Floating particle and multi-floating-particle high-precision positioning method

Abstract

The invention relates to a floating particle and a multi-floating-particle high-precision positioning method, and the floating particle comprises a signal collection module, an environment monitoring module, a positioning module, a communication module, a particle signal forwarding module and an energy management module which are integrated in a housing. By designing a multi-buoy cooperative positioning system and combining an environment adaptive calibration mechanism and a low-power-consumption communication technology, efficient and reliable floating particle real-time monitoring is realized in a complex dynamic marine environment, the requirements of multiple fields such as environmental protection, scientific research and disaster emergency are met, and high-precision real-time positioning of floating particles in the complex marine environment is realized.

Description

Technical Field

[0001] The invention relates to a floating particle and a high-precision positioning method for multiple floating particles, and belongs to the field of marine wireless sensor network node positioning. Background Art

[0002] With the growing demand for marine ecological protection and environmental governance, marine floating particle positioning technology has important applications in many fields, including tracking the distribution of plastic waste, controlling the spread of oil spills, monitoring marine ecology, collecting buoy observation data, protecting fishery resources, and accurately recovering floating instruments. These scenarios place higher demands on real-time, high-precision positioning of the distribution of floating objects, especially in long-term monitoring, dynamic tracking, and emergency response.

[0003] At present, the methods for locating floating particles mainly include the following: satellite-based ocean monitoring and positioning technology, drone visual monitoring, single floating node monitoring technology, acoustic positioning technology, and radar monitoring technology. Regarding satellite-based ocean monitoring and positioning technology, satellite monitoring can cover a wide area of ​​the sea and is suitable for large-scale floating object distribution statistics. However, due to resolution limitations, it is difficult to accurately locate small floating particles, and the real-time performance is poor, which cannot meet the needs of dynamic tracking. Regarding drone visual monitoring, drones carry cameras or sensors, so floating objects in local areas can be identified and tracked. However, due to battery life and environmental conditions (such as wind speed and weather), it is difficult to operate for a long time and over a large range. The monitoring technology of a single floating node uses a single float to monitor floating objects through GPS or signal ranging. Due to the lack of multi-point coordination, its measurement results are easily affected by multipath effects and wave interference, and the positioning accuracy is low. Acoustic positioning technology uses the propagation characteristics of sound waves to locate underwater or surface objects. In complex marine environments, sound wave propagation is greatly affected by factors such as temperature, salinity and water depth, which can easily cause ranging errors; in addition, acoustic equipment has high power consumption and is not suitable for long-term operation. Radar monitoring technology uses high-frequency radar to detect the location and distribution of floating objects through reflected signals. This technology has insufficient detection sensitivity for smaller floating particles (such as plastic microparticles) and has high equipment costs, making it unsuitable for large-scale marine applications.

[0004] Therefore, it is necessary to provide a new method for locating floating particles to overcome the problems in the existing technology such as "insufficient positioning accuracy, especially for small floating particles", "unstable monitoring results affected by the complex marine environment", "high cost, high power consumption, and difficulty in meeting long-term operation requirements", "large communication delay and poor real-time performance" and "lack of large-scale collaborative monitoring capabilities". Summary of the invention

[0005] The present invention provides a floating particle and a high-precision positioning method for multiple floating particles, which realizes high-precision real-time positioning of floating particles in a complex marine environment through the collaborative work of multiple floats, environmental adaptive calibration and low-power communication technology.

[0006] The technical solution adopted by the present invention to solve its technical problem is: A floating particle, comprising a signal acquisition module, an environment monitoring module, a positioning module, a communication module, a particle signal forwarding module and an energy management module, all of which are integrated in a housing; The signal acquisition module includes an optical sensor, an ultrasonic sensor, and an electromagnetic sensor, which are used to detect signals of floating particles and transmit the signals to a central processing unit deployed at a coastal station or on a vessel; The environmental monitoring module includes an anemometer, a turbidity sensor, an acceleration sensor, and a temperature sensor, which are used to collect environmental data in real time and assist in positioning error correction; The positioning module is used to measure the distance or position relationship between floating particles; The communication module is based on the LoRa communication protocol to perform long-distance data transmission; The particle signal forwarding module expands the coverage of floating particle signals through a mutual forwarding mechanism; The energy management module includes an integrated solar or wave energy generation system; A method for high-precision positioning of multiple floating particles comprises the following steps: Step S1, placing a number of floating particles into the ocean; Step S2, starting the floating particles, and the signal acquisition module sends the floating particle signals and the collected environmental data to the central processing unit through the communication module; Step S3, constructing a multi-objective processing model, separating the features of different floating particles from the signals of a number of floating particles, assigning an identifier to each floating particle, and outputting the predicted coordinates of a single floating particle; Step S4, narrowing the target position range of the floating particles based on the predicted coordinates obtained in step S3; Step S5, within the target position range, a floating particle is selected as a target floating node, a signal propagation model of a single floating particle is established through the time difference of the signal of the target floating node reaching different floating particles, and the position of the target floating node is determined; Step S6, based on the environmental data collected in step S2, using a neural network to dynamically adjust the propagation speed in the signal propagation model established in step S5 to correct errors caused by environmental changes; Step S7, obtaining a corrected signal propagation model by using the corrected propagation velocity in step S6, and combining it with the target floating node position determined in step S5 to construct an optimized single floating particle trajectory prediction model; Furthermore, in step S3, the steps of constructing a multi-objective processing model to output the predicted coordinates of a single floating particle are specifically as follows: Step S31, constructing a model architecture of a multi-objective processing model, including an input layer, a feature extraction layer, and a fusion classification layer; Step S32, the input layer converts the original signal x ( t ) is converted into a spectrum by fast Fourier transform X ( F ), the feature extraction layer extracts the frequency feature vector and the amplitude feature vector respectively through the dual-branch convolutional network, the fusion classification layer concatenates the dual-branch features, and outputs the classification results through the fully connected layer. The classification result output formula is:

[0007] In the formula, y is the classification result, is the frequency eigenvector, is the amplitude eigenvector, W is the weight, b is bias; Step S33, each floating particle is regarded as a target, defined as k , generate a unique identifier for a single floating particle based on the classification results ,Right now

[0008] In the formula, For identification, For the goal k The frequency, For the goal k The amplitude of For the goal k Timestamp of first occurrence, Hash The function ensures uniqueness; Step S34, using a recursive neural network to process the normalized frequency and amplitude to obtain the time series data hidden at the current moment, the hidden state update formula is:

[0009] In the formula, h t For the goal k The hidden state at the current moment, h t-1 For the goal k The hidden state at the previous moment,x t For the goal k Current moment input, i.e. normalized frequency F t and amplitude A t , W h is the weight at the current moment, b h is the bias at the current moment, σ is the Sigmoid activation function; Step S35, output the predicted coordinates of a single floating particle, the prediction formula is:

[0010] In the formula, The output contains the predicted coordinates for the next moment ( x ′, y ′, z ′), W y is the weight containing the next moment, h t For the goal k The hidden state at the current moment, b y is the bias that contains the next moment; Furthermore, in step S5, the steps of constructing the signal propagation speed model are specifically as follows: Step S511, select a floating particle from a number of floating particles, define it as a target floating node, and calculate the time difference between the other floating particles and the target floating node.

[0011] In the formula, is the time difference, is the time of other floating particles, is the time of the target floating node, N is the total number of floating particles; Step S512: construct a signal propagation model:

[0012] In the formula, is the distance difference corresponding to the time difference, is the propagation speed, is the time difference; Furthermore, in step S5, the step of determining the position of the target floating node is specifically as follows: Step S521, constructing a hyperbolic surface equation corresponding to the time difference between every two floating particles, specifically:

[0013] In the formula, All are obtained through the predicted coordinates of the single floating particle output in step S35; like N ≥4, N The floating particles form multiple hyperbolic equations, and an overdetermined system of equations is established: ; Step S522, optimizing the constructed hyperbolic surface equation to determine the target floating node position, the optimization formula is:

[0014] In the formula, , M is the total number of equations; Furthermore, in step S6, the specific steps of dynamically adjusting the propagation speed in the signal propagation model are: Step S61, construct a signal propagation speed model, and the correction formula is:

[0015] In the formula, is the propagation speed correction value, is the standard propagation speed, E is an environmental factor, wherein the environmental factor is temperature, humidity or wind speed, is a nonlinear function used for environmental correction; Step S62, evaluate the accuracy of the correction formula prediction, adjust the parameters of the correction formula according to the evaluation results, take the environmental factors as input, and finally output the propagation speed correction value, where the evaluation formula is

[0016] In the formula, is the corrected value of the propagation speed predicted by the correction formula, is the corrected value of the actual propagation speed obtained by offline data collection, ζ is the loss function of the correction formula; Furthermore, in step S7, the specific steps of constructing the optimized single floating particle trajectory prediction model are: Step S71, using a convolutional neural network, extracts multidimensional features from the time difference matrix and the signal matrix generated when determining the position of the target floating node, and fuses the extracted features to output an accurate position estimate. The fusion formula is:

[0017] In the formula, is the predicted position vector, T is the time difference matrix, R is the signal matrix; Step S72, evaluate the difference between the predicted position and the actual position, adjust the parameters of the single floating particle trajectory prediction model, and finally accurately locate the floating particle position, where the evaluation formula is:

[0018] In the formula, To evaluate the loss function of the formula, is the predicted position vector, is the known true position, is the Euclidean distance.

[0019] Through the above technical solution, compared with the prior art, the present invention has the following beneficial effects: 1. The floating particles provided by the present invention utilize low-power LoRa communication and solar power supply technology, are easy to produce and deploy, can be covered by a large-scale floating network, and realize the long-term monitoring capability of the equipment; 2. The high-precision positioning method for multiple floating particles provided by the present invention meets the real-time monitoring requirements of wide sea areas through a distributed network structure and real-time algorithm; 3. The multi-floating particle high-precision positioning method provided by the present invention solves the signal interference problem in complex marine environments and realizes high-precision particle positioning through multi-floating collaborative positioning and dynamic error correction; 4. The high-precision positioning method for multiple floating particles provided by the present invention is a positioning solution with low cost, strong adaptability and high reliability, which meets the application requirements of multiple scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0021] Figure 1 This is a comparison example of the root mean square error between the multi-objective processing method of the present invention and the Kalman filtering method; Figure 2 This is a comparison example of the false alarm rate between the multi-objective processing method of the present invention and the Kalman filtering method; Figure 3 This is an example diagram of the distributed computing method provided by the present invention. DETAILED DESCRIPTION

[0022] The present invention will now be described in further detail with reference to the accompanying drawings.

[0023] As described in the background technology, due to the complex and dynamic characteristics of the marine environment (such as waves, wind speed, tides, etc.), existing positioning methods are usually difficult to ensure reliability and accuracy under these conditions, and there are significant technical limitations. For example, satellite-based positioning technology has high costs, poor real-time performance, and its accuracy is greatly affected by the weather. The visual monitoring of floating objects by drones carrying cameras has a short battery life and cannot achieve long-term monitoring. A single float monitors floating objects through GPS or signal ranging technology, lacks multi-point coordination, has large positioning errors, and is easily affected by environmental interference.

[0024] Therefore, the present application provides a method for high-precision positioning of multiple floating particles, aiming to solve the technical problem of high-precision real-time positioning and monitoring of floating particles in the marine environment, and realize high-precision real-time positioning of multiple floating particles in a complex marine environment. If a number of floating particles are to be positioned in real time with high precision, the structure of the floating particles is also the key. The floating particles provided by the present application include a signal acquisition module, an environmental monitoring module, a positioning module, a communication module, a particle signal forwarding module, and an energy management module, and all the aforementioned modules are integrated inside a housing. The signal acquisition module includes an optical sensor, an ultrasonic sensor, an electromagnetic sensor or other applicable detection device, which is used to detect the signal of the floating particles and transmit the signal to a central processing unit deployed at a coastal station or on a ship. The environmental monitoring module includes an anemometer, a turbidity sensor, an acceleration sensor, and a temperature sensor, which are used to collect environmental data (such as wind speed, current speed, temperature, wave height, wave period, seawater turbidity) in real time to assist in positioning error correction. The positioning module supports GPS positioning and wireless signal ranging (such as RSSI, TOA) functions, which are used to measure the distance or position relationship between floating particles. The communication module is based on the LoRa communication protocol and supports low-power, long-distance data transmission. The particle signal forwarding module enhances the communication capability between particles, expands the signal coverage through the mutual forwarding mechanism, and improves the reliability of data transmission. The energy management module includes an integrated solar or wave power generation system to ensure long-term power supply capability.

[0025] Regarding the central processing unit, it is responsible for receiving data from floating particles (radio signals actively transmitted), performing positioning algorithms, and outputting the real-time position of floating particles.

[0026] Next is a method for high-precision positioning of multiple floating particles provided by the present application, which specifically includes the following steps: Step S1, placing a number of the floating particles into the ocean, each floating particle is a basic unit in the distributed network, responsible for collecting signals, processing and transmitting.

[0027] Step S2, after starting the floating particles, they perform self-positioning and establish a distributed LoRa communication network; the signal acquisition module sends the floating particle signals and the collected environmental data to the central processing unit through the communication module.

[0028] Since there are many floating particles on the sea, resulting in complex signals, we need to separate the characteristics of different floating particles from the mixed signals to preliminarily predict the position of a single floating particle and roughly narrow the position of the floating particle to a certain range, which can avoid signal loss, improve the calculation speed, and facilitate the tracking of floating particles. That is, step S3, the central processing unit receives the data of multiple floating particles, builds a multi-target processing model, separates the characteristics of different floating particles from the signals of several floating particles, assigns an identifier to each floating particle, and outputs the predicted coordinates of a single floating particle.

[0029] The specific steps are: Step S31, constructing a model architecture of a multi-objective processing model, including an input layer, a feature extraction layer, and a fusion classification layer; Step S32, the input layer converts the original signal x(t) Converted into spectrum by fast Fourier transform X(F) The feature extraction layer extracts the frequency feature vector and the amplitude feature vector respectively through a dual-branch convolutional network (CNN). Preferably, the frequency branch: 1D convolution kernel (size = 5, channels = 32) → extracts the local features in the frequency domain. The amplitude branch: 1D convolution kernel (size = 3, channels = 16) → extracts the amplitude change features. The fusion classification layer concatenates the dual-branch features and outputs the classification results through the fully connected layer. The classification result output formula is:

[0030] In the formula, y is the classification result, is the frequency eigenvector, is the amplitude eigenvector, W is the weight, b is bias; Step S33, each floating particle is regarded as a target, defined as k , generate a unique identifier for a single floating particle based on the classification results ,Right now

[0031] In the formula, For identification, For the goal k The frequency, For the goal k The amplitude of For the goal k Timestamp of first occurrence, Hash The function ensures uniqueness; Next, a recurrent neural network (RNN) is used to process the time series signal characteristics, track the dynamic changes of floating particles, and improve the accuracy of target differentiation and recognition. The process is as follows: Step S34, a recurrent neural network is used to process the normalized frequency and amplitude to obtain the hidden time series data at the current moment. The hidden state update formula is:

[0032] In the formula, h t For the goal k The hidden state at the current moment, h t-1 For the goal k The hidden state at the previous moment, For the goal k Current moment input, i.e. normalized frequency and amplitude , is the weight at the current moment, is the bias at the current moment, is the Sigmoid activation function; Step S35, output the predicted coordinates of a single floating particle, the prediction formula is:

[0033] In the formula, The output contains the predicted coordinates for the next moment , is the weight containing the next moment, h t For the goal k The hidden state at the current moment, b y Contains the bias for the next moment.

[0034] Regarding the benefits of using this method, it is compared here with the currently commonly used Kalman filtering method, and the condition settings are shown in Table 1.

[0035] Table 1

[0036] from Figure 1 The root mean square error is Figure 2 From the comparison of false alarm rates, it can be seen that under different sea conditions, the performance obtained by the multi-target processing method adopted in this application is obviously higher than that obtained by the Kalman filtering method.

[0037] Thus, in step S4, the target position range of the floating particles is narrowed down based on the predicted coordinates obtained in step S3.

[0038] After narrowing the target range, the floating particles are accurately located based on the TDOA algorithm. The position of the floating particles is determined by measuring the time difference between the target floating node signal reaching different floating particles.

[0039] Specifically, in step S5, firstly, a signal propagation speed model is constructed, and the specific steps are as follows: Step S511, select a floating particle from a number of floating particles, define it as a target floating node, and calculate the time difference between the other floating particles and the target floating node.

[0040] In the formula, is the time difference, is the time of other floating particles, is the time of the target floating node, N is the total number of floating particles; Step S512: construct a signal propagation model:

[0041] In the formula, is the distance difference corresponding to the time difference, is the propagation speed, For the time difference.

[0042] Continue to determine the location of the target floating node. The specific steps are as follows: Step S521: The time difference between each two floating particles corresponds to a hyperbolic surface equation. Taking the target floating node and another floating particle as an example, a hyperbolic surface equation corresponding to the time difference is constructed. Specifically,

[0043] In the formula, All are obtained through the predicted coordinates of the single floating particle output in step S35; N floating particles (numbered from 1 to N) can form multiple double-sided curves. The above is the general form. N ≥4, N A floating particle can form multiple hyperbolic equations, so an overdetermined system of equations is established: ; Step S522, using the least square method and neural network method to process noise and environmental errors, that is, optimizing the constructed hyperbolic equation to determine the target floating node position, the optimization formula is:

[0045] In the formula, , M is the total number of equations.

[0046] Next is the key technical details of this application. In order to enhance positioning accuracy and reliability, it is necessary to add an optimization strategy to the provided TDOA algorithm, namely step S6, using environmental data such as temperature, humidity, wind speed, and ocean current collected by environmental sensors, and using a neural network to dynamically adjust the propagation speed in the signal propagation model established in step S5 to correct errors caused by environmental changes. This design is an environmental adaptive calibration mechanism that dynamically adjusts the positioning algorithm by collecting data such as wind speed, ocean current, and temperature in real time to eliminate interference from environmental changes on signal propagation.

[0047] The specific steps are: Step S61, construct a signal propagation speed model, and the correction formula is:

[0048] In the formula, is the propagation speed correction value, is the standard propagation speed, E is an environmental factor, wherein the environmental factor is temperature, humidity or wind speed, is a nonlinear function used for environmental correction, i.e., a neural network training model; the standard propagation speed here is It assumes that known speeds (such as the speed of wireless signals is close to the speed of light and the speed of sound waves is about 1500 m / s) are propagated in a uniform environment.

[0049] Construct a multi-layer feedforward neural network (MLP) based on environmental factors is the input and output propagation velocity correction value .

[0050] Step S62, evaluate the accuracy of the prediction of the modified formula, and adjust the parameters of the neural network accordingly to make its prediction closer to the true value, where the evaluation formula is:

[0051] In the formula, is the corrected value of the propagation speed predicted by the correction formula, is the corrected value of the actual propagation speed obtained by offline data collection, and ζ is the loss function of the correction formula.

[0052] After training is completed, the model is deployed to floating particles, which can infer the propagation speed correction results in real time, improving the efficiency and accuracy of the correction.

[0053] Step S7, constructing an optimized single floating particle trajectory prediction model based on the signal propagation model corrected by the propagation speed correction formula in step S6 and the target float node position determined in step S5.

[0054] This step is actually a data fusion process, combining RSSI (received signal strength indication) and multimodal sensor data to further optimize the TDOA calculation results through a deep neural network model. The specific steps are: Step S71, using a convolutional neural network (CNN), extract multi-dimensional features from the time difference matrix and signal matrix generated when determining the position of the target floating node, and fuse the extracted features to output an accurate position estimate. The fusion formula is:

[0055] In the formula, is the predicted position vector, T is the time difference matrix, R is the signal matrix; The network model adopts supervised learning method and is trained based on a data set with location labels, which can effectively reduce the interference of environmental noise and sensor error on the positioning results. Step S72, evaluate the gap between the predicted position and the actual position, adjust the parameters of the single floating particle trajectory prediction model, and finally accurately locate the floating particle position, where the evaluation formula is

[0056] In the formula, To evaluate the loss function of the formula, is the predicted position vector, is the known true position, is the Euclidean distance. The final output is the predicted and more accurate floating particle position.

[0057] In the above description of the positioning method, when it is necessary to separate the characteristics of different floating particles from the mixed signal, some floating particles float far away, and their signals cannot be directly transmitted to the control center. Therefore, it is necessary to use the floating particles closer to themselves to forward the signals through the particle signal forwarding module. However, there are many ways of signal forwarding, which need to be distinguished and determined. Therefore, this application continues to provide a distributed computing method, in which the floating nodes run a lightweight neural network (such as the CNN model mentioned above) to pre-process data and reduce the amount of data transmission.

[0058] A dynamic routing selection algorithm is added to floating particles, and a reinforcement learning model (such as DQN) is used to dynamically optimize the signal forwarding path, with the goal of minimizing communication delay and energy consumption:

[0059] In the formula, and is the weight parameter.

[0060] The Federated Learning framework is used to jointly train the neural network models of multiple floating particles. Each floating particle updates the model weight locally to improve the model generalization ability while ensuring data privacy:

[0061] In the formula, It is The weight parameter of each floating particle.

[0062] Figure 3 As shown, assuming that floating nodes A, floating nodes B and floating nodes C are far away from the selected positioning particles, after the above distributed computing method is implemented, a suitable link is successfully selected to form signal transmission with the positioning particles.

[0063] Through the above optimization, this application has achieved significant improvements in environmental adaptability, complex target processing capabilities and computing efficiency, and is particularly suitable for high-precision TDOA positioning scenarios in marine environments.

[0064] In summary, the floating particles provided by this application support long-term monitoring through LoRa communication and solar power supply, achieving the purpose of low power consumption. The multi-floating particle high-precision positioning method provided, multi-floating coordination and dynamic calibration meet the requirements of high precision, the distributed network and efficient algorithm meet the needs of real-time positioning of floating particles, the environmental adaptive calibration mechanism overcomes the interference of complex marine environment, and has strong adaptability. At the same time, the floating particles realize dynamic relay communication through the signal forwarding function, expand the signal coverage range, and can effectively transmit the signal to the floating node or Internet access device even in a long distance or complex marine environment.

[0065] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0066] The meaning of "and / or" described in this application means that the situations where each exists alone or both exist at the same time are included.

[0067] The term “connection” as used in this application may mean a direct connection between components or an indirect connection between components via other components.

[0068] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A floating particle, characterized in that: It includes a signal acquisition module, an environment monitoring module, a positioning module, a communication module, a particle signal forwarding module and an energy management module, and all of the above modules are integrated in one housing; The signal acquisition module includes an optical sensor, an ultrasonic sensor, and an electromagnetic sensor, which are used to detect signals of floating particles and transmit the signals to a central processing unit deployed at a coastal station or on a vessel; The environmental monitoring module includes an anemometer, a turbidity sensor, an acceleration sensor, and a temperature sensor, which are used to collect environmental data in real time and assist in positioning error correction; The positioning module is used to measure the distance or position relationship between floating particles; The communication module is based on the LoRa communication protocol to perform long-distance data transmission; The particle signal forwarding module expands the coverage of floating particle signals through a mutual forwarding mechanism; The energy management module includes an integrated solar or wave power generation system.

2. A high-precision positioning method for multiple floating particles, characterized in that: The specific steps include: Step S1, placing a plurality of floating particles as claimed in claim 1 into the ocean; Step S2, starting the floating particles, and the signal acquisition module sends the floating particle signals and the collected environmental data to the central processing unit through the communication module; Step S3, constructing a multi-objective processing model, separating the features of different floating particles from the signals of a number of floating particles, assigning an identifier to each floating particle, and outputting the predicted coordinates of a single floating particle; Step S4, narrowing the target position range of the floating particles based on the predicted coordinates obtained in step S3; Step S5, within the target position range, a floating particle is selected as a target floating node, a signal propagation model of a single floating particle is established through the time difference of the signal of the target floating node reaching different floating particles, and the position of the target floating node is determined; Step S6, based on the environmental data collected in step S2, using a neural network to dynamically adjust the propagation speed in the signal propagation model established in step S5 to correct errors caused by environmental changes; In step S7, the propagation speed corrected in step S6 is used to obtain a corrected signal propagation model, and combined with the target float node position determined in step S5, an optimized single floating particle trajectory prediction model is constructed.

3. The high-precision positioning method for multiple floating particles according to claim 2 is characterized in that: In step S3, the steps of constructing a multi-objective processing model to output the predicted coordinates of a single floating particle are specifically as follows: Step S31, constructing a model architecture of a multi-objective processing model, including an input layer, a feature extraction layer, and a fusion classification layer; Step S32, the input layer converts the original signal x ( t ) is converted into a spectrum by fast Fourier transform X ( F ), the feature extraction layer extracts the frequency feature vector and the amplitude feature vector respectively through the dual-branch convolutional network, the fusion classification layer concatenates the dual-branch features, and outputs the classification results through the fully connected layer. The classification result output formula is: ； In the formula, y is the classification result, is the frequency eigenvector, is the amplitude eigenvector, W is the weight, is bias; Step S33, each floating particle is regarded as a target, defined as , generate a unique identifier for a single floating particle based on the classification results ,Right now ； In the formula, For identification, For the goal k The frequency, For the goal k The amplitude of For the goal k Timestamp of first occurrence, Hash The function ensures uniqueness; Step S34, using a recursive neural network to process the normalized frequency and amplitude to obtain the time series data hidden at the current moment, the hidden state update formula is: ； In the formula, For the goal k The hidden state at the current moment, For the goal k The hidden state at the previous moment, For the goal k Current moment input, i.e. normalized frequency and amplitude , is the weight at the current moment, is the bias at the current moment, is the Sigmoid activation function; Step S35, output the predicted coordinates of a single floating particle, the prediction formula is: ； In the formula, The output contains the predicted coordinates for the next moment , is the weight containing the next moment, For the goal k The hidden state at the current moment, Contains the bias for the next moment.

4. The high-precision positioning method for multiple floating particles according to claim 3 is characterized in that: In step S5, the steps of constructing the signal propagation speed model are specifically as follows: Step S511, select a floating particle from a number of floating particles, define it as a target floating node, and calculate the time difference between the other floating particles and the target floating node. ； In the formula, is the time difference, is the time of other floating particles, is the time of the target floating node, N is the total number of floating particles; Step S512: construct a signal propagation model: ； In the formula, is the distance difference corresponding to the time difference, is the propagation speed, For the time difference.

5. The high-precision positioning method for multiple floating particles according to claim 4 is characterized in that: In step S5, the steps of determining the position of the target floating node are specifically as follows: Step S521, constructing a hyperbolic surface equation corresponding to the time difference between every two floating particles, specifically: ； In the formula, All are obtained through the predicted coordinates of the single floating particle output in step S35; like N ≥4, N The floating particles form multiple hyperbolic equations, and an overdetermined system of equations is established: ； Step S522, optimizing the constructed hyperbolic surface equation to determine the target floating node position, the optimization formula is: ； In the formula, , M is the total number of equations.

6. The high-precision positioning method for multiple floating particles according to claim 5, characterized in that: In step S6, the specific steps of dynamically adjusting the propagation speed in the signal propagation model are: Step S61, construct a signal propagation speed model, and the correction formula is: ； In the formula, is the propagation speed correction value, is the standard propagation speed, E is an environmental factor, wherein the environmental factor is temperature, humidity or wind speed, is a nonlinear function used for environmental correction; Step S62, evaluate the accuracy of the correction formula prediction, adjust the parameters of the correction formula according to the evaluation results, take the environmental factors as input, and finally output the propagation speed correction value, where the evaluation formula is ； In the formula, is the corrected value of the propagation speed predicted by the correction formula, is the corrected value of the actual propagation speed obtained by offline data collection, and ζ is the loss function of the correction formula.

7. The high-precision positioning method for multiple floating particles according to claim 6, characterized in that: In step S7, the specific steps of constructing the optimized single floating particle trajectory prediction model are: Step S71, using a convolutional neural network, extracts multidimensional features from the time difference matrix and the signal matrix generated when determining the position of the target floating node, and fuses the extracted features to output an accurate position estimate. The fusion formula is: ； In the formula, is the predicted position vector, is the time difference matrix, is the signal matrix; Step S72, evaluate the difference between the predicted position and the actual position, adjust the parameters of the single floating particle trajectory prediction model, and finally accurately locate the floating particle position, where the evaluation formula is: ； In the formula, To evaluate the loss function of the formula, is the predicted position vector, is the known true position, is the Euclidean distance.

Images