Adaptive machine learning-based system and method for monitoring marine observation device having abnormality detection function

Abstract

The present invention generally relates to a marine observation device and a sensor device and, more particularly, to a method for monitoring, predicting, and diagnosing the performance and abnormal signs of a sensor device in a marine observation device by using an advanced data analysis technique, such as a machine learning algorithm, and integrating a power circulation mechanism and an improved communication method with an external device, wherein a system for monitoring a marine observation device, according to an embodiment, may include: a microcurrent collection unit which collects microcurrents transmitted from a sensor to a measuring device; a monitoring unit that identifies an abnormal value among data points of the collected microcurrent by using an intelligent model; a diagnosis unit that generates device state information by determining deterioration or malfunction on the basis of a pattern of the identified abnormal value; and a communication unit for transmitting the generated device state information to an external system.

Description

Marine observation equipment monitoring system and method with adaptive machine learning-based anomaly detection function

[0001] The present invention relates generally to marine observation equipment and sensor devices, and more particularly to a method for monitoring, predicting and diagnosing the performance and anomalies of sensor devices in marine observation equipment by using advanced data analysis techniques such as machine learning algorithms and incorporating power cycling mechanisms and improved communication methods with external equipment.

[0002] Marine observation equipment can consist of various types of sensors attached to buoys.

[0003] These sensors measure the quality of ocean water through water quality sensors, monitoring temperature, salinity, and oxygen concentration. They also measure meteorological conditions in the marine environment through weather sensors, measuring wind speed, wind direction, atmospheric pressure, and precipitation. Furthermore, marine biology sensors investigate marine biodiversity, measuring the dynamics of algae, seaweed, and aquatic animals.

[0004] These sensors can be attached to buoys to measure and record ocean weather conditions, providing real-time weather information by measuring wave height, wind speed, and direction. They can also measure ocean currents, currents, and water temperature, and record current strength and direction, water temperature, and other data.

[0005] These marine observation devices are crucial for monitoring and preserving the health of the ocean and marine ecosystems. They detect environmental changes such as pollution, climate change, and sea level rise, and provide data for sustainable resource management.

[0006] Marine observation equipment can help predict and respond to natural disasters such as marine weather and tsunamis. Real-time data enables rapid and effective responses, contributing to safe navigation and coastal area management.

[0007] Additionally, ocean observation equipment can support climate change research by monitoring ocean temperature, salinity, and ocean currents. This will help increase understanding of global climate change and improve prediction models.

[0008] Additionally, it is used to assess and monitor the health of marine ecosystems. This can contribute to the sustainable management of fisheries and aquatic resources by investigating factors such as fish migration patterns and seaweed distribution.

[0009] Sensors used in marine observation equipment are subject to deterioration and failure, and technologies for monitoring the operating status of such equipment in real time face several challenges:

[0010] First, its high price makes it difficult for budget-constrained organizations or groups to purchase and deploy it for ocean observation purposes.

[0011] Additionally, many existing devices are designed for grid power, making them unsuitable for use in remote marine environments where alternative power sources such as solar or battery power are more common.

[0012] Existing devices also lacked communication capabilities. Typically, they only supported USB connections, making interfacing with external systems and devices difficult. This limited critical functions needed to manage and monitor equipment from remote locations.

[0013] Another issue is the lack of advanced diagnostic capabilities. Existing devices primarily focus on basic current and voltage measurements and lack advanced analysis or predictive capabilities, making them incapable of effectively identifying potential malfunctions or predicting performance degradation in marine observation equipment.

[0014] Furthermore, existing devices do not account for the unique challenges of the marine environment, such as corrosion, biological contamination, or changes in equipment performance due to extreme weather conditions. This may result in inaccurate representation of the actual operating conditions of marine observation equipment.

[0015] Against this backdrop, technological advancements are highlighting the need for cost-effective and adaptable monitoring systems.

[0016] The present invention is applicable to demanding marine environments where data reliability and continuity are essential, and aims to minimize the need for frequent maintenance visits.

[0017] The present invention provides advanced data analysis techniques, such as machine learning algorithms including recurrent neural networks or long short-term memory networks, to accurately predict the performance of a sensor device based on measured current values ​​and classified equipment status and to detect potential defects or abnormalities at an early stage.

[0018] The present invention aims to enable a more comprehensive understanding of the operating status of a sensor device by taking into account various equipment states (idle, power saving, normal operation, maximum operation, off) while measuring the supplied current value.

[0019] The present invention aims to simulate a reboot situation and ensure the safety and reliability of equipment.

[0020] The present invention aims to overcome the limitations of existing technologies that support limited communication methods such as USB.

[0021] The present invention aims to present a new standard for marine observation equipment by improving connectivity and compatibility with external systems or devices.

[0022] The present invention aims to provide a user-friendly interface that displays real-time monitoring data, predicted sensor device performance, and detected anomalies.

[0023] A marine observation equipment monitoring system according to one embodiment may include a microcurrent collection unit that collects microcurrents transmitted from a sensor to a measuring device, a monitoring unit that identifies abnormal values ​​for data points of the collected microcurrents using an intelligent model, a diagnostic unit that determines whether there is deterioration or failure based on a pattern of the identified abnormal values ​​and generates equipment status information, and a communication unit that transmits the generated equipment status information to an external system.

[0024] According to one embodiment, the monitoring unit can identify an outlier using a local outlier factor (LOF) for the collected microcurrent data points.

[0025] According to an embodiment, the monitoring unit may identify surrounding neighbors using an intelligent model for the data point, calculate a density ratio based on a distance from the identified surrounding neighbors, calculate a local outlier factor (LOF) based on the calculated density ratio, and identify an outlier if the calculated local outlier factor (LOF) is greater than a threshold value.

[0026] According to one embodiment, the monitoring unit may activate a new power cycling mechanism that temporarily interrupts and then reconnects the supplied voltage and current when the abnormal value is identified.

[0027] According to one embodiment, the monitoring unit can learn the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

[0028] According to one embodiment, the diagnostic unit can determine whether the device is deteriorated or broken by considering at least one of idle, power saving, normal, maximum operation, and off as the equipment status while measuring the microcurrent value and the pattern of the identified abnormal value.

[0029] According to one embodiment, the diagnostic unit can diagnose deterioration or failure of the sensor by using at least one of a pattern, a trend, and machine learning for the abnormal value.

[0030] According to one embodiment, the diagnostic unit can diagnose deterioration or failure of the sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0031] A marine observation equipment monitoring system according to an embodiment may include a sensor status analysis device that collects microcurrents transmitted from a sensor to a measuring device and transmits data points corresponding to the collected microcurrents to an external system, a sensor management server that collects data points of the microcurrents, identifies abnormal values, and determines whether there is deterioration or failure based on the pattern of the identified abnormal values.

[0032] According to one embodiment, the sensor management server can identify outliers using a local outlier factor (LOF) for the collected microcurrent data points.

[0033] The sensor management server according to one embodiment may identify surrounding neighbors using an intelligent model for the data point, calculate a density ratio based on the distance to the identified surrounding neighbors, calculate a local outlier factor (LOF) based on the calculated density ratio, and identify an outlier if the calculated local outlier factor (LOF) is greater than a threshold value.

[0034] According to one embodiment, the sensor management server may operate a new power cycling mechanism that temporarily interrupts and then reconnects the voltage and current supplied to the sensor status analysis device when the abnormal value is identified.

[0035] The sensor management server according to one embodiment can learn the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

[0036] The sensor management server according to one embodiment can determine whether there is deterioration or failure by considering at least one of idle, power saving, normal, maximum operation, and off as the equipment status while measuring the microcurrent value and the pattern of the identified abnormal value.

[0037] The sensor management server according to one embodiment can diagnose deterioration or failure of the sensor by using at least one of a pattern, trend, and machine learning for the abnormal value, or can diagnose deterioration or failure of the sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0038] An operating method of a marine observation equipment monitoring system according to an embodiment may include a step of collecting microcurrents transmitted from a sensor to a measuring device in a microcurrent collection unit, a step of identifying abnormal values ​​for data points of the collected microcurrents using an intelligent model in a monitoring unit, a step of determining whether there is deterioration or failure based on a pattern of the identified abnormal values ​​in a diagnosis unit and generating equipment status information, and a step of transmitting the generated equipment status information to an external system in a communication unit.

[0039] The step of identifying an outlier for the collected microcurrent data point using the intelligent model according to an embodiment may include the step of identifying a surrounding neighbor for the data point using the intelligent model, calculating a density ratio based on a distance from the identified surrounding neighbor, calculating a local outlier factor (LOF) based on the calculated density ratio, and identifying the data point as an outlier if the calculated local outlier factor (LOF) is greater than or equal to a threshold value.

[0040] The step of generating equipment status information by determining whether there is deterioration or failure based on the pattern of the identified abnormal value according to one embodiment may include a step of distinguishing and diagnosing deterioration or failure of the sensor by using at least one of a pattern, a trend, and machine learning for the abnormal value, and a step of distinguishing and diagnosing deterioration or failure of the sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0041] The method of operating a marine observation equipment monitoring system according to one embodiment may further include a step of learning the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

[0042] In one embodiment, by addressing the shortcomings of existing technologies and providing advanced and differentiated features, the monitoring, maintenance, and overall performance of marine observation equipment can be significantly improved, ultimately enhancing the quality and continuity of marine data.

[0043] In one embodiment, it provides a more affordable microcurrent sensing and measurement device, making it accessible to a wide range of organizations involved in ocean observation.

[0044] According to one embodiment, it is designed to be used with a variety of power sources, including batteries, solar power, and grid power, making it suitable for a variety of marine environments and power configurations.

[0045] According to one embodiment, it supports low-speed data methods such as RS-485 and RS-232, making it easy to integrate with external equipment and enabling remote monitoring and maintenance of marine observation equipment.

[0046] In one embodiment, machine learning algorithms such as the GRU model can be integrated to analyze current consumption data and equipment status to accurately diagnose potential malfunctions and predict equipment deterioration.

[0047] In one embodiment, it can provide more accurate monitoring and maintenance of marine observation equipment by taking into account the unique challenges of the marine environment such as corrosion, biofouling and extreme weather conditions.

[0048] In one embodiment, improved management and maintenance of marine observation equipment allows for more consistent and reliable data collection in remote and challenging environments.

[0049] In one embodiment, it is applicable to challenging marine environments where data reliability and continuity are essential, minimizing the need for frequent maintenance visits.

[0050] According to one embodiment, advanced data analysis techniques, such as machine learning algorithms including recurrent neural networks or long short-term memory networks, can be provided to accurately predict the performance of sensor devices based on measured current values ​​and classified equipment status, and to detect potential defects or abnormalities at an early stage.

[0051] According to one embodiment, the operating status of the sensor device can be more comprehensively determined by taking into account various equipment states (idle, power saving, normal operation, full operation, off) while measuring the supplied current value.

[0052] According to one embodiment, it can simulate reboot situations, ensure the safety and reliability of the equipment, overcome the limitations of existing technologies that support limited communication methods such as USB, and improve connectivity and compatibility with external systems or devices, thereby presenting a new standard for marine observation equipment.

[0053] Figure 1 is a drawing illustrating a marine observation equipment monitoring system according to an embodiment.

[0054] Figure 2 is a drawing illustrating a marine observation equipment monitoring system according to another embodiment.

[0055] Figure 3 is a drawing illustrating the overall operation of the marine observation equipment monitoring system.

[0056] Figure 4 is a diagram explaining the operation of the intelligent model.

[0057] Figure 5 is a diagram showing LOF, which depicts outlier scores for collected data points.

[0058] Specific structural or functional descriptions of embodiments according to the concept of the present invention disclosed in this specification are merely illustrative for the purpose of explaining embodiments according to the concept of the present invention, and embodiments according to the concept of the present invention may be implemented in various forms and are not limited to the embodiments described in this specification.

[0059] Embodiments according to the concept of the present invention may have various modifications and take various forms, and thus, embodiments are illustrated in the drawings and described in detail in this specification. However, this is not intended to limit embodiments according to the concept of the present invention to specific disclosed forms, but rather includes modifications, equivalents, or alternatives that fall within the spirit and technical scope of the present invention.

[0060] While terms such as "first" or "second" may be used to describe various components, these components should not be limited by these terms. These terms are intended solely to distinguish one component from another. For example, a first component may be referred to as a "second component," and similarly, a second component may also be referred to as a "first component," without departing from the scope of the invention.

[0061] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components in between. Conversely, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Expressions that describe relationships between components, such as "between," "immediately between," or "directly adjacent to," should be interpreted similarly.

[0062] The terminology used herein is for the purpose of describing specific embodiments only and is not intended to limit the present invention. The singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, it should be understood that the terms "comprises" or "has" are intended to specify the presence of a described feature, number, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0063] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0064]

[0065] Hereinafter, embodiments will be described in detail with reference to the attached drawings. However, the scope of the patent application is not limited or restricted by these embodiments. The same reference numerals in each drawing represent the same components.

[0066] FIG. 1 is a drawing illustrating a marine observation equipment monitoring system (100) according to one embodiment.

[0067] The marine observation equipment monitoring system (100) can diagnose deterioration or failure of a sensor based on the microcurrent transmitted from the sensor mounted on the buoy to the measuring device.

[0068] For this purpose, the system can be divided into a type that includes an analysis engine and a type in which the analysis engine is included in a remote server.

[0069] The marine observation equipment monitoring system (100) shows a form in which an analysis engine is included, and the marine observation equipment monitoring system (200) of FIG. 2 shows a form in which the analysis engine is included in a remote server.

[0070] First, a marine observation equipment monitoring system (100) according to an embodiment may include a microcurrent collection unit (110), a monitoring unit (120), a diagnosis unit (130), a communication unit (140), and a control unit (150).

[0071] A microcurrent collection unit (110) according to one embodiment can collect microcurrent transmitted from a sensor to a measuring device.

[0072] Devices that measure voltage may not be able to effectively identify potential malfunctions or predict performance degradation of marine observation equipment, leading to data loss or equipment failure.

[0073] The monitoring unit (120) according to one embodiment can identify abnormal values ​​for the collected microcurrent data points using an intelligent model.

[0074] Intelligent models, related to artificial intelligence or machine learning, are models that analyze and learn data to recognize and understand patterns. Examples of intelligent models include neural networks, decision trees, and ensemble models.

[0075] An outlier is a value that deviates from the normal pattern of a given data point, and identifying it can help predict or respond to abnormal behavior or failures in a system.

[0076] The monitoring unit (120) can identify outliers using the local outlier factor (LOF) for the collected microcurrent data points. In particular, the monitoring unit uses an intelligent model to identify surrounding neighbors for the data point, calculates a density ratio based on the distance to the identified surrounding neighbors, calculates a local outlier factor (LOF) based on the calculated density ratio, and if the calculated local outlier factor (LOF) is greater than a threshold value, the data point can be identified as an outlier.

[0077] The monitoring unit (120) identifies the surrounding neighbors of a data point. That is, it identifies the surrounding neighbors for each data point by utilizing an intelligent model. This intelligent model can accurately identify the surrounding neighbors by understanding the relationships between data points and determining connectivity.

[0078] The monitoring unit (120) calculates the density ratio based on the distance to surrounding neighbors by measuring the distance to the surrounding neighbors and calculating the surrounding density of the data point. This is necessary to quantify the cluster or density environment to which the data point belongs.

[0079] The monitoring unit (120) utilizes the calculated surrounding density to calculate the LOF, which is a value indicating how unusual a data point is compared to its neighbors. A high LOF value may indicate that the data point is relatively more unusual compared to other data points.

[0080] The monitoring unit (120) compares the calculated LOF value with a predefined threshold value. If the LOF value is greater than the threshold value, the corresponding data point can be identified as an outlier. Through this, the monitoring unit (120) can detect abnormal operation or problems in the system among the microcurrent data and take action accordingly.

[0081] Through this procedure, the monitoring unit (120) can detect abnormalities among data points and perform additional investigation and action on data determined to be abnormal values.

[0082] The monitoring unit (120) can drive a new power cycling mechanism that temporarily interrupts and then reconnects the supplied voltage and current when an abnormal value is identified.

[0083] The monitoring unit (120) according to one embodiment can learn the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

[0084] The monitoring unit (120) collects microcurrent data when the sensor is operating in a normal state, and this data reflects the microcurrent characteristics when the sensor is operating properly.

[0085] Meanwhile, the monitoring unit (120) collects microcurrent data when the sensor is operating in a faulty state, and this data represents the microcurrent pattern when a problem occurs in the sensor.

[0086] The monitoring unit (120) learns an intelligent model by utilizing the collected microcurrent data in normal and fault states. The intelligent model can identify patterns in the microcurrent data in these two states and learn the characteristics of each to acquire the ability to distinguish the sensor state.

[0087] More specifically, the monitoring unit (120) collects microcurrent data in normal and fault conditions and then appropriately preprocesses the data. Preprocessing includes steps such as matching the data size, removing noise, and extracting features as needed. Next, the monitoring unit (120) divides the collected data into training and validation data. The training data is used to train the model, and the validation data is used to evaluate the model's performance.

[0088] The monitoring unit (120) can select an appropriate intelligent model to learn and classify sensor status. For example, various models such as neural networks, decision trees, and random forests can be used.

[0089] Next, the monitoring unit (120) can define and set the structure of the selected model. This includes the structure of the input layer, hidden layer, and output layer, activation function, optimization algorithm, etc., and determines how the model will be trained.

[0090] The monitoring unit (120) can train a model using learning data. The model can determine the sensor status by learning microcurrent patterns in normal and fault states.

[0091] The monitoring unit (120) can use validation data to evaluate the model's performance. It can determine how accurately the model predicts new data and, if necessary, adjust the model to improve performance.

[0092] The monitoring unit (120) according to one embodiment can adjust or optimize hyperparameters to improve model performance. This can be implemented by adjusting model settings such as learning rate, batch size, number of hidden layers, etc.

[0093] A diagnostic unit (130) according to one embodiment can generate equipment status information by determining whether there is deterioration or failure based on the pattern of identified abnormal values.

[0094] According to one embodiment, the monitoring unit (120) analyzes the measured microcurrent data and transmits the pattern of identified abnormal values ​​to the diagnostic unit (130). The abnormal values ​​are the result of the model detecting microcurrents that deviate from the normal pattern, and the diagnostic unit (130) analyzes the pattern of the received abnormal values.

[0095] The diagnostic unit (130) determines the sensor's status as deteriorated or faulty based on the analyzed abnormal value pattern. This process interprets the characteristics indicated by the abnormal value and determines the sensor's current status accordingly. To this end, the diagnostic unit (130) can determine the current status by comparing it to a previously deteriorated or faulty microcurrent pattern.

[0096] The diagnostic unit (130) that determines whether a sensor is deteriorated or faulty generates equipment status information for the sensor. This information may include the sensor's current status, possible causes of failure, and recommended actions.

[0097] According to an embodiment, the diagnostic unit (130) can determine whether there is deterioration or failure by considering at least one of the patterns of identified abnormal values ​​and the equipment status while measuring the microcurrent value, namely idle, power saving, normal, maximum operation, and off. In addition, the deterioration or failure of the sensor can be distinguished and diagnosed by using at least one of the patterns, trends, and machine learning for the abnormal values, or the deterioration or failure of the sensor can be distinguished and diagnosed by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0098] According to one embodiment, a communication unit (140) can transmit the generated equipment status information to an external system.

[0099] Equipment status information transmitted to an external system can be displayed on an output device as needed or transmitted to a management system to notify users or maintenance personnel in real time.

[0100] Accordingly, maintenance notifications or automated actions can be triggered based on the generated equipment status information. For example, if a critical failure is detected, the equipment can be automatically disabled or a maintenance request can be generated.

[0101] The control unit (150) according to one embodiment can be interpreted as a central processing unit (CPU) and can perform various operations and process data within the system.

[0102] In particular, the control unit (150) can read commands from memory, interpret and execute the commands, and can also perform arithmetic operations such as addition, subtraction, multiplication, and division.

[0103] In addition, the control unit (150) can handle data storage and retrieval, and can also perform the function of reading data from memory and storing the results of performing operations back into memory.

[0104] In addition, the control unit (150) can manage the execution flow of the program, and in particular, can control the flow of the program using commands such as conditional statements (if-else) or iterative statements (for, while). In addition, the control unit (150) can have a small and fast memory device called a register placed inside, and this register can be used to temporarily store data or perform operations.

[0105] The control unit (150) can process and take appropriate action when an external event or exceptional situation occurs, and can quickly access data and instructions by using cache memory that is faster than the main memory.

[0106] In addition, the control unit (150) can use a system bus to communicate with memory or input / output devices, and can provide various power management functions to minimize power consumption.

[0107] FIG. 2 is a drawing illustrating a marine observation equipment monitoring system (200) according to another embodiment.

[0108] The present invention provides an advanced and differentiated method for monitoring, predicting and diagnosing the performance and abnormal signs of sensor devices of marine observation equipment.

[0109] Another example of a marine observation equipment monitoring system (200) is an example of managing an intelligent model through an external server.

[0110] A marine observation equipment monitoring system (200) according to one embodiment includes a sensor status analysis device (110) located on the sensor side and a sensor management server (120) located at a remote location.

[0111] The sensor status analysis device (110) can collect microcurrents transmitted from the sensor to the measuring device and transmit data points corresponding to the collected microcurrents.

[0112] The sensor management server (120) can collect data points of the transmitted microcurrent, identify abnormal values, and determine whether there is deterioration or failure based on the pattern of the identified abnormal values.

[0113] The sensor status analysis device (110) and the sensor management server (120) can communicate through various communication methods.

[0114] For this purpose, HTTP / HTTPS protocols, MQTT (Message Queuing Telemetry Transport), CoAP (Constrained Application Protocol), WebSocket, etc. can be used.

[0115] The sensor management server (120) can identify outliers for the collected microcurrent data points using the local outlier factor (LOF).

[0116] The present invention uses a local outlier factor (Local Outlier Factor KNN) derived from a KNN intelligent model.

[0117] LOF is a machine learning model that can be used to identify outliers in a data set. This model measures the local density of each data point, and data points with low local density are more likely to be outliers. The basic method is similar to the distance measurement of KNN, but it applies the concept of density to determine outliers and separate them.

[0118] The present invention allows for a more comprehensive understanding of the operating status of a sensor device by considering various equipment states (idle, power-saving, normal operation, full operation, and off) while measuring the supplied microcurrent value. This adaptive monitoring approach enables more accurate prediction and anomaly detection, setting it apart from existing technologies.

[0119] The present invention introduces a novel power cycling mechanism that temporarily interrupts and then reconnects the voltage and current supplied to equipment when a defect or abnormality is detected. This feature simulates a reboot situation and ensures equipment safety and reliability, representing a significant advancement over existing methods.

[0120] These aspects of the present invention, combined with the aforementioned drawings and detailed description, demonstrate the progressive and distinctive nature of the present invention in the context of marine observation equipment and sensor devices. By incorporating advanced data analysis techniques, comprehensive monitoring, an innovative power cycle mechanism, improved communication methods, and an intuitive user interface, the present invention addresses challenges faced by existing technologies and offers a significant leap forward in ensuring data reliability and continuity in marine environments.

[0121] According to an embodiment, a sensor management server (120) can identify surrounding neighbors for a data point using an intelligent model, and calculate a density ratio based on the distance to the identified surrounding neighbors. In addition, a local outlier factor (LOF) can be calculated based on the calculated density ratio, and if the calculated local outlier factor (LOF) is greater than a threshold value, the sensor can be identified as an outlier.

[0122] Additionally, the sensor management server (120) can learn the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

[0123] In addition, the sensor management server (120) considers at least one of idle, power saving, normal, maximum operation, and off as the equipment status while measuring the microcurrent value and the pattern of the identified abnormal value to determine whether there is deterioration or failure, and diagnoses the deterioration or failure of the sensor by using at least one of the pattern, trend, and machine learning for the abnormal value, or diagnoses the deterioration or failure of the sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0124] The present invention can significantly improve the monitoring, maintenance, and overall performance of marine observation equipment by addressing the shortcomings of existing technologies and providing advanced and differentiated features, ultimately contributing to improving the quality and continuity of marine data.

[0125] Additionally, the present invention provides a more affordable microcurrent sensing and measurement device, making it accessible to a wide range of organizations involved in ocean observation, and is designed to be used with a variety of power supplies, including batteries, solar, and grid power, making it suitable for a variety of marine environments and power configurations.

[0126] The present invention supports low-speed data methods such as RS-485 and RS-232, making integration with external equipment easy and enabling remote monitoring and maintenance of marine observation equipment.

[0127] The present invention integrates machine learning algorithms, such as the GRU model, to analyze current consumption data and equipment status, enabling accurate diagnosis of potential malfunctions and prediction of equipment deterioration. These capabilities can improve the overall reliability and performance of marine observation equipment.

[0128] Figure 3 is a drawing (300) explaining the overall operation of the marine observation equipment monitoring system.

[0129] In the process of the sensor (310) transmitting (301) a microcurrent to the measuring device (320), the marine observation equipment monitoring system (330) can collect the transmitted microcurrent.

[0130] In addition, the marine observation equipment monitoring system (330) can identify abnormal values ​​for the collected microcurrent data points using an intelligent model through step 303, and determine whether there is deterioration or failure based on the pattern of the identified abnormal values ​​to generate equipment status information.

[0131] Additionally, the marine observation equipment monitoring system (330) can transmit the equipment status information generated through step 304 to an external system (340).

[0132] The marine observation equipment monitoring system (330) can identify surrounding neighbors using an intelligent model for data points in step 303, and calculate a density ratio based on the distance to the identified surrounding neighbors. In addition, a local outlier factor (LOF) can be calculated based on the calculated density ratio, and if the calculated local outlier factor (LOF) is greater than a threshold value, the system can identify the point as an outlier.

[0133] In addition, in order to determine whether there is deterioration or failure based on the pattern of the identified outliers and generate equipment status information, deterioration or failure of the sensor can be distinguished and diagnosed using at least one of a pattern, trend, and machine learning for the outliers, and deterioration or failure of the sensor can be distinguished and diagnosed using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

[0134] Figure 4 is a diagram explaining the operation of the intelligent model.

[0135] The marine observation equipment monitoring system (330) may further include a process of learning (402) the intelligent model using microcurrent data (401) when the sensor is operating in a normal state and microcurrent data when the sensor is operating in a fault state.

[0136] Afterwards, by measuring the real-time microcurrent value (403), the score of the abnormal value can be calculated (404).

[0137] Additionally, based on the calculated score, it can be determined whether the distance score or density is greater than a threshold value (step 405).

[0138] Next, according to the operation of the intelligent model, if the distance score or density is below a threshold, it is determined as a normal operating state (406), and if the distance score or density is above a threshold, it can be determined as an abnormal operating state. In addition, idle, power saving, maximum operation, and off selection can be performed by selecting (407) from among the learned outlier data (408).

[0139] The Local Outlier Counting (LOF) algorithm is used for anomaly detection, identifying data points that differ significantly from the rest of the data points. In the context of equipment condition analysis, LOF can be used to identify deviations from normal conditions that may indicate equipment problems or abnormal operating conditions.

[0140] The following is a brief overview of how to use LOF in this context.

[0141] Data on the current and known states of the equipment can be collected and used to train a LOF model. The data should be labeled with known states (idle, power-saving, normal operation, full operation, off).

[0142] Next, the model is trained. This data is fed into the LOF algorithm, which allows the model to learn normal patterns for each state and calculate an "outlier score" indicating how different each data point is from its neighbors.

[0143] In real-time tasks, current values ​​can be input into the LOF model. The model can then calculate an outlier score for these values ​​based on learned patterns.

[0144] Next, the equipment's condition can be diagnosed based on the outlier score. A low score indicates the equipment is likely in good condition, while a high score indicates the equipment may be in an abnormal or malfunctioning state.

[0145] According to the present invention, very small current values ​​can be continuously collected from a microcurrent collection unit to implement a local outlier counting (LOF) algorithm.

[0146] Next, the monitoring unit normalizes or standardizes the collected current value data and processes it into a form suitable for the algorithm.

[0147] The monitoring unit can calculate LOF using [Mathematical Equation 1], and to construct the LOF model, the neighbors parameter is set to determine the number of neighbors for each point. The number of neighbors can affect the sensitivity of the model.

[0148]

[0149] [Mathematical Formula 1]

[0150]

[0151]

[0152] In [Mathematical Equation 1], LOF(k) is the local outlier coefficient for a data point x with k neighbors, Nk(x) are the k nearest neighbors of the data point x, and lrd(x) is the local reachability density of the data point x. This is calculated as the average of the reciprocals of the distances between x and its neighbors. In addition, lrd(Nk(x)) corresponds to the local reachability density of the neighbors of x.

[0153] The monitoring unit uses the LOF algorithm to calculate an outlier score for each data point and can set a threshold for the score to be considered an outlier. This threshold is crucial for determining the condition of the equipment.

[0154] Next, if the outlier score is higher than the threshold, it may indicate that the equipment is in an abnormal state (e.g., broken).

[0155] If the score is low, it may indicate that the equipment is in a normal state (e.g., power saving, idle, or full operation).

[0156] FIG. 5 is a diagram (500) showing LOF, which depicts outlier scores for collected data points.

[0157] LOF (Local Outlier Factor) is an outlier detection algorithm. It calculates the local abnormality of a data point and uses it to identify outliers. In other words, it can identify outliers in microcurrent output from a sensor.

[0158] LOF generates an outlier score for each data point, which can indicate how unusual that data point is compared to its neighbors.

[0159] Data points represent individual observations, or data points, in a dataset. Each data point consists of multiple characteristics (dimensions). For example, in sensor data, a data point might be composed of characteristics such as microcurrent or temperature.

[0160] Outlier Scores: The outlier score for each data point in LOF indicates how unusual that point is compared to its neighbors. A high outlier score indicates that the data point is relatively dissimilar to its neighbors.

[0161] LOF can identify outliers by considering the neighbors surrounding each data point. Neighbors are defined based on distance and represent other points close to a given data point. The number of neighbors can be set as one of LOF's hyperparameters.

[0162] LOF can be used to measure the local anomaly of a data point. It can be calculated by considering the relative density of that data point relative to its neighbors.

[0163] If LOF is greater than 1, it means that the data point has a lower density than its surroundings, and may be considered an outlier.

[0164] For each data point, neighbors can be identified and the relative density between neighbors can be calculated to generate a LOF. If the outlier score exceeds a certain threshold, the data point can be identified as an outlier, indicating deterioration or failure of the sensor outputting the microcurrent.

[0165] Ultimately, the present invention addresses the shortcomings of existing technologies and offers advanced, differentiated features, significantly improving the monitoring, maintenance, and overall performance of marine observation equipment, ultimately enhancing the quality and continuity of marine data. Furthermore, it provides a more affordable microcurrent detection and measurement device, making it accessible to a wide range of organizations involved in marine observation.

[0166]

[0167] The devices described above may be implemented as hardware components, software components, and / or a combination of hardware components and software components. For example, the devices and components described in the embodiments may be implemented using one or more general-purpose computers or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing instructions and responding to them. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing device is sometimes described as being used alone; however, one of ordinary skill in the art will recognize that the processing device may include multiple processing elements and / or multiple types of processing elements. For example, a processing unit may include multiple processors, or a processor and a controller. Other processing configurations, such as parallel processors, are also possible.

[0168] Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing device to perform a desired operation or may, independently or collectively, command the processing device. The software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave, for interpretation by the processing device or for providing instructions or data to the processing device. The software may also be distributed over networked computer systems and stored or executed in a distributed manner. The software and data may be stored on one or more computer-readable recording media.

[0169] The method according to the embodiment may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., alone or in combination. The program commands recorded on the medium may be those specially designed and configured for the embodiment or may be those known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of the program commands include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiment, and vice versa.

[0170] Although the embodiments described above have been described with limited drawings, those skilled in the art will recognize that various modifications and variations can be made based on the above description. For example, appropriate results can still be achieved even if the described techniques are performed in a different order than described, and / or components of the described systems, structures, devices, circuits, etc. are combined or combined in a different manner than described, or are replaced or substituted with other components or equivalents.

[0171] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims described below.

Claims

1. A microcurrent collection unit that collects microcurrent transmitted from the sensor to the measuring device; A monitoring unit that identifies outliers in the collected microcurrent data points using an intelligent model; A diagnostic unit that generates equipment status information by determining whether there is deterioration or failure based on the pattern of the identified abnormal values; and Communication unit that transmits the above generated equipment status information to an external system A marine observation equipment monitoring system characterized by including:

2. In paragraph 1, The above monitoring unit, A marine observation equipment monitoring system characterized in that it identifies outliers using a local outlier factor (LOF) for the collected microcurrent data points.

3. In paragraph 2, The above monitoring unit, Using an intelligent model for the above data points, we identify surrounding neighbors, and calculate a density ratio based on the distance to the identified surrounding neighbors. A marine observation equipment monitoring system characterized in that a local outlier factor (LOF) is calculated based on a calculated density ratio, and if the calculated local outlier factor (LOF) is greater than a threshold value, the system is identified as an outlier.

4. In paragraph 2, The above monitoring unit, A marine observation equipment monitoring system characterized by initiating a novel power cycling mechanism that temporarily interrupts and then reconnects the supplied voltage and current when the above-mentioned outliers are identified.

5. In paragraph 1, The above monitoring unit, A marine observation equipment monitoring system characterized in that the intelligent model is learned using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

6. In paragraph 1, The above diagnostic section, A marine observation equipment monitoring system characterized in that it determines whether there is deterioration or failure by considering at least one of idle, power saving, normal, maximum operation, and off as the equipment status while measuring the microcurrent value and the pattern of the identified abnormal values.

7. In paragraph 1, The above diagnostic section, A marine observation equipment monitoring system characterized in that it diagnoses and distinguishes deterioration or failure of the sensor by using at least one of a pattern, trend, and machine learning for the above outliers.

8. In paragraph 1, The above diagnostic section, A marine observation equipment monitoring system characterized in that it diagnoses and distinguishes deterioration or failure of a sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

9. A sensor status analysis device that collects microcurrents transmitted from a sensor to a measuring device and transmits data points corresponding to the collected microcurrents; A sensor management server that collects data points of the microcurrent transmitted above, identifies abnormal values, and determines whether there is deterioration or failure based on the pattern of the identified abnormal values. A marine observation equipment monitoring system characterized by including:

10. In paragraph 9, The above sensor management server, A marine observation equipment monitoring system characterized in that it identifies outliers using a local outlier factor (LOF) for the collected microcurrent data points.

11. In paragraph 10, The above sensor management server, Using an intelligent model for the above data points, we identify surrounding neighbors, and calculate a density ratio based on the distance to the identified surrounding neighbors. A marine observation equipment monitoring system characterized in that a local outlier factor (LOF) is calculated based on the calculated density ratio, and if the calculated local outlier factor (LOF) is greater than a threshold value, the system is identified as an outlier.

12. In paragraph 10, The above sensor management server, A marine observation equipment monitoring system characterized by driving a new power cycling mechanism that temporarily interrupts and then reconnects the voltage and current supplied to the sensor status analysis device when the above abnormal values ​​are identified.

13. In paragraph 9, The above sensor management server, A marine observation equipment monitoring system characterized in that the intelligent model is learned using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state.

14. In paragraph 9, The above sensor management server, A marine observation equipment monitoring system characterized in that it determines whether there is deterioration or failure by considering at least one of idle, power saving, normal, maximum operation, and off as the equipment status while measuring the microcurrent value and the pattern of the identified abnormal values.

15. In paragraph 9, The above sensor management server, A marine observation equipment monitoring system characterized in that it diagnoses and distinguishes deterioration or failure of the sensor by using at least one of a pattern, trend, and machine learning for the above-mentioned outliers, or it diagnoses and distinguishes deterioration or failure of the sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor.

16. A step of collecting microcurrent transmitted from a sensor to a measuring device in a microcurrent collection unit; In the monitoring unit, a step of identifying outliers for the collected microcurrent data points using an intelligent model; In the diagnostic section, a step of generating equipment status information by determining whether there is deterioration or failure based on the pattern of the identified abnormal values; and In the communication department, a step of transmitting the generated equipment status information to an external system A method of operating a marine observation equipment monitoring system, characterized in that it includes a.

17. In paragraph 16, The step of identifying outliers for the collected microcurrent data points using the above intelligent model is as follows: A step of identifying surrounding neighbors using an intelligent model for the above data points and calculating a density ratio based on the distance to the identified surrounding neighbors; A step of calculating a local outlier factor (LOF) based on the calculated density ratio; Step for identifying an outlier if the above calculated Local Outlier Factor (LOF) is greater than the threshold value A method of operating a marine observation equipment monitoring system, characterized in that it includes a.

18. In paragraph 16, The step of generating equipment status information by determining whether there is deterioration or failure based on the pattern of the identified abnormal values ​​is as follows. A step of diagnosing and distinguishing deterioration or failure of the sensor by using at least one of a pattern, trend, and machine learning for the above outlier; and A step of diagnosing and distinguishing deterioration or failure of a sensor by using at least one additional sensor among a temperature sensor, a vibration sensor, and a noise sensor. A method of operating a marine observation equipment monitoring system, characterized in that it includes a.

19. In Article 16, A step of learning the intelligent model using microcurrent data when the sensor operates in a normal state and microcurrent data when the sensor operates in a fault state. A method of operating a marine observation equipment monitoring system, characterized in that it further includes:

Images