A cross-water-air medium communication method based on data fusion

Abstract

This invention discloses a cross-water-air medium communication method based on data fusion. The transmitting end is located underwater, and the receiving end is located above water, forming a multi-sensor system composed of millimeter-wave radar, lidar, and terahertz radar placed at equal intervals. The cross-water-air medium communication method receives vibration signals from the water surface through the millimeter-wave radar, lidar, and terahertz radar, and constructs the state vector and observation model required by the Kalman filter algorithm. Then, the water surface displacement and error covariance matrix are initialized, and the water surface displacement and covariance matrix at the next moment are predicted through the state transition model. The weights of each radar are calculated based on the radar's signal-to-noise ratio, bit error rate, and measurement error standard deviation. The water surface displacement and error covariance matrix are updated using the measurement data of each radar through the Kalman filter algorithm. Finally, the updated results of all radars are combined to obtain the final water surface displacement and error covariance matrix, thereby improving the quality of cross-water-air medium communication.

Description

Technical Field

[0001] This invention relates to the field of communications, and more particularly to a cross-water and air medium communication method based on data fusion. Background Technology

[0002] Direct communication across water and air media has broad potential applications. In the civilian sector, direct communication across water and air media is a crucial component of air-space-ground-sea observation networks, but it also constrains the real-time performance, convenience, and cost-effectiveness of data transmission from underwater to surface. In the military sector, the demand for information transmission across water and air media is increasingly strong, becoming a key issue restricting the improvement of the combat effectiveness of underwater vehicles. An emerging method for cross-water-air media communication provides an effective approach by detecting surface wave vibration signals excited by a sound source and inverting the information transmitted by the underwater sound source. This communication technology combines underwater acoustics and surface radio technology, and its feasibility for achieving direct communication across water and air media has been fully demonstrated, showing broad application prospects.

[0003] In cross-water / air communication combining underwater acoustics and electromagnetic waves, the complex marine environment makes it impossible to accurately and completely capture water surface vibrations with a single sensor, resulting in a decline in communication quality. For example... Figure 1 As shown, the communication quality of cross-water-air medium communication using a single sensor deteriorates sharply with increasing detection altitude and the influence of the marine environment. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a cross-water-air medium communication method based on data fusion. By employing three different sensors—millimeter-wave radar, lidar, and terahertz radar—to detect water surface vibrations, data with varying accuracy, integrity, and noise levels are obtained. The data fusion method is then used to improve the accuracy of vibration displacement measurement, thereby enhancing the accuracy of subsequent demodulation and enabling reliable communication in complex marine environments.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A cross-water-air medium communication method based on data fusion is proposed. This method employs a communication system consisting of a transmitter and a receiver. The transmitter is located underwater and consists of an underwater sound source array composed of multiple high-power sound sources. All sound sources transmit sound wave signals, and the sound waves impact the water surface, generating micron-level vibrations. The receiver is located above the water and is a multi-sensor system consisting of millimeter-wave radar, lidar, and terahertz radar placed at equal intervals to form a receiver array.

[0007] The cross-water-air medium communication method includes the following steps:

[0008] Step 1: Obtain the measured values ​​of the received signals using millimeter-wave radar, lidar, and terahertz radar respectively;

[0009] Step 2: Construct the state vector and observation model required for the Kalman filter algorithm; the state vector is the displacement in the received signal measurement; the observation model is the observation equation defining the relationship between the received signal measurement in Step 1 and the state vector:

[0010] z t (1) =H t (1) x t +v t (1)

[0011] z t (2) =H t (2) x t +v t (2)

[0012] z t (3) =H t (3) x t +v t (3)

[0013] Among them, H t (1) H t (2) and H t (3) These are, respectively, the observation matrices between the water surface displacement values ​​measured by millimeter-wave radar and the actual water surface displacement, the observation matrices between the water surface displacement values ​​measured by lidar and the actual water surface displacement, and the observation matrices between the water surface displacement values ​​measured by terahertz radar and the actual water surface displacement; v t (1) v t (2) and v t (3) The measurement noises for millimeter-wave radar, lidar, and terahertz radar are respectively; x t Let represent the state vector, and let represent the water surface displacement at time t.

[0014] Step 3: Initialize the water surface displacement and error covariance matrix; the error covariance matrix is ​​used to characterize the deviation between the estimated and actual values ​​of the water surface displacement.

[0015] Step 4: Predict the water surface displacement and covariance matrix at the next time step using the state transition model;

[0016] Step 5: Calculate the weight of each radar based on its signal-to-noise ratio, bit error rate, and measurement error standard deviation;

[0017] Step 6: Update the water surface displacement and error covariance matrix using the measurement data from each radar through the Kalman filter algorithm;

[0018] Step 7: Combine the update results of all radars to obtain the final water surface displacement and error covariance matrix.

[0019] Furthermore, in step four, the water surface displacement and covariance matrix at the next moment are predicted using a state transition model, specifically as follows:

[0020] Based on the estimated water surface displacement at the previous moment, predict the water surface displacement at the current moment:

[0021]

[0022] in, Let F be the predicted value of the water surface displacement at the current moment, F be the state transition matrix, and is the element matrix; This is the estimated water surface displacement from the previous moment;

[0023] Predict the current state estimation error covariance matrix based on the updated state estimation error covariance matrix from the previous time step:

[0024]

[0025] Among them, P t+1|t Let P be the state estimation error covariance matrix at the current time. t|t Let be the error covariance matrix of the updated state estimation at the previous time step, and Q be the process noise covariance matrix.

[0026] Furthermore, in step five, the weight w of each radar sensor is calculated based on the signal-to-noise ratio, bit error rate, and measurement error standard deviation of each radar sensor. i The specific calculation formula is as follows:

[0027]

[0028] Among them, SNR i The signal-to-noise ratio (BER) of a radar sensor. i σ represents the bit error rate. i It represents the standard deviation of the measurement error.

[0029] Furthermore, step six specifically includes the following sub-steps:

[0030] S6.1: Calculate the covariance matrix R1 and Kalman gain of the millimeter-wave radar respectively.

[0031]

[0032] Where σ1 represents the standard deviation of the measurement error of the millimeter-wave radar, P t+1|t The state estimation error covariance matrix at the current moment is calculated in step four, and w1 is the weight of the millimeter-wave radar calculated in step five.

[0033] S6.2: Calculate the water surface displacement estimated from the water surface displacement measured by millimeter-wave radar at the current moment. And update the state estimation error covariance matrix of the millimeter-wave radar.

[0034]

[0035] in, This is the predicted value of the water surface displacement at the current moment, calculated from step four. It is the water surface displacement value measured by millimeter-wave radar at the current moment; P t+1|t The state estimation error covariance matrix at the current time is calculated in step four;

[0036] S6.3: Calculate the covariance matrix R² and Kalman gain of the lidar respectively.

[0037]

[0038] Where σ² represents the standard deviation of the measurement error of the lidar. w1 is the state estimation error covariance matrix obtained at the last moment in the update step of the millimeter-wave radar; w2 is the weight of the lidar calculated in step five.

[0039] S6.4: Calculate the water surface displacement estimated from the water surface displacement measured by lidar at the current moment. And update the state estimation error covariance matrix of the lidar.

[0040]

[0041] in, The value is the predicted value of the water surface displacement at the current moment, estimated from measurements taken by millimeter-wave radar. It is the water surface displacement value measured by the lidar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the millimeter-wave radar update step.

[0042] S6.5: Calculate the covariance matrix R3 and Kalman gain of the terahertz radar respectively.

[0043]

[0044] Where σ3 represents the standard deviation of the measurement error of the terahertz radar. w3 is the state estimation error covariance matrix obtained at the last moment in the update step of the lidar, and w3 is the weight of the terahertz radar calculated in step five.

[0045] S6.6: Calculate the water surface displacement estimated from the water surface displacement measured by the terahertz radar at the current moment. And update the state estimation error covariance matrix of the terahertz radar.

[0046]

[0047] in, This is the predicted value of the water surface displacement at the current moment, estimated from the measurements taken by lidar. It is the water surface displacement value measured by the terahertz radar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the lidar update step.

[0048] Furthermore, step seven specifically includes: estimating the water surface displacement obtained from the water surface displacement measured by terahertz radar at the current time, as calculated in step six. The water surface displacement estimate obtained from the final fusion The final updated state estimation error covariance matrix of the terahertz radar The covariance matrix P at the final current moment t+1|t+1 .

[0049] The beneficial effects of this invention are as follows:

[0050] This invention fully utilizes the advantages of three different sensors: millimeter-wave radar, lidar, and terahertz radar. By fusing data from multiple sensors, it achieves higher precision in measuring water surface vibration displacement, improves the signal-to-noise ratio and integrity of the received signal, and overcomes the shortcomings of single-sensor systems that are susceptible to the influence of sea surface waves. This provides a reliable solution for the application of cross-water-air medium communication in actual marine environments. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of a cross-medium communication method that combines underwater acoustics and electromagnetic waves in the prior art.

[0052] Figure 2 A schematic diagram of the communication system structure used in the high-precision cross-water-air medium communication method based on data fusion technology to realize the method of the present invention.

[0053] Figure 3 This is a flowchart of a cross-water-air medium communication method based on data fusion, according to an embodiment of the present invention. Detailed Implementation

[0054] The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments. The purpose and effects of the present invention will become clearer. It should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0055] Millimeter-wave radar, lidar, and terahertz radar each have their own advantages and disadvantages. Lidar offers high accuracy and a small beam, performing well in conditions with low waves, no heavy fog, and no strong light, but it is susceptible to ambient light interference. Millimeter-wave radar, when detecting waves under wavy conditions, follows the principle of diffuse reflection, resulting in better signal integrity than lidar. Terahertz radar operates in a band between microwaves, millimeter waves, and infrared light, offering improved accuracy compared to millimeter-wave radar and higher stability than lidar. When all three sensors are used simultaneously to detect vibrations on the water surface, they can yield data with varying levels of accuracy, integrity, and noise.

[0056] like Figure 2 As shown, the cross-water / air medium communication method based on data fusion of the present invention employs a communication system including a transmitter and a receiver. The transmitter is located underwater and consists of an underwater sound source array composed of multiple high-power sound sources. All sound sources transmit sound wave signals, which impact the water surface, generating micron-level vibrations. The receiver is located above the water and is a multi-sensor system that simultaneously receives the vibration signals from the water surface. These signals are then processed using a Kalman filter data fusion algorithm to obtain more accurate vibration displacement measurements.

[0057] The multi-sensor system on water consists of high-precision millimeter-wave radar, lidar, and terahertz radar, which are placed at equal intervals to form a receiving array. Due to the presence of waves on the sea surface, the direction of the reflected signal received by each radar changes with the sea surface, resulting in each radar receiving reflected signals of random intensity and timing at different times.

[0058] like Figure 3 As shown, the cross-water-air medium communication method based on data fusion in this embodiment includes the following steps:

[0059] Step 1: Obtain the measured values ​​of the received signals using millimeter-wave radar, lidar, and terahertz radar respectively;

[0060] Among them, the millimeter-wave radar measurement value is LiDAR measurement values Terahertz radar measurements The subscript t indicates time t, that is, and These represent the water surface displacement values ​​measured by millimeter-wave radar, lidar, and terahertz radar at time t, respectively.

[0061] Step 2: Construct the state vector and observation model required for the Kalman filter algorithm;

[0062] Define the state vector x t Let x be the water surface displacement at time t. t :

[0063] x t =x t

[0064] The relationship between the state vector and the measured value in step one is defined by the observation equation. For millimeter-wave radar, lidar, and terahertz radar, there exists a corresponding observation equation:

[0065] Z t (1) =H t (1) x t +v t (1)

[0066] z t (2) =H t (2) x t +v t (2)

[0067] z t (3) =H t (3) +v t (3)

[0068] Among them, H t (1) H t (2) and H t (3) This is the observation matrix, which represents the relationship between the water surface displacement values ​​measured by millimeter-wave radar, lidar, and terahertz radar and the actual water surface displacement; v t (1) v t (2) and v t (3) These are the measurement noises for millimeter-wave radar, lidar, and terahertz radar, respectively.

[0069] The state transition matrix F describes how the system transitions from one time step to the next without considering external inputs. In the system described in this embodiment, the system state variable is the water surface displacement; therefore, without considering external inputs, the system state will retain the value of the previous time step. Thus, the state transition matrix is ​​a unit matrix, i.e.:

[0070] F = 1

[0071] The observation matrix describes the relationship between the sensor and the state vector, that is, the relationship between the water surface displacement measured by millimeter-wave radar, lidar, and terahertz radar and the actual water surface displacement. Ideally, for all sensors, without considering noise, the sensor should accurately reflect the actual water surface vibration without scaling. Therefore, the observation matrix is ​​also fixed as a unit matrix, i.e.:

[0072]

[0073] Step 3: Initialize the water surface displacement and error covariance matrix; the error covariance matrix is ​​used to characterize the deviation between the estimated value and the actual value of the water surface displacement.

[0074] State estimation is an estimate of the displacement of the water surface. Initialize to manually set an initial water surface displacement value without the presence of millimeter-wave radar, lidar, and terahertz radar measurements.

[0075] Error covariance reflects the degree of deviation between the estimated and actual state values, that is, the deviation between the estimated and actual values ​​of water surface displacement, expressed as the state vector x. t and estimated value The expected value of the difference between them is calculated. In the initialization phase of data fusion, the initial error covariance P0 is set manually.

[0076] Step 4: Predict the water surface displacement and covariance matrix at the next moment using the state transition model.

[0077] (1) State prediction: In the state prediction stage, the water surface displacement at the current moment is predicted based on the water surface displacement estimated at the previous moment. The calculation method is to input the water surface displacement estimated at the previous moment. Multiplying the value by the state transition matrix F yields the predicted value of the water surface displacement at the current moment.

[0078]

[0079] in, Let F be the predicted value of the water surface displacement at the current moment, F be the state transition matrix, and is the element matrix; This is the estimated water surface displacement at the previous moment.

[0080] (2) Covariance prediction: Estimate the error covariance matrix P based on the updated state at the previous time step. t|t Predicting the state estimation error covariance matrix P at the current moment t+1|t The calculation method involves inputting the updated state estimation error covariance matrix P from the previous time step. t|t The state estimation error covariance matrix P at the current time is obtained by performing the following operations with the state transition matrix F and the process noise covariance matrix Q. t+1|t :

[0081]

[0082] Where Q is the process noise covariance matrix, which is obtained through theoretical modeling or empirical estimation.

[0083] Step 5: Calculate the weight of each radar based on its signal-to-noise ratio, bit error rate, and measurement error standard deviation.

[0084] The weight w for each sensor is calculated based on its signal-to-noise ratio (SNR), bit error rate (BER), and standard deviation of measurement error (σ). i :

[0085]

[0086] Step Six: Update the water surface displacement and error covariance matrix using the Kalman filter algorithm with the measurement data from each radar. Specifically, for each radar sensor, update the state estimate and covariance.

[0087] Step six specifically includes the following sub-steps:

[0088] S6.1: Calculate the covariance matrix R1 and Kalman gain of the millimeter-wave radar respectively.

[0089]

[0090] Where σ1 represents the standard deviation of the measurement error of the millimeter-wave radar, P t+1|t The state estimation error covariance matrix at the current moment is calculated in step four, and w1 is the weight of the millimeter-wave radar calculated in step five.

[0091] S6.2: Calculate the water surface displacement estimated from the water surface displacement measured by millimeter-wave radar at the current moment. And update the state estimation error covariance matrix of the millimeter-wave radar.

[0092]

[0093] in, This is the predicted value of the water surface displacement at the current moment, calculated from step four. It is the water surface displacement value measured by millimeter-wave radar at the current moment; P t+1|t The state estimation error covariance matrix at the current time is calculated in step four;

[0094] S6.3: Calculate the covariance matrix R² and Kalman gain of the lidar respectively.

[0095]

[0096] Where σ² represents the standard deviation of the measurement error of the lidar. w1 is the state estimation error covariance matrix obtained at the last moment in the update step of the millimeter-wave radar; w2 is the weight of the lidar calculated in step five.

[0097] S6.4: Calculate the water surface displacement estimated from the water surface displacement measured by lidar at the current moment. And update the state estimation error covariance matrix of the lidar.

[0098]

[0099] in, The value is the predicted value of the water surface displacement at the current moment, estimated from measurements taken by millimeter-wave radar. It is the water surface displacement value measured by the lidar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the millimeter-wave radar update step.

[0100] S6.5: Calculate the covariance matrix R3 and Kalman gain of the terahertz radar respectively.

[0101]

[0102] Where σ3 represents the standard deviation of the measurement error of the terahertz radar. w3 is the state estimation error covariance matrix obtained at the last moment in the update step of the lidar, and w3 is the weight of the terahertz radar calculated in step five.

[0103] S6.6: Calculate the water surface displacement estimated from the water surface displacement measured by the terahertz radar at the current moment. And update the state estimation error covariance matrix of the terahertz radar.

[0104]

[0105] in, This is the predicted value of the water surface displacement at the current moment, estimated from the measurements taken by lidar. It is the water surface displacement value measured by the terahertz radar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the lidar update step.

[0106] Step 7: Combine the update results of all radars to obtain the final water surface displacement and error covariance matrix.

[0107] The final fusion result is obtained from the update result of the third sensor, specifically:

[0108] Final state estimation: The estimated water surface displacement obtained from the final fusion. The water surface displacement estimated from the terahertz radar measurement at the current time, calculated in step six.

[0109]

[0110] Final covariance matrix: The final covariance matrix P at the current time step. t+1|t+1 The final updated state estimation error covariance matrix for the terahertz radar

[0111]

[0112] By using the above method and fusing measurement data from millimeter-wave radar, lidar, and terahertz radar with an improved Kalman filter, the measurement accuracy of water surface vibration displacement can be effectively improved, thereby enhancing the accuracy and efficiency of cross-water-air communication.

[0113] It will be understood by those skilled in the art that the above descriptions are merely preferred examples of the invention and are not intended to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the invention should be included within the scope of protection of the invention.

Claims

1. A cross-water / air medium communication method based on data fusion, characterized in that, This method employs a communication system consisting of a transmitter and a receiver. The transmitter is located underwater and consists of an underwater sound source array composed of multiple high-power sound sources. All sound sources transmit sound wave signals, and the sound waves impact the water surface, generating micron-level vibrations. The receiver is located above the water and is a multi-sensor system consisting of millimeter-wave radar, lidar, and terahertz radar placed at equal intervals to form a receiver array. The cross-water-air medium communication method includes the following steps: Step 1: Obtain the measured values ​​of the received signals using millimeter-wave radar, lidar, and terahertz radar respectively; Step 2: Construct the state vector and observation model required for the Kalman filter algorithm; the state vector is the displacement in the received signal measurement; the observation model is the observation equation defining the relationship between the received signal measurement in Step 1 and the state vector: ； ； ； in, , and These are, respectively, the observation matrix between the water surface displacement value measured by millimeter-wave radar and the actual water surface displacement, the observation matrix between the water surface displacement value measured by lidar and the actual water surface displacement, and the observation matrix between the water surface displacement value measured by terahertz radar and the actual water surface displacement. , and The measurement noises are respectively those of millimeter-wave radar, lidar, and terahertz radar; Let represent the state vector, and let represent the water surface displacement at time t. Step 3: Initialize the water surface displacement and error covariance matrix; the error covariance matrix is ​​used to characterize the deviation between the estimated and actual values ​​of the water surface displacement. Step 4: Predict the water surface displacement and covariance matrix at the next time step using the state transition model; Step 5: Calculate the weight of each radar based on its signal-to-noise ratio, bit error rate, and measurement error standard deviation; Step 6: Update the water surface displacement and error covariance matrix using the measurement data from each radar through the Kalman filter algorithm; Step 7: Combine the update results from all radars to obtain the final water surface displacement and error covariance matrix; Step six specifically includes the following sub-steps: S6.1: Calculate the covariance matrix of the millimeter-wave radar respectively. and Kalman gain : ； ； Where σ1 represents the standard deviation of the measurement error of the millimeter-wave radar. It is the state estimation error covariance matrix at the current moment, calculated in step four. The weights of the millimeter-wave radar are calculated in step five. S6.2: Calculate the water surface displacement estimated from the water surface displacement measured by millimeter-wave radar at the current moment. And update the state estimation error covariance matrix of the millimeter-wave radar. : ； ； in, This is the predicted value of the water surface displacement at the current moment, calculated from step four. It is the water surface displacement value measured by millimeter-wave radar at the current moment; The state estimation error covariance matrix at the current time is calculated in step four; S6.3: Calculate the covariance matrix R² and Kalman gain of the lidar respectively. : ； ； Where σ² represents the standard deviation of the measurement error of the lidar. It is the state estimation error covariance matrix obtained at the last moment in the update step of the millimeter-wave radar; These are the weights of the lidar calculated in step five; S6.4: Calculate the water surface displacement estimated from the water surface displacement measured by lidar at the current moment. And update the state estimation error covariance matrix of the lidar. : ； ； in, The value is the predicted value of the water surface displacement at the current moment, estimated from measurements taken by millimeter-wave radar. It is the water surface displacement value measured by the lidar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the millimeter-wave radar update step. S6.5: Calculate the covariance matrix R3 and Kalman gain of the terahertz radar respectively. : ； ； Where σ3 represents the standard deviation of the measurement error of the terahertz radar. It is the state estimation error covariance matrix obtained at the last moment in the lidar update step. These are the weights of the terahertz radar calculated in step five; S6.6: Calculate the water surface displacement estimated from the water surface displacement measured by the terahertz radar at the current moment. And update the state estimation error covariance matrix of the terahertz radar. : ； ； in, This is the predicted value of the water surface displacement at the current moment, estimated from the measurements taken by lidar. It is the water surface displacement value measured by the terahertz radar at the current moment; This is the state estimation error covariance matrix obtained at the current moment in the lidar update step.

2. The cross-water / air medium communication method based on data fusion according to claim 1, characterized in that, In step four, the water surface displacement and covariance matrix at the next moment are predicted using a state transition model, specifically as follows: Based on the estimated water surface displacement at the previous moment, predict the water surface displacement at the current moment: ； in, Let F be the predicted value of the water surface displacement at the current moment, F be the state transition matrix, and is the element matrix; This is the estimated water surface displacement from the previous moment; Predict the current state estimation error covariance matrix based on the updated state estimation error covariance matrix from the previous time step: ； in, The state estimation error covariance matrix at the current moment. Let be the error covariance matrix of the updated state estimation at the previous time step, and Q be the process noise covariance matrix.

3. The cross-water / air medium communication method based on data fusion according to claim 2, characterized in that, In step five, the weight of each radar sensor is calculated based on its signal-to-noise ratio, bit error rate, and measurement error standard deviation. The specific calculation formula is as follows: ； Among them, SNR i The signal-to-noise ratio (BER) of a radar sensor. i σ represents the bit error rate. i It represents the standard deviation of the measurement error.

4. The cross-water / air medium communication method based on data fusion according to claim 1, characterized in that, Step seven specifically includes: estimating the water surface displacement obtained from the water surface displacement measured by terahertz radar at the current moment, which was calculated in step six. The water surface displacement estimate obtained from the final fusion The final updated state estimation error covariance matrix of the terahertz radar The covariance matrix at the final current moment .

Images