A kayak hull posture monitoring method and monitoring system

Abstract

The present invention discloses a kayak hull posture monitoring method and monitoring system, which specifically relates to the field of kayak posture monitoring, and is used to solve the accuracy and reliability problems of kayak sensor installation looseness and structural health monitoring. The method is to install sensors at key positions of kayaks, collect status data in real time, use filtering and navigation technology to monitor sensor position and distance change trends, and use linear regression and abnormal score analysis to warn of looseness and structural changes. Construct a feature matrix and factor decomposition machine model, dynamically adjust the noise suppression level, optimize model parameters, generate and compare the vibration stability index and threshold, evaluate the looseness of the sensor and provide an early warning. After confirming that the installation is stable, record and preprocess the impact data, construct a finite element model, calculate the elastic deformation of the hull, analyze the deformation mode and change rate, evaluate the structural health status and provide timely early warning to ensure the safety and service life of the kayak.

Description

Technical Field

[0001] The present invention relates to the field of kayak posture monitoring, and more specifically, to a kayak hull posture monitoring method and a monitoring system. Background Art

[0002] Kayak posture monitoring refers to the real-time monitoring of the kayak's posture changes on the water surface by installing multiple sensors such as gyroscopes, accelerometers and GPS at specific locations on the kayak. These sensors construct a kayak posture diagram through a virtual coordinate system to provide accurate real-time data to help users better understand the kayak's motion status. Through such a posture monitoring system, kayakers can obtain detailed data about their motion posture and kayak stability during training and competitions, so as to make targeted adjustments and improvements to improve their athletic performance.

[0003] However, the existing kayak attitude monitoring system has some shortcomings in practical applications. First, in an outdoor rafting environment, due to the unpredictability of the natural environment, kayaks are often subjected to various impacts and shocks. These impacts not only easily cause the installation of sensors to loosen, resulting in distortion of attitude monitoring data, but may also cause structural damage to the kayak itself. In addition, the current attitude monitoring system mainly focuses on monitoring the attitude changes of the kayak, and the monitoring of the sensor installation quality and the early warning function of the hull damage are relatively weak. This means that in the event of an unexpected collision, the system cannot detect and feedback the status of the sensor and the damage to the hull in a timely manner, which may cause the user to be unable to take necessary repair and adjustment measures in time, thereby increasing safety hazards. Therefore, it is necessary to develop a comprehensive monitoring system that can simultaneously monitor the kayak attitude and hull damage based on the existing attitude monitoring system to improve the safety and reliability of use.

[0004] In order to solve the above problems, a technical solution is now provided. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a kayak hull posture monitoring method and monitoring system, by installing sensors at multiple key positions of the kayak, collecting status data in real time, and using extended Kalman filtering and inertial navigation systems, combined with time window and sliding window technology, accurately monitoring the three-dimensional position and distance change trend of the sensor. By linear regression analysis of the slope and calculation of the abnormal score, the loose installation and structural changes of the sensor are detected and warned in time. Based on the abnormal data, a feature matrix is ​​constructed and the factor decomposition machine model parameters are initialized to capture high-order feature interactions, dynamically adjust the noise suppression level, optimize the model parameters, generate a vibration stability index, and evaluate the overall looseness of the sensor. By comparing the vibration stability index with the steady-state threshold, loose sensor installation or structural problems are identified, and looseness warnings are provided to ensure the reliability of the posture monitoring system. After confirming that the sensor is installed firmly, the impact data is recorded and preprocessed in real time, a finite element model is constructed, the elastic deformation of the hull is calculated through numerical simulation, the deformation pattern is extracted and time series analysis is performed, the deformation pattern change rate and acceleration are predicted, the structural deterioration trend index is calculated and compared with the structural damage threshold, the structural health status is evaluated, and structural problems are warned in time to ensure the safety and service life of the kayak, so as to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] S1, fixed sensors are installed at multiple key locations of the kayak. The sensors collect data in real time, and after preprocessing and verification, a benchmark model is constructed as a reference standard for posture monitoring and damage warning;

[0008] S2, after the benchmark model is built, each sensor collects the kayak status data in real time and pre-processes it, then converts the data into a unified coordinate system, and generates comprehensive posture and motion status information through a multi-sensor data fusion algorithm;

[0009] S3, using the data preprocessed by S2, calculates the three-dimensional position coordinates of each sensor in real time, uses the time window technology to monitor the distance change rate between sensors, and uses the sliding window technology to analyze the distance change trend. Combined with the slope and anomaly score, it is determined whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly;

[0010] S4, based on the abnormal data obtained in step S3, construct a feature matrix, initialize the factor decomposition machine model parameters, analyze the data by calculating feature interaction terms and adaptive filters, optimize the model parameters and calculate the parameter volatility, and judge the installation stability of the sensor by combining the slope and the overall distance change amplitude;

[0011] S5, after confirming that the sensor is firmly installed, starts to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes and uses wavelet transform analysis; simulates elastic deformation through finite element model, extracts deformation mode through principal component analysis, calculates the change rate and prediction error through time series analysis, judges the health status of the kayak structure and triggers an early warning.

[0012] In a preferred embodiment, step S3 includes the following contents:

[0013] S3.1, obtain the real-time position coordinates of each sensor (x i ,y i ,z i ), the acquisition process is as follows:

[0014] S3.1.1, determining the initial position coordinates of each sensor according to the reference model of step S1;

[0015] S3.1.2, each sensor sends real-time data to the central processing unit via wireless transmission;

[0016] S3.1.3, after receiving the sensor data, the central processing unit updates the three-dimensional position coordinates of each sensor in real time through the data transmission link between sensors and its own data calculation;

[0017] Using the real-time data from the gyroscope and accelerometer, the relative motion trajectory of each sensor is calculated through the inertial navigation system;

[0018] Using the relative distance data between sensors, the position of each sensor is further corrected through triangulation and real-time calibration algorithms;

[0019] Using the distance information between multiple sensors, nonlinear least square method is applied to perform position calibration to ensure the accuracy of position calculation;

[0020] S3.1.4, through the sensor data transmission link, sensors regularly exchange position information and motion status;

[0021] S3.1.5, using extended Kalman filtering for real-time error correction.

[0022] In a preferred embodiment, S3.2, according to step S3.1, the real-time position coordinates (x i ,y i ,z i ), where i represents the sensor number;

[0023] S3.3, initialize the time window T, which is used to store the sensor position data within a period of time. Let the time window be [t0,t n ];

[0024] S3.4, within the time window T, use the Euclidean distance to calculate each moment t k (t0≤t k ≤t n ) Calculate the distance between sensor i and sensor j;

[0025] S3.5, calculate the distance change rate Δd within the time window ij (t k ):

[0026]

[0027] Store the distance change rate at each moment in the time series array {Δdij(tk)};

[0028] S3.6, within the time window T, a sliding window technique is used, each sliding window contains M time points, and a statistical analysis is performed on the distance change rate within each sliding window; where p represents the sequence number of the sliding window;

[0029] Perform linear regression on the distance change rate {Δdij(tk)} within each sliding window to obtain the slope m and intercept b. The slope represents the trend of distance change.

[0030] S3.7, calculate each time point t k The anomaly score α(t k ):

[0031] Among them, μ Δd and σ Δd are the mean and standard deviation of the distance change rate, respectively;

[0032] If the anomaly score exceeds the anomaly threshold, it is determined to be an abnormal change and the abnormal data is recorded.

[0033] In a preferred embodiment, step S4 includes the following contents:

[0034] S4.1, based on the abnormal data obtained in step S3, obtain the actual location data set and historical data of each sensor to form a feature matrix X;

[0035] Initialize the factorization machine model parameters, including feature weights ω and factor matrix V;

[0036] S4.2, for each sensor data point xi, calculate the feature interaction term:

[0037]

[0038] Where: ω0 represents the bias term of the model; ωo represents the linear weight of feature o; o represents the index of the oth feature; v represents the index of the vth feature; <V o ,V v > represents the factor vector V o and V v The inner product of

[0039] S4.3 uses the noise suppression factor η to dynamically adjust the noise suppression level by analyzing the time series of sensor data and using adaptive filtering to estimate and eliminate noise:

[0040]

[0041] Where: m″ represents the order of the noise suppression filter; k represents the number of time delay steps or the index of past data points, k traverses from 1 to m″; η(t) is the noise suppression factor at the current time t; η0 is the initial noise suppression factor, which is a constant used to provide basic noise suppression effect;

[0042] αk is the filter coefficient, which represents the contribution weight of the previous k historical data points to the current noise suppression factor; x(tk) represents the sensor data value at the previous k moments;

[0043] S4.4, use stochastic gradient descent to optimize the model parameters, with the goal of minimizing the loss function.

[0044] In a preferred embodiment, S4.5, the vibration stability index is obtained by combining the slope, the characteristics of the parameter volatility and the amplitude of the overall distance change:

[0045]

[0046] Where: S represents the vibration stability index; |g| represents the absolute value of the slope, indicating the trend strength of the sensor distance change; σ ω represents the standard deviation of the feature weight ω, indicating the volatility of the weight; σ V represents the standard deviation of the factor vector in the factor matrix V, indicating the volatility of the factor; f1 represents the weight coefficient of |g|; f2 represents The weight coefficient of The weight coefficient of

[0047] The vibration steady-state index is compared with the steady-state threshold: if the vibration steady-state index is greater than or equal to the steady-state threshold, there is a significant loose installation problem and a deterioration trend signal is generated; conversely, if the vibration steady-state index is less than the steady-state threshold, there is no significant loose problem and an installation stability signal is generated.

[0048] In a preferred embodiment, step S5 includes the following contents:

[0049] S5.1, after obtaining the installation stability signal, when the sensor detects the impact, the relative distance change data between the sensors is recorded in real time, including the sensor number and time information;

[0050] S5.2, preprocessing the recorded sensor data, and then using wavelet transform technology to decompose the data into detail signals and approximate signals of multiple different scales;

[0051] S5.3, construct a finite element model of the kayak based on its physical characteristics, and divide the kayak into a number of finite element units; use the real-time recorded sensor data as boundary conditions to input into the finite element model, and calculate the elastic deformation of the hull after the impact through numerical simulation;

[0052] S5.4, calculate the dynamic response of the kayak after the impact, perform dimensionality reduction processing through principal component analysis method, and extract the main deformation mode:

[0053] Calculate the dynamic response of the kayak after being hit and obtain the deformation field at each moment;

[0054] Extract the main modes in the deformation field and obtain the main deformation modes through dimensionality reduction processing through principal component analysis;

[0055] S5.5, perform time series analysis on the extracted deformation pattern, calculate its rate of change and acceleration, predict the deformation pattern through an autoregressive moving average model, and calculate the prediction error;

[0056] S5.6, calculating the cumulative change of the deformation mode and the standard deviation of the prediction error, and calculating the structural deterioration trend index by the ratio of the cumulative change to the standard deviation of the prediction error;

[0057] S5.7, compare the structural deterioration trend index with the structural damage threshold: if the structural deterioration trend index is greater than or equal to the structural damage threshold, an early warning reminder is issued; conversely, if the structural deterioration trend index is less than the structural damage threshold, it indicates that the structural health condition is good.

[0058] A kayak hull attitude monitoring system, comprising: an initial benchmark modeling module, a multi-dimensional attitude fusion module, a time sequence position analysis module, a parameter steady-state evaluation module and a dynamic deformation warning module;

[0059] The initial benchmark modeling module installs fixed sensors at multiple key locations on the kayak. The sensors collect data in real time and build a benchmark model after preprocessing and verification. This model serves as a reference standard for attitude monitoring and damage warning. The initial position and state data of the sensors are transmitted to the multi-dimensional attitude fusion module.

[0060] After the benchmark model is built, the multi-dimensional posture fusion module collects the kayak status data in real time and pre-processes it. The data is then converted into a unified coordinate system. The comprehensive posture and motion status information is generated through the multi-sensor data fusion algorithm. The comprehensive posture and motion status information is transmitted to the time series position analysis module.

[0061] The time series position analysis module calculates the three-dimensional position coordinates of each sensor in real time through the preprocessed data, monitors the distance change rate between sensors using the time window technology, and analyzes the distance change trend through the sliding window technology. It combines the slope and the anomaly score to determine whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly, and the recorded abnormal data is passed to the parameter steady-state evaluation module;

[0062] The parameter steady-state assessment module constructs a feature matrix based on the acquired abnormal data, initializes the factor decomposition machine model parameters, analyzes the data by calculating feature interaction terms and adaptive filters, optimizes the model parameters and calculates the parameter volatility, and judges the installation stability of the sensor by combining the slope and the overall distance change amplitude. The installation stability of the sensor is transmitted to the dynamic deformation warning module;

[0063] After confirming that the sensor is firmly installed, the dynamic deformation warning module begins to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes it and uses wavelet transform analysis; it simulates elastic deformation through the finite element model, extracts deformation patterns using principal component analysis, calculates the rate of change and prediction error through time series analysis, determines the health of the kayak structure and triggers an early warning.

[0064] The technical effects and advantages of a kayak hull posture monitoring method and monitoring system of the present invention are as follows:

[0065] 1. The present invention accurately monitors the real-time three-dimensional position coordinates and status information of the sensor by installing sensors at multiple key positions of the kayak and collecting status data in real time. By using the extended Kalman filter and inertial navigation system, combined with the time window and sliding window technology, the distance change trend between sensors is monitored in real time, and the slope is analyzed by linear regression to determine whether the distance change is abnormal. By calculating the abnormal score and recording the abnormal data, the loose installation of the sensor and the structural changes can be detected and warned in time. It provides effective support for the posture monitoring and structural health assessment of the kayak. It can reflect the dynamic response of the kayak in a complex environment in real time, ensure the accuracy and reliability of the monitoring results, and thus improve the safety and service life of the kayak. It can not only optimize the installation stability of the sensor, but also effectively warn of structural hidden dangers, avoid misjudgment caused by distortion of sensor data, and ensure the safe operation of users in various drifting environments.

[0066] 2. The present invention can accurately capture high-order feature interactions in sensor data by constructing a feature matrix based on abnormal data and initializing the factor decomposition machine model parameters. By calculating the feature interaction terms and using an adaptive filter to analyze the data, the noise suppression level is dynamically adjusted to ensure the accuracy and consistency of the data. The model parameters are optimized using the stochastic gradient descent method, and the parameter volatility is calculated to determine the stability of the sensor installation position. Combining the slope, parameter volatility and overall distance change amplitude, a vibration stability index is generated to comprehensively evaluate the overall looseness of the sensor. The vibration stability index can effectively identify loose installation or structural problems of the sensor by comparing it with the steady-state threshold, and timely generate a deterioration trend signal or a stable installation signal. Thereby, not only can the installation stability of the sensor be comprehensively analyzed, an effective loosening warning be provided, but also the reliability and accuracy of the kayak posture monitoring system can be ensured. In practical applications, by dynamically monitoring and evaluating the state of the sensor, potential risks can be identified in advance, monitoring errors caused by loose sensors or inaccurate data can be avoided, and the safety of kayaking and the accuracy of operation can be improved. It provides a solid technical guarantee for the safe monitoring of kayaks in complex environments.

[0067] 3. Under the premise of judging that the installation of the sensor is stable, the present invention analyzes the dynamic response of the kayak after the impact by real-time recording and preprocessing the sensor data when the sensor detects the impact. After decomposing the signal using wavelet transform technology, a finite element model based on the physical characteristics of the kayak is constructed, and the elastic deformation of the hull is calculated by numerical simulation, the main deformation mode is extracted and time series analysis is performed. The rate of change and acceleration of the deformation mode are predicted by the autoregressive sliding average model, the structural deterioration trend index is calculated, and compared with the structural damage threshold to judge the structural health of the kayak. Then, the stability and consistency of the relative distance change between the sensors are evaluated, the risk of structural damage is effectively identified, and the accuracy and reliability of the monitoring system are ensured. Through real-time monitoring of the structural deterioration trend index, structural problems of the kayak can be warned in time, potential safety hazards can be avoided, and the safety and life of the kayak can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A schematic diagram of a process of monitoring a kayak hull posture according to the present invention;

[0069] Figure 2 The present invention is a structural schematic diagram of a kayak hull posture monitoring system. DETAILED DESCRIPTION

[0070] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0071] Example 1

[0072] Figure 1 The present invention provides a kayak hull posture monitoring method, comprising:

[0073] S1, fixed sensors are installed at multiple key locations of the kayak. The sensors collect data in real time, and after preprocessing and verification, a benchmark model is constructed as a reference standard for posture monitoring and damage warning;

[0074] S2, after the benchmark model is built, each sensor collects the kayak status data in real time and pre-processes it, then converts the data into a unified coordinate system, and generates comprehensive posture and motion status information through a multi-sensor data fusion algorithm;

[0075] S3, using the data preprocessed by S2, calculates the three-dimensional position coordinates of each sensor in real time, uses the time window technology to monitor the distance change rate between sensors, and uses the sliding window technology to analyze the distance change trend. Combined with the slope and anomaly score, it is determined whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly;

[0076] S4, based on the abnormal data obtained in step S3, construct a feature matrix, initialize the factor decomposition machine model parameters, analyze the data by calculating feature interaction terms and adaptive filters, optimize the model parameters and calculate the parameter volatility, and judge the installation stability of the sensor by combining the slope and the overall distance change amplitude;

[0077] S5, after confirming that the sensor is firmly installed, starts to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes and uses wavelet transform analysis; simulates elastic deformation through finite element model, extracts deformation mode by principal component analysis, calculates change rate and prediction error through time series analysis, obtains structural deterioration trend index and compares it with structural damage threshold, judges the health status of kayak structure and triggers warning.

[0078] Step S1 includes the following contents:

[0079] S1.1, install fixed attitude monitoring sensors at multiple key locations of the kayak (such as the bow, stern, left and right sides, etc.). These sensors include but are not limited to:

[0080] Gyroscope: Used to measure the angular velocity of the kayak.

[0081] Accelerometer: Used to measure the acceleration of the kayak.

[0082] GPS module: used to provide the kayak's geographic location and movement speed data.

[0083] Position sensor: used to monitor the relative position and distance between sensors.

[0084] During the installation of the sensor, it is necessary to ensure that the position of each sensor is fixed and stable to avoid loosening or displacement. Use waterproof and vibration-proof materials for protection during installation to adapt to the harsh conditions of the outdoor drifting environment.

[0085] S1.2, after the sensor is installed, the data collection process is started. Each sensor collects the status data of its location in real time, including location, posture and environmental parameters. The data of each sensor will be transmitted to the central processing unit through the low-power Bluetooth (BLE) transmission module.

[0086] S1.3, build a benchmark model. The benchmark model is the reference standard of the entire monitoring system and is used for subsequent posture monitoring and damage warning. The construction process of the benchmark model is as follows:

[0087] S1.3.1, the system performs initial data collection in a static state after the sensor is installed. The initial data collected by each sensor includes its relative position, posture data and respective status information. These data are transmitted to the central processing unit via the BLE module;

[0088] S1.3.2, the central processing unit pre-processes the received initial data, including data denoising, filtering and correction. The denoising process uses a signal processing algorithm (such as the Kalman filter algorithm) to reduce the interference of environmental noise on the data and ensure the accuracy of the data;

[0089] S1.3.3, using the pre-processed data, the central processing unit calculates the relative position and distance between each sensor. The relative coordinate position of each sensor is determined by a triangulation positioning algorithm (such as ranging triangulation positioning method);

[0090] S1.3.4, based on the calculated relative position and distance data, the central processing unit constructs a three-dimensional virtual coordinate system as a reference model of the kayak. The core elements of the reference model include:

[0091] Sensor position vector: the initial position of each sensor in the three-dimensional coordinate system.

[0092] Sensor state matrix: includes initial state information such as the sensor's attitude, angular velocity, and acceleration.

[0093] Environmental parameters: Environmental factors such as temperature and humidity that may affect sensor data.

[0094] S1.5, to ensure the accuracy and reliability of the benchmark model, multiple static and dynamic verifications will be conducted. Static verification includes repeatedly collecting initial data under different environmental conditions and comparing it with the benchmark model; dynamic verification includes collecting real-time data during the actual drift process and dynamically comparing it with the benchmark model to verify the stability of the model.

[0095] S1.6, after verification, the benchmark model will be stored in the system's storage module as a reference standard for subsequent posture monitoring and damage warning. In actual application, the system will continuously update and optimize the benchmark model to adapt to environmental changes and changes in sensor status.

[0096] Step S2 includes the following contents:

[0097] After the benchmark model is built, each sensor starts to collect the kayak's status data in real time, including the angular velocity data of the gyroscope, the linear acceleration data of the accelerometer, the geographic location and motion speed data of the GPS module, and the relative position and distance change data of the position sensor. Each sensor sends the real-time data to the central processing unit through the wireless transmission module. After receiving the data, the central processing unit first performs data verification to ensure the integrity and time synchronization of the data packet; then performs data denoising, and removes environmental noise and high-frequency interference signals through low-pass filters and Kalman filters; then performs data correction to correct the sensor's system error and zero offset, and converts the data into a unified virtual coordinate system; then performs multi-sensor data fusion, using the extended Kalman filter (EKF) or unscented Kalman filter (UKF) algorithm to fuse the data of different sensors to generate comprehensive posture and motion status information; finally, the preprocessed data is stored in the temporary buffer and historical data storage module to ensure the real-time, continuity and convenience of subsequent playback analysis of the data. Through these preprocessing steps, the central processing unit can provide accurate, reliable and consistent data, providing a basis for posture monitoring and damage warning.

[0098] Step S3 includes the following contents:

[0099] S3.1, in order to realize the dynamic monitoring of the distance change trend between sensors, it is necessary to first obtain the real-time position coordinates (x i ,y i ,z i ), the acquisition process is as follows:

[0100] S3.1.1, determine the initial position coordinates of each sensor based on the reference model of step S1, and store them in the reference model of the system. Initial position coordinates (xi0 ,y i0 ,z i0 ) is obtained through precise measurement and calibration.

[0101] S3.1.2, each sensor sends real-time data to the central processing unit through the wireless transmission module. These data include the current acceleration, angular velocity and other necessary environmental parameters of each sensor.

[0102] S3.1.3, after receiving the sensor data, the central processing unit adopts a multi-sensor data fusion algorithm to update the three-dimensional position coordinates of each sensor in real time through the data transmission link between sensors and its own data calculation;

[0103] Using the real-time data from the gyroscope and accelerometer, the relative motion trajectory of each sensor is calculated through the inertial navigation system (INS). The core of the inertial navigation system is to integrate acceleration and angular velocity to obtain displacement and attitude changes.

[0104]

[0105] Among them, vi(t) is the velocity; ai(t) is the acceleration; xi(tk) is the position at time tk.

[0106] Using the relative distance data between sensors, the position of each sensor is further corrected through triangulation and real-time calibration algorithm. Assuming the distance between sensor i and sensor j is dij, then:

[0107] In order to utilize the distance information between multiple sensors in the real-time calculation process, the nonlinear least square method is applied to perform position calibration to ensure the accuracy of position calculation.

[0108] S3.1.4, by establishing a sensor data transmission link, sensors regularly exchange position information and motion status. Each sensor not only transmits its own data, but also transmits data received from other sensors to form a chain network. When receiving data, the central processing unit uses the information in the link to perform comprehensive calculations to enhance the reliability of position calculations.

[0109] By establishing a sensor data transmission link, sensors regularly exchange position information and motion status. Each sensor not only transmits its own data, but also transmits data received from other sensors to form a chain network. When receiving data, the central processing unit first performs time synchronization and integrity verification on the data from each sensor to ensure that the data is consistent on the same time basis. Then, the central processing unit uses a multi-sensor fusion algorithm to integrate the position information of each sensor, fuses the data through a Kalman filter, eliminates noise and errors, and uses a nonlinear least squares method to accurately calibrate the relative distance between each sensor. The central processing unit also integrates the motion trajectory of the sensor in combination with the data of the inertial navigation system to further optimize the real-time position coordinates of each sensor. Through the above comprehensive calculation process, the central processing unit can accurately determine the real-time three-dimensional position coordinates of each sensor, improving the accuracy and reliability of position calculation.

[0110] S3.1.5, in order to reduce the cumulative error in position calculation, the central processing unit uses the extended Kalman filter (EKF) for real-time error correction. EKF can effectively integrate inertial navigation data and relative position data, correct position errors, and improve calculation accuracy.

[0111] Through steps S3.1.1 to S3.1.5, the real-time position coordinates of each sensor can be calculated by the inertial navigation system, the relative position calibration between sensors, and the error correction and filtering to obtain the data transmission link.

[0112] S3.2, according to step S3.1, obtain the real-time position coordinates (x i ,y i ,z i ), where i represents the sensor number.

[0113] S3.3, initialize the time window T, which is used to store the sensor position data within a period of time. Let the time window be [t0,t n ];

[0114] S3.4, within the time window T, use the Euclidean distance to calculate each moment t k (t0≤t k ≤t n ) Calculate the distance d between sensor i and sensor j ij (t k ).

[0115] S3.5, calculate the distance change rate Δd within the time window ij (t k ):

[0116]

[0117] The distance change rate at each moment is stored in the time series array {Δdij(tk)}.

[0118] S3.6, within the time window T, a sliding window technique is used, each sliding window contains M time points, and a statistical analysis is performed on the distance change rate within each sliding window.

[0119] Where p represents the sequence number of the sliding window.

[0120] Linear regression is performed on the distance change rate {Δdij(tk)} within each sliding window to obtain the slope m and intercept b. The slope represents the trend of distance change.

[0121] Based on the linear regression results, determine whether the slope deviates significantly from zero. If the slope exceeds the trend threshold, it indicates that there is a trend of worsening distance change.

[0122] If the slope is greater than zero, the distance between the sensors is increasing. This could indicate that the sensor mounting has become loose or that the kayak structure has expanded under the forces.

[0123] If the slope is less than zero, it means the distance between the sensors is decreasing. This could mean the kayak structure is contracting under the forces or the sensors are moving closer together.

[0124] If the slope is approximately zero, it means that the distance between the sensors has not changed significantly and the distance has remained relatively stable, which means that the sensor mounting locations and the kayak structure have not changed significantly.

[0125] S3.7, calculate each time point t k The anomaly score α(t k ):

[0126] Among them, μ Δd and σ Δd are the mean and standard deviation of the distance change rate, respectively.

[0127] If the anomaly score exceeds the anomaly threshold, it is determined to be an abnormal change and the abnormal data is recorded.

[0128] For each time t when an abnormal change is detected k ,The sensor’s real-time position, distance change rate, anomaly score and other data are stored in the anomaly record database.

[0129] The present invention accurately monitors the real-time three-dimensional position coordinates and status information of the sensor by installing sensors at multiple key positions of the kayak and collecting status data in real time. By using the extended Kalman filter and the inertial navigation system, combined with the time window and sliding window technology, the distance change trend between the sensors is monitored in real time, and the slope is analyzed by linear regression to determine whether the distance change is abnormal. By calculating the abnormal score and recording the abnormal data, the loose installation of the sensor and the structural change can be detected and warned in time. It provides effective support for the posture monitoring and structural health assessment of the kayak. It can reflect the dynamic response of the kayak in a complex environment in real time, ensure the accuracy and reliability of the monitoring results, and thus improve the safety and service life of the kayak. It can not only optimize the installation stability of the sensor, but also effectively warn of structural hidden dangers, avoid misjudgment caused by distortion of sensor data, and ensure the safe operation of users in various drifting environments.

[0130] Step S4 includes the following contents:

[0131] Analyzing the overall looseness of the sensors in the kayak is of great significance for kayak posture monitoring, because the looseness of the sensors will cause distortion of the posture monitoring data, affecting the analysis and control of the kayak's motion state on the water. If the looseness of the sensors is not discovered and repaired in time, it may cause erroneous data transmission, which in turn affects the judgment of the kayak's motion trajectory and stability, and increases the risk of accidents in complex water environments. In addition, the analysis of the overall looseness can provide early warning to prevent the sensors from completely failing during intense exercise or impact, thereby ensuring that the monitoring system can continue to provide accurate posture data, helping users to adjust the kayak's operating strategy in a timely manner and improve sports performance and safety. By comprehensively evaluating and monitoring the stability of all sensors, the reliability and accuracy of the kayak posture monitoring system can be ensured, providing users with a more efficient and safe operating experience.

[0132] S4.1, based on the abnormal data obtained in step S3, obtain the actual location data set and historical data of each sensor to form a feature matrix X.

[0133] Initialize the factorization machine model parameters, including feature weights ω and factor matrix V.

[0134] Factorization machine is a model suitable for sparse data and can capture high-order feature interactions. Based on FM, this algorithm introduces an adaptive noise suppression mechanism to reduce the errors caused by sensor noise and environmental changes through iterative optimization, thereby more accurately analyzing the stability of the sensor installation position.

[0135] S4.2, for each sensor data point xi, calculate the feature interaction term:

[0136]

[0137] in:

[0138] ω0 represents the bias term of the model;

[0139] ωo represents the linear weight of feature o;

[0140] o represents the index of the oth feature, and feature o is a specific feature in the sensor data, such as the position information, velocity information, or acceleration information of a sensor;

[0141] v represents the index of the vth feature. Feature v is also a specific feature in the sensor data, which is different from but related to feature o.

[0142] <V o ,V v > represents the factor vector V o and V v The inner product of .

[0143] S4.3, using the noise suppression factor η, dynamically adjusts the noise suppression level by analyzing the time series of sensor data. Use an adaptive filter (such as an LMS filter) to estimate and eliminate noise:

[0144]

[0145] in:

[0146] m″ represents the order of the noise suppression filter, or the number of historical data points used to calculate the noise suppression factor;

[0147] k represents the number of time delay steps or the index of past data points, and k traverses from 1 to m″;

[0148] η(t) is the noise suppression factor at the current time t;

[0149] η0 is the initial noise suppression factor, which is a constant used to provide basic noise suppression effect;

[0150] αk is the filter coefficient, which represents the contribution weight of the first k historical data points to the current noise suppression factor;

[0151] x(tk) represents the sensor data value at the previous k moments. These data are used to calculate the noise suppression factor at the current moment, reflecting the impact of historical data on the current noise level.

[0152] S4.4, use stochastic gradient descent (SGD) to optimize the model parameters, with the goal of minimizing the loss function.

[0153] By analyzing the optimized model parameters and calculating the parameter volatility, that is, the changes in the feature weight ω and the factor matrix V, the stability of the sensor installation position can be determined. If the model parameters fluctuate greatly over a period of time, there may be a risk of sensor loosening.

[0154] S4.5, in order to clarify whether the position of the sensor is at risk of loosening, the vibration stability index is obtained by combining the slope, parameter volatility characteristics and the amplitude of the overall distance change:

[0155]

[0156] The slope, parameter volatility characteristics, and the amplitude of the overall distance change respectively indicate different directions of the looseness of the sensor installation position: the slope reflects the trend of the distance change between sensors, indicating the displacement direction of the sensor in space, that is, the looseness of the sensor in a certain direction; the parameter volatility characteristics, including the standard deviation of the feature weight and the factor matrix, reflect the stability and consistency of the sensor data, indicating the amplitude of the frequent changes of the sensor in a short time, that is, the dynamic stability of the installation position; the amplitude of the overall distance change combines the cumulative effect of the distance change between all sensors, reflecting the overall intensity of the distance change between sensors, that is, the comprehensive looseness of the sensor in different directions. Through the combined analysis of these three parameters, the overall looseness of the kayak sensor can be comprehensively evaluated: the slope provides directional displacement information, the parameter volatility provides the stability information of the installation position, and the amplitude of the overall distance change provides the information of the comprehensive change intensity. This multi-dimensional analysis method is highly reasonable because it can comprehensively capture the stability and change trend of the sensor installation position from different angles, ensuring that the evaluation of the sensor looseness is both accurate and comprehensive, thereby effectively ensuring the reliability and accuracy of the kayak posture monitoring system.

[0157] in:

[0158] S represents the vibration stability index;

[0159] g| represents the absolute value of the slope, indicating the trend strength of the sensor distance change;

[0160] σ ω represents the standard deviation of the feature weight ω, indicating the volatility of the weight;

[0161] σ V represents the standard deviation of the factor vector in the factor matrix V, indicating the factor volatility;

[0162] It represents the absolute value and the rate of change of distance between all sensors, indicating the magnitude of the overall distance change;

[0163] n represents the number of sensors;

[0164] f1 represents the weight coefficient of |g|, which is used to adjust the contribution of the distance change trend (slope) between sensors to the vibration stability index;

[0165] f2 means The weight coefficient is used to adjust the contribution of feature weight and factor matrix volatility to the vibration stability index;

[0166] f3 means The weight coefficient is used to adjust the contribution of the comprehensive amplitude of the distance change rate between all sensors to the vibration stability index;

[0167] |g| reflects the trend strength of the distance change between sensors. The absolute value means that both increasing and decreasing trends are taken into account.

[0168] It reflects the standard deviation of feature weights and factor matrices, indicating the volatility of sensor data. The larger the standard deviation, the more unstable the sensor installation position.

[0169] It represents the comprehensive amplitude of the distance change between all sensors. The normalization process ensures the comparability of sensors with different numbers.

[0170] The vibration steady-state index is used to indicate the overall looseness of all sensors in a kayak. A larger index indicates more serious overall looseness of the sensors, reflecting a significant trend in the distance change between sensors, greater parameter volatility, and a larger overall change range. On the contrary, a smaller index indicates that the overall installation of the sensors is relatively stable, with an insignificant trend in the distance change between sensors, smaller parameter volatility, and a smaller overall change range. This index can be used to comprehensively evaluate the installation stability of all sensors in a kayak and provide an effective looseness warning.

[0171] Compare the vibration steady-state index to the steady-state threshold:

[0172] If the vibration steady-state index is greater than or equal to the steady-state threshold, it means that the overall looseness of the sensor is serious, and there may be significant installation looseness or structural problems. It is necessary to check and maintain the sensor installation position in time, and generate a deterioration trend signal; on the contrary, if the vibration steady-state index is less than the steady-state threshold, it means that the overall sensor is installed firmly, there is no significant looseness problem, the system can continue to operate stably, and no immediate maintenance measures are required, and a stable installation signal is generated. Through this comparison, the installation status of the sensor can be effectively judged to ensure the reliability and accuracy of the kayak attitude monitoring system.

[0173] After the sensor is firmly installed, it is necessary to analyze the structural deterioration trend of the kayak sensor, because even if the sensor is firmly installed, the kayak will still be subject to various external impacts and environmental influences during use, which may cause gradual damage and deformation of the hull structure. By analyzing the structural deterioration trend, potential problems of the kayak structure can be detected and warned in time, ensuring that the data provided by the sensor is accurate and reliable, thereby ensuring the normal operation of the kayak attitude monitoring system and the safety of users. Understanding the structural deterioration trend not only helps maintain and maintain the kayak and extend its service life, but also can detect and deal with structural hidden dangers in the early stage, avoid serious accidents, and improve overall safety and stability.

[0174] The present invention can accurately capture high-order feature interactions in sensor data by constructing a feature matrix based on abnormal data and initializing the factor decomposition machine model parameters. By calculating feature interaction terms and using an adaptive filter to analyze data, the noise suppression level is dynamically adjusted to ensure the accuracy and consistency of the data. The model parameters are optimized using the stochastic gradient descent method, and the parameter volatility is calculated to determine the stability of the sensor installation position. Combining the slope, parameter volatility and overall distance change amplitude, a vibration stability index is generated to comprehensively evaluate the overall looseness of the sensor. The vibration stability index can effectively identify loose installation or structural problems of the sensor by comparing it with the steady-state threshold, and timely generate a deterioration trend signal or a stable installation signal. Thereby, not only can the installation stability of the sensor be comprehensively analyzed, an effective loosening warning is provided, but also the reliability and accuracy of the kayak posture monitoring system can be ensured. In practical applications, by dynamically monitoring and evaluating the state of the sensor, potential risks can be identified in advance, monitoring errors caused by loose sensors or inaccurate data can be avoided, and the safety of kayaking and the accuracy of operation can be improved. It provides a solid technical guarantee for the safe monitoring of kayaks in complex environments.

[0175] Step S5 includes the following contents:

[0176] The reason why the structural safety of the kayak can only be analyzed after obtaining the signal that the sensor is firmly installed is that the stability of the sensor installation is a prerequisite for ensuring the accuracy and reliability of the data. If the sensor is not firmly installed, the sensor data will be affected by the change of the installation position, resulting in inaccurate measurement results and failure to reflect the actual structural state of the kayak. Only when the sensor is firmly installed can the sensor data truly reflect the external force and deformation of the kayak during drifting, so as to accurately analyze the structural safety of the kayak. Doing so can avoid misjudgment caused by loose sensors, ensure the effectiveness and reliability of the monitoring system, and provide accurate structural safety assessments to ensure the safety of kayak use and the accuracy of posture monitoring.

[0177] S5.1, after obtaining the installation stability signal, when the sensor detects the impact, the relative distance change data between the sensors is recorded in real time, including information such as sensor number and time;

[0178] S5.2, pre-process the recorded sensor data to remove noise and interference signals. Then use wavelet transform technology to decompose the data into multiple detail signals and approximate signals of different scales. Wavelet transform can decompose complex signals into multiple levels, making the signal analysis more refined and accurate;

[0179] S5.3, based on the physical characteristics of the kayak, construct its finite element model, divide the kayak into several finite element units, each unit has specific physical properties, such as elastic modulus and Poisson's ratio. Use the real-time recorded sensor data as boundary conditions to input into the finite element model, and calculate the elastic deformation of the hull after the impact through numerical simulation. Use explicit finite element method for calculation to improve calculation efficiency and accuracy;

[0180] S5.4, calculate the dynamic response of the kayak after the impact, perform dimensionality reduction processing through principal component analysis method, and extract the main deformation mode. Principal component analysis can convert high-dimensional data into low-dimensional data while retaining the main deformation information;

[0181] Calculate the dynamic response of the kayak after being hit and obtain the deformation field at each moment;

[0182] The main modes in the deformation field are extracted and the main deformation modes are obtained through dimensionality reduction processing using principal component analysis (PCA).

[0183] S5.5, perform time series analysis on the extracted deformation patterns and calculate their rate of change and acceleration. Use the autoregressive moving average model to predict the deformation patterns and calculate the prediction error. The autoregressive moving average model can capture the trends and periodic changes in the time series data, thus making accurate predictions;

[0184] S5.6, calculate the cumulative change of the deformation mode and the standard deviation of the prediction error, and calculate the structural deterioration trend index by the ratio of the cumulative change and the standard deviation of the prediction error. The cumulative change represents the overall change of the deformation mode over a period of time, and the standard deviation of the prediction error reflects the accuracy of the prediction. If the deterioration trend index exceeds the preset threshold, it is determined that there is a deterioration trend.

[0185] This structural deterioration trend index is used to indicate the structural health of the kayak after being impacted. It is specifically reflected in the stability and consistency of the change in the relative distance between sensors. The larger the index, the more serious the deterioration of the kayak structure, and there is a significant risk of deformation or damage; the smaller the index, the relatively stable kayak structure, without obvious deformation or damage, and the better the structural health.

[0186] S5.7, compare the structural deterioration trend index with the structural damage threshold:

[0187] If the structural deterioration trend index is greater than or equal to the structural damage threshold, it means that the structural health of the kayak after the impact has reached or exceeded the damage threshold, which means that the stability and consistency of the relative distance changes between sensors are seriously affected, and the kayak structure is at significant risk of deformation or damage, requiring immediate inspection and maintenance to ensure safety and normal use, and an early warning reminder will be issued; conversely, if the structural deterioration trend index is less than the structural damage threshold, it means that the kayak structure is relatively stable, without obvious deformation or damage, and the structural health is good. The relative distance changes reflected by the sensor data are stable and consistent, and the kayak can continue to be used safely and normally.

[0188] The present invention, under the premise of judging that the installation of the sensor is stable, records and pre-processes the sensor data in real time when the sensor detects the impact, and analyzes the dynamic response of the kayak after the impact. After decomposing the signal using wavelet transform technology, a finite element model based on the physical characteristics of the kayak is constructed, and the elastic deformation of the hull is calculated by numerical simulation, the main deformation mode is extracted and time series analysis is performed. The rate of change and acceleration of the deformation mode are predicted by the autoregressive sliding average model, the structural deterioration trend index is calculated, and compared with the structural damage threshold to judge the structural health of the kayak. Then, the stability and consistency of the relative distance change between the sensors are evaluated, the risk of structural damage is effectively identified, and the accuracy and reliability of the monitoring system are ensured. Through real-time monitoring of the structural deterioration trend index, structural problems of the kayak can be warned in time, potential safety hazards can be avoided, and the safety and life of the kayak can be improved.

[0189] Example 2

[0190] Figure 2 The present invention provides a kayak hull posture monitoring system, comprising: an initial benchmark modeling module, a multi-dimensional posture fusion module, a time sequence position analysis module, a parameter steady-state evaluation module and a dynamic deformation early warning module;

[0191] The initial benchmark modeling module installs fixed sensors at multiple key locations on the kayak. The sensors collect data in real time and build a benchmark model after preprocessing and verification. This model serves as a reference standard for attitude monitoring and damage warning. The initial position and state data of the sensors are transmitted to the multi-dimensional attitude fusion module.

[0192] After the benchmark model is built, the multi-dimensional posture fusion module collects the kayak status data in real time and pre-processes it. The data is then converted into a unified coordinate system. The comprehensive posture and motion status information is generated through the multi-sensor data fusion algorithm. The comprehensive posture and motion status information is transmitted to the time series position analysis module.

[0193] The time series position analysis module calculates the three-dimensional position coordinates of each sensor in real time through the preprocessed data, monitors the distance change rate between sensors using the time window technology, and analyzes the distance change trend through the sliding window technology. It combines the slope and the anomaly score to determine whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly, and the recorded abnormal data is passed to the parameter steady-state evaluation module;

[0194] The parameter steady-state assessment module constructs a feature matrix based on the acquired abnormal data, initializes the factor decomposition machine model parameters, analyzes the data by calculating feature interaction terms and adaptive filters, optimizes the model parameters and calculates the parameter volatility, and judges the installation stability of the sensor by combining the slope and the overall distance change amplitude. The installation stability of the sensor is transmitted to the dynamic deformation warning module;

[0195] After confirming that the sensor is firmly installed, the dynamic deformation warning module begins to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes it and uses wavelet transform analysis; it simulates elastic deformation through the finite element model, extracts deformation patterns using principal component analysis, calculates the rate of change and prediction error through time series analysis, determines the health of the kayak structure and triggers an early warning.

[0196] 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 and thresholds in the formula are set by technicians in this field according to actual conditions.

[0197] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0198] 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.

[0199] In the several embodiments provided in the present application, it should be understood that the disclosed systems and devices can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0200] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0201] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0202] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0203] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present 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. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0204] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring the posture of a kayak, characterized in that: S1, fixed sensors are installed at multiple key locations of the kayak. The sensors collect data in real time, and after preprocessing and verification, a benchmark model is constructed as a reference standard for posture monitoring and damage warning; S2, after the benchmark model is built, each sensor collects the kayak status data in real time and pre-processes it, then converts the data into a unified coordinate system, and generates comprehensive posture and motion status information through a multi-sensor data fusion algorithm; S3, using the data preprocessed by S2, calculates the three-dimensional position coordinates of each sensor in real time, uses the time window technology to monitor the distance change rate between sensors, and uses the sliding window technology to analyze the distance change trend. Combined with the slope and anomaly score, it is determined whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly; S4, based on the abnormal data obtained in step S3, construct a feature matrix, initialize the factor decomposition machine model parameters, analyze the data by calculating feature interaction terms and adaptive filters, optimize the model parameters and calculate the parameter volatility, and judge the installation stability of the sensor by combining the slope and the overall distance change amplitude; S5, after confirming that the sensor is firmly installed, it starts to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes it and uses wavelet transform analysis; simulates elastic deformation through finite element model, extracts deformation mode through principal component analysis, calculates change rate and prediction error through time series analysis, judges the health status of the kayak structure and triggers an early warning; Step S3 includes the following contents: S3.1, obtain the real-time position coordinates of each sensor (x i ,y i ,z i ), the acquisition process is as follows: S3.1.1, determining the initial position coordinates of each sensor according to the reference model of step S1; S3.1.2, each sensor sends real-time data to the central processing unit via wireless transmission; S3.1.3, after receiving the sensor data, the central processing unit updates the three-dimensional position coordinates of each sensor in real time through the data transmission link between sensors and its own data calculation; Using the real-time data from the gyroscope and accelerometer, the relative motion trajectory of each sensor is calculated through the inertial navigation system; Using the relative distance data between sensors, the position of each sensor is further corrected through triangulation and real-time calibration algorithms; Using the distance information between multiple sensors, nonlinear least square method is applied to perform position calibration to ensure the accuracy of position calculation; S3.1.4, through the sensor data transmission link, sensors regularly exchange position information and motion status; S3.1.5, using extended Kalman filtering for real-time error correction; S3.2, according to step S3.1, obtain the real-time position coordinates (x i ,y i ,z i ), where i represents the sensor number; S3.3, initialize the time window T, which is used to store the sensor position data within a period of time. Let the time window be [t0,t n ]; S3.4, within the time window T, use the Euclidean distance formula to calculate the distance at each moment t k (t0≤t k ≤t n ) The distance between sensor i and sensor j; S3.5, calculate the distance change rate Δd within the time window ij (t k ): The distance change rate at each moment is stored in the time series array {Δd ij (t k )}; S3.6, within the time window T, a sliding window technique is used, each sliding window contains M time points, and a statistical analysis is performed on the distance change rate within each sliding window; where p represents the sequence number of the sliding window; For each distance change rate {Δd ij (t k )} Perform linear regression to obtain the slope m and intercept b. The slope represents the trend of distance change; S3.7, calculate each time point t k The anomaly score α(t k ): Among them, μ Δd and σ Δd are the mean and standard deviation of the distance change rate, respectively; If the anomaly score exceeds the anomaly threshold, it is considered an abnormal change and the abnormal data is recorded; Step S4 includes the following contents: S4.1, based on the abnormal data obtained in step S3, obtain the actual location data set and historical data of each sensor to form a feature matrix X; Initialize the factorization machine model parameters, including feature weights ω and factor matrix V; S4.2, for each sensor data point x i , calculate the feature interaction term: Where: ω0 represents the bias term of the model; ω o Represents feature x o The linear weight of ; o represents the index of the oth feature; v represents the index of the vth feature; <V o ,V v > represents the factor vector V o and V v The inner product of S4.3 uses the noise suppression factor η to dynamically adjust the noise suppression level by analyzing the time series of sensor data and using adaptive filtering to estimate and eliminate noise: Where: m″ represents the order of the noise suppression filter; k represents the number of time delay steps or the index of past data points, k traverses from 1 to m″; η(t) is the noise suppression factor at the current time t; η0 is the initial noise suppression factor, which is a constant used to provide basic noise suppression effect; α k is the filter coefficient, which represents the contribution weight of the previous k historical data points to the current noise suppression factor; x(tk) represents the sensor data value at the previous k moments; S4.4, use stochastic gradient descent to optimize the model parameters, with the goal of minimizing the loss function; S4.5, combining the slope, the characteristics of parameter volatility and the amplitude of the overall distance change, the vibration stability index is obtained: Where: S represents the vibration stability index; |g| is the absolute value of the slope, indicating the trend strength of the sensor distance change; σ ω represents the standard deviation of the feature weight ω, indicating the volatility of the weight; σ V represents the standard deviation of the factor vector in the factor matrix V, indicating the volatility of the factor; f1 represents the weight coefficient of |g|; f2 represents The weight coefficient of The weight coefficient of The vibration steady-state index is compared with the steady-state threshold: if the vibration steady-state index is greater than or equal to the steady-state threshold, there is a significant loose installation problem and a deterioration trend signal is generated; conversely, if the vibration steady-state index is less than the steady-state threshold, there is no significant loose problem and an installation stability signal is generated.

2. A kayak hull posture monitoring method according to claim 1, characterized in that: Step S5 includes the following contents: S5.1, after obtaining the installation stability signal, when the sensor detects the impact, the relative distance change data between the sensors is recorded in real time, including the sensor number and time information; S5.2, preprocessing the recorded sensor data, and then using wavelet transform technology to decompose the data into detail signals and approximate signals of multiple different scales; S5.3, construct a finite element model of the kayak based on its physical characteristics, and divide the kayak into a number of finite element units; use the real-time recorded sensor data as boundary conditions to input into the finite element model, and calculate the elastic deformation of the hull after the impact through numerical simulation; S5.4, calculate the dynamic response of the kayak after the impact, perform dimensionality reduction processing through principal component analysis method, and extract the main deformation mode: Calculate the dynamic response of the kayak after being hit and obtain the deformation field at each moment; Extract the main modes in the deformation field and obtain the main deformation modes through dimensionality reduction processing through principal component analysis; S5.5, perform time series analysis on the extracted deformation pattern, calculate its rate of change and acceleration, predict the deformation pattern through an autoregressive moving average model, and calculate the prediction error; S5.6, calculating the cumulative change of the deformation mode and the standard deviation of the prediction error, and calculating the structural deterioration trend index by the ratio of the cumulative change to the standard deviation of the prediction error; S5.7, comparing the structural deterioration trend index with the structural damage threshold: if the structural deterioration trend index is greater than or equal to the structural damage threshold, issuing an early warning reminder; On the contrary, if the structural deterioration trend index is less than the structural damage threshold, it means that the structural health condition is good.

3. A kayak hull posture monitoring system, used to implement the kayak hull posture monitoring method according to any one of claims 1-2, comprising: Initial benchmark modeling module, multi-dimensional attitude fusion module, time series position analysis module, parameter steady-state evaluation module and dynamic deformation warning module; The initial benchmark modeling module installs fixed sensors at multiple key locations on the kayak. The sensors collect data in real time and build a benchmark model after preprocessing and verification. This model serves as a reference standard for attitude monitoring and damage warning. The initial position and state data of the sensors are transmitted to the multi-dimensional attitude fusion module. After the benchmark model is built, the multi-dimensional posture fusion module collects the kayak status data in real time and pre-processes it. The data is then converted into a unified coordinate system. The comprehensive posture and motion status information is generated through the multi-sensor data fusion algorithm. The comprehensive posture and motion status information is transmitted to the time series position analysis module. The time series position analysis module calculates the three-dimensional position coordinates of each sensor in real time through the preprocessed data, monitors the distance change rate between sensors using the time window technology, and analyzes the distance change trend through the sliding window technology. It combines the slope and the anomaly score to determine whether the distance change between sensors is abnormal. For abnormal changes, the abnormal data is recorded accordingly, and the recorded abnormal data is passed to the parameter steady-state evaluation module; The parameter steady-state assessment module constructs a feature matrix based on the acquired abnormal data, initializes the factor decomposition machine model parameters, analyzes the data by calculating feature interaction terms and adaptive filters, optimizes the model parameters and calculates the parameter volatility, and judges the installation stability of the sensor by combining the slope and the overall distance change amplitude. The installation stability of the sensor is transmitted to the dynamic deformation warning module; After confirming that the sensor is firmly installed, the dynamic deformation warning module begins to monitor the impact of the kayak during drifting. When the sensor detects an impact, it records the relative distance change data in real time, pre-processes it and uses wavelet transform analysis; it simulates elastic deformation through the finite element model, extracts deformation patterns using principal component analysis, calculates the rate of change and prediction error through time series analysis, determines the health of the kayak structure and triggers an early warning.

Images