AI-POWERED AUTONOMOUS NAVIGATION AND OBSTACLE DETECTION METHOD FOR UNMANNED MARINE VEHICLES

Abstract

This invention relates to an AI-powered autonomous navigation and obstacle detection method for unmanned marine vehicles that can be used in industrial fields such as maritime, defense, environmental and disaster management, and its features include: at least one data structures and algorithm knowledge module structured for data processing processes (1), at least one programming module that enables the execution, management and integration of AI algorithms (2), at least one data acquisition module that obtains environmental data from sensors, datasets or real-time data monitoring (3), a data preparation module that applies data cleaning, data normalization, noise reduction or Kalman filter operations on data obtained from sensors (4), at least one model creation module that creates an autonomous decision model using AI and / or machine learning algorithms (5), and at least one training initiation module that automatically starts the training process using datasets on the created model (6).The system must consist of at least one detection module (7) which analyzes the model's accuracy, error rate, or performance level after training; at least one evaluation module (8) which performs a performance evaluation of the model outputs after training; at least one graph module (9) which generates the success, accuracy, and error data during the training process graphically; at least one difference module (10) which performs a model performance analysis by calculating the difference between the actual data and the data predicted by the model; and at least one error margin module (11) which calculates the percentage error margin of the system.

Description

1 TARIFF AI-POWERED AUTONOMOUS FOR UNMANNED MARINE VEHICLES NAVIGATION AND OBSTACLE DETECTION METHOD Technological Field: This invention has applications in industrial fields such as maritime, defense, environmental, and disaster management. AI-powered autonomous navigation for usable unmanned marine vehicles and It is related to the obstacle detection method. 10 State of the Art: Unmanned marine vehicles (UMVs) generally operate based on remote control principles. and appear as systems with limited automation capabilities. There are 15 such systems. Vehicle navigation, mission planning, and environmental aspects in current systems Evaluation processes are largely performed by human operators. These structures, used within the scope of known techniques, process sensor data directly. lacking the ability to make independent decisions and requiring constant intervention from the operator. He hears. 20 When the operating principles of existing systems are examined, the sensors located on the vehicle The data obtained through this process is generally transmitted to the operator for decision-making. It appears that the mechanism has been left to human control. This situation, especially System performance in complex mission scenarios and varying environmental conditions is 25 It is limiting. Dynamic obstacle detection, autonomous route planning and real-time Advanced functions like adaptation are not sufficiently developed in known techniques. This Therefore, existing systems need to respond quickly and effectively to changing environmental parameters. is unable to provide. However, one of the most significant drawbacks of current systems is the human element. They are highly dependent on intervention. Operator-related delays, 2 Incorrect directions and communication breakdowns, especially in challenging sea conditions. This can seriously jeopardize operational safety. Furthermore, these systems can affect the performance of tasks. It is failing to ensure continuity and efficiency in long-term operations. This leads to losses. These limitations, which are due to the human factor, increase costs, This leads to negative consequences such as wasted time and increased operational risks. 5 On the other hand, most of the DDAs used in known techniques are specific to a particular application area. It is specifically designed and has limited adaptability to different sectors. It offers flexibility. This makes the multidisciplinary use of systems difficult. and reveals the need for restructuring in different mission scenarios. It also involves collecting and analyzing environmental data and making decisions based on this data. 10 Since production processes are not considered within an integrated structure, the efficiency of the systems... and its accuracy remains limited. In conclusion, current systems suffer from insufficient autonomous movement capabilities, and in reality... underdeveloped timely decision-making mechanisms, adaptation to environmental conditions 15 inability to provide, perceive dynamic obstacles and respond effectively to them It has significant shortcomings. These shortcomings result in operations being slower, more... This leads to it being carried out in a costly and higher-risk manner. This highlights the need to improve existing techniques. Description of the invention: This invention is an unmanned marine system that can overcome the disadvantages mentioned above. It is an AI-powered autonomous navigation and obstacle detection method for vehicles, Its feature is real-time decision-making that enhances autonomous movement capabilities. mechanisms that develop, adapt to environmental conditions, and overcome dynamic obstacles. Detecting and effectively responding to these detections makes operations faster and more cost-effective. It is a technology that enables the process to be carried out with a low and reduced risk. The invention has a multidisciplinary structure and is flexible enough to be used in different sectors. It has expertise primarily in maritime affairs, environmental engineering, defense industry, and disaster management. and can be used in sectors such as fishing. Marine pollution for environmental analysis. 3 AI-powered applications in areas such as measurements, water quality assessment, and ecosystem monitoring. Disaster response systems (DRSs) are capable of collecting and analyzing environmental data. They are used in post-disaster search and rescue operations. In their operations, they work in areas where human access is risky, carrying out imaging and It is capable of transmitting data. In military applications, it performs reconnaissance and surveillance tasks. It can be carried out without human intervention. It is also used in academic research and autonomous 5 It can also be used in system development projects. From a technical disciplines perspective, artificial intelligence, machine learning, robotics, marine technologies, electronics, mechatronics and This invention has direct applications in fields such as software engineering. The aim of the invention is to create a vehicle that can move autonomously on the sea surface, monitoring environmental data. an AI-powered unmanned marine capable of making real-time decisions by analyzing data. They have developed the tool. Existing systems generally operate with manual control and are limited. While offering automation, this invention enables tasks to be performed without the need for human intervention. with its features of route planning, obstacle avoidance and adaptation to environmental conditions It stands out. The system, which offers advantages in terms of time, cost and security, completes tasks in 15 by delivering it faster, at lower cost and without risk compared to existing solutions. It offers a significant innovation. This invention develops an artificial intelligence-powered autonomous unmanned marine vehicle (UMV) system. It includes. The developed tool analyzes environmental data on the sea surface and 20 capable of independently determining routes, avoiding obstacles, and working in a task-oriented manner. It is integrated with an artificial intelligence module. Especially dangerous or difficult to access. in areas (for example, in contaminated waters, storm aftermath areas or disaster zones) Manual control carries a high risk. Therefore, a safer, more efficient and A sustainable method of marine exploration and surveillance is required. The invention's core 25 Its aim is to solve these problems using a system that works with machine learning algorithms. to minimize and create a sea capable of performing tasks without human intervention. to bring the vehicle to life. The invention reduces human intervention by moving autonomously with artificial intelligence, saving costs and 30 It reduces travel time. Its unique feature is that it performs real-time data analysis, identifies obstacles, and determines the route. 4 It is an optimization. Its ability to work in challenging areas without human intervention is important. It is an advantage. The invention is easy to assemble thanks to the fact that its components can be easily fastened together. It is being installed, and thanks to the short assembly time, the costs are low. 5 Furthermore, the invention has a robust structure. Explaining the Figures: The invention will be described by referring to the attached figures, so that the features of the invention are outlined in 10 It will be understood and appreciated more clearly, but the purpose of this invention is this obvious It is not about limiting it with regulations. On the contrary, the invention is defined by the accompanying claims. all alternatives, modifications, and options that could be included within the defined area The aim is to cover their equivalences. The details shown are only for the present invention. It was shown to illustrate the preferred arrangements and both methods are 15 shaping, as well as the rules and conceptual features of the invention, in the most useful way. It should be understood that they are presented to provide a readily understandable definition. This in the drawings; Figure 1 shows a schematic view of the system. 20 Illustrations that will help understand this invention are shown in the attached image. They are numbered and their names are given below. References Explanation: 25 1. Data Structures and Algorithm Knowledge Module 2. Programming Module 3. Data Collection (Data Set or Data Tracking) 4. Data Preparation Module 30 5. Model Creation Module 6. Training Initiation Module 7. Detection Module 8. Evaluation Module 9. Graphics Module 10. Difference Module 11. Error Margin Module 5 Description of the Invention: The invention provides at least one structured data structure and algorithm for data processing processes. Information module (1), running, managing and 10 artificial intelligence algorithms at least one programming module (2) that enables integration, from sensors, data at least one data source that obtains environmental data from datasets or real-time data monitoring. data collection module (3), data cleaning on data obtained from sensors, data data that has undergone normalization, noise reduction, or Kalman filter processing. preparation module (4), using artificial intelligence and / or machine learning algorithms 15 At least one model creation module (5) that creates an autonomous decision model, created at least one model that automatically initiates the training process using datasets. The training initiation module (6) checks the model accuracy, error rate or after training. At least one detection module (7) that analyzes the performance level, model after training at least one assessment that evaluates the performance of its outputs 20 module (8) graphically displays success, accuracy and error data in the training process. at least one graphical module (9) formed by the model with real data and estimated by the model At least one person performs model performance analysis by calculating the difference between the data. difference modulus (10), at least one error margin that calculates the percentage error margin of the system. It consists of module (11). 25 The invention utilizes artificial neural networks, deep learning algorithms, and machine learning in decision-making processes. Data structures that run learning algorithms or optimization algorithms and It includes the algorithm information module (1). The invention involves processing data from sensors and managing an artificial intelligence model. and includes the programming module (2) which performs autonomous control operations. 6 The invention collects real-time data from LiDAR, sonar, radar, camera, and GPS sensors. It includes the data collection module (3). The invention relates to processes such as object detection, environmental analysis, route planning, or obstacle detection. It includes the model creation module (5) which creates an artificial intelligence model for the purpose. 5 Faster R-CNN, YOLOv3, deep learning or image processing based algorithms It uses. The invention demonstrates the environmental suitability and operational success of the model outputs. It includes the evaluation module (8) which evaluates the system's percentage error. calculates the ratio and determines the error margin module (11) which determines the system reliability. It includes. The invention enables autonomous navigation of unmanned marine vehicles by analyzing environmental data. obstacle avoidance, target identification, border security, environmental monitoring or maritime surveillance 15 They perform their tasks without human intervention. The invention enables multi-decision making that generates alternative decision outcomes based on different operational scenarios. It includes the mechanism. The invention involves capturing, digitizing, processing, and... environmental images. It includes an image processing-based analysis structure that enables improvement. The invention is used for coastal monitoring, environmental analysis, sea surface change detection, or orthophoto. It performs image analysis. 25 The invention enables remote control, task management, or operational guidance from land. It includes the remote access infrastructure that enables this. The invention integrates with Geographic Information Systems (GIS) data to perform spatial analysis and 30 It performs route optimization. 7 The invention aims to improve operational safety by mitigating traffic congestion, environmental risks, or It generates safe navigation output by analyzing mission scenarios. Detailed Description of the Invention: The invention involves using a server system that includes at least one processor and one memory unit. working on AI-powered autonomous navigation and environmental data in unmanned marine vehicles. analysis, obstacle detection and decision-making processes It is structured as a method. The server system processes the artificial intelligence algorithms used within the system. the operation of processes, model building procedures and decision-making mechanisms This is provided by the processor located within the server system, which processes the data. It performs the execution of algorithms. Located within the server system. The memory unit stores data sets, model parameters, analysis results, and 15 It enables the storage of operational data. Within the scope of the invention, the process was first implemented with the data structures and algorithm knowledge module (1). The Data Structures and Algorithm Knowledge Module (1) is initiated by the system. Processing, classifying, sorting, and using artificial intelligence for the data to be used. 20 the necessary algorithmic infrastructure for management in accordance with its algorithms It provides. Then the programming module (2) is activated. Programming module (2), artificial intelligence, machine learning, data processing and autonomous 25 running on the server system the execution, management, and integration of control algorithms It enables the transaction. Data collection module (3) after the programming infrastructure is created It is operated. Data acquisition module (3), LiDAR located on the unmanned marine vehicle, 30 It collects environmental data from sensors such as sonar, radar, cameras, and GPS. 8 Data collection module (3) can also collect data from datasets or through real-time data monitoring. It transfers the collected information to the server system. The collected data is then sent to the data preparation module (4). preparation module (4), raw data obtained from sensors or data sets 5 performing data cleaning, data normalization, noise reduction, and filtering operations on it. It performs. The data preparation module (4) is responsible for the sea conditions. by organizing noisy, incomplete, or inconsistent data, the artificial intelligence model It creates a suitable data structure that it can use. The prepared data is transferred to the model creation module (5). Model creation module (5), artificial intelligence and machine learning running on server system using algorithms to enable the unmanned underwater vehicle to sense environmental conditions and objects its ability to recognize obstacles, plan routes, and make autonomous decisions. It creates the model that enables its production. 15 Model creation module (5) can create vehicles, obstacles on the sea surface or underwater, It creates an artificial intelligence structure for detecting targets and dangerous areas. It brings. Model creation module (5), image processing, deep learning, object Environmental analysis processes using sensing and route planning algorithms 20 It is carrying out. After the model is created, the training start module (6) is activated. Training The startup module (6) launches the generated artificial intelligence model using datasets or real-time It starts training based on the data collected. The training start module (6), 25 The model's ability to recognize environmental data, distinguish obstacles, and create a safe route and helps improve decision-making skills. After the training process is completed, the detection module (7) is operational. The detection module (7), After training, the model's error rate, accuracy value, and performance level were analyzed. It analyzes the extent to which the model responds to environmental data. The detection module (7) analyzes the extent to which the model responds to environmental data. It determines that it produces the correct result. 9 After the detection process, the evaluation module (8) is activated. Evaluation module (8) evaluates the outputs produced by the model of the unmanned marine vehicle. suitability to job requirements, environmental conditions and operational needs The evaluation module (8) evaluates the model's autonomous navigation, obstacle avoidance. It analyzes avoidance, target setting, and safe movement capabilities. 5 Success, accuracy, and error data obtained during and after the training process. The graphics module (9) processes the model's training. graphically displaying success rates, error variations, and accuracy levels throughout the process. It creates the system performance visually. The graphics module (9) 10 It enables monitoring. The prediction results produced by the model are analyzed by the difference modulus (10). Difference module (10) distinguishes between the actual data and the data predicted by the model. It calculates the difference modulus (10), the model's 15 based on the calculated difference values. It determines the prediction success and performance level. The results obtained by the difference module (10) are converted to the margin of error module (11) It is transferred. The error margin module (11) calculates the percentage error rate of the system. The margin of error module (11) sets the model's reliability level to 20 based on the calculated error rate. It determines. The system transmits environmental data via a server system that includes a processor and memory unit. By processing the data, the system analyzes the environment of the unmanned marine vehicle. Based on the data, it detects obstacles, creates safe routes, and performs navigation operations. It performs and makes autonomous decisions without human intervention. 25 Thanks to this design, the unmanned marine vehicle can operate safely, efficiently, and independently in the marine environment. It works in this way.

Claims

REQUESTS 1- The invention describes a server system that includes at least one processor and one memory unit. AI-powered autonomous navigation and obstacle detection for unmanned marine vehicles. It is related to the method and its characteristic is; 5  At least one structured data structure and algorithm for data processing processes information module (1),  Running, managing, and integrating artificial intelligence algorithms at least one programming module (2) that provides  Environmental 10 from sensors, datasets or real-time data monitoring at least one data collection module that obtains data (3),  Data cleaning on data obtained from sensors, data normalization, noise reduction or Kalman filter processes data preparation module (4),  Autonomous decision-making using artificial intelligence and / or machine learning algorithms 15 At least one model creation module (5) that creates the model  Automating the training process using datasets on the created model at least one training initiation module (6) that starts as  Model accuracy, error rate, or performance after training at least one detection module (7) that analyzes the level, 20  Performance evaluation of model outputs after training at least one evaluation module (8) that is carried out  Graphically displaying success, accuracy, and error data from the training process. at least one graphics module (9), which constitutes  The difference between the actual data and the data predicted by the model is 25 at least one difference that performs model performance analysis by calculating module (10),  at least one error margin module that calculates the percentage error margin of the system (11) is the formation. 2- AI-powered autonomous operation for unmanned marine vehicles mentioned in Claim 1. It is a navigation and obstacle detection method whose characteristic is that it uses artificial intelligence in decision-making processes. 11 neural networks, deep learning algorithms, machine learning algorithms, or data structures and algorithm knowledge module that runs optimization algorithms (1) It is characterized by its inclusion. 3- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its feature is that it uses data obtained from sensors. processing, managing the artificial intelligence model and autonomous control operations It is characterized by containing the programming module (2) which performs the programming. 4- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its features include LiDAR, sonar, radar, and camera. It includes a data acquisition module (3) that collects real-time data from GPS sensors. It is characterized by... 5- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its features include object detection and environmental analysis. A model that creates an artificial intelligence model for route planning or obstacle detection processes. It is characterized by containing the creation module (5). 6- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, featuring Faster R-CNN, YOLOv3, and deep navigation. model building module using learning or image processing-based algorithms (5) is characterized by its inclusion. 7- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, characterized by its ability to integrate model outputs with the environment. an evaluation assessing compliance with conditions and operational success It is characterized by containing module (8). 8- AI-powered autonomous 30 for unmanned marine vehicles mentioned in Claim 1. It is a navigation and obstacle detection method, and its feature is that it reduces the system's percentage error rate. 12 It includes the error margin module (11) which calculates and determines the system reliability. It is the characterization of the situation. 9- AI-powered autonomous operation for unmanned marine vehicles mentioned in Claim 1. It is a navigation and obstacle detection method, and its feature is that it analyzes environmental data. 5 Autonomous navigation, obstacle avoidance, target acquisition, and boundary demarcation of the unmanned marine vehicle. security, environmental monitoring or marine surveillance tasks without human intervention It is characterized by its accomplishment. 10- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its feature is that it is suitable for different operational scenarios. It is characterized by its inclusion of a multiple decision-making mechanism that generates alternative decision outcomes. It is done. 11- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its feature is the capture of environmental images. image processing-based systems that perform digitization, processing, and enhancement. It is characterized by its inclusion of an analytical structure. 12- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, and its features include coastal area monitoring and environmental monitoring. analysis, sea surface change detection, or orthophoto image analysis. It is the characterization of the situation. 13- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, its feature is remote control from land, mission. remote access infrastructure that performs management or operational guidance It is characterized by its inclusion. 14- AI-powered autonomous 30 for unmanned marine vehicles mentioned in Claim 1. It is a navigation and obstacle detection method, and its characteristic feature is Geographic Information Systems (GIS). 13 by working in integration with data to perform spatial analysis and route optimization. It is the characterization of the situation. 15- AI-powered autonomous systems for unmanned marine vessels mentioned in Claim 1. It is a navigation and obstacle detection method, the feature of which is to increase operational safety. 5 In order to ensure safety, we analyze traffic density, environmental risks, or mission scenarios. It is characterized by its ability to generate navigation output. 15 25

Images