Maritime unmanned platform satellite-facing communication antenna system adapted to high sea conditions and beam forming method

Abstract

The invention discloses a maritime unmanned platform satellite communication antenna system suitable for high sea conditions and a beam forming method. The system comprises an array antenna, a beam forming controller and an unmanned platform attitude sensor. The method comprises the steps of rolling angle sensing, prediction, beam alignment and optimization, and load data transmission and feedback. In order to solve the problem that a conventional fixed beam forming scheme is easy to fail due to frequent rolling of an offshore unmanned platform influenced by sea waves, the invention provides an uplink transmission frame and a rolling angle prediction model by compensating rolling angle change through adaptive beam forming. The channel state is calculated based on prediction information, a joint beam forming algorithm is designed, effective beam alignment and optimization under the high sea condition are achieved, and the stability of a maritime satellite communication link is ensured.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to an unmanned offshore platform satellite communication antenna system and a beam forming method adapted to high sea conditions. Background Art

[0002] With the sharp increase in maritime activities such as ship navigation, maritime rescue, environmental monitoring and marine mineral exploration, the importance of maritime communications has become increasingly prominent. However, due to the lack of base station deployment, 5G wireless networks are difficult to use for ocean maritime communications. Its seamless global coverage and powerful backhaul transmission capabilities make it an ideal choice to meet the growing demand for broadband communications in the marine field. Therefore, low-orbit satellite-assisted maritime communication systems, especially satellite-assisted maritime Internet of Things systems, have become a research hotspot in academia and industry. Unmanned platforms are usually deployed at sea to upload payload data collected from buoys and sensors to satellites.

[0003] Due to the influence of wave fluctuations in the marine environment, the unmanned platform at sea will produce six degrees of freedom movement, including linear movement and rotational movement. Therefore, the height and angle of the shipboard antenna change with the changes in the waves. Due to the long communication distance between the satellite and the unmanned platform at sea, the impact of antenna tilt on the communication link is particularly significant, which may lead to weakened signal strength, increased channel fading, and unstable and frequent interruptions of the communication link.

[0004] Traditional solutions usually design wide antennas for unmanned platforms to improve the robustness of the beam pattern or use high-gain antennas with pointing mechanisms. Wide beamwidth antennas help achieve acceptable gain attenuation within the range of angle variations. However, increasing the beamwidth inevitably leads to a decrease in antenna gain, making it difficult to cope with severe path fading due to long-distance communications with satellites. On the other hand, for high-gain directional antenna solutions, a mechanically controlled pointing mechanism must be used to compensate for the platform angle tilt so that the antenna is pointed in the direction of the satellite. However, such directional antennas with pointing mechanisms are bulky and expensive. Balancing beamwidth and antenna gain has always been a challenging task in the design of maritime communication antennas. In addition, once the antenna is designed, its radiation pattern is fixed, making it difficult to flexibly cope with the high dynamic scenarios of satellite-assisted maritime communications. Summary of the invention

[0005] In view of the problems existing in the prior art, the present invention provides an unmanned offshore platform satellite communication antenna system and a beamforming method that are adaptable to high sea conditions.

[0006] The satellite communication antenna system of the unmanned offshore platform includes: array antenna, attitude sensor and beamforming controller.

[0007] A beamforming controller, including unmanned platform control hardware consisting of one or more microprocessors, memory, interface circuits, etc., and unmanned platform antenna control software running on the unmanned platform controller;

[0008] The unmanned platform antenna control software includes:

[0009] Angle perception module, used to receive and process the roll angle information of the attitude sensor on the unmanned platform;

[0010] An angle prediction module, used to predict the roll angles of the next few time slots based on historical roll angle information;

[0011] A control calculation module is used to calculate the beamforming vector of the antenna array according to the acquired roll state (perceived or predicted information) to realize beam alignment with the satellite;

[0012] The collaboration module is used for signaling interaction and channel feedback with the satellite.

[0013] Furthermore, in order to cope with the impact of unmanned platform antenna angle deviation caused by wave motion under high sea conditions on the marine satellite communication system, the present invention proposes an adaptive dynamic beamforming implementation method based on unmanned platform roll angle prediction for the proposed satellite communication antenna system. The designed transmission frame consists of several beam tracking cycles, each of which consists of 1 prediction time slot and T n The prediction time slot consists of three parts: roll angle measurement of the unmanned platform, online prediction, and channel state information (CSI) feedback. The unmanned platform can sense and store its own instantaneous roll angle information with the help of attitude sensors, and then feed it back to the satellite in the form of control signaling. Based on the current and historical roll angle information extracted from the initial training data set, a triple-layer long short-term memory (TL-LSTM) model is used to predict the future T n The roll angle of each tracking time slot is calculated. The tracking time slot includes roll angle measurement, adaptive beamforming, and payload data transmission. Using the roll angle information obtained in the prediction time slot, the corresponding beamforming vector is designed for beam alignment to achieve effective payload data transmission. The specific steps are as follows:

[0014] Step 1: Unmanned platform roll angle perception:

[0015] In the marine environment, the shaking of the unmanned platform will cause frequent beam misalignment between the unmanned platform and the satellite, thus affecting the stability of the link. Traditional beamforming based on the position information of the transmitter and receiver cannot effectively cope with the angle changes caused by the shaking of non-particle motion.

[0016] Therefore, the present invention deploys a N t In addition, the satellite is equipped with a phased array with N r A uniform planar array of antennas is used for directional reception. For simplicity, the present invention focuses on studying the rolling motion of the unmanned platform caused by wave motion, and performs beam alignment and adaptive beamforming. However, it is worth noting that the corresponding method can be easily extended to handle other spatial angle changes of the unmanned platform.

[0017] Specifically, the real-time roll angle information of the unmanned platform can be obtained through the attitude sensor on the unmanned platform, and the current roll angle information is recorded for future roll angle prediction, and shared with the satellite through the control link to achieve joint beamforming of the transmitter and receiver.

[0018] Step 2, roll angle prediction:

[0019] Considering the transmission delay between the unmanned platform and the satellite, the roll angle information feedback from the unmanned platform to the satellite may easily expire. Therefore, the corresponding channel state information may be outdated, and the beam design based on this will lead to system performance loss. In addition, sufficient time needs to be reserved for the calculation and configuration of beam alignment. To this end, the present invention proposes a roll angle prediction model based on TL-LSTM, which replaces the instantaneous measurement value with the predicted roll angle value.

[0020] The roll angle prediction scheme framework proposed in this invention consists of two parts: offline training and online prediction. The purpose of offline training is to improve the accuracy of prediction through the historical data of the roll angle of the unmanned platform, so that it can learn the law and trend of the roll angle change of the unmanned platform under the motion of waves. When a prediction request arrives, the current channel state information will be used as input data for online prediction of future channel states.

[0021] Specifically, the structure of the roll angle prediction scheme consists of an input layer, an output layer, and multiple hidden layers, which reasonably learn the correlation of the roll angle changes and predict the corresponding output results.

[0022] (2-1) Input layer

[0023] The present invention establishes a sequence-to-sequence model to make predictions based on the past L steps of angle measurements as input.

[0024] (2-2) LSTM layer

[0025] A single LSTM unit includes three sigmoid functions, namely the forget gate f, the input gate i, and the output gate o. The forget gate is the core of the LSTM unit because its output f tWill determine the current cell state c t How much of the previous cell state c to keep t-1 The input gate adds some new information x t To the previous cell state c t-1 To update the cell state c t The final output h of the LSTM unit is t By output gate o t and the cell state c t Joint decision.

[0026] In summary, the gate operations performed by the LSTM unit are summarized as follows:

[0027] f t =σ(W i ·[h t-1 ,x t ]+b f ),

[0028] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0029] c t =f t ⊙c t-1 +i t ⊙tanh(W c ·[h t-1 ,x t ]+b c ),

[0030] o t =σ(W o ·[h t-1 ,x t ]+b o ),

[0031] h t =o t ⊙tanh(c t ),

[0032] Where tanh represents the hyperbolic tangent function, σ(·) represents the sigmoid function, ⊙ is the Hadamard product, and W f ,W i ,W c ,W o represents the weight matrix, b f ,b i ,b c ,b o is the bias vector.

[0033] The outputs of multiple LSTM units are connected to the inputs to form an LSTM layer. The prediction scheme of the present invention includes 3 LSTM layers. By increasing the depth of the network, the representation and learning capabilities of the model can be improved, and the complex patterns in the time-varying roll angle can be better captured.

[0034] (2-3) Dropout layer

[0035] Each LSTM layer is followed by a Dropout layer, which provides an effective regularization method by randomly dropping some neurons to reduce overfitting problems and improve generalization ability.

[0036] (2-4) Fully connected layer

[0037] In addition, a dense layer is deployed before the output layer to provide more sensitive activation and avoid saturation. The Adam optimizer is implemented to dynamically adjust the learning rate to obtain higher model accuracy and fast convergence.

[0038] (2-5) Output layer

[0039] Output for future T n The predicted value of the roll angle in each time slot.

[0040] In the process of model learning, the root mean-square error (RMSE) is selected to measure the prediction accuracy, which is expressed as

[0041]

[0042] Among them, N is the total number of samples in the test set, γ is the parameter set of the prediction model, and the model will continuously update the parameters in the parameter set to minimize RMSE.

[0043] Step 3, beam alignment and optimization:

[0044] The present invention is based on the predicted roll angle The predicted channel matrix under a fixed time slot can be obtained The predicted channel matrix H US is the true pitch angle The actual pitch angle is given by the position-based pitch angle φ e and roll angle By setting the coordinate system position reasonably, the pitch angle and roll angle based on the position in the system can be linearly added.

[0045] The present invention designs adaptive beamforming vectors based on the predicted channel state information to cope with the angle changes in different time slots and maximize the uplink transmission rate. This rate maximization problem is equivalently converted into a weighted mean square error minimization problem, as shown below

[0046]

[0047] st||w|| 2 ≤P,

[0048] Where w is the transmit beam vector at the unmanned platform, P is the transmit power limit, U is the linear receive beam vector at the satellite, q is an auxiliary variable, and e is the mean square error of the estimated signal at the satellite, which is given by

[0049]

[0050] Among them, H US is the predicted channel matrix for the current time slot Abbreviation of Represents the conjugate transpose of a matrix.

[0051] Note that by introducing auxiliary variables and fixing two of the variables w, U, q, the optimization problem of the remaining variables is simplified to a convex optimization problem. Therefore, the block coordinate descent method can be used to iteratively solve it:

[0052] (3-1) Given w and q, the optimal value of U is obtained by setting the first-order derivative of the objective function with respect to U to 0, as shown in the following formula:

[0053]

[0054] (3-2) Similarly, the optimal auxiliary variable q is given by the following formula

[0055]

[0056] (3-3) Given U and q, we now focus on the transmit beam w in the optimization problem. Ignoring the irrelevant polynomials in the objective function, the original optimization problem is transformed into

[0057]

[0058] st||w|| 2 ≤P.

[0059] In order to reduce the complexity caused by using the CVX solver, the Lagrange multiplier method is used to obtain a closed-form expression for w. Specifically, the Lagrange multiplier μ is introduced for the power constraint to obtain the following Lagrangian function:

[0060]

[0061] make For the first-order derivative of w to be zero, the optimal expression of w can be obtained as

[0062]

[0063] Substituting the above formula into the power constraint, we can get the value of μ. Substituting the above formula into the power constraint, we can get

[0064]

[0065] (3-4) Solve the optimization problem by iteratively optimizing w, U, q until the convergence condition is met.

[0066] Step 4: Payload data transmission and feedback:

[0067] After the beam alignment scheme is implemented, the payload data of the unmanned platform is transmitted. The unmanned platform uploads the payload data collected from the surrounding buoys or pre-cached to the satellite, and feeds back the angle information to the satellite for effective reception.

[0068] Compared with the prior art, the technical effects of the present invention are as follows:

[0069] 1) Unlike the existing work that considers the beam alignment scenarios of mobile users on the ground or at sea and mobile satellites, the present invention focuses on the roll motion of unmanned platforms in satellite-assisted maritime communication systems affected by waves, which is a new and more challenging problem. Traditional position-based beam alignment is not suitable for the angle tilt caused by waves. To this end, the present invention proposes to deploy array antennas on unmanned platforms, which can effectively compensate for the change in roll angle through adaptive beamforming on the unmanned platform side.

[0070] 2) The uplink transmission system of the offshore unmanned platform and the satellite communication and the rolling model of the unmanned platform were studied.

[0071] Since the unmanned platform is subjected to the rolling motion caused by the waves under high sea conditions, the channel between the unmanned platform and the satellite often undergoes frequent and rapid changes. In order to obtain the instantaneous channel state information more accurately, the present invention proposes a corresponding transmission framework and an unmanned platform roll angle prediction model, and designs an adaptive beam alignment scheme based on this. Specifically, a part of the total time slot is divided as a prediction time slot, and the TL-LSTM model is used to predict the roll angle of subsequent time slots based on the historical observation of the roll angle information to predict the channel state. Based on the predicted channel state information, a joint beamforming algorithm is designed for the unmanned platform and the satellite to achieve adaptive beam alignment and optimization under high sea conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1A schematic diagram of a satellite communication antenna system for an unmanned offshore platform according to an embodiment of the present invention;

[0073] Figure 2 The adaptive beamforming transmission framework of the embodiment of the present invention;

[0074] Figure 3 This is an unmanned platform roll angle prediction solution based on TL-LSTM in an embodiment of the present invention;

[0075] Figure 4 The following is a comparison of prediction performances of different prediction schemes according to the embodiments of the present invention;

[0076] Figure 5 The figure is a comparison of the uplink transmission rate performance of the systems according to the embodiments of the present invention. DETAILED DESCRIPTION

[0077] The present invention is further explained below in conjunction with the accompanying drawings and implementation examples, but this should not limit the protection scope of the present invention.

[0078] This embodiment provides a satellite communication antenna system and beamforming method for an unmanned platform at sea in high sea conditions. The system combines array antenna technology and an intelligent beamforming controller to achieve stable and efficient maintenance of satellite communication links. By predicting the roll angle of the unmanned platform at sea, the system can dynamically adjust the beam direction to ensure reliable data transmission in harsh marine environments.

[0079] See also Figure 1 , Figure 1 1 is a schematic diagram of a satellite communication antenna system for an unmanned platform at sea according to an embodiment of the present invention. As shown in the figure, the antenna system includes:

[0080] Array antenna: A phased array of 16 transmitting antennas deployed on the unmanned platform to control the beam direction.

[0081] A beamforming controller, including unmanned platform control hardware consisting of one or more microprocessors, memory, interface circuits, etc., and unmanned platform antenna control software running on the unmanned platform controller;

[0082] Among them, the unmanned platform antenna control software includes:

[0083] Angle perception module, which receives and processes the roll angle information of the angle sensor on the unmanned platform in real time;

[0084] The angle prediction module uses the TL-LSTM model to predict the roll angles of the next few time slots based on the historical roll angle information to cope with the needs of transmission delay and beam alignment.

[0085] The control calculation module calculates and adjusts the beamforming vector of the antenna array based on the predicted roll angle (perceived or predicted information) to ensure that the beam is always aimed at the satellite;

[0086] The collaborative module interacts with the satellite through signaling, including feedback of roll angle information and sharing of channel state information, to achieve joint beamforming and optimization.

[0087] In view of the proposed satellite communication antenna system, the present invention proposes an adaptive dynamic beamforming implementation method based on the roll angle prediction of the unmanned platform. Figure 2 As shown, it consists of several beam tracking cycles, each of which consists of 1 prediction time slot and T n The prediction time slot consists of three parts: the measurement of the roll angle of the unmanned platform, online prediction and feedback of channel state information (CSI); the tracking time slot is used for re-measurement of the roll angle, adaptive beamforming and transmission of payload data. The unmanned platform can sense and store its own instantaneous roll angle information with the help of attitude sensors, and then feed it back to the satellite in the form of control signaling. Based on the current and historical roll angle information extracted from the initial training data set, the TL-LSTM model is used to predict the future T n The roll angle of each tracking time slot is calculated. The roll angle information obtained in the predicted time slot is used to design the corresponding beamforming vector for beam alignment to achieve effective payload data transmission. The specific steps are as follows:

[0088] Step 1: Unmanned platform roll angle perception:

[0089] In the marine environment, the shaking of the unmanned platform will cause frequent beam misalignment between the unmanned platform and the satellite, thus affecting the stability of the link. Traditional beamforming based on the position information of the transmitter and receiver cannot effectively cope with the angle changes caused by the shaking of non-particle motion.

[0090] Therefore, the present invention deploys a phased array consisting of 16 transmitting antennas on the unmanned platform. In addition, the satellite is equipped with a uniform planar array with 256 antennas for directional reception. For simplicity, the present invention focuses on studying the rolling motion of the unmanned platform caused by fluctuations, and performs beam alignment and adaptive beamforming. However, it is worth noting that the corresponding method can be easily extended to handle other spatial angle changes of the unmanned platform.

[0091] Specifically, the real-time roll angle information of the unmanned platform can be obtained through the attitude sensor on the unmanned platform, and the current roll angle information is recorded for future roll angle prediction, and shared with the satellite through the control link to achieve joint beamforming of the transmitter and receiver.

[0092] Step 2, roll angle prediction:

[0093] Considering the transmission delay between the unmanned platform and the satellite, the roll angle information feedback from the unmanned platform to the satellite may easily expire. Therefore, the corresponding channel state information may be outdated, and the beam design based on this will lead to system performance loss. In addition, sufficient time needs to be reserved for the calculation and configuration of beam alignment. To this end, the present invention proposes a roll angle prediction model based on TL-LSTM, which replaces the instantaneous measurement value with the predicted roll angle value.

[0094] like Figure 3 As shown in FIG. 1 , the roll angle prediction scheme framework proposed in the present invention consists of two parts: offline training and online prediction. The purpose of offline training is to improve the accuracy of prediction through the historical data of the roll angle of the unmanned platform, so that it can learn the law and trend of the roll angle change of the unmanned platform under the motion of waves. When a prediction request arrives, the current channel state information will be used as input data for online prediction of the future channel state.

[0095] Specifically, the structure of the roll angle prediction scheme consists of an input layer, an output layer, and multiple hidden layers, which reasonably learn the correlation of the roll angle changes and predict the corresponding output results.

[0096] (2-1) Input layer

[0097] The present invention establishes a sequence-to-sequence model to make predictions based on the past L steps of angle measurements as input.

[0098] (2-2) LSTM layer

[0099] A single LSTM unit includes three sigmoid functions, namely the forget gate f, the input gate i, and the output gate o. The forget gate is the core of the LSTM unit because its output f t Will determine the current cell state c t How much of the previous cell state c to keep t-1 The input gate adds some new information x t To the previous cell state c t-1 To update the cell state c t The final output h of the LSTM unit is t By output gate o t and the cell state c t Joint decision.

[0100] In summary, the gate operations performed by the LSTM unit are summarized as follows:

[0101] f t =σ(W f ·[h t-1 ,xt ]+b f ),

[0102] i t =σ(W i ·[h t-1 ,x t ]+b i ),

[0103] c t =f t ⊙c t-1 +i t ⊙tanh(W c ·[h t-1 ,x t ]+b c ),

[0104] o t =σ(W o ·[h t-1 ,x t ]+b o ),

[0105] h t =o t ⊙tanh(c t ),

[0106] Where tanh represents the hyperbolic tangent function, σ(·) represents the sigmoid function, ⊙ is the Hadamard product, and W f ,W i ,W c ,W o represents the weight matrix, b f ,b i ,b c ,b o is the bias vector.

[0107] The outputs of multiple LSTM units are connected to the inputs to form an LSTM layer. The prediction scheme of the present invention includes 3 LSTM layers. By increasing the depth of the network, the representation and learning capabilities of the model can be improved, and the complex patterns in the time-varying roll angle can be better captured.

[0108] (2-3) Dropout layer

[0109] Each LSTM layer is followed by a Dropout layer, which provides an effective regularization method by randomly dropping some neurons to reduce overfitting problems and improve generalization ability.

[0110] (2-4) Fully connected layer

[0111] In addition, a dense layer is deployed before the output layer to provide more sensitive activation and avoid saturation. The Adam optimizer is implemented to dynamically adjust the learning rate to obtain higher model accuracy and fast convergence.

[0112] (2-5) Output layer

[0113] Output for future T n The predicted value of the roll angle in each time slot.

[0114] In the process of model learning, the root mean square error (RMSE) is selected to measure the prediction accuracy, which is expressed as

[0115]

[0116] Wherein, N is the total number of samples in the test set. The present invention collects roll angle change data within 200s, the sampling time interval is 0.1s, and 20% of the data is selected as the test set. Therefore, the total number of samples of the test machine N=400, and γ is the parameter set of the prediction model. The model will continuously update the parameters in the parameter set to minimize RMSE.

[0117] Step 3, beam alignment and optimization:

[0118] The present invention is based on the predicted roll angle The predicted channel matrix under a fixed time slot can be obtained The predicted channel matrix H US is the true pitch angle The actual pitch angle is given by the position-based pitch angle φ e and roll angle By setting the coordinate system position reasonably, the pitch angle and roll angle based on the position in the system can be linearly added.

[0119] The present invention designs adaptive beamforming vectors based on the predicted channel state information to cope with the angle changes in different time slots and maximize the uplink transmission rate. This rate maximization problem is equivalently converted into a weighted mean square error minimization problem, as shown below

[0120]

[0121] st||w|| 2 ≤P,

[0122] Where w is the transmit beam vector at the unmanned platform, P is the transmit power limit, U is the linear receive beam vector at the satellite, q is an auxiliary variable, and e is the mean square error of the estimated signal at the satellite, which is given by

[0123]

[0124] Among them, H US is the predicted channel matrix for the current time slot Abbreviation of Represents the conjugate transpose of a matrix.

[0125] Note that by introducing auxiliary variables and fixing two of the variables w, U, q, the optimization problem of the remaining variables is simplified to a convex optimization problem. Therefore, the block coordinate descent method can be used to iteratively solve it:

[0126] (3-1) Given w and q, the optimal value of U is obtained by setting the first-order derivative of the objective function with respect to U to 0, as shown in the following formula:

[0127]

[0128] (3-2) Similarly, the optimal auxiliary variable q is given by the following formula

[0129]

[0130] (3-3) Given U and q, we now focus on the transmit beam w in the optimization problem. Ignoring the irrelevant polynomials in the objective function, the original optimization problem is transformed into

[0131]

[0132] st||w|| 2 ≤P.

[0133] In order to reduce the complexity caused by using the CVX solver, the Lagrange multiplier method is used to obtain a closed-form expression for w. Specifically, the Lagrange multiplier μ is introduced for the power constraint to obtain the following Lagrangian function:

[0134]

[0135] make For the first-order derivative of w to be zero, the optimal expression of w can be obtained as

[0136]

[0137] Substituting the above formula into the power constraint, we can get the value of μ. Substituting the above formula into the power constraint, we can get

[0138]

[0139] (3-4) Solve the optimization problem by iteratively optimizing w, U, q until the convergence condition is met.

[0140] Step 4: Payload data transmission and feedback:

[0141] After the beam alignment scheme is implemented, the payload data of the unmanned platform is transmitted. The unmanned platform uploads the payload data collected from the surrounding buoys or pre-cached to the satellite, and feeds back the angle information to the satellite for effective reception.

[0142] Figure 4 The prediction performance of different prediction schemes is shown. We compare the proposed TL-LSTM scheme with the popular RNN and GRU methods. It is noted that in the RNN scheme, there is a significant time delay between the predicted data and the actual data. In addition, the early prediction results of the GRU scheme are unstable. The RMSE performance of the proposed TL-LSTM scheme is the best.

[0143] Figure 5 The average uplink transmission rate of different beamforming schemes within 40 seconds is shown. Based on different roll angle information, three adaptive joint beamforming (AJBF) schemes are studied, including real channel state information (iCSI), predicted channel state information (pCSI) and outdated channel state information (oCSI). Compared with the other three adaptive schemes, the position-based fixed beamforming scheme (FJBF-lCSI) will cause serious loss of transmission performance due to beam misalignment in harsh marine environments. For the adaptive beamforming scheme, the performance of the AJBF-pCSI scheme is close to that of the AJBF-iCSI scheme, indicating the accuracy of the angle prediction method proposed in the present invention. From the performance of the AJBF-oCSI scheme, it can be seen that due to the long distance between the unmanned platform and the satellite, the roll angle information such as the channel state information obtained on the satellite is easily outdated, which will cause a certain degree of transmission performance loss. Therefore, it can be concluded that the roll angle prediction and transmission framework design proposed in the present invention are very necessary and effective for maritime satellite communications. In addition, the distributed beamforming of the ASBF-pCSI scheme can achieve similar performance as joint beamforming, which proves that adaptive beamforming on the unmanned platform side is more critical.

Claims

1. An unmanned platform satellite communication antenna system for offshore that is adaptable to high sea conditions, characterized in that: include: array antennas, used to communicate with satellites; An attitude sensor connected to the beamforming controller, used to provide attitude data of the unmanned platform, including roll angle, pitch angle, orientation and acceleration; A beamforming controller, connected to the array antenna and the attitude sensor respectively, comprises: An angle sensing module, connected to the attitude sensor, for sensing the roll angle of the unmanned platform in real time; An angle prediction module, connected to the angle perception module, for predicting the roll angle at a future moment according to the historical roll angle data; A control calculation module, connected to the angle sensing module, the angle prediction module and the coordination module respectively, for calculating channel state information according to the predicted roll angle and designing a corresponding beamforming strategy; The coordination module is connected to the array antenna and is used to coordinate the beamforming operation of the array antenna to ensure that the beam is always aligned, thereby maintaining a stable communication link.

2. The satellite communication antenna system for unmanned offshore platforms adapted to high sea conditions according to claim 1 is characterized in that: The array antenna is a phased array antenna that can dynamically adjust the beam direction to follow the rolling motion of the unmanned platform.

3. The satellite communication antenna system for unmanned offshore platforms adapted to high sea conditions according to claim 2 is characterized in that: The phased array antenna is adapted to a uniform planar array of multiple antennas equipped on the satellite, and the uniform planar array is used for directionally receiving signals from an unmanned platform.

4. An adaptive dynamic beamforming method using the satellite communication antenna system of an unmanned offshore platform according to any one of claims 1 to 3, characterized in that: The following steps are involved: Step 1. Obtain the current posture data of the unmanned platform in real time through the posture sensor; Step 2. Use the angle prediction module to predict the roll angle change in the future based on the historical roll angle in the attitude data; Step 3. The control calculation module calculates the channel state information according to the predicted roll angle, and adjusts the beam direction of the array antenna through the coordination module to achieve continuous alignment and optimization between the beam and the satellite; Step 4. After beam alignment and optimization, data is transmitted through the antenna system and feedback data is collected to further adjust and optimize the beamforming strategy.

5. The adaptive dynamic beamforming method according to claim 4, characterized in that: The step 2 specifically includes: Offline training: Using the historical data of the roll angle of the unmanned platform, the roll angle prediction model is trained to improve the prediction accuracy; Online prediction: When a prediction request arrives, the current channel state information is used as input data, and the future channel state, that is, the roll angle, is predicted through the trained roll angle prediction model.

6. The adaptive dynamic beamforming method according to claim 5, characterized in that: The roll angle prediction model includes an input layer, an output layer and multiple hidden layers, and selects the root mean square error (RMSE) as an indicator to measure the prediction accuracy; the RMSE is minimized by continuously updating the parameter set of the model, thereby improving the accuracy of the prediction; The Adam optimizer is used to dynamically adjust the learning rate to obtain higher model accuracy and fast convergence.

7. The adaptive dynamic beamforming method according to claim 5, characterized in that: The step 3 specifically includes: The predicted roll angle is used to construct the predicted channel matrix at a fixed time slot; Predict the future channel matrix by predicting the roll angle and the known pitch angle based on the position; Using the predicted channel state information, adaptive beamforming vectors are designed to cope with the angle changes in different time slots.

8. The adaptive dynamic beamforming method according to any one of claims 5 to 7, characterized in that: It also includes steps to dynamically adjust the beamforming strategy based on quality indicators of payload data transmission, such as bit error rate or throughput, to adapt to different sea conditions and communication requirements.

9. A maritime satellite communication system, characterized in that: It comprises an unmanned offshore platform satellite communication antenna system as described in any one of claims 1-3, and a satellite for communicating therewith, wherein the system adopts a beamforming method as described in any one of claims 4 to 9 for communication.

Images