Method and system for predicting drift trajectory of man overboard at sea

Abstract

The present invention relates to the technical field of predicting drift trajectories of targets at sea, and in particular to a method and system for predicting the drift trajectory of a man overboard at sea. The method comprises: reading training set data and test set data from drift data; performing time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data; substituting the training set time series data and the test set time series data into an LSTM model to obtain a training model; and testing the test set data by means of the training model to obtain a trajectory prediction model. The LSTM model can capture a long-term dependency relationship in time series data and is suitable for various sequence prediction tasks, thereby accurately simulating the actual drift of a man overboard under different sea conditions.

Description

A method and system for predicting the drifting trajectory of a person falling into the sea TECHNICAL FIELD

[0001] The present application relates to the technical field of predicting the drifting trajectory of a target object at sea, and in particular to a method and system for predicting the drifting trajectory of a person falling into the sea. BACKGROUND

[0002] With the expansion of ocean development and the frequent occurrence of maritime activities, maritime accidents have attracted more and more attention. Maritime accidents can bring great disaster to the world's coastal countries in politics, economy and military, not only in terms of personnel casualties and property losses, but also in terms of adverse effects on social development. Therefore, maritime search and rescue work is increasingly valued by coastal countries. For the rapidly developing maritime transportation industry and fishing industry, maritime search and rescue operations can provide irreplaceable protection for personnel and property safety.

[0003] Currently, the search and rescue command coordination work in China mainly relies on the experience and judgment of search and rescue command personnel, and cannot quickly and efficiently predict the drifting trajectory of the accident personnel according to the hydrological and meteorological conditions of the accident site, which greatly affects the speed and accuracy of the command and coordination work, often causing the search and rescue operation to miss the opportunity. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a method and system for predicting the drifting trajectory of a person falling into the sea, which can accurately simulate the real situation of the person drifting after falling into the sea under different sea conditions.

[0005] To solve the above technical problems, the first technical solution adopted by the present application is:

[0006] A method for predicting the drifting trajectory of a person falling into the sea, comprising the following steps:

[0007] S1, obtaining drifting data;

[0008] S2, reading training set data and test set data from the drifting data;

[0009] S3, performing time series processing operation on the training set data and test set data to obtain training set time series data and test set time series data;

[0010] S4, establishing an LSTM model;

[0011] S5, substituting the training set time series data and test set time series data into the LSTM model to obtain a training model;

[0012] S6, testing the test set data by the training model to obtain a trajectory prediction model.

[0013] The second technical solution adopted by the present application is:

[0014] A sea personnel falling into water drift trajectory prediction system comprises one or more processors and a memory, and the memory stores a program which, when executed by the processor, implements the following steps:

[0015] S1, obtain drift data;

[0016] S2, read training set data and test set data from the drift data;

[0017] S3, perform time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data;

[0018] S4, establish an LSTM model;

[0019] S5, substitute the training set time series data and the test set time series data into the LSTM model to obtain a training model;

[0020] S6, test the test set data through the training model to obtain a trajectory prediction model.

[0021] The present application has the following beneficial effects:

[0022] The present application reads training set data and test set data from the drift data, performs time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data, substitutes the training set time series data and the test set time series data into the LSTM model to obtain a training model, tests the test set data through the training model to obtain a trajectory prediction model, the LSTM model can capture long-term dependence in time series data and is suitable for various sequence prediction tasks, so that the real situation of personnel drifting after falling into water under different sea conditions can be accurately simulated. BRIEF DESCRIPTION OF DRAWINGS

[0023] Fig. 1 is a step flow chart of the sea personnel falling into water drift trajectory prediction method of the present application;

[0024] Fig. 2 is a network structure diagram of the LSTM model of the sea personnel falling into water drift trajectory prediction method of the present application;

[0025] Fig. 3 is a structural block diagram of the sea personnel falling into water drift trajectory prediction system of the present application;

[0026] Label explanation:

[0027] 1, processor; 2, memory. DETAILED DESCRIPTION

[0028] In order to explain the technical content of the present application, the purposes and effects achieved, the following will be described in conjunction with the embodiments and the accompanying drawings.

[0029] Please refer to Figure 1, the first technical solution adopted by the present application is:

[0030] A method for predicting the drifting trajectory of a person falling into the sea, comprising the following steps:

[0031] S1, obtaining drift data;

[0032] S2, reading training set data and test set data from the drift data;

[0033] S3, performing time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data;

[0034] S4, establishing an LSTM model;

[0035] S5, substituting the training set time series data and the test set time series data into the LSTM model to obtain a training model;

[0036] S6, testing the test set data by the training model to obtain a trajectory prediction model.

[0037] From the above description, the beneficial effects of the present application are:

[0038] The present application reads training set data and test set data from drift data, performs time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data, substitutes the training set time series data and the test set time series data into the LSTM model to obtain a training model, tests the test set data by the training model to obtain a trajectory prediction model, and the LSTM model can capture long-term dependencies in time series data and is suitable for various sequence prediction tasks, so as to accurately simulate the real situation of the person drifting after falling into the sea under different sea conditions.

[0039] Further, step S6 further comprises the following steps:

[0040] S7, obtaining predicted trajectory data according to the trajectory prediction model;

[0041] S8, substituting the predicted trajectory data into a Kalman filter to obtain an optimized trajectory prediction model.

[0042] From the above description, the Kalman filter can further optimize the trajectory prediction model and provide more accurate predicted trajectory.

[0043] Further, between step S2 and step S3, the following steps are further included:

[0044] The training set data and the test set data are normalized.

[0045] From the above description, by normalizing the training set data and the test set data, the model can better learn the relationship between the features.

[0046] Further, step S1 specifically comprises the following steps:

[0047] The floating ball is thrown at a designated position, real-time positioning is performed by the positioning module on the floating ball, and the drifting data of the floating ball is recorded.

[0048] Please refer to FIG. 3, the second technical solution adopted by the present application is:

[0049] A sea personnel falling into water drifting trajectory prediction system comprises one or more processors and a memory, the memory stores a program, and the program is executed by the processor to realize the following steps:

[0050] S1, obtain the drifting data;

[0051] S2, read the training set data and the test set data from the drifting data;

[0052] S3, perform time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data;

[0053] S4, establish an LSTM model;

[0054] S5, substitute the training set time series data and the test set time series data into the LSTM model to obtain a training model;

[0055] S6, test the test set data by the training model to obtain a trajectory prediction model.

[0056] Further, the program is executed by the processor to realize the following steps:

[0057] S7, obtain predicted trajectory data according to the trajectory prediction model;

[0058] S8, substitute the predicted trajectory data into a Kalman filter to obtain an optimized trajectory prediction model.

[0059] Further, the program is executed by the processor to realize the following steps:

[0060] The training set data and the test set data are normalized.

[0061] Further, the program is executed by the processor to realize the following steps:

[0062] The floating ball is thrown at a specified location, real-time positioning is performed through the positioning module on the floating ball, and the drifting data of the floating ball is recorded.

[0063] Please refer to FIG. 1 and FIG. 2, the embodiment one of the present application is:

[0064] Please refer to FIG. 1, a method for predicting the drifting trajectory of a person falling into the sea, comprising the following steps:

[0065] S1, obtaining the drifting data; step S1 specifically comprises the following steps:

[0066] The floating ball is thrown at a specified location, real-time positioning is performed through the positioning module on the floating ball, and the drifting data of the floating ball is recorded.

[0067] S2, reading training set data and test set data from the drifting data; using the pd.read_excel function to read the training set data and the test set data; the floating ball data set numbered 17 is used as the training set data, and the floating ball data set numbered 20 is used as the test set data;

[0068] S3, performing time series processing operation on the training set data and the test set data to obtain training set time series data and test set time series data; in the code, the time step is set to 3 through the variable time_steps (referring to the time step), and the model will use the data of the past three time points to predict the value of the next time point; two for loop functions are used to perform time series processing on the training set data and the test set data; in each loop, continuous time_steps data points are extracted through slicing operation, and they are stored as a sequence in the list X_train_sequence or X_test_sequence (referring to the training set); at the same time, the first data point after the sequence is stored as the target value in y_train_next_step or y_test_next_step (referring to the test set); the list X_train_sequence and X_test_sequence are converted into NumPy arrays for model training and testing;

[0069] Between step S2 and step S3, the following steps are further included:

[0070] Normalization operation is performed on the training set data and test set data; first, the MinMaxScaler class function is imported, a normalization object is created through the MinMaxScaler function, and the scaler. fit_transform(X_train) function is used to fit and transform the training set feature data X_train; this step will calculate the minimum and maximum values of the training set data, and convert the data to the range of 0 to 1 according to these values; the scaler. transform(X_test) function is used to convert the test set feature data X_test to the same scale as the training set; the converted data is stored in new variables X_train_scaled and X_test_scaled (referring to the scale of the training set) for subsequent model training and testing;

[0071] S4, establish an LSTM model;

[0072] The LSTM (Long Short-Term Memory) model is a special type of recurrent neural network that is suitable for processing time series data. It can learn long-term dependencies in time series data. Compared with traditional RNN (Recurrent Neural Network), LSTM solves the problem of gradient vanishing and gradient explosion by introducing a gating mechanism, making it possible to process longer sequences of data.

[0073] The LSTM network is composed of multiple LSTM units, each containing three gates: forget gate, input gate and output gate, as shown in Figure 2.

[0074] The forget gate determines which information in the cell state should be forgotten or discarded. It contains a sigmoid activation function that outputs a value between 0 and 1 based on the input data and the previous time step's hidden state.

[0075] The input gate controls the degree of influence of new input information on the current cell state. It includes a sigmoid activation function to generate a value between 0 and 1, indicating which values in each cell state should be updated.

[0076] The output gate controls how the current hidden state and the next cell state affect the final output at the current time. It contains a sigmoid activation function to determine which parts of the output state will be activated.

[0077] These gate structures enable LSTM to selectively retain or forget information during training, effectively addressing the problems of gradient vanishing and gradient exploding; specifically, the forget gate is responsible for deciding which information to discard from the cell state, the input gate is responsible for deciding which new information to add to the cell state, and the output gate is responsible for deciding which information to output from the cell state; through this mechanism, LSTM can capture long-term dependencies in time series data and is suitable for various sequence prediction tasks such as language models, time series prediction, and bioinformatics.

[0078] The fit function is used to train the model, which accepts the time series training set data X_train_sequence and target data y_train_next_step, as well as the validation data X_test_sequence and y_test_next_step; during training, the LSTM model will learn how to predict the value at the next time point based on the input time series data; in the fit function, the LSTM model will perform multiple epochs of training; in each epoch, the LSTM model will iterate through the entire training set data and update the model weights based on the loss function; this process is achieved through the backpropagation algorithm;

[0079] S5, the training set time series data and test set time series data are substituted into the LSTM model to obtain a training model;

[0080] S6, test the test set data through the training model to obtain a trajectory prediction model.

[0081] Step S6 also includes the following steps:

[0082] S7, according to the trajectory prediction model, obtain predicted trajectory data;

[0083] S8, substitute the predicted trajectory data into the Kalman filter to obtain an optimized trajectory prediction model.

[0084] First, the test data needs to go through the same preprocessing steps as the training set data, including feature selection, normalization, etc.; the test data also needs to be converted into time series format; this is the same as the processing method of the training set data, that is, the data is sliced into sequences with time_steps consecutive time points; use the trained model to predict the time series test data; the trained model will output the predicted value of the next time point for each sequence; compare the model's predicted value with the actual value in the test data to evaluate the model's performance; mean squared error MSE and determination coefficient R 2 are used as evaluation indicators; finally, visualize the actual value and predicted value to visually demonstrate the model's prediction effect.

[0085] The core of the Kalman filter lies in its recursive estimation process, which includes a prediction step and an update step. Through these two steps, prior knowledge and observation data are effectively fused to estimate the system state.

[0086] In the prediction step (time update), the Kalman filter predicts the current state based on the previous state estimate and control input, and estimates the covariance of the prediction error.

[0087] Prediction state equation:

[0088] Prediction covariance equation:

[0089] In the above equations, is the state estimate at time k; A is the state transition matrix, describing the transition of system state; B is the control input matrix, describing the effect of control input on state; u k is the control input vector; Q is the process noise covariance matrix, describing the unpredictable part of the system model;

[0090] In the measurement update step, the Kalman filter combines the observation data and the prediction error to adjust the predicted state through the Kalman gain to obtain a more accurate estimate.

[0091] Kalman gain equation:

[0092] Updated prediction state equation:

[0093] Updated prediction covariance equation:

[0094] In the above equations, the state vector x k = Ax k-1 + Bu k + w k , where u k is the control input vector, w k is the process noise, usually assumed to be Gaussian white noise with mean zero and covariance Q.

[0095] Measurement model z k = Hx k + v k , where H is the measurement matrix, v k is the measurement noise, usually assumed to be Gaussian white noise with mean zero and covariance R.

[0096] The Kalman filter obtains the optimal estimation of the system state by weighing the past state estimation and the current measurement value to minimize the mean square error of the estimation error; by recursively performing prediction and measurement update, continuous tracking and adjustment of the system state are realized.

[0097] The prediction trajectory data generated by the trajectory prediction model is used as the initial prediction of the Kalman filter, and the measurement value (i.e. the trajectory predicted by the trajectory prediction model) and the noise covariance (measurement noise covariance R and process noise covariance Q) are used as input to optimize the estimation of the trajectory through prediction and update steps. The initial state estimation value The last time step position predicted by the trajectory prediction model is used as the initial position estimation, and the unit matrix multiplied by 0.4 is used as the initial covariance P0.

[0098] The mean square error MSE and the determination coefficient R 2 between the Kalman filter output result and the trajectory prediction model direct output result are calculated. The specific experimental data is shown in Table 1.

[0099] Mean square error MSE: used to measure the average of the square difference between the predicted value and the actual value, the smaller the value, the more accurate the model prediction.

[0100]

[0101] In the formula, n is the sample number, y i is the i-th actual value, is the i-th predicted value.

[0102] Determination coefficient R 2 : a statistical quantity that measures the fitting of the model data, between 0 and 1, the closer to 1, the better the model fitting effect.

[0103]

[0104] In the formula, n is the sample number, y i is the i-th actual value, is the i-th predicted value, is the average of the actual value.

[0105] The specific simulation results and experimental comparisons are as follows:

[0106]

[0107] Table 1

[0108] As can be seen from Table 1, the trajectory prediction after Kalman filter post-processing has lower mean square error MSE and higher determination coefficient R 2, which indicates that the Kalman filter can effectively reduce the prediction error and improve the accuracy of trajectory prediction.

[0109] To further enhance the predictive ability of the model, the following improvements can be considered:

[0110] Adjusting the hyperparameters of the LSTM model, such as increasing the number of LSTM units, adjusting the learning rate, etc., to explore more suitable model structures; the application of the Kalman filter improves the prediction performance of the LSTM model, but its performance is also limited by the selection of filter parameters and the accuracy of the system model; future work will focus on further optimizing the Kalman filter parameters and exploring the combination of other filtering techniques with the LSTM model to achieve more accurate offshore personnel fall drift trajectory prediction; using more features or introducing additional environmental variables (such as ocean environmental parameters) to enrich the input information of the model; conducting more detailed analysis and preprocessing of the data set to improve the understanding and capture ability of the model on the drift characteristics of the buoy.

[0111] These improvements can help optimize the performance of the model, making it more accurate in predicting fall drift trajectories in different sea conditions, and improving the reliability and effectiveness of practical applications.

[0112] Please refer to FIG. 3, the second embodiment of the present application is:

[0113] A system for predicting the fall drift trajectory of offshore personnel, comprising one or more processors 1 and a memory 2, the memory 2 storing a program that, when executed by the processor 1, implements the following steps:

[0114] S1, obtaining drift data;

[0115] S2, reading training set data and test set data from the drift data;

[0116] S3, performing time series processing operation on the training set data and test set data to obtain training set time series data and test set time series data;

[0117] S4, establishing an LSTM model;

[0118] S5, substituting the training set time series data and test set time series data into the LSTM model to obtain a training model;

[0119] S6, testing the test set data through the training model to obtain a trajectory prediction model.

[0120] The program is executed by the processor 1 to implement the following steps:

[0121] S7, obtaining predicted trajectory data according to the trajectory prediction model;

[0122] S8, the predicted trajectory data is substituted into a Kalman filter to obtain an optimized trajectory prediction model.

[0123] The program is executed by the processor 1 to realize the following steps:

[0124] The training set data and the test set data are normalized.

[0125] The program is executed by the processor 1 to realize the following steps:

[0126] The floating ball is thrown at a specified position, real-time positioning is performed through a positioning module on the floating ball, and the drifting data of the floating ball is recorded.

[0127] In summary, the present application provides a kind of offshore personnel falls into water and drifts trajectory prediction method and system, training set data and test set data are read from drifting data, training set data and test set data are time-sequenced, training set time series data and test set time series data are obtained, training set time series data and test set time series data are substituted into LSTM model, training model is obtained, test set data is tested by training model, trajectory prediction model is obtained, LSTM model can capture long-term dependence in time series data, suitable for various sequence prediction tasks, so as to accurately simulate the real situation of personnel drifting after falling into water under different sea conditions.

[0128] The above is only an embodiment of the present application, and does not limit the patent scope of the present application, any equivalent transformation or direct or indirect application in related technical fields using the content of the present application specification and drawings is also included in the patent protection scope of the present application.

Claims

1. A method of predicting a drifting trajectory of a person in the water at sea, characterized in that, The method comprises the following steps: S1, obtaining drift data; S2, reading training set data and test set data from the drift data; S3, performing time series processing operation on the training set data and test set data to obtain training set time series data and test set time series data; S4, establishing an LSTM model; S5, substituting the training set time series data and test set time series data into the LSTM model to obtain a training model; S6, testing the test set data through the training model to obtain a trajectory prediction model.

2. The method of claim 1, wherein, After step S6, the following steps are further included: S7, obtaining predicted trajectory data according to the trajectory prediction model; S8, substituting the predicted trajectory data into a Kalman filter to obtain an optimized trajectory prediction model.

3. The method of claim 1, wherein, Between step S2 and step S3, the following step is further included: Performing normalization processing operation on the training set data and test set data.

4. The method of claim 1, wherein, Step S1 specifically comprises the following steps: Throwing the floating ball at a specified position, performing real-time positioning through a positioning module on the floating ball, and recording the drift data of the floating ball.

5. A system for predicting a drift trajectory of a person in the water at sea, characterized by The device comprises one or more processors and a memory, and the memory stores a program which, when executed by the processor, implements the following steps: S1, obtaining drift data; S2, reading training set data and test set data from the drift data; S3, performing time series processing operation on the training set data and test set data to obtain training set time series data and test set time series data; S4, establishing an LSTM model; S5, substituting the training set time series data and test set time series data into the LSTM model to obtain a training model; S6, testing the test set data through the training model to obtain a trajectory prediction model.

6. The offshore man-overboard drift trajectory prediction system of claim 5, wherein, The program, when executed by the processor, implements the following steps: S7, obtaining predicted trajectory data according to the trajectory prediction model; S8, substituting the predicted trajectory data into a Kalman filter to obtain an optimized trajectory prediction model.

7. The offshore man-overboard drift trajectory prediction system of claim 5, wherein, The program, when executed by the processor, implements the following steps: Performing normalization processing operation on the training set data and test set data.

8. The offshore man-overboard drift trajectory prediction system of claim 5, wherein, The program, when executed by the processor, implements the following steps: Throwing the floating ball at a specified position, performing real-time positioning through a positioning module on the floating ball, and recording the drift data of the floating ball.

Images