent integration of all hardware and software elements, including various sensors and actuators installed on the ship, and ultimately plays a role in supporting the achievement of the goals of ensuring the safety, efficiency, and eco-friendliness of autonomous navigation.

[0158] The specific details of the embodiments of the present invention are as follows.

[0159] One embodiment relates to a digital twin unit (900) that performs simulations in a virtual environment based on actual ship operation data. This module contributes to enhancing the safety and reliability of the autonomous navigation system by precisely simulating the dynamic behavior of the ship through physics-based modeling and performing virtual verification of various operation scenarios based on this.

[0160] The core components of the digital twin section (900) are as follows:

[0161] 1. The 3D ship modeling module (901) generates a 3D model that accurately reproduces the shape, dimensions, structure, etc. of an actual ship. It models the shapes of major parts such as the hull, superstructure, propeller, and rudder using CAD data, 3D scanning data, etc. The level of detail of the model can be appropriately adjusted according to the purpose of the simulation and the required accuracy.

[0162] 2. Physics Engine (902): The physics engine is a core module that calculates the dynamic behavior of the ship. It includes hull motion equations, thrust and moment models for the propeller and rudder, and environmental load models such as waves and wind, and through this, simulates the 6-degrees-of-freedom motion of the ship. In addition, through fluid-structure coupled analysis, the effect of hull deformation on motion can be considered.

[0163] 3. The sensor and equipment model (903) is a module that simulates the operation of various sensors and equipment installed on an actual ship. It implements in detail the operating characteristics and control logic of navigation sensors such as GPS, IMU, radar, and cameras, as well as engine equipment such as engines, generators, and pumps. Through this, the dynamic characteristics of the equipment and the impact of sensor noise on the autonomous navigation system can be identified in advance.

[0164] 4. The virtual environment generator (904) creates a virtual environment in which a ship operates. It includes various terrain and water depth information, such as ports, canals, and straits, and realistically models weather conditions (wind, waves, currents, etc.). It also provides a traffic generation function and a collision detection function to simulate the movement of surrounding ships.

[0165] 5. The scenario editor (905) is an editing tool that helps users easily configure desired scenarios. It allows for fine adjustment of key environmental conditions such as terrain, weather, and traffic, as well as the initial state of the vessel, target route, and control strategy. It provides predefined scenario templates to enable the rapid configuration of general operating situations.

[0166] 6. The simulation controller (906) is a module that controls the execution of the simulation and analyzes the results. Along with basic execution control functions such as starting, pausing, and ending the simulation, it provides convenience functions such as time reduction / extension and saving / loading intermediate states. In addition, the simulation results can be saved and visualized in various formats (graphs, videos, log data, etc.).

[0167] 7. The data linkage module (907) is a module that links the actual ship's operation data with the simulation model. It can receive measurement data from sensors and equipment in real time to update the state of the virtual model, or conversely, transmit simulation results to the actual ship to use as control inputs. Through this, the performance of the autonomous navigation system can be effectively verified and improved by organically linking the verification of the actual ship and the verification of the virtual ship.

[0168] The above modules are organically linked, and the digital twin unit (900) operates in the following procedure:

[0169] 1. Construct a virtual model of the ship based on the actual ship's design data, measurement data, etc.

[0170] 2. Create a virtual environment according to the flight scenario and set the simulation conditions.

[0171] 3. Simulate the behavior of the ship model through a physics engine, and in this process, generate outputs for the sensors and equipment models.

[0172] 4. Perform a closed-loop simulation by integrating the situation awareness, decision-making, and control logic of the autonomous navigation system into a simulation loop.

[0173] 5. Analyze the simulation results to evaluate the performance of the autonomous navigation system and identify improvements if necessary.

[0174] 6. Update the actual system to reflect the derived improvements, and synchronize the changes with the virtual model.

[0175] By repeating this series of processes, the digital twin (900) enables continuous verification and improvement of the autonomous navigation system. Its value is particularly great in that it allows for the verification and strengthening of response capabilities for extreme situations or emergency scenarios that are difficult to verify in reality.

[0176] Example 7 relates to a communication unit (800) responsible for real-time data exchange with a land control center. The communication unit (800) contributes to enhancing the safety and efficiency of autonomous navigation by enabling close interaction between land and ships, such as remote monitoring, control, and support, through smooth and stable data transmission and reception.

[0177] The communication unit (800) includes the following main components:

[0178] The multiple communication channels (801) support various communication channels to select the optimal communication means according to the communication environment and data characteristics. Mobile communication technologies such as LTEM and 5G are used for short distances, and satellite communication (VSAT, Iridium, etc.) is used for long distances. It also includes maritime communication channels such as VHF and AIS for direct communication between ships.

[0179] The communication protocol stack (802) supports protocols optimized for each communication channel to ensure efficient and reliable data transmission. It enables stable communication even in environments with limited bandwidth and power by utilizing not only general-purpose protocols such as TCP / IP and UDP, but also lightweight IoT protocols such as MQTT and CoAP.

[0180] The data compression and optimization module (803) minimizes the size of the transmitted data to efficiently use limited communication resources. It reduces the amount of data by adjusting the sampling rate of sensor data or by applying lossless compression algorithms such as delta encoding and Huffman coding. It also provides a QoS management function that dynamically adjusts the transmission priority according to the importance of the data.

[0181] The security and authentication module (804) provides security functions to ensure the confidentiality, integrity, and availability of data transmitted through a communication channel. It ensures robustness against cyber attacks by applying various encryption technologies, such as data encryption, digital signatures, and certificate-based mutual authentication. Additionally, it can ensure the imperviousness and transparency of data by utilizing blockchain technology.

[0182] The network management module (805) provides the function of monitoring and diagnosing the status of the communication channel in real time. It automatically performs network topology search, link quality measurement, and fault detection, and increases communication stability by dynamically switching channels or reconfiguring relay paths when necessary. In addition, it can flexibly control network resources by introducing SoftwareDefined Networking (SDN) technology.

[0183] The application interface (806) provides a standardized interface to support various application services between land and ships. Through RESTful APIs, gRPC, etc., it is possible to call various functions such as data transmission, remote control, software updates, and log downloads. In addition, interoperability between heterogeneous systems is ensured by applying a common data model that defines data formats and semantics.

[0184] The communication unit (800) follows the following operation procedure:

[0185] Data collected from various sensors and equipment is normalized and packaged according to a common data model.

[0186] Select the optimal communication method by considering the characteristics of the data and channel conditions, and encode the data using the corresponding protocol.

[0187] Data is encrypted in accordance with security policies, and after verifying integrity, it is transmitted to land.

[0188] Control commands, route information, etc. received on land are transmitted back to the vessel.

[0189] Periodically diagnose the status of the communication channel and take appropriate recovery measures if abnormal signs are detected.

[0190] By ensuring the communication unit (800) operates stably, the ship can always perform autonomous navigation based on the latest information. In particular, rapid dissemination of information and response are possible in the event of an emergency, which can significantly contribute to accident prevention. Furthermore, as a vast amount of navigation data is accumulated, it is expected that a foundation will be laid for the advancement of navigation technology through big data analysis, machine learning, etc.

[0191] Finally, Example 8 relates to a condition monitoring unit (1000) that monitors the condition of ship equipment in real time, and a predictive maintenance unit (1100) that performs predictive maintenance based on this. These two modules contribute to securing safety and economic efficiency of autonomous operation by maximizing the availability and reliability of ship equipment.

[0192] The status monitoring unit (1000) includes the following components:

[0193] The sensor network (1001) consists of various sensors attached to major equipment such as engines, generators, and pumps. It measures physical quantities indicating the condition of the equipment, such as temperature, pressure, vibration, and current, and transmits data through a wired or wireless network. In the case of analog sensors, the data is converted into digital data through appropriate signal conversion and filtering.

[0194] The data collection and preprocessing module (1002) processes raw data collected from the sensor and converts it into a form suitable for analysis. It performs data cleaning such as noise removal, outlier detection, and missing value interpolation, and adjusts the temporal and spatial resolution of the data through downsampling, smoothing, etc., if necessary. It also normalizes the data or performs feature extraction to increase the efficiency of subsequent analysis steps.

[0195] The anomaly detection module (1003) analyzes collected data to identify signs of abnormality in the equipment in real time. It detects abnormal patterns that deviate from the normal state using statistical methods (e.g., 3sigma rule, Mahalanobis distance, etc.) or machine learning-based methods (e.g., OneClass SVM, Autoencoders, etc.). The detection results are used for generating warning messages, updating status indicators, etc.

[0196] The condition diagnosis module (1004) performs the role of identifying the cause of an anomaly and determining its severity when an anomaly is detected. It secures a high level of diagnostic accuracy by appropriately combining methods utilizing expert knowledge, such as rule-based reasoning, decision trees, and Bayesian networks, with deep learning-based methods such as CNN and LSTM. The diagnosis results include failure modes, severity grades, and supporting information.

[0197] The state prediction module (1005) predicts future state changes based on the current state of the equipment. It estimates the remaining useful life (RUL), failure probability, etc. using time series prediction models (e.g., ARIMA, Prophet, etc.) and survival analysis techniques (e.g., Cox PH model, Weibull AFT model, etc.). This is used as key input information for decision-making by the predictive maintenance unit (1100).

[0198] The HMI module (1006) provides an interface for effectively delivering status monitoring results to the operator. It enables intuitive information delivery through various forms of visualization tools, such as dashboards, alarms, and reports. It also supports interactions such as the operator's information search and analysis requests, thereby enabling active response to situations.

[0199] The operation procedure of the status monitoring unit (1000) is as follows:

[0200] Status data of key equipment is collected in real time through a sensor network (1001).

[0201] The collected data undergoes quality control and processing in the data collection and preprocessing module (1002).

[0202] For the processed data, an anomaly detection module (1003) examines abnormal patterns, and for the detected anomalies, a state diagnosis module (1004) performs cause identification and severity determination.

[0203] The state prediction module (1005) predicts the future state based on the pattern of state change up to the present.

[0204] The HMI module (1006) processes the status monitoring results into an appropriate form so that they can be used for the user's decision-making.

[0205] The role of the predictive maintenance department (1100) is to perform preemptive maintenance activities using this condition monitoring information. The specific components are as follows.

[0206] The maintenance policy optimization module (1101) establishes an optimal maintenance policy by considering constraints such as condition monitoring results, operational plans, and maintenance resources. It primarily applies Condition-Based Maintenance (CBM) and Predictive Maintenance (PdM) policies, and determines the maintenance priority and timing by comprehensively considering maintenance effects, costs, and risks. To this end, reinforcement learning-based techniques such as POMDP and Multi-Armed Bandit, or heuristic optimization techniques such as PSO and GA, are utilized.

[0207] The maintenance work planning and scheduling module (1102) establishes a specific maintenance work plan and coordinates the schedule according to the maintenance policy. It details the work content by considering the equipment to be maintained, the type of maintenance (prevention / breakdown / improvement), necessary parts and personnel, and the time required, and generates an optimal schedule that satisfies the order of events and resource constraints. Various scheduling techniques such as PERT / CPM, genetic algorithms, and reinforcement learning may be applied.

[0208] The maintenance work management module (1103) manages the execution of planned maintenance work and monitors progress. It supports all necessary aspects of maintenance activities, such as issuing work orders, manpower allocation, and material management, and evaluates the level of implementation compared to the plan by collecting and analyzing information from the work site. In addition, it ensures maintenance quality by promptly identifying and responding to issues that arise during the work process.

[0209] The maintenance performance analysis module (1104) analyzes and evaluates the performance of completed maintenance work from various angles. It calculates key performance indicators such as work time, input costs, and equipment utilization rates, and derives areas for improvement in the maintenance process by comparing and benchmarking with prior plans and similar cases. In addition, it verifies and optimizes the effectiveness of the maintenance policy by analyzing failure trends and life extension effects.

[0210] The maintenance knowledge management module (1105) systematically manages know-how and lessons learned through maintenance activities. It databases maintenance history, defect cases, and best practices, and increases their utility value through searching for similar cases and statistical analysis. In addition, it aims to standardize and improve the efficiency of maintenance work by digitizing maintenance manuals and technical documents and continuously updating them.

[0211] The predictive maintenance unit (1100) operates in the following procedure:

[0212] Based on the status diagnosis / prediction information transmitted from the status monitoring unit (1000), the maintenance policy optimization module (1101) establishes short-term and long-term maintenance plans.

[0213] The maintenance work planning and scheduling module (1102) plans detailed maintenance work and coordinates the schedule.

[0214] Maintenance work is performed on-site under the control of the maintenance work management module (1103), and work progress information is monitored / feedback in real time.

[0215] After the maintenance project is completed, the maintenance performance analysis module (1104) performs a comprehensive performance evaluation and derives improvement plans.

[0216] The maintenance knowledge management module (1105) manages and utilizes accumulated maintenance data and knowledge.

[0217] Through organic collaboration between the condition monitoring unit (1000) and the predictive maintenance unit (1100), the ship equipment can always be maintained in the best condition. This directly leads to the enhancement of the autonomous ship's ability to achieve its mission, and furthermore, is expected to lead to innovative achievements such as the prevention of marine accidents and the reduction of ship operating costs.

Claims

1. In an integrated control system for safe navigation and carbon emission reduction of autonomous vessels, A multi-sensor unit (100) for collecting environmental information around the ship; A collision risk analysis unit (200) that analyzes data collected from the above multi-sensor unit (100) to evaluate the risk of collision with surrounding vessels and obstacles; A path optimization unit (300) that calculates the optimal path by considering the current location, destination, and weather information; A ship control unit (400) that directly controls the ship's engine, rudder, and propulsion system; A carbon emission monitoring unit (500) that calculates and monitors the real-time carbon emissions of a ship; An integrated AI control unit (600) that establishes an optimal control strategy by comprehensively analyzing data from the above-mentioned collision risk analysis unit (200), path optimization unit (300), ship control unit (400), and carbon emission monitoring unit (500); A central control unit (700) that coordinates and controls each subsystem according to the command of the integrated AI control unit (600) above; A communication unit (800) responsible for real-time data transmission and reception with a ground control center; A digital twin unit (900) that creates a virtual model of an actual ship and performs a simulation; A status monitoring unit (1000) that monitors the status of the ship's main equipment in real time; and An integrated control system for safe navigation and carbon emission reduction of an autonomous vessel, characterized by including a predictive maintenance unit (1100) that analyzes equipment status data to predict failures and optimizes maintenance schedules.

2. In Claim 1, The above multi-sensor unit (100) includes a radar, lidar, AIS receiver, camera, GPS, and weather sensor, and The above collision risk analysis unit (200) includes an object detection and tracking module, a CPA / TCPA calculation module, and a collision risk evaluation module, and The above-mentioned path optimization unit (300) includes a nautical chart and route information database, a weather and ocean current information collection module, and a multi-purpose optimization algorithm. The above-mentioned ship control unit (400) includes an engine control module, a rudder control module, and a propulsion system control module, and The above carbon emission monitoring unit (500) includes an engine performance data collection module, a carbon emission calculation module, and an emission visualization module, and The above integrated AI control unit (600) includes a data preprocessing module, a deep learning-based situation recognition module, a reinforcement learning-based decision module, and a control command generation module, and The above communication unit (800) includes an LTEM module, a satellite communication module, and a data encryption and security module, and The above digital twin unit (900) includes a 3D ship modeling module, a real-time data synchronization module, a simulation engine, and a result analysis and feedback module, and The above-mentioned state monitoring unit (1000) includes a sensor data collection module, a data analysis and diagnosis module, and The above predictive maintenance unit (1100) includes a failure prediction algorithm and a maintenance schedule optimization module, and The integrated AI control unit (600) establishes a control strategy that simultaneously optimizes the safety, efficiency, and environmental friendliness of a ship by comprehensively considering the collision risk assessment result of the collision risk analysis unit (200), the optimal path calculation result of the path optimization unit (300), the carbon emission data of the carbon emission monitoring unit (500), the equipment status information of the status monitoring unit (1000), and the maintenance schedule information of the predictive maintenance unit (1100). The digital twin unit (900) simulates the control strategy established by the integrated AI control unit (600) in a virtual environment to verify its safety and efficiency, and feeds the verification results back to the integrated AI control unit (600) to continuously improve the control strategy. The central control unit (700) controls the ship's engine, rudder, and propulsion system through the ship control unit (400) according to the control strategy established by the integrated AI control unit (600). The communication unit (800) transmits ship operation data, carbon emission status, and equipment status information to the land control center in real time, receives updated weather information, traffic information, and emergency control commands from the land control center, and transmits them to the integrated AI control unit (600). An integrated control system for safe navigation and carbon emission reduction of an autonomous vessel, characterized in that the integrated AI control unit (600) stores control strategies including engine speed control, optimal propulsion setting, and route adjustment to minimize fuel consumption of the vessel based on optimal route information provided by the route optimization unit (300).

3. In Claim 1, The digital twin unit (900) performs a simulation to predict changes in the state and potential for failure of the ship's main equipment, analyzes the expected lifespan and potential for failure of the equipment to optimize the maintenance plan based on the predicted results, and analyzes the condition and expected usage time of the equipment to optimize the maintenance timing to efficiently adjust the maintenance timing, thereby creating an integrated control system for safe navigation and carbon emission reduction of an autonomous ship.

4. In Claim 1, The above central control unit (700) detects the risk of collision between ships, obstacles, and worsening weather conditions during the navigation of ships. An integrated control system for safe navigation and carbon emission reduction of an autonomous vessel, characterized in that, in the event of an emergency situation including the above, a route change process is executed in accordance with an emergency control command received from a land control center through the communication unit (800) to immediately change the vessel's route, and at this time, the route change process analyzes the vessel's current location, surrounding vessels, and obstacle information to select a safe route.

5. An integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, A step of collecting data through a multi-sensor unit (100) for collecting environmental information around a ship; A step of analyzing collision risk through a collision risk analysis unit (200) that analyzes data collected from the multi-sensor unit (100) and evaluates the risk of collision with surrounding vessels and obstacles; A step of calculating an optimal path through a path optimization unit (300) that calculates an optimal path by considering current location, destination, and weather information; A step of executing control commands through a ship control unit (400) to control the ship's engine, rudder, and propulsion system; A step of monitoring carbon emissions through a carbon emission monitoring unit (500) that calculates and monitors the real-time carbon emissions of a ship; A step of establishing a control strategy through an integrated AI control unit (600) that establishes an optimal control strategy by comprehensively analyzing data from the above-mentioned collision risk analysis unit (200), path optimization unit (300), ship control unit (400), and carbon emission monitoring unit (500); A step of controlling the entire system through a central control unit (700) that coordinates and controls each subsystem according to the command of the integrated AI control unit (600); A step of performing a simulation through a digital twin unit (900) that creates a virtual model of an actual ship and performs the simulation, and providing feedback on the results; A step of monitoring the equipment status through a status monitoring unit (1000) that monitors the status of the ship's main equipment in real time; and An integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, characterized by including a step of establishing a maintenance plan through a predictive maintenance unit (1100) that analyzes the status data of the equipment to predict failures and optimizes the maintenance schedule.

6. In Claim 5, An integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, characterized in that the above-mentioned path optimization unit (300) considers weather changes, ocean current direction and speed, and fuel consumption efficiency of the vessel along the path to additionally calculate an optimal fuel-saving path by reflecting weather information and ocean current information.

7. In Claim 5, An integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, characterized in that the digital twin unit (900) generates a virtual model including a digital model that simulates the operation of all major equipment of the vessel in 3D, predicts changes in the state of the equipment through the virtual model to evaluate the need for maintenance, and evaluates the need for maintenance based on expected usage time, failure history, and simulation results.

8. In Claim 5, An integrated control method for safe navigation and carbon emission reduction of an autonomous vessel, characterized in that the central control unit (700) executes a route change process in accordance with an emergency control command received from a land control center via a communication unit (800) when an emergency occurs, thereby changing the route to an optimal path by considering the vessel's current location, destination, surrounding obstacles, and weather conditions.

Images