Unmanned ship navigation path intelligent optimization system and method based on ocean data

Abstract

The invention discloses an unmanned ship navigation path intelligent optimization system and method based on ocean data, and particularly relates to the technical field of unmanned ship navigation. The accuracy of monitoring data of the sensor in a marine environment is continuously evaluated through a series of steps of preliminary calibration, data acquisition, filtering processing, abnormal marking, feature extraction and the like of the unmanned ship sensor, and real-time evaluation and time period division of the quality of the monitoring data are realized through comprehensive analysis of abnormal features of sound velocity and conductivity. The abnormal degree is predicted in the abnormal monitoring time period, and for the condition of high abnormal degree, a safety avoidance strategy is automatically set through a path planning system, and the route and the navigational speed of the unmanned ship are dynamically optimized, so that a risk area is avoided, the navigation precision and safety are improved, the reliability of sensor data is remarkably improved, and the safety of the unmanned ship is improved. The influence of data deviation on path planning is reduced, and the task success rate and the navigation safety of the unmanned ship in a complex marine environment are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of unmanned boat navigation technology, and in particular to an unmanned boat navigation path intelligent optimization system and method based on ocean data. Background Art

[0002] Intelligent optimization of unmanned boat navigation paths based on ocean data refers to the use of ocean environment data to intelligently plan and adjust the driving paths of unmanned boats, thereby improving navigation efficiency and safety. Ocean data includes information such as wind speed, wind direction, ocean currents, wave height, temperature, etc., which are crucial to the navigation of unmanned boats. By analyzing and utilizing this data, the system can calculate the best route to help unmanned boats avoid adverse navigation conditions (such as strong wind areas and areas with high waves), thereby saving fuel, reducing navigation time and ensuring the safety of unmanned boats.

[0003] Specifically, this intelligent optimization system usually consists of sensors, data analysis algorithms, and path planning algorithms. First, the unmanned boat collects real-time ocean data around it through sensors and transmits the data back to the processing center or onboard computing system. Next, the data analysis algorithm evaluates the pros and cons of different paths based on preset goals. Finally, the path planning algorithm dynamically adjusts the navigation route of the unmanned boat based on the analysis results to avoid obstacles or dangerous areas, and continuously updates and optimizes the route during navigation so that the unmanned boat can reach its destination smoothly.

[0004] The prior art has the following deficiencies: The marine environment places extremely high demands on equipment reliability. Long-term seawater immersion, corrosive salt spray, temperature changes, etc. may cause sensor failure or reduced data accuracy. If key sensors such as positioning, meteorology, and wave height measurement fail, the collected data will deviate from the actual situation, leading to incorrect path planning and decision-making. At the same time, erroneous data will cause the system to make inaccurate judgments about the marine environment, causing the unmanned boat to misjudge the surrounding risk areas, resulting in deviation from the scheduled route or entry into dangerous waters. In addition, long-term sensor errors may accumulate into large deviations, causing the unmanned boat's path optimization algorithm to run based on distorted data, ultimately posing a great threat to navigation safety and mission success rate. Summary of the invention

[0005] The purpose of the present invention is to provide a system and method for intelligently optimizing the navigation path of an unmanned boat based on ocean data, so as to solve the shortcomings of the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: an intelligent optimization method for unmanned boat navigation path based on ocean data, comprising the following steps: S1: Perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over several time periods; S2: filtering the collected sound velocity distribution data and conductivity data and comparing them with the benchmark data, marking abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; S3: extracting features from the marked abnormal sound velocity distribution data and abnormal conductivity data, respectively extracting abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; S4: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; S5: according to the accuracy of the monitoring data of the sensor in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; S6: For abnormal monitoring time periods, the degree of abnormality in the accuracy of the sensor's monitoring data in the marine environment is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

[0007] Preferably, in S2, a low-pass filter is applied to smooth the sound velocity and conductivity data, the sound velocity and conductivity data after filtering are normalized and converted into relative values ​​relative to the benchmark data, and the allowable normal fluctuation range is set according to the distribution characteristics of the benchmark data; for the sound velocity and conductivity data at each time point, the deviation value relative to the benchmark data is calculated, and the expression is: ; Compare the deviation value with the preset normal fluctuation range, mark all time point data that exceed the normal range, and mark the data points that deviate from the normal range as abnormal data.

[0008] Preferably, in S3, after analyzing the change rate drift of the abnormal sound velocity value in the extracted sound velocity distribution abnormality feature, a sound velocity value change rate drift index is generated, and the method for obtaining the sound velocity value change rate drift index is: Select the sound speed change rate sequence under normal conditions as the reference sequence , the sound velocity change rate data collected in the current time period is used as the sequence to be tested , define the distance between any two points in the reference sequence and the sequence to be tested, use the Euclidean distance to measure the difference between the two points, and calculate the distance between each pair of points , the expression is: ; Construct an N×M matrix D, where the element D(i,j) in the matrix represents the distance between the i-th point of the reference sequence and the j-th point of the sequence to be tested, that is: D(i,j)=d(i,j); Construct a cumulative distance matrix C to store the minimum path distance from the beginning of the sequence to the current point; Set the cumulative distance of the starting point to the starting point distance, that is, C(1,1)=D(1,1); Starting from the upper left corner of the matrix, gradually calculate the minimum cumulative distance of each position. The element C(i,j) in the cumulative distance matrix represents the minimum cumulative distance from the starting point (1,1) to the point (i,j), which is calculated by the recursive formula: ; The last element C(N,M) of the cumulative distance matrix C represents the total cumulative distance between the reference sequence and the sequence to be tested, that is, the drift index of the sound speed change rate is calculated, and the expression is: ; Where HK is the drift index of the sound velocity change rate.

[0009] Preferably, in S3, after analyzing the deviation amplitude of the conductivity abnormal value in the extracted conductivity fluctuation feature relative to the reference conductivity value, a conductivity fluctuation amplitude deviation index is generated, and the method for obtaining the conductivity fluctuation amplitude deviation index is: Select the conductivity time series under normal conditions As a benchmark, collect the conductivity time series for the current time period ; The Bayesian change point detection probability model is selected to describe the data distribution of the conductivity time series, and the conductivity fluctuation is assumed to obey the Gaussian distribution. The expression is: ;in, and are the mean and variance of the conductivity values, respectively. A prior distribution is set for the change point position. The likelihood function is calculated based on the currently observed conductivity data to update the probability distribution of the change point. According to the Bayesian formula, the probability of the change point at time point i is calculated. , for each time point i, the updated change point probability is: ;in, represents the time series subset from the starting point to i, P(T) is the normalization constant, and the significant change point positions are screened out ; According to the detected change points, the sequence T to be tested is divided into k+1 sub-intervals, and the average fluctuation amplitude of the conductivity value is calculated in each sub-interval , the expression is: ;in, represents the length of subinterval j, and They represent the change point positions respectively. Using the segmented results of the change point detection, the average values ​​of the fluctuation amplitudes of all segments are summarized to obtain the conductivity fluctuation amplitude deviation index, which is expressed as: ; GH is the conductivity fluctuation amplitude deviation index.

[0010] Preferably, in S4, a comprehensive analysis is performed on the extracted abnormal characteristics of the sound velocity distribution and the conductivity fluctuation characteristics to evaluate the accuracy of the monitoring data of the sensor in the marine environment in each time period, specifically: The sound velocity change rate drift index and the conductivity fluctuation amplitude deviation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the sensor's monitoring data in the marine environment in each time period as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the sensor's monitoring data in the marine environment in all time periods as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy value of the sensor's monitoring data in the marine environment in each time period is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0011] Preferably, in S5, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods according to the accuracy of the monitoring data of the sensor in the marine environment in each time period; The acquired accuracy value of the sensor's monitoring data in the marine environment in each time period is compared with the reference threshold of the accuracy value under normal conditions set according to historical data. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is greater than or equal to the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is high. In this case, no warning signal is generated, and the corresponding time period is divided into a normal monitoring time period. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is less than the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is low. In this case, a warning signal is generated, and the corresponding time period is divided into an abnormal monitoring time period.

[0012] Preferably, in S6, for the abnormal monitoring time period, the abnormal degree of the accuracy of the sensor monitoring data in the marine environment is predicted. If the abnormal degree is high, a safety avoidance strategy is set by the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path, specifically: For abnormal monitoring time periods, that is, the accuracy value of the sensor's monitoring data in the marine environment during the corresponding time period is less than the reference threshold of the accuracy value, the accuracy values ​​of the monitoring data that are less than the reference threshold of the accuracy value in the subsequent time period are collected, and a data set is established. The mean and standard deviation in the data set are calculated, and after analysis, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted based on the analysis results.

[0013] Preferably, if the mean of the accuracy values ​​in the data set is greater than or equal to the reference threshold of the mean of the accuracy values, and the standard deviation of the accuracy values ​​is less than the reference threshold of the standard deviation of the accuracy values, it means that the overall deviation of the monitoring data of the sensor is small and the fluctuation is low, indicating that the sensor has returned to a stable state, and the avoidance buffer distance of the unmanned boat is gradually reduced and the normal speed is restored; If the mean accuracy value is greater than or equal to the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, it means that although the average accuracy of the data is high, the fluctuation is large, indicating that the sensor is in a recovery state but unstable, and it is necessary to continue to maintain the avoidance strategy and closely monitor the data fluctuation of the sensor; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, the overall accuracy of the sensor data is low and the fluctuation range is large. A high buffer distance and low speed should be maintained, risk avoidance should be prioritized, and the sensor should be inspected and replaced; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is less than the reference threshold of the standard deviation of the accuracy value, it means that the average accuracy of the data is low but the fluctuation is small. Maintain the avoidance strategy and implement emergency maintenance or stop using the sensor data.

[0014] The present invention also provides an unmanned boat navigation path intelligent optimization system based on ocean data, including a data acquisition module, an anomaly detection module, a feature extraction module, an accuracy evaluation module, a time period classification module and a path optimization module; Data acquisition module: perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over a certain period of time; Abnormal detection module: Filter the collected sound velocity distribution data and conductivity data and compare them with the benchmark data, and mark the abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; Feature extraction module: extract features from the marked abnormal sound velocity distribution data and abnormal conductivity data, and respectively extract abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; Accuracy evaluation module: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; Time period classification module: according to the accuracy of the sensor monitoring data in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; Path optimization module: For abnormal monitoring time periods, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted. If the abnormal degree is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

[0015] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention significantly improves the data reliability of sensors in complex marine environments and the system's adaptability to the environment by systematically calibrating, processing data, analyzing anomalies, and optimizing paths for unmanned boat sensors. The accuracy assessment is performed using the sound velocity change rate drift index and the conductivity fluctuation amplitude deviation index in combination with a polynomial regression model. The accuracy value of the sensor monitoring data can be quantified in real time, and the time period can be divided into normal and abnormal monitoring time periods by setting reference thresholds based on historical data, ensuring that the system can dynamically respond to complex changes in the marine environment. By predicting the degree of abnormality in the accuracy of sensor data within the abnormal monitoring time period, the route buffer distance and speed of the unmanned boat can be adjusted in a timely manner to reduce the risk of deviation and navigation losses.

[0016] 2. This invention effectively solves the problem of path planning errors caused by sensor failure or reduced accuracy in traditional technologies. Through refined feature analysis and dynamic optimization strategies, it achieves accurate avoidance of risk areas, improving the navigation safety and mission success rate of unmanned boats. Through comprehensive data monitoring, accuracy assessment and path optimization, this system can adapt to long-term mission requirements, reduce the impact of the marine environment on equipment reliability, and provide strong technical support for the stable operation of unmanned boats under extreme conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0018] Figure 1 The figure is a flow chart of the method of the present invention.

[0019] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 and Figure 2 As shown, the method for intelligently optimizing the navigation path of an unmanned boat based on ocean data described in this embodiment includes the following steps: S1: Perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over several time periods; S2: filtering the collected sound velocity distribution data and conductivity data and comparing them with the benchmark data, marking abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; S3: extracting features from the marked abnormal sound velocity distribution data and abnormal conductivity data, respectively extracting abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; S4: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; S5: according to the accuracy of the monitoring data of the sensor in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; S6: For abnormal monitoring time periods, the degree of abnormality in the accuracy of the sensor's monitoring data in the marine environment is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

[0022] In S1, all relevant sensors on the unmanned boat are preliminarily calibrated to ensure that their initial status is normal, and the baseline data is recorded. The sound velocity distribution data and conductivity data collected by the unmanned boat in real time in the designated sea area in several time periods are collected, specifically: Determine the types of sensors that need to be calibrated, including key sensors such as sound velocity meters and conductivity meters, to ensure that these sensors can accurately capture the sound velocity distribution and conductivity in the marine environment. Anchor the unmanned boat on shore or in calm waters, prepare standard reference equipment (such as standard sound velocity sensors and calibrated conductivity meters) for comparison of measurements, and prepare recording equipment to facilitate storage and subsequent retrieval of benchmark data.

[0023] Zero-point calibration is performed on each sensor to ensure that the sensor reading is the standard zero value in the initial state and eliminate the influence of environmental factors on the initial reading. The sensor is placed in a standard solution with known sound velocity and conductivity or in a measured sea area, and the output value of the sensor is recorded and compared with the standard value to calibrate its accuracy. After multiple tests, the sensor parameters are adjusted according to the deviation to ensure that its measured value is within the acceptable error range.

[0024] After calibration, use the sensor to collect sound velocity distribution data and conductivity data in an undisturbed sea area to obtain initial "baseline data". The environmental data during this period should be data in a normal, stable state. The baseline data is stored in the database with a timestamp and used for comparison with real-time data in subsequent analysis. This data will become the basis for judging the performance fluctuations of the sensor under different marine environmental conditions.

[0025] After the unmanned boat is launched in the designated sea area, it collects sound velocity distribution and conductivity data in real time at preset time intervals (for example, every 1 minute or less) to ensure that enough continuous data points are obtained. Each collected data needs to be accurately timestamped and stored in sequence in the shipboard computing system or transmitted to the shore-based processing center. In this way, the time point of the data can correspond to the actual environmental changes in subsequent analysis.

[0026] At the same time, auxiliary environmental data such as temperature and pressure are recorded to facilitate multi-dimensional analysis of the sound velocity distribution and conductivity changes in subsequent analysis. The quality of the initially collected sound velocity distribution and conductivity data is checked to detect whether there is obvious noise, sudden changes or deviations. If any abnormality is found, the sensor is adjusted or recalibrated in a timely manner.

[0027] S2: The collected sound velocity distribution data and conductivity data are filtered and compared with the benchmark data, and abnormal sound velocity distribution data and abnormal conductivity data that deviate from the normal range in each time period are marked.

[0028] Synchronize the timestamps of the acoustic velocity distribution data and the conductivity data to ensure that the data in the same time period can be accurately compared to avoid data errors caused by time deviation. Check whether there are missing values ​​in the data. For small-scale missing values, linear interpolation can be performed using the data of the previous and next time points; if the missing amount is large, mark it and exclude it in subsequent analysis.

[0029] Apply a low-pass filter (such as Kalman filter or mean filter) to smooth the sound velocity and conductivity data to remove short-term noise caused by sensor jitter or external interference during the measurement process. Use a sliding window technique (such as a 3-5 minute window) to calculate the mean and standard deviation of the data in each time period. Mark and remove data points that deviate from the mean by a certain multiple (such as 3 times the standard deviation) to further improve the stability of the data.

[0030] Normalize the sound velocity and conductivity data after filtering and convert them into relative values ​​relative to the benchmark data for subsequent comparison. Set the allowable normal fluctuation range based on the distribution characteristics of the benchmark data (such as mean and standard deviation). Usually set to ± a certain percentage or multiple of the standard deviation, such as ± 2 times the standard deviation, as the normal range limit.

[0031] For the sound velocity and conductivity data at each time point, calculate the deviation value relative to the baseline data. The expression is: ; Compare the deviation value with the preset normal fluctuation range and mark all time point data that exceed the normal range. Data points that deviate from the normal range are marked as abnormal data for further feature extraction and analysis.

[0032] The data points marked as abnormal are recorded with details such as timestamp, deviation value and sensor information for subsequent feature extraction and analysis. The data is classified and stored into normal and abnormal categories. Normal data is used for path optimization and system decision-making, while abnormal data is used for further analysis to help the unmanned boat determine whether path adjustment or sensor calibration is needed.

[0033] Perform a quality check on the marked abnormal data to ensure that the marking of the data point meets the preset threshold standard to avoid misjudgment. The rationality of the marking can be confirmed by statistically analyzing the deviation distribution of various types of data. After completing the marking and classification, output the marked abnormal data and generate the corresponding abnormal report for feature extraction and comprehensive analysis in the subsequent steps.

[0034] S3: extracting features from the marked abnormal sound velocity distribution data and abnormal conductivity data, respectively extracting abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data.

[0035] Feature extraction is performed on the marked abnormal sound velocity distribution data and abnormal conductivity data in order to further analyze the laws and patterns of the abnormal data for subsequent comprehensive analysis and path optimization adjustment.

[0036] The extraction of abnormal sound velocity distribution features specifically includes: Abnormal amplitude: Calculate the deviation amplitude of the sound velocity value at each abnormal time point relative to the reference sound velocity to quantify the severity of the abnormality. The calculation formula is: ; Abnormal duration: detect the time interval between adjacent abnormal points, identify the time period of continuous abnormalities, record the start and end time of each abnormality, and calculate the total duration of the abnormality. Abnormalities with long duration often indicate drastic changes in the marine environment.

[0037] Spatial distribution characteristics: If the unmanned boat has a geographic location record, the abnormal sound speed data is associated with the geographic location to determine the spatial distribution of the abnormal data in different areas and analyze whether the anomalies are concentrated in a specific area.

[0038] Abnormal change rate: Calculate the change rate of abnormal sound velocity values, that is, the increment of sound velocity values ​​at each time point, to identify sudden change anomalies or slow drift anomalies. The change rate can help determine the impact of environmental factors (such as temperature and depth changes) on the sound velocity. The calculation expression is: .

[0039] Frequency characteristics: Spectral analysis is performed on the abnormal sound velocity distribution data to extract the main frequency components. This can help identify periodic fluctuation anomalies (such as sea wave effects) and random anomaly characteristics.

[0040] The conductivity fluctuation feature extraction specifically includes: Fluctuation amplitude: Calculate the deviation amplitude of the conductivity anomaly value relative to the baseline conductivity value to quantify the intensity of the anomaly fluctuation: .

[0041] Fluctuation frequency: Through Fourier transform or wavelet analysis, the conductivity anomaly data is subjected to spectrum analysis to extract the main fluctuation frequency. This feature can be used to determine the periodicity (such as periodic changes in salinity) and sudden changes in conductivity anomalies, especially the influence of tides or underwater activities that may occur in specific sea areas.

[0042] Fluctuation Trend: Calculate the temporal trend of conductivity anomalies to identify long-term fluctuations (such as gradual increases or decreases) or short-term abrupt changes. This helps distinguish different patterns of environmental change, such as sudden pollutant impacts or the long-term effects of flow rate changes on conductivity.

[0043] Anomaly Cumulative Effects: Calculates the cumulative value of conductivity fluctuations to identify the cumulative effects of abnormal conductivity changes. This feature is suitable for identifying the accumulation of conductivity fluctuations caused by long-term changes in seawater composition.

[0044] Fluctuation area characteristics: Combined with the geographic location data of the unmanned boat, the conductivity anomaly data is mapped to the specific area, and the spatial distribution of abnormal fluctuations is analyzed. This helps to identify high-salinity or low-salinity areas and analyze the potential impact of the area on the navigation path of the unmanned boat.

[0045] The abnormal characteristics of the sound velocity distribution and the conductivity fluctuation characteristics are summarized into the feature database, and a detailed feature description is established for each time period, including the abnormal amplitude, duration, spatial distribution, change rate, frequency and other characteristics. The extracted features are output as a structured data format (such as JSON or CSV file) so that they can be used for sensor data accuracy analysis and path optimization algorithm adjustment in subsequent steps.

[0046] After analyzing the change rate drift of the abnormal sound velocity value in the extracted sound velocity distribution abnormality feature, a sound velocity value change rate drift index is generated. The method for obtaining the sound velocity value change rate drift index is as follows: Select the sound speed change rate sequence under normal conditions as the reference sequence , the sound velocity change rate data collected in the current time period is used as the sequence to be tested , define the distance between any two points in the reference sequence and the sequence to be tested, use the Euclidean distance to measure the difference between the two points, and calculate the distance between each pair of points , the expression is: ; Construct an N×M matrix D, where the element D(i,j) in the matrix represents the distance between the i-th point of the reference sequence and the j-th point of the sequence to be tested, that is: D(i,j)=d(i,j); Construct a cumulative distance matrix C to store the minimum path distance from the beginning of the sequence to the current point. Set the cumulative distance of the starting point to the starting point distance, that is, C(1,1)=D(1,1); Starting from the upper left corner of the matrix, gradually calculate the minimum cumulative distance of each position. The element C(i,j) in the cumulative distance matrix represents the minimum cumulative distance from the starting point (1,1) to the point (i,j), which is calculated by the recursive formula: ; In the calculation of C(i,j), the minimum cumulative distance path above, to the left or to the upper left is selected to ensure that the obtained path is the shortest, thus achieving the best match. The last element C(N,M) of the cumulative distance matrix C represents the total cumulative distance between the reference sequence and the sequence to be tested, that is, the drift index of the sound speed change rate is calculated, and the expression is: ; Where HK is the drift index of the sound velocity change rate.

[0047] The larger the drift index (DTW distance) of the sound speed value change rate, the more significant the deviation between the current sound speed change rate and the baseline normal sequence. This usually indicates that the monitoring data of the sensor is greatly affected by the environment or that there is a problem with the accuracy of the sensor itself. A large drift index may be caused by severe ocean conditions (such as drastic changes in temperature, pressure, and salinity), resulting in the accumulation of deviations in sound speed measurements; it may also be due to sensor aging, failure, or interference from marine organisms, causing the measured value to deviate from the actual value. In general, the larger the drift index, the more serious the data deviation, the lower the measurement accuracy of the sensor, and the worse the reliability and effectiveness of the monitoring data.

[0048] On the contrary, the smaller the drift index of the sound speed value change rate, the smaller the difference between the measured sequence and the reference normal sequence, indicating that the sensor can stably and accurately monitor the sound speed change rate in the current ocean environment. A smaller drift index usually indicates that the ocean environment is relatively stable, the sensor can operate within the predetermined accuracy range, and the data accuracy and reliability are high. Therefore, when the drift index is close to zero, it can be considered that the data monitored by the sensor basically conforms to the actual ocean conditions, and the system's judgment of the environment is more accurate, which can support the unmanned boat to perform more accurate path optimization and risk avoidance.

[0049] After analyzing the deviation amplitude of the conductivity abnormal value in the extracted conductivity fluctuation feature relative to the reference conductivity value, a conductivity fluctuation amplitude deviation index is generated. The method for obtaining the conductivity fluctuation amplitude deviation index is as follows: Select the conductivity time series under normal conditions As a benchmark, collect the conductivity time series for the current time period ; The Bayesian change point detection probability model is selected to describe the data distribution of the conductivity time series, and the conductivity fluctuation is assumed to obey the Gaussian distribution. The expression is: in, and are the mean and variance of the conductivity values, respectively. Set a prior distribution for the change point position. The commonly used prior is uniform distribution to indicate that the change point may appear at any position in the initial state. Calculate the likelihood function based on the currently observed conductivity data to update the probability distribution of the change point. According to the Bayesian formula, calculate the probability of the change point at time point i , for each time point i, the updated change point probability is: ;in, represents the time series subset from the starting point to i, P(T) is the normalization constant, and the significant change point positions are screened out by setting the threshold of the change point probability (for example, 0.9). ; According to the detected change points, the sequence T to be tested is divided into k+1 sub-intervals, and the average fluctuation amplitude of the conductivity value is calculated in each sub-interval , the expression is: ;in, represents the length of subinterval j, and They represent the change point positions respectively. Using the segmented results of the change point detection, the average values ​​of the fluctuation amplitudes of all segments are summarized to obtain the conductivity fluctuation amplitude deviation index, which is expressed as: ; GH is the conductivity fluctuation amplitude deviation index.

[0050] The larger the conductivity fluctuation deviation index, the more significant the fluctuation of the conductivity data to be measured is compared with the baseline value. This usually means that the monitoring data of the sensor is strongly affected by the environment or the sensor itself has errors. A larger deviation index may be caused by drastic changes in ocean conditions (such as sharp changes in salinity, temperature or pressure), or due to sensor aging, failure, or interference from pollutants or marine life. A larger deviation index indicates that the instability of the conductivity data increases, and the accuracy and reliability of the data decrease, which will provide poor support for subsequent tasks such as path optimization and environmental judgment.

[0051] On the contrary, the smaller the conductivity fluctuation deviation index, the more stable the conductivity data fluctuation is, and it remains relatively consistent with the baseline value. This shows that the sensor can work stably in the current marine environment and the monitoring data is more accurate. A low deviation index usually means that the marine environmental conditions are stable, and the conductivity data measured by the sensor is within a reasonable fluctuation range, reflecting the real marine conditions. At this time, the unmanned boat can rely on higher-quality data to make environmental judgments and path adjustments, improving the safety and efficiency of navigation.

[0052] S4: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period.

[0053] The sound velocity change rate drift index and the conductivity fluctuation amplitude deviation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the sensor's monitoring data in the marine environment in each time period as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the sensor's monitoring data in the marine environment in all time periods as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy value of the sensor's monitoring data in the marine environment in each time period is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0054] The method for obtaining the accuracy value of the sensor monitoring data in the marine environment in each time period is: from the comprehensive feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, HK is the drift index of the sound velocity value change rate, GH is the conductivity fluctuation amplitude deviation index, It is the accuracy value of the monitoring data of the sensor in the marine environment in each time period.

[0055] S5: according to the accuracy of the monitoring data of the sensor in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; The acquired accuracy value of the sensor's monitoring data in the marine environment in each time period is compared with the reference threshold of the accuracy value under normal conditions set according to historical data. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is greater than or equal to the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is high. In this case, no warning signal is generated, and the corresponding time period is divided into a normal monitoring time period. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is less than the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is low. In this case, a warning signal is generated, and the corresponding time period is divided into an abnormal monitoring time period.

[0056] The data within the normal monitoring period is stored in the database and archived as high-quality benchmark data. The original data including the sound velocity value, the rate of change of sound velocity, the conductivity and its fluctuation amplitude and the calculated accuracy index are saved. The stored normal monitoring data can be used as a comparison benchmark for future abnormal conditions, which is convenient for subsequent comparative analysis under different regions and seasonal conditions in the marine environment, and improves data traceability.

[0057] Use data from the normal monitoring period to further optimize the machine learning model. In subsequent time periods, gradually update the model training data, and use the new normal period data to fine-tune the model weights to maintain the model's sensitivity to environmental changes. This enables the model to continuously learn the latest normal environmental patterns, thereby improving accuracy and adaptability in real-time monitoring. Continuous optimization of normal data helps the model cope with different seasonal changes and sudden environmental disturbances.

[0058] After accumulating enough high-accuracy data during the normal monitoring period, the reference threshold can be dynamically adjusted to adapt to the long-term changes in the marine environment. New upper and lower limits can be set based on the normal accuracy value distribution over a period of time to refine the classification criteria for subsequent time periods.

[0059] S6: For abnormal monitoring time periods, the degree of abnormality in the accuracy of the sensor's monitoring data in the marine environment is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

[0060] For abnormal monitoring time periods, that is, the accuracy value of the sensor's monitoring data in the marine environment during the corresponding time period is less than the reference threshold of the accuracy value, the accuracy values ​​of the monitoring data that are less than the reference threshold of the accuracy value in the subsequent time period are collected, and a data set is established. The mean and standard deviation in the data set are calculated, and after analysis, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted based on the analysis results.

[0061] If the mean accuracy value in the data set is greater than or equal to the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is less than the reference threshold of the standard deviation of the accuracy value, it means that the overall deviation of the sensor's monitoring data is small and the fluctuation is low, indicating that the sensor has returned to a relatively stable state, the degree of abnormality is low, and the avoidance buffer distance of the unmanned boat is gradually reduced and the normal speed is restored; If the mean accuracy value is greater than or equal to the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, it means that although the average accuracy of the data is high, the fluctuation is large, indicating that the sensor is in a recovery state but unstable, and may be intermittently disturbed by environmental factors. It is necessary to continue to maintain a certain avoidance strategy, and at the same time closely monitor the data fluctuations of the sensor to evaluate whether the route needs to be further adjusted; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, the overall accuracy of the sensor data is low and the fluctuation range is large, indicating that the sensor may be in a serious abnormal state and the reliability of the monitoring data is extremely low. A high buffer distance and low speed should be maintained, risk avoidance should be prioritized, and the sensor should be inspected or replaced; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is less than the reference threshold of the standard deviation of the accuracy value, it means that the average accuracy of the data is low but the fluctuation is small, indicating that the sensor may have persistent failures or systematic deviations. Maintain an avoidance strategy and implement emergency maintenance or stop using the sensor's data when possible to avoid systematic errors affecting path optimization.

[0062] In this embodiment, the data accuracy and navigation safety in a complex marine environment are ensured by systematically monitoring and optimizing the sensors of the unmanned boat. First, all sensors are preliminarily calibrated, the baseline data is recorded, and the sound velocity distribution and conductivity data are collected in real time; then, the collected data is filtered and compared with the baseline data, and the abnormal sound velocity distribution and conductivity data are marked; then, the marked abnormal data is feature extracted, and the abnormal sound velocity distribution features and conductivity fluctuation features are extracted respectively; then, the extracted features are comprehensively analyzed, the accuracy of the sensor monitoring data in each time period is evaluated, and the time period is divided into a normal monitoring time period and an abnormal monitoring time period, and the normal and abnormal time periods are processed respectively according to the comparison of the accuracy value with the reference threshold; in the abnormal time period, the degree of abnormality is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system, and the deviation buffer distance and speed of the unmanned boat are continuously optimized to maximize the safety and mission completion rate of the unmanned boat in a complex marine environment.

[0063] Embodiment 2, the unmanned boat navigation path intelligent optimization system based on ocean data described in this embodiment includes a data acquisition module, an anomaly detection module, a feature extraction module, an accuracy assessment module, a time period classification module and a path optimization module; Data acquisition module: perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over a certain period of time; Abnormal detection module: Filter the collected sound velocity distribution data and conductivity data and compare them with the benchmark data, and mark the abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; Feature extraction module: extract features from the marked abnormal sound velocity distribution data and abnormal conductivity data, and respectively extract abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; Accuracy evaluation module: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; Time period classification module: according to the accuracy of the sensor monitoring data in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; Path optimization module: For abnormal monitoring time periods, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted. If the abnormal degree is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

[0064] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0065] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0066] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0067] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An intelligent optimization method for unmanned boat navigation paths based on ocean data, characterized in that: The following steps are involved: S1: Perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over several time periods; S2: filtering the collected sound velocity distribution data and conductivity data and comparing them with the benchmark data, marking abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; S3: extracting features from the marked abnormal sound velocity distribution data and abnormal conductivity data, respectively extracting abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; S4: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; S5: according to the accuracy of the monitoring data of the sensor in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; S6: For abnormal monitoring time periods, the degree of abnormality in the accuracy of the sensor's monitoring data in the marine environment is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

2. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 1 is characterized in that: In S2, a low-pass filter is applied to smooth the sound velocity and conductivity data, and the sound velocity and conductivity data after filtering are normalized and converted into relative values ​​relative to the benchmark data. According to the distribution characteristics of the benchmark data, the allowable normal fluctuation range is set; for the sound velocity and conductivity data at each time point, the deviation value relative to the benchmark data is calculated, and the expression is: ; Compare the deviation value with the preset normal fluctuation range, mark all time point data that exceed the normal range, and mark the data points that deviate from the normal range as abnormal data.

3. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 1 is characterized in that: In S3, after analyzing the change rate drift of the abnormal sound velocity value in the extracted sound velocity distribution abnormality feature, a sound velocity value change rate drift index is generated. The method for obtaining the sound velocity value change rate drift index is: Select the sound speed change rate sequence under normal conditions as the reference sequence , the sound velocity change rate data collected in the current time period is used as the sequence to be tested , define the distance between any two points in the reference sequence and the sequence to be tested, use the Euclidean distance to measure the difference between the two points, and calculate the distance between each pair of points , the expression is: ; Construct an N×M matrix D, where the element D(i,j) in the matrix represents the distance between the i-th point of the reference sequence and the j-th point of the sequence to be tested, that is: D(i,j)=d(i,j); construct a cumulative distance matrix C to store the minimum path distance from the beginning of the sequence to the current point; set the cumulative distance of the starting point to the starting point distance, that is, C(1,1)=D(1,1); starting from the upper left corner of the matrix, gradually calculate the minimum cumulative distance of each position. The element C(i,j) in the cumulative distance matrix represents the minimum cumulative distance from the starting point (1,1) to the point (i,j), which is calculated by the recursive formula: ; The last element C(N,M) of the cumulative distance matrix C represents the total cumulative distance between the reference sequence and the sequence to be tested, that is, the drift index of the sound speed change rate is calculated, and the expression is: ; Where HK is the drift index of the sound velocity change rate.

4. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 3 is characterized in that: In S3, after analyzing the deviation amplitude of the conductivity abnormal value in the extracted conductivity fluctuation feature relative to the reference conductivity value, a conductivity fluctuation amplitude deviation index is generated. The method for obtaining the conductivity fluctuation amplitude deviation index is: Select the conductivity time series under normal conditions As a benchmark, collect the conductivity time series for the current time period ; The Bayesian change point detection probability model is selected to describe the data distribution of the conductivity time series, and the conductivity fluctuation is assumed to obey the Gaussian distribution. The expression is: ;in, and are the mean and variance of the conductivity values, respectively. A prior distribution is set for the change point position. The likelihood function is calculated based on the currently observed conductivity data to update the probability distribution of the change point. According to the Bayesian formula, the probability of the change point at time point i is calculated. , for each time point i, the updated change point probability is: ;in, represents the time series subset from the starting point to i, P(T) is the normalization constant, and the significant change point positions are screened out ; According to the detected change points, the sequence T to be tested is divided into k+1 sub-intervals, and the average fluctuation amplitude of the conductivity value is calculated in each sub-interval , the expression is: ;in, represents the length of subinterval j, and They represent the change point positions respectively. Using the segmented results of the change point detection, the average values ​​of the fluctuation amplitudes of all segments are summarized to obtain the conductivity fluctuation amplitude deviation index, which is expressed as: ; GH is the conductivity fluctuation amplitude deviation index.

5. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 4 is characterized in that: In S4, the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics are comprehensively analyzed to evaluate the accuracy of the sensor monitoring data in the marine environment in each time period, specifically: The sound velocity change rate drift index and the conductivity fluctuation amplitude deviation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy value label of the sensor's monitoring data in the marine environment in each time period as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy value labels of the sensor's monitoring data in the marine environment in all time periods as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy value of the sensor's monitoring data in the marine environment in each time period is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

6. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 5 is characterized in that: In S5, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods according to the accuracy of the monitoring data of the sensor in the marine environment in each time period; The acquired accuracy value of the sensor's monitoring data in the marine environment in each time period is compared with the reference threshold of the accuracy value under normal conditions set according to historical data. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is greater than or equal to the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is high. In this case, no warning signal is generated, and the corresponding time period is divided into a normal monitoring time period. If the accuracy value of the sensor's monitoring data in the marine environment in the time period is less than the reference threshold of the accuracy value, it means that the accuracy of the sensor's monitoring data in the marine environment in the time period is low. In this case, a warning signal is generated, and the corresponding time period is divided into an abnormal monitoring time period.

7. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 1 is characterized in that: In S6, for the abnormal monitoring period, the degree of abnormality of the accuracy of the sensor's monitoring data in the marine environment is predicted. If the degree of abnormality is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path, specifically: For abnormal monitoring time periods, that is, the accuracy value of the sensor's monitoring data in the marine environment during the corresponding time period is less than the reference threshold of the accuracy value, the accuracy values ​​of the monitoring data that are less than the reference threshold of the accuracy value in the subsequent time period are collected, and a data set is established. The mean and standard deviation in the data set are calculated, and after analysis, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted based on the analysis results.

8. The method for intelligent optimization of navigation paths of unmanned boats based on ocean data according to claim 7 is characterized in that: If the mean accuracy value in the data set is greater than or equal to the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is less than the reference threshold of the standard deviation of the accuracy value, it means that the overall deviation of the sensor's monitoring data is small and the fluctuation is low, indicating that the sensor has returned to a stable state, and the avoidance buffer distance of the unmanned boat is gradually reduced and the normal speed is restored; If the mean accuracy value is greater than or equal to the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, it means that although the average accuracy of the data is high, the fluctuation is large, indicating that the sensor is in a recovery state but unstable, and it is necessary to continue to maintain the avoidance strategy and closely monitor the data fluctuation of the sensor; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is greater than or equal to the reference threshold of the standard deviation of the accuracy value, the overall accuracy of the sensor data is low and the fluctuation range is large. A high buffer distance and low speed should be maintained, risk avoidance should be prioritized, and the sensor should be inspected and replaced; If the mean accuracy value is less than the reference threshold of the mean accuracy value, and the standard deviation of the accuracy value is less than the reference threshold of the standard deviation of the accuracy value, it means that the average accuracy of the data is low but the fluctuation is small. Maintain the avoidance strategy and implement emergency maintenance or stop using the sensor data.

9. An unmanned boat navigation path intelligent optimization system based on ocean data, used to implement the unmanned boat navigation path intelligent optimization method based on ocean data according to any one of claims 1 to 8, characterized in that: It includes data collection module, anomaly detection module, feature extraction module, accuracy assessment module, time period classification module and path optimization module; Data acquisition module: perform preliminary calibration on all relevant sensors on the unmanned boat to ensure that their initial status is normal, record baseline data, and collect sound velocity distribution data and conductivity data collected in real time by the unmanned boat in the designated sea area over a certain period of time; Abnormal detection module: Filter the collected sound velocity distribution data and conductivity data and compare them with the benchmark data, and mark the abnormal sound velocity distribution data and conductivity data that deviate from the normal range in each time period; Feature extraction module: extract features from the marked abnormal sound velocity distribution data and abnormal conductivity data, and respectively extract abnormal sound velocity distribution features from the abnormal sound velocity distribution data and conductivity fluctuation features from the abnormal conductivity data; Accuracy evaluation module: Comprehensively analyze the extracted sound velocity distribution anomaly characteristics and conductivity fluctuation characteristics to evaluate the accuracy of the sensor's monitoring data in the marine environment during each time period; Time period classification module: according to the accuracy of the sensor monitoring data in the marine environment in each time period, all time periods are divided into normal monitoring time periods and abnormal monitoring time periods; Path optimization module: For abnormal monitoring time periods, the abnormal degree of the accuracy of the sensor's monitoring data in the marine environment is predicted. If the abnormal degree is high, a safety avoidance strategy is set through the path planning system to continuously optimize the buffer distance and speed of the unmanned boat deviating from the original path.

Images