A positioning method and device applied to a ship environment

Abstract

The application provides a positioning method and device applied to a ship environment. The method provided by the application comprises the following steps: collecting a group of positioning related data when a current sampling period arrives; determining a joint fluctuation index of an IQ sample according to a magnitude variation and a phase variation of the IQ sample, and determining an updated sampling period of next sampling according to the joint fluctuation index, a preset minimum sampling period and a preset standard sampling period; performing next round of data collection when the updated sampling period arrives, and repeating the above steps; screening and preprocessing a plurality of groups of continuously collected positioning related data to obtain a training sample set; training an initial positioning model based on the training sample set to obtain a trained positioning model; inputting a group of to-be-tested data collected in real time into the trained positioning model, so that the trained positioning model outputs a predicted position of a transmitter corresponding to real-time IQ samples in the group of to-be-tested data based on the group of to-be-tested data.

Description

Technical Field

[0001] This application relates to the field of wireless positioning technology, and in particular to a positioning method and apparatus for use in a marine environment. Background Technology

[0002] Indoor positioning is becoming increasingly important in marine environments, playing a crucial role in emergency response, logistics, and industrial automation. While traditional indoor positioning technologies such as Wi-Fi and RFID can successfully achieve the required positioning functions in normal environments, their positioning accuracy is limited and the positioning effect is poor in the specific environment of a ship. Due to the ship's metal structure and its dynamic environment, traditional indoor positioning technologies are used for positioning. Summary of the Invention

[0003] In view of this, this application provides a positioning method and apparatus for use in a ship environment, for achieving accurate positioning in a ship environment.

[0004] Specifically, this application is implemented through the following technical solution:

[0005] The first aspect of this application provides a positioning method applied to a ship environment, the positioning method comprising:

[0006] When the current sampling period arrives, a set of positioning-related data is collected; wherein, the set of positioning-related data includes IQ samples obtained by the receiver in the process of receiving wireless signals transmitted from the transmitter, the current attitude of the hull, the current position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0007] Based on the amplitude and phase changes of the IQ samples, the joint fluctuation index of the IQ samples is determined, and based on the joint fluctuation index, the preset minimum sampling period and the preset standard sampling period, the update sampling period corresponding to the next sampling is determined.

[0008] When the update sampling period arrives, the next round of data acquisition is performed, and the above steps are repeated;

[0009] Multiple sets of continuously collected positioning-related data are filtered and preprocessed to obtain a training sample set; each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0010] The initial localization model is trained based on the training sample set to obtain the trained localization model.

[0011] A set of real-time collected test data is input into the trained positioning model, so that the trained positioning model outputs the predicted position of the transmitter corresponding to the real-time IQ sample in the set of test data; wherein, the set of test data includes the real-time attitude of the ship, the real-time position offset of the ship, and the real-time IQ sample.

[0012] A second aspect of this application provides a positioning device for use in a ship environment. The device includes a data acquisition module, a determination module, a processing module, a training module, and a prediction module. The data acquisition module is used to acquire a set of positioning-related data when the current sampling period arrives. This set of positioning-related data includes an IQ sample obtained by the receiver during the reception of a wireless signal transmitted from a transmitter, the current attitude of the ship, the current position offset of the ship, and the location of the transmitter corresponding to the IQ sample.

[0013] The determining module is used to determine the joint fluctuation index of the IQ sample based on the amplitude change and phase change of the IQ sample, and to determine the update sampling period corresponding to the next sampling based on the joint fluctuation index, as well as the preset minimum sampling period and the preset standard sampling period.

[0014] The acquisition module is used to perform the next round of data acquisition when the update sampling period arrives, and repeat the above steps.

[0015] The processing module is used to filter and preprocess multiple sets of continuously collected positioning-related data to obtain a training sample set; each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0016] The training module is used to train an initial localization model based on the training sample set to obtain a trained localization model.

[0017] The prediction module is used to input a set of real-time collected test data into the trained positioning model, so that the trained positioning model can output the predicted position of the transmitter corresponding to the real-time IQ sample in the set of test data based on the set of test data; wherein, the set of test data includes the real-time attitude of the ship, the real-time position offset of the ship, and the real-time IQ sample.

[0018] The positioning method and apparatus provided in this application for use in ship environments, based on the joint amplitude and phase fluctuation index of IQ samples, adaptively shortens or extends the sampling period using exponential nonlinear and linear strategies. When the signal fluctuates drastically, the sampling frequency is increased to enhance data sampling density and capture transient characteristics of the samples. When the signal is stable, the sampling frequency is reduced to decrease redundant data and computational overhead. This adaptive sampling strategy not only significantly improves sampling efficiency but, more importantly, enhances the timeliness and representativeness of the model training samples. In the dynamically changing ship operating environment, this adaptive sampling strategy effectively enhances the coverage of key disturbance states by the training samples, enabling the model to capture more comprehensive environmental interference patterns and spatial distribution characteristics during the training phase, thereby improving the model's prediction accuracy and generalization ability in complex environments. Furthermore, by incorporating real-time attitude information and position offset of the ship into the positioning modeling, more physically interpretable high-dimensional training samples are constructed. This allows the model to learn the interference patterns of attitude and offset on the signal waveform, more accurately reconstructing the relationship between IQ samples and the transmitter's true position, significantly improving the model's generalization ability and positioning accuracy in dynamic environments. In summary, by combining adaptive sampling mechanisms with attitude and displacement offset fusion modeling, the localization model can more sensitively perceive dynamic environmental changes, ultimately achieving a more stable and accurate localization effect. Attached Figure Description

[0019] Figure 1 A flowchart illustrating an embodiment of the positioning method for use in a marine environment provided in this application;

[0020] Figure 2 A schematic diagram illustrating the implementation principle of collecting positioning-related data, as shown in an exemplary embodiment of this application;

[0021] Figure 3 A flowchart of Embodiment 2 of the positioning method applied to a ship environment provided in this application;

[0022] Figure 4 This is a structural schematic diagram of the positioning device applied to a marine environment provided in this application. Detailed Implementation

[0023] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0024] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0025] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0026] The following specific embodiments are given to illustrate the technical solution of this application in detail.

[0027] Figure 1 This is a flowchart of an embodiment of the positioning method applied to a ship environment provided in this application. Please refer to... Figure 1 The method provided in this embodiment may include:

[0028] S101. When the current sampling period arrives, collect a set of positioning-related data.

[0029] Figure 2 This is a schematic diagram illustrating the implementation principle of collecting positioning-related data, as shown in an exemplary embodiment of this application. Please refer to... Figure 2 Multiple reference positions (transmitting positions) and one receiving position are pre-set on the hull. The receiver is positioned in the receiving position, and the transmitter is positioned in any of the reference positions. The transmitter sends a wireless signal (Bluetooth 5.1 AoA message), and the receiver receives the wireless signal to generate IQ samples.

[0030] Specifically, when the current sampling period arrives, a set of positioning-related data is collected. It should be noted that the set of positioning-related data includes the IQ sample obtained by the receiver during the process of receiving the wireless signal transmitted from the transmitter, the current attitude of the hull, the current position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0031] Specifically, an IQ sample is complex-form sampling data obtained by the receiver during the demodulation of the in-phase and quadrature components of the radio frequency signal when receiving the wireless signal transmitted from the transmitter. This IQ sample characterizes the amplitude and phase features of the received wireless signal in time and space, and can be used as the data basis for subsequent transmitter positioning. For example, an IQ sample can be in the form of i n +j·q n , where i n For in-phase components, q n These are orthogonal components.

[0032] Furthermore, the current attitude of the hull reflects the ship's tilt, rotation, and other states, and can be measured by a gyroscope or attitude sensor; while the current position offset of the hull is the amount of movement of the hull's current position relative to the reference position, which can be obtained by GPS or other positioning systems, and is not limited in this application; finally, the location of the transmitter corresponding to the IQ sample is the actual location of the transmitter.

[0033] In summary, in this step, when the current sampling period arrives, a set of positioning-related data is collected using various data acquisition methods.

[0034] S102. Based on the amplitude change and phase change of the IQ sample, determine the joint fluctuation index of the IQ sample, and based on the joint fluctuation index, the preset minimum sampling period and the preset standard sampling period, determine the update sampling period corresponding to the next sampling.

[0035] In this application, the method for collecting positioning-related data is adaptive sampling. Based on the internal environment of the ship, the collection cycle of positioning-related data is adaptively updated through adaptive sampling, thereby improving the quality of the collected positioning-related data. In addition, it can also improve the environmental adaptability of the positioning method for the ship environment provided in this application, ensuring that data collection and accurate positioning can still be successfully completed in complex ship environments.

[0036] Specifically, the sampling period for each set of positioning-related data in this application is dynamically adjusted. As the complex environment inside the ship changes, the update sampling period corresponding to the next sampling is determined based on the joint fluctuation index, as well as the preset minimum sampling period and the preset standard sampling period.

[0037] Optionally, in one possible implementation, determining the joint volatility index of the IQ sample based on the amplitude change and phase change of the IQ sample includes:

[0038] Step 1: Based on the amplitude change and the preset amplitude change threshold, determine the first normalized fluctuation index corresponding to the amplitude change.

[0039] In this step, the first normalized volatility index corresponding to the amplitude change can be determined according to the following formula:

[0040]

[0041] in, It is the first normalized volatility indicator; τ is the change in amplitude. A This is the threshold for amplitude change.

[0042] Step 2: Based on the phase change amount and the preset phase change threshold, determine the second normalized fluctuation index corresponding to the phase change amount.

[0043] In this step, the second normalized fluctuation index corresponding to the phase change can be determined according to the following formula:

[0044]

[0045] in, This is the second normalized volatility indicator; τ is the phase change. A This is the phase change threshold.

[0046] Step 3: Based on the first normalized volatility index and the second normalized volatility index, determine the first weight corresponding to the phase and the second weight corresponding to the amplitude.

[0047] Specifically, in one possible implementation, based on the first normalized volatility index and the second normalized volatility index, a first weight corresponding to the phase and a second weight corresponding to the amplitude are determined, including:

[0048] The first weight is calculated according to either the third or fourth formula, and the second weight is determined based on the first weight; wherein the third formula is:

[0049]

[0050] Among them, a k It is the first weight; It is the first normalized volatility indicator; ε is the second normalized volatility index; ε is the perturbation term, preset to a positive value.

[0051] The fourth formula is:

[0052]

[0053] Among them, a k As the first weight; a k-1λ is the phase weight of the previous sampling point; λ is the smoothing factor, with a value range of [0, 1]. It is the first normalized volatility indicator; ε is the second normalized volatility index; ε is the perturbation term, preset to a positive value.

[0054] Furthermore, the first weight 'a' corresponding to the phase is calculated using the third or fourth formula. k Then, since the sum of the first and second weights is 1, let b k =1-a k b k This is the second weight.

[0055] In this step, by introducing a first weight corresponding to the phase and a second weight corresponding to the amplitude, a dynamic balance assessment of the phase and amplitude characteristics of the IQ sample is achieved. In this way, the first and second weights can adaptively reflect the stability difference between the phase and amplitude characteristics, and the first and second weights are determined by the stability difference between the phase and amplitude characteristics. For example, when the phase characteristic is more stable, that is, when the second normalized volatility index is smaller, the first normalized volatility index is larger. At this time, the contribution of the phase to the joint volatility index is emphasized. When the amplitude characteristic is more stable, that is, when the second normalized volatility index is larger, the contribution of the amplitude to the joint volatility index is emphasized.

[0056] Referring to the preceding description, by dynamically weighting the phase and amplitude characteristics of the IQ samples, we can more effectively address multipath effects, phase jumps, and dynamic disturbances in complex ship environments. Furthermore, we can prioritize the use of features with higher stability to optimize the calculation accuracy of the joint fluctuation index, thereby dynamically adjusting the adaptive sampling period.

[0057] Step 4: Weight the first normalized volatility index and the second normalized volatility index according to the first weight and the second weight to obtain the joint volatility index.

[0058] In this step, the first normalized volatility index and the second normalized volatility index can be weighted according to the first weight and the second weight using the following formula to obtain the joint volatility index:

[0059]

[0060] Among them, S k For joint volatility indicators; a k b is the first weight; k As the second weight, b k =1-a k ; This is the second normalized volatility indicator; It is the first normalized volatility indicator.

[0061] In summary, this embodiment introduces a first weight corresponding to the phase and a second weight corresponding to the amplitude to achieve a dynamic balance evaluation of the phase and amplitude fluctuation characteristics of the IQ sample, adaptively reflecting the stability difference between the two. Furthermore, based on the allocation of the weights of the phase and amplitude characteristics, features with higher stability are preferentially utilized, thereby optimizing the accuracy of the joint fluctuation index and adjusting the adaptive sampling period. This improves the efficiency of acquiring positioning-related data in complex ship environments, thereby enhancing the quality of the acquired positioning-related data and the robustness of the positioning model.

[0062] Furthermore, in this step, after determining the joint volatility index of the IQ samples, the update sampling period corresponding to the next sampling can be determined based on the linked volatility index, the preset minimum sampling period, and the preset standard sampling period. Specifically, in one possible implementation, the process of determining the update sampling period corresponding to the next sampling may include:

[0063] Step 1: Determine the target adjustment strategy corresponding to the joint volatility index based on the joint volatility index.

[0064] It should be noted that different joint volatility index ranges correspond to different adjustment strategies; the different adjustment strategies include exponential nonlinear adjustment strategies and linear adjustment strategies. When the joint volatility index is greater than a preset threshold, the target adjustment strategy corresponding to the joint volatility index is the exponential nonlinear adjustment strategy; otherwise, the target adjustment strategy corresponding to the joint volatility index is the linear adjustment strategy.

[0065] Specifically, the specific value of the preset threshold is set according to actual needs, and is not limited in this application. For example, in one embodiment, the preset threshold is set to 0.5, and the joint fluctuation index range includes a low fluctuation range [0, 0.5] and a high fluctuation range [0.5, 1]. When the joint fluctuation index is in the low fluctuation range, the IQ sample signal is relatively stable. At this time, a linear adjustment strategy is selected to smoothly extend the sampling period, reduce data redundancy, and thus save computing and storage resources. When the joint fluctuation index is in the high fluctuation range, the IQ sample signal fluctuates violently. At this time, an exponential nonlinear adjustment strategy is selected to shorten the sampling period in order to capture the rapid changes of phase jump or multipath effect.

[0066] Step 2: Determine the update sampling period based on the target adjustment strategy, the minimum sampling period, and the standard sampling period.

[0067] It should be noted that the minimum sampling period is a preset minimum sampling time interval, corresponding to the highest sampling frequency, and can be used in scenarios with drastic signal fluctuations; the standard sampling period is a preset normal sampling time period, corresponding to the standard sampling frequency, and is used in scenarios with stable signals. The minimum sampling period and the standard sampling period can be set according to actual needs, and are not limited in this application.

[0068] Specifically, in one possible implementation, determining the update sampling period based on the target adjustment strategy, the minimum sampling period, and the standard sampling period includes:

[0069] When the target adjustment strategy is an exponential nonlinear adjustment strategy, the update sampling period is determined according to the first formula; otherwise, the update sampling period is determined according to the second formula.

[0070] The first formula is:

[0071]

[0072] Where, Δt k The update sampling period corresponding to the k-th sampling; γ is the sensitivity factor, γ > 0; Δt min The minimum sampling period; Δt nom The standard sampling period; S k For joint volatility indicators;

[0073] The second formula is:

[0074] Δt k =Δt min +(1-S k )·(Δt nom -Δt min );

[0075] Where, Δt k The update sampling period corresponding to the k-th sampling; Δt min The minimum sampling period; Δt nom The standard sampling period; S k This is a combined volatility indicator.

[0076] The positioning method for ship environments provided in this embodiment first normalizes the changes in amplitude and phase characteristics to obtain a first normalized fluctuation index and a second normalized fluctuation index. Then, based on the stability difference between the two, a first weight corresponding to the phase characteristic and a second weight corresponding to the amplitude are dynamically allocated. The first and second weights are then used to weight the first and second normalized fluctuation indices to construct a joint fluctuation index. Subsequently, when the joint fluctuation index is greater than a preset threshold, an exponential nonlinear adjustment strategy is used to determine the sampling period according to a first formula; when the joint fluctuation index is less than the preset threshold, a linear adjustment strategy is used to determine the sampling period according to a second formula. This effectively balances the quality and efficiency of the acquired data within the acquisition period. It can not only reduce the sampling frequency to decrease the volatility of the acquired data when the signal fluctuation is severe, but also increase the sampling frequency to improve the quality of the acquired data when the signal is stable, thus effectively improving the robustness of data acquisition.

[0077] S103. When the update sampling period arrives, perform the next round of data acquisition and repeat the above steps.

[0078] It should be noted that after completing the data acquisition steps for the current sampling period, and following the update sampling period confirmation method in step S102, the update sampling period corresponding to the next sampling is determined based on the signal quality under the current environment. In this way, when the update sampling period arrives, the next round of data acquisition can be carried out according to the update sampling period, and steps S101 and S102 can be repeated again until the sampling ends. This ensures the quality of the data acquired in multiple sampling periods.

[0079] S104. Filter and preprocess multiple sets of continuously collected positioning-related data to obtain a training sample set.

[0080] In this step, after continuously updating the sampling period based on the current ship environment and collecting multiple sets of positioning-related data according to the updated sampling period, it is necessary to filter and preprocess the collected multiple sets of positioning-related data. This can remove noise interference and outliers in the multiple sets of positioning-related data, improve the quality of positioning-related data, and reduce the negative impact of low-quality IQ samples on positioning accuracy and system performance.

[0081] It should be noted that by filtering and preprocessing multiple sets of continuously collected positioning-related data, a training sample set consisting of multiple training samples can be obtained. Each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0082] The screening and preprocessing process will be described in detail in the following examples, and will not be repeated here.

[0083] S105. Train the initial localization model based on the training sample set to obtain the trained localization model.

[0084] In this step, an initial localization model is established. This initial localization model is a three-layer fully connected DNN model, which includes an input layer, a hidden layer, and an output layer.

[0085] Specifically, when training the initial localization model using training samples, the training sample set is first input into the input layer. In the input layer, the IQ samples in the training sample set are split into real and imaginary numbers to form a feature vector of length 2N, where N is the length of the sample. The input layer can split the IQ samples into real and imaginary numbers to form a feature vector of length 2N according to the following formula:

[0086] x=[Re(V1),Im(V1),…,Re(V N ),Im(V N )] T ;

[0087] Where x is the feature vector formed by the input layer; Re(V i ) represents the real part of the i-th IQ sample; Im(V1) represents the imaginary part of the i-th IQ sample.

[0088] Furthermore, the pose (pose is 3-dimensional) and position offset (position offset is 3-dimensional) are concatenated to obtain the final feature vector, which has a dimension of 2N+6.

[0089] Furthermore, the initial localization model includes multiple hidden layers, each containing a non-linear activation function. These non-linear activation functions non-linearly process the feature vector output from the input layer, and the output of each upper hidden layer serves as the input to the next hidden layer, thus progressively extracting high-level features from the IQ samples. The output of each input layer can be calculated using the following formula:

[0090] h i =f(W i h i-1 +b i );

[0091] Among them, h i W is the output of the i-th hidden layer; i b is the weight matrix of the i-th hidden layer; i h is the bias term of the i-th hidden layer; f(·) is the non-linear activation function; i-1 This is the output of the (i-1)th hidden layer.

[0092] It should be noted that the nonlinear activation function can be selected according to actual needs, and this application does not limit it. For example, in one embodiment, the nonlinear activation function can be the ReLU activation function; in another embodiment, the nonlinear activation function can be the Sigmoid activation function.

[0093] Furthermore, the output of the last hidden layer is input into the output layer, which then uses a linear activation function to generate the positioning coordinates according to the following formula:

[0094]

[0095] in, The output of the initial localization model; W o b represents the weights of the output layer. o For the bias of the output layer; h L This is the output of the last hidden layer.

[0096] In addition, the initial localization model also has a loss function. During training, the initial localization model evaluates the difference between the output predicted location and the actual location based on the loss function:

[0097]

[0098] Among them, y i It is the actual location corresponding to the i-th prediction. B is the output of the initial localization model in the i-th prediction, and B is the batch size.

[0099] Furthermore, after obtaining the difference between the output predicted location and the actual location, the weights and biases of each layer of the initial localization model are updated based on the results of the loss function through the backpropagation algorithm. In this way, the trained localization model can be obtained.

[0100] It should be noted that in this embodiment, the positioning model obtains the real-time attitude (roll, pitch, yaw) and position offset of the ship and performs coordinate transformation or feature-level correction on the IQ samples, thereby reducing the positioning error caused by attitude and improving the positioning stability under dynamic sea conditions.

[0101] S106. Input a set of real-time collected test data into the trained positioning model, so that the trained positioning model outputs the predicted location of the transmitter corresponding to the real-time IQ sample in the set of test data based on the set of test data.

[0102] The set of data to be tested includes the ship's real-time attitude, the ship's real-time position offset, and real-time IQ samples.

[0103] In this step, a set of test data collected in real time is input into the trained localization model, which can then output the predicted location of the transmitter corresponding to the real-time IQ sample in the test data based on the set of test data.

[0104] The positioning method for ship environments provided in this embodiment adaptively shortens or extends the sampling period using exponential nonlinear and linear strategies based on the combined amplitude and phase fluctuation index of IQ samples. When the signal fluctuates drastically, the sampling frequency is increased to enhance data sampling density and capture transient characteristics of the samples. When the signal is stable, the sampling frequency is reduced to decrease redundant data and computational overhead. This adaptive sampling strategy not only significantly improves sampling efficiency but, more importantly, enhances the timeliness and representativeness of the model training samples. In the dynamically changing ship operating environment, this adaptive sampling strategy effectively enhances the coverage of key disturbance states by the training samples, enabling the model to capture more comprehensive environmental interference patterns and spatial distribution characteristics during the training phase. This improves the model's prediction accuracy and generalization ability in complex environments. Furthermore, by incorporating real-time attitude information and position offset of the ship into the positioning modeling, more physically interpretable high-dimensional training samples are constructed. This allows the model to learn the interference patterns of attitude and offset on the signal waveform, more accurately reconstructing the relationship between IQ samples and the transmitter's true position, significantly improving the model's generalization ability and positioning accuracy in dynamic environments. In summary, by combining adaptive sampling mechanisms with attitude and displacement offset fusion modeling, the localization model can more sensitively perceive dynamic environmental changes, ultimately achieving a more stable and accurate localization effect.

[0105] Subsequently, low-quality data is eliminated by combining a phase stability index screening mechanism. Multi-level preprocessing, including bandpass filtering, phase normalization, amplitude shaping, and data augmentation, effectively suppresses ship multipath effects and noise interference. Then, an initial positioning model is established and trained using a training sample set. The weights and biases of each layer in the initial positioning model are optimized to obtain a well-trained positioning model. Finally, the real-time acquired IQ samples, combined with the ship's attitude and position offset, are input into the well-trained positioning model, which can output a high-precision transmitter predicted position. In this way, the positioning method applied to the ship environment can not only significantly improve the robustness, data validity, and real-time positioning accuracy of the positioning model in complex ship environments, but also optimize the utilization of computational resources of the positioning model.

[0106] Specifically, Figure 3 This is a flowchart of Embodiment 2 of the positioning method applied to a ship environment provided in this application. Please refer to... Figure 3 Based on the above embodiments, the filtering and preprocessing of multiple sets of continuously collected positioning-related data includes:

[0107] S301. For each set of positioning-related data, determine the phase stability index of the positioning-related data based on the phase value of the IQ sample in the positioning-related data and the phase value of the IQ sample in all positioning-related data preceding the positioning-related data.

[0108] The phase stability index is used to characterize the degree of phase change of the IQ samples in the positioning-related data.

[0109] It should be noted that the phase stability index in a set of positioning-related data can be determined by calculating the degree of phase change between adjacent IQ samples. For example, the stability of the wireless signal received by the receiver from the transmitter can be measured by the degree of phase change between the nth IQ sample and the (n-1)th IQ sample. A larger phase change indicates poor stability of the wireless signal received by the receiver, while a smaller phase change indicates high stability.

[0110] Furthermore, in one possible implementation, the phase stability index of this set of positioning-related data can be calculated according to the following formula:

[0111]

[0112] Wherein, PSI is the phase stability index of the Mth group of positioning-related data; φ m Let m be the phase value of the IQ sample. The phase mean of the M IQ samples preceding the location-related data of the Mth group; denoted as the phase deviation of the m-th IQ sample relative to the mean; j is the imaginary unit.

[0113] S302. Based on the phase stability index of each set of positioning-related data, select target positioning-related data from multiple sets of continuously collected positioning-related data whose phase stability index is less than or equal to a preset threshold.

[0114] Specifically, when the phase stability index of a set of location-related data is less than or equal to a preset threshold, it indicates that the phase change of the IQ sample in that set of location-related data is large, and the quality of the wireless signal transmitted by the transmitter corresponding to the IQ sample is poor, possibly due to severe distortion. Therefore, the IQ sample needs to be deleted. The location-related data can be filtered according to the following rules:

[0115] if PSI≤τ φ Discard the frame;

[0116] Where, τ φThe preset threshold is defined as ; PSI is the phase stability index of a set of positioning-related data.

[0117] In this step, the relevant stability index evaluation mechanism can effectively screen out low-quality IQ samples, thus avoiding inaccurate data from affecting subsequent positioning accuracy.

[0118] S303. For each set of target positioning related data, perform bandpass complex filtering on the IQ samples in the set of target positioning related data to obtain the bandpass complex processing result corresponding to the IQ sample.

[0119] It should be noted that during signal transmission in a ship environment, multipath effect is often one of the main causes of signal distortion. Due to the multipath propagation of the transmitted signal in the ship environment, multiple copies of the transmitted signal will arrive at the receiving antenna at different times, causing signal delay, attenuation, reflection and interference. In this case, the spectrum of the transmitted signal may be affected by low-frequency or high-frequency noise, thereby affecting the quality of the IQ sample in the transmitted signal.

[0120] In this step, to eliminate these interferences, the signal is preliminarily processed using bandpass filtering. This suppresses low-frequency and high-frequency signal components in the transmitted signal, filters out noise and invalid signals outside the desired frequency band, and improves the quality of the transmitted signal. Specifically, bandpass filtering of the transmitted signal can be performed according to the following formula:

[0121] V (1) (n)=V(n)*h BP (n);

[0122] Among them, V (1) (n) represents the result of bandpass complex number processing; V(n) represents the original IQ sample at the nth sampling point; h BP (n) represents the impulse response of the bandpass filter; V(n)*h BP (n) represents the convolution operation.

[0123] S304. Perform phase normalization processing on the bandpass complex number processing result to obtain the phase normalization processing result.

[0124] It should be noted that the metallic structures and dynamic changes in the ship's environment can cause phase shifts in the signal. Therefore, to eliminate these phase shifts caused by environmental factors within the ship, the IQ samples undergo phase normalization to align the signal phase to a standard reference value. Specifically, the bandpass complex number processing result can be phase normalized according to the following formula:

[0125]

[0126] Among them, V (2) (n) represents the phase normalization result of the nth sampling point; V (1) (n) represents the bandpass complex number processing result of the nth sampling point; ∑ n V (1) (n) is the sum of the bandpass complex number processing results of all sampling points before the nth sampling point; j is the imaginary unit.

[0127] S305. Perform amplitude shaping on the phase normalization result to obtain the amplitude shaping result.

[0128] In this step, the signal strength may be too low during signal transmission within the ship, affecting the stability of subsequent processing. Therefore, it is necessary to perform amplitude shaping on the transmitted signal to adjust the amplitude of the transmitted signal to a uniform range.

[0129] Specifically, in one possible implementation, the amplitude shaping of the phase normalization result can be performed according to the following formula:

[0130]

[0131] Among them, V (3) (n) represents the amplitude shaping result of the nth sampling point; V (2) (n) represents the phase normalization result of the nth sampling point; |V (2) (n)| represents the amplitude of the phase-normalized result of the nth sampling point; μ r σ is the mean amplitude of the phase normalization result of n sampling points; r denoted as the standard deviation of the amplitude of the phase normalization result of n sampling points; j is the imaginary unit.

[0132] S306. Perform data augmentation on the amplitude shaping result to obtain the data augmentation result, and replace the IQ samples in the target positioning related data with the data augmentation result to obtain the initial training samples corresponding to the target positioning related data.

[0133] Specifically, in one possible implementation, the data augmentation steps include:

[0134] (1) Based on the amplitude shaping results of the IQ data in the target positioning related data and the amplitude shaping results of the IQ data in the previous target positioning related data before the target positioning related data, determine the sliding differential processing results of the IQ data in the target positioning related data.

[0135] It should be noted that in the process of data augmentation of amplitude shaping results, sliding differential augmentation is required first to enhance the instantaneous features in amplitude shaping results, highlight signal changes related to the desired location, and suppress signal components that are static or slowly changing.

[0136] Specifically, in one possible implementation, the amplitude shaping result can be subjected to a sliding difference operation according to the following formula:

[0137] V (4) (n)=V (3) (n)-V (3) (n-1);

[0138] Among them, V (4) (n) represents the sliding difference processing result of the nth sampling point; V (3) (n) represents the amplitude shaping result of the nth sampling point; V (3) (n-1) represents the magnitude shaping result at the (n-1)th point.

[0139] (2) The amplitude jittering and phase rotation processing are performed on the sliding differential processing result to obtain the data enhancement processing result.

[0140] In this step, amplitude jitter and phase rotation processing can be performed on the sliding differential processing results according to the following formulas:

[0141]

[0142] in, V (4) This is the result of the sliding difference processing; 'a' is the amplitude jitter factor, which follows a mean of 1 and a variance of . The normal distribution θ is the phase rotation angle, which follows a uniform distribution in [-α, α]. j is the imaginary unit.

[0143] Furthermore, by replacing the IQ samples in the target localization related data with the data augmentation results, the initial training samples corresponding to the target localization related data are obtained.

[0144] S307. Randomly discard some sampling points in the initial training samples to obtain the training sample set.

[0145] In this step, some sampling points in the initial training sample are randomly dropped to simulate different multipath distortions and channel interferences in a ship environment, and finally a training sample set is obtained.

[0146] It should be noted that the number of randomly lost sample points can be selected according to actual needs, and this application does not limit it. For example, in one embodiment, the number of randomly lost sample points accounts for 5-10% of the initial training samples.

[0147] The ship environment positioning method provided in this embodiment can eliminate the distortion of IQ samples caused by multipath effects, signal attenuation and dynamic changes of the ship by screening and preprocessing multiple sets of positioning-related data in the whole chain. This improves the quality of IQ samples in multiple sets of positioning-related data, generates a high-quality training sample set, and provides a reliable data foundation for subsequent training of the initial positioning model based on the training sample set.

[0148] Corresponding to the aforementioned embodiment of a positioning method applied to a ship environment, this application also provides an embodiment of a positioning device applied to a ship environment.

[0149] Figure 4 This is a structural schematic diagram of the positioning device applied to a marine environment provided in this application. Please refer to... Figure 4 The apparatus provided in this embodiment includes a data acquisition module 410, a determination module 420, a processing module 430, a training module 440, and a prediction module 450. The data acquisition module 410 is used to acquire a set of positioning-related data when the current sampling period arrives. This set of positioning-related data includes IQ samples obtained by the receiver during the reception of wireless signals transmitted from the transmitter, the current attitude of the hull, the current position offset of the hull, and the location of the transmitter corresponding to the IQ samples.

[0150] The determining module 420 is used to determine the joint fluctuation index of the IQ sample based on the amplitude change and phase change of the IQ sample, and to determine the update sampling period corresponding to the next sampling based on the joint fluctuation index, as well as the preset minimum sampling period and the preset standard sampling period.

[0151] The acquisition module 410 is used to perform the next round of data acquisition when the update sampling period arrives, and repeat the above steps.

[0152] The processing module 430 is used to filter and preprocess multiple sets of continuously collected positioning-related data to obtain a training sample set; wherein each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample.

[0153] The training module 440 is used to train an initial localization model based on the training sample set to obtain a trained localization model.

[0154] The prediction module 450 is used to input a set of real-time collected test data into the trained positioning model, so that the trained positioning model can output the predicted position of the transmitter corresponding to the real-time IQ sample in the set of test data based on the set of test data; wherein, the set of test data includes the real-time attitude of the hull, the real-time position offset of the hull, and the real-time IQ sample.

[0155] The apparatus of this embodiment can be used to perform... Figure 1 The steps of the method embodiment shown are similar in principle and process, and will not be repeated here.

[0156] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0157] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0158] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A positioning method applied to a shipboard environment, characterized in that, The positioning method includes: When the current sampling period arrives, a set of positioning-related data is collected; wherein, the set of positioning-related data includes IQ samples obtained by the receiver in the process of receiving wireless signals transmitted from the transmitter, the current attitude of the hull, the current position offset of the hull, and the location of the transmitter corresponding to the IQ sample. Based on the amplitude and phase changes of the IQ samples, the joint fluctuation index of the IQ samples is determined, and based on the joint fluctuation index, the preset minimum sampling period and the preset standard sampling period, the update sampling period corresponding to the next sampling is determined. When the update sampling period arrives, the next round of data acquisition is performed, and the above steps are repeated; Multiple sets of continuously collected positioning-related data are filtered and preprocessed to obtain a training sample set; each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample. The initial localization model is trained based on the training sample set to obtain the trained localization model. A set of real-time collected test data is input into the trained positioning model, so that the trained positioning model outputs the predicted position of the transmitter corresponding to the real-time IQ sample in the set of test data; wherein, the set of test data includes the real-time attitude of the ship, the real-time position offset of the ship, and the real-time IQ sample.

2. The method according to claim 1, characterized in that, The filtering and preprocessing of multiple sets of continuously collected positioning-related data includes: For each set of positioning-related data, a phase stability index is determined based on the phase value of the IQ sample in the positioning-related data and the phase values ​​of the IQ samples in all positioning-related data preceding the positioning-related data; wherein, the phase stability index is used to characterize the degree of phase change of the IQ sample in the positioning-related data. Based on the phase stability index of each set of positioning-related data, target positioning-related data with a phase stability index less than or equal to a preset threshold are selected from multiple sets of continuously collected positioning-related data. For each set of target localization related data, bandpass complex filtering is performed on the IQ samples in the set of target localization related data to obtain the bandpass complex processing result corresponding to the IQ sample; The bandpass complex number processing result is subjected to phase normalization processing to obtain the phase normalization processing result; The phase normalization result is then subjected to amplitude shaping to obtain the amplitude-shaping result. Data augmentation is performed on the amplitude shaping result to obtain the data augmentation result, and the IQ samples in the target localization related data are replaced with the data augmentation result to obtain the initial training samples corresponding to the target localization related data. The training sample set is obtained by randomly discarding some sampling points from the initial training sample set.

3. The method according to claim 1, characterized in that, The step of determining the joint volatility index of the IQ sample based on the amplitude and phase changes of the IQ sample includes: Based on the amplitude change and the preset amplitude change threshold, a first normalized fluctuation index corresponding to the amplitude change is determined. Based on the phase change amount and the preset phase change threshold, a second normalized fluctuation index corresponding to the phase change amount is determined. Based on the first normalized volatility index and the second normalized volatility index, the first weight corresponding to the phase and the second weight corresponding to the amplitude are determined. The first normalized volatility index and the second normalized volatility index are weighted according to the first weight and the second weight to obtain the joint volatility index.

4. The method according to claim 1, characterized in that, The step of determining the update sampling period corresponding to the next sampling based on the joint fluctuation index, as well as the preset minimum sampling period and the preset standard sampling period, includes: The target adjustment strategy corresponding to the joint volatility index is determined based on the joint volatility index; wherein, different joint volatility index ranges correspond to different adjustment strategies; the different adjustment strategies include an exponential nonlinear adjustment strategy and a linear adjustment strategy. When the joint volatility index is greater than a preset threshold, the target adjustment strategy corresponding to the joint volatility index is the exponential nonlinear adjustment strategy; otherwise, the target adjustment strategy corresponding to the joint volatility index is the linear adjustment strategy. The update sampling period is determined based on the target adjustment strategy, the minimum sampling period, and the standard sampling period.

5. The method according to claim 4, characterized in that, The step of determining the updated sampling period based on the target adjustment strategy, the minimum sampling period, and the standard sampling period includes: When the target adjustment strategy is an exponential nonlinear adjustment strategy, the update sampling period is determined according to the first formula; otherwise, the update sampling period is determined according to the second formula. The first formula is: Where, Δt k The update sampling period corresponds to the k-th sampling; γ is the sensitivity factor, γ > 0; Δt min The minimum sampling period; Δt nom The standard sampling period; S k For joint volatility indicators; The second formula is: Δt k =Δt min +(1-S k )·(Δt nom -Δt min ); Where, Δt k The update sampling period corresponding to the k-th sampling; Δt min The minimum sampling period; Δt nom The standard sampling period; S k This is a combined volatility indicator.

6. The method according to claim 2, characterized in that, The data augmentation steps include: Based on the amplitude shaping results of the IQ data in this set of target positioning related data, and the amplitude shaping results of the IQ data in the previous set of target positioning related data, the sliding differential processing results of the IQ data in this set of target positioning related data are determined. The amplitude jittering and phase rotation processing are applied to the sliding differential processing result to obtain the data augmentation processing result.

7. The method according to claim 3, characterized in that, Based on the first normalized volatility index and the second normalized volatility index, the first weight corresponding to the phase and the second weight corresponding to the amplitude are determined, including: The first weight is calculated according to either the third or fourth formula, and the second weight is determined based on the first weight; wherein the third formula is: Among them, a k It is the first weight; It is the first normalized volatility indicator; This is the second normalized volatility index; ε is the perturbation term, preset to a positive value. The fourth formula is: Among them, a k As the first weight; a k-1 λ is the phase weight of the previous sampling point; λ is the smoothing factor, with a value range of [0, 1]. It is the first normalized volatility indicator; ε is the second normalized volatility index; ε is the perturbation term, preset to a positive value.

8. The method according to claim 2, characterized in that, The phase normalization process includes: Phase normalization is performed according to the following formula: Among them, V (2) (n) is the IQ sample vector of the nth sampling point after phase normalization; V (1) (n) represents the bandpass complex number processing result of the nth sampling point; ∑ n V (1) (n) represents the summation of the bandpass complex number processing results for n sampling points.

9. The method according to claim 6, characterized in that, The process of performing amplitude jittering and phase rotation processing on the sliding differential processing result to obtain the data augmentation result includes: Perform amplitude jittering and phase rotation processing according to the following formulas: in, To augment the data results; V (4) This is the result of the sliding difference processing; 'a' is the amplitude jitter factor, which follows a mean of 1 and a variance of . The normal distribution θ is the phase rotation angle, which follows a uniform distribution in [-α, α].

10. A positioning device for use in a marine environment, characterized in that, The device includes an acquisition module, a determination module, a processing module, a training module, and a prediction module; wherein, The acquisition module is used to acquire a set of positioning-related data when the current sampling period arrives; wherein, the set of positioning-related data includes IQ samples obtained by the receiver in the process of receiving wireless signals transmitted from the transmitter, the current attitude of the hull, the current position offset of the hull, and the location of the transmitter corresponding to the IQ sample. The determining module is used to determine the joint fluctuation index of the IQ sample based on the amplitude change and phase change of the IQ sample, and to determine the update sampling period corresponding to the next sampling based on the joint fluctuation index, as well as the preset minimum sampling period and the preset standard sampling period. The acquisition module is used to perform the next round of data acquisition when the update sampling period arrives, and repeat the above steps. The processing module is used to filter and preprocess multiple sets of continuously collected positioning-related data to obtain a training sample set; wherein each training sample includes an IQ sample, the attitude of the hull, the position offset of the hull, and the location of the transmitter corresponding to the IQ sample. The training module is used to train an initial localization model based on the training sample set to obtain a trained localization model. The prediction module is used to input a set of real-time collected test data into the trained positioning model, so that the trained positioning model can output the predicted position of the transmitter corresponding to the real-time IQ sample in the set of test data based on the set of test data; wherein, the set of test data includes the real-time attitude of the ship, the real-time position offset of the ship, and the real-time IQ sample.

Images