High-precision satellite tracking method for moving target indication
By employing communication carrier energy detection and BP neural network training, the problem of insufficient satellite signal tracking accuracy in complex environments for mobile communication systems was solved, achieving high-precision satellite tracking and broadband communication, simplifying system design, and reducing costs.
Patent Information
- Application Number
- CN202310882497.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-18
- Publication Date
- 2026-04-24
- Estimated Expiration
- 2043-07-18
AI Technical Summary
Traditional mobile communication systems lack sufficient satellite signal tracking accuracy in complex environments. In particular, the electronically scanned phased array antenna cannot simultaneously receive communication and beacon signals, and the combined navigation error suppression effect is poor, resulting in limited or interrupted communication bandwidth.
A method combining communication carrier energy detection with BP neural network training is adopted. By detecting the communication carrier signal, the optimal beam direction is found, and the neural network is used for error suppression to achieve high-precision satellite tracking.
It improves satellite dynamic tracking accuracy, simplifies system design, reduces costs, maintains high-bandwidth communication in complex environments, and avoids dependence on beacon signals.
Smart Images

Figure CN116931018B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of wireless communication technology, specifically relating to an error suppression method based on communication carrier detection and neural networks to obtain the optimal communication beam direction, thereby achieving higher precision satellite dynamic tracking in "communication on the move" application scenarios. Background Technology
[0002] With the rapid development of information technology, the Internet of Things, and artificial intelligence, the amount of data that mobile communications need to transmit is increasing dramatically. People urgently need to send and receive large amounts of multimedia information such as voice, data, images, and video in real time in various environments. Satellite communication, as an effective supplement and extension of traditional terrestrial communication, plays an irreplaceable role in scenarios such as oceans, deserts, and remote mountainous areas where traditional terrestrial communication networks are difficult to cover. It can achieve seamless global network information coverage and is widely used in aviation, maritime, land, and rail transportation. To meet the needs of satellite communication in these application scenarios, satellite communication equipment with "satellite communications on the move" (SOTM) functionality has become an inevitable choice. Facing various harsh operating environments such as bumpy roads, rough seas, and unstable air currents, the design of SOTM systems faces severe challenges.
[0003] Traditional mobile communication systems mainly consist of a parabolic antenna, a mechanical servo mechanism, and a modem. They employ satellite navigation combined with inertial navigation for positioning and attitude measurement, and achieve continuous dynamic tracking via satellite beacon signal tracking. Connecting to a satellite modem, they enable broadband communication while continuously aligned with the satellite. With the rapid development of phased array mobile communication systems, low-profile, miniaturized mobile communication antennas have broken free from the constraints of traditional three-axis motor servo systems, making mobile communication systems more compact and easier to install and use, thus injecting strong momentum into the development of satellite mobile communication.
[0004] However, whether using a traditional mechanical on-the-move antenna or an electronically scanned phased array antenna, achieving real-time and precise satellite alignment requires frequency conversion, sampling, and scanning of the satellite beacon signal to find the optimal signal level. Since the frequencies of the beacon signal and the satellite communication carrier signal are generally far apart, a beacon receiver needs to be built for traditional mechanical on-the-move antennas to collect the beacon signal. However, for electronically scanned phased array antennas, because the array elements, element spacing, and phase shift are often designed and calculated for the communication carrier frequency, it is impossible to simultaneously receive both the communication and beacon signals when their frequencies are far apart. If dynamic beacon signal tracking is still required, a separate antenna must be designed to receive the beacon signal, and then connected to a beacon receiver to complete the frequency conversion and sampling of the beacon signal.
[0005] The performance of integrated navigation (satellite navigation + inertial navigation) determines the accuracy of the on-the-move communication system's self-positioning and attitude measurement. When the satellite navigation system is interfered with or receives abnormal satellite navigation information signals, and when the inertial navigation system experiences increasing errors such as drift over time, both lead to position and attitude measurement errors. For some satellite communications with high tracking accuracy requirements, this may cause the antenna to remain in a low-gain alignment state for extended periods, thus limiting communication bandwidth or even causing communication interruption. Therefore, error suppression in integrated navigation is a key technology for improving satellite dynamic tracking. Currently, common integrated navigation error suppression methods mainly employ Kalman filtering, which requires accurate error modeling. Furthermore, due to the nonlinear characteristics of system errors, the error sources included in the Kalman filter error model cannot adequately describe the error propagation mechanism, resulting in a large computational load. Summary of the Invention
[0006] The technical problem to be solved by this invention is:
[0007] To address the shortcomings of mobile satellite tracking antennas, such as reliance on beacon receivers and the need for dedicated beacon receiving antennas when the communication carrier and beacon frequencies are far apart, this invention provides a high-precision mobile satellite tracking method. It utilizes communication carrier energy detection combined with backpropagation (BP) neural network training to suppress errors, thereby improving satellite dynamic tracking performance. This method eliminates the reliance on beacon signals for tracking, using communication carrier signal detection as the data source for neural network training, thus improving error suppression capabilities for integrated navigation and achieving high-precision satellite tracking and broadband communication.
[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:
[0009] A high-precision satellite tracking method while in motion, characterized by the following steps:
[0010] The azimuth and elevation angles of the satellite tracking beam, obtained by transforming the attitude values of the integrated navigation measurement vehicle into coordinates, and the acceleration measurements from the inertial navigation system output are used as inputs to the neural network. The azimuth and elevation angles of the optimal communication beam, detected by the communication carrier energy, are used as the outputs of the neural network for online training. Before completing the online training of the neural network, the azimuth and elevation angles of the satellite tracking beam are calculated from the attitude values of the measurement vehicle output by the integrated navigation system for low-precision satellite tracking.
[0011] When the deviation between the optimal beam position obtained from carrier energy detection and the satellite beam position calculated by the neural network is within a threshold range, the neural network model is deemed to have reached a practical level and training is stopped. Subsequently, the azimuth and elevation angles of the satellite beam based on the phased array plane and the acceleration measurement values output by the inertial navigation system, obtained by coordinate transformation of the attitude values of the integrated navigation measurement vehicle, are directly input into the neural network model to obtain the corrected optimal azimuth and elevation angles of the satellite beam, thereby suppressing integrated navigation errors and completing accurate satellite tracking.
[0012] A further technical solution of the present invention: the lower-precision star tracking includes the following steps:
[0013] After the phased array mobile communication antenna completes initial positioning and carrier attitude measurement through integrated navigation, it calculates the azimuth and elevation angles of the satellite beam based on the carrier coordinates through coordinate transformation according to the location of the communication satellite.
[0014] The phased array antenna forms a satellite beam based on the calculated azimuth and elevation angles of the satellite beam, completing the initial satellite alignment and network integration. It continuously updates the direction of the satellite beam in the carrier coordinate system by periodically outputting the measured heading, elevation, and roll angles of the integrated navigation system, and performs satellite tracking with lower precision.
[0015] A further technical solution of the present invention: the calculation based on the azimuth and elevation angles of the star beam relative to the carrier coordinates specifically includes:
[0016] Let the attitude measurement information provided by the inertial navigation system be (θ, ρ, ψ), where θ represents the heading angle, ρ represents the pitch angle, and ψ represents the roll angle. Then the transformation matrix from the geographical coordinate system where the satellite is located to the carrier coordinate system is:
[0017]
[0018] Let the coordinates of the satellite's location in the carrier coordinate system be obtained after transformation using a transformation matrix. Then the azimuth angle α0 and elevation angle β0 of the star beam based on the phased array plane can be obtained as follows:
[0019]
[0020]
[0021] A further technical solution of the present invention: the step of detecting the azimuth and elevation angles of the optimal communication beam based on the communication carrier energy includes the following steps:
[0022] The communication carrier signal received by the antenna is split into two paths by a 1-to-2 power divider, one path of which is sent to the carrier signal energy detection circuit, and the other path is connected to the modem for satellite communication.
[0023] Design an RF filter circuit to constrain the detection signal within the bandwidth of the communication carrier. Use an RF logarithmic detector circuit to detect the carrier energy and convert it into a level value. Combined with a beam scanning optimization process, the level value is sampled multiple times by a multi-channel high-speed AD converter and the final sampled value is obtained after averaging.
[0024] Carrier energy detection is performed by scanning the beam with small step size n times within a fixed interval period. The beam azimuth and elevation angles of the point with the strongest signal energy are marked as α1 and β1.
[0025] A further technical solution of the present invention: the neural network is a BP neural network, which includes three layers: an input layer, a hidden layer and an output layer; wherein the input layer has 8 nodes, the hidden layer has 9 nodes and the output layer has 2 nodes.
[0026] A further technical solution of the present invention includes: after the azimuth and elevation angles of the satellite beam calculated by the neural network have accumulated for a certain period of time, the system returns to the carrier signal energy measurement mode to find the optimal communication beam position. When the deviation between the optimal beam position obtained by the neural network and the carrier signal energy measurement reaches 0.8 degrees, the system re-enters the neural network training mode. When the deviation of the training result is within the threshold range, the system returns to the error correction mode based on the neural network and enters the precise tracking mode.
[0027] A further technical solution of the present invention includes comparing the deviation between the neural network calculated correction value and the integrated navigation measurement calculated output value each time the azimuth angle and elevation angle of the star-guided beam are output after correction. When the deviation reaches 1 / 2 angle of the 3dB beam of the phased array antenna, the neural network online training mode correction is performed again.
[0028] A computer system is characterized by comprising: one or more processors, and a computer-readable storage medium for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method described above.
[0029] A computer-readable storage medium is characterized by storing computer-executable instructions, which, when executed, are used to implement the above-described method.
[0030] The beneficial effects of this invention are as follows:
[0031] This invention provides a high-precision satellite tracking method for mobile communication. By optimizing communication carrier signal detection and combining it with BP neural network training, it aims to suppress integrated navigation errors, improve satellite alignment and dynamic tracking accuracy, and enable the system to operate at higher gain of the phased array antenna for extended periods, achieving higher bandwidth and higher-speed communication. It has the following advantages:
[0032] First, the advantages of rapidly changing the antenna beam direction by using phased array antennas are proposed. The communication beam direction with the highest star gain is found by detecting the communication carrier signal, and this is used as a precise tracking target for training the neural network.
[0033] Secondly, due to the complex and ever-changing working environment, mobile communication systems inevitably contain various types of noise, making it difficult to establish accurate error models. Neural networks, however, do not require precise mathematical models and possess inherent advantages in handling nonlinear system errors due to their nonlinearity, self-organization, and self-learning capabilities. Using the coordinate transformation of the attitude values of the integrated navigation measurement vehicle, the azimuth and elevation angles of the satellite tracking beam based on the phased array plane, and the angular velocity and acceleration measurements from the inertial navigation system, as inputs to the neural network, and the azimuth and elevation angles of the optimal communication beam detected by the communication carrier energy as the outputs, a three-layer BP neural network is trained online. After training, the optimal communication beam direction can be provided in real time. Unlike previous methods that estimated errors by subtracting the velocity and position outputs from inertial navigation and satellite navigation, using Kalman filtering or combining neural networks, this invention does not require a precise mathematical model of error. Instead, it uses communication carrier signal detection to find the optimal communication direction as training samples, ultimately directly obtaining the corrected azimuth and elevation angles of the satellite tracking beam for higher-precision satellite tracking. The phased array antenna beam can always operate at a high gain, enabling the system to achieve higher bandwidth communication.
[0034] Finally, this invention eliminates the dependence of phased array antenna satellite tracking on beacon signals, eliminating the need for dedicated beacon receiving antennas and beacon tracking equipment, thus simplifying system design. It is particularly suitable for applications where the communication carrier and beacon frequencies are far apart or even where there are no beacon signals. Furthermore, it allows for the selection of lower-precision integrated navigation systems for system design, achieving better error suppression through this invention. These features all contribute to reducing system costs and enabling the integrated and miniaturized design of phased array antennas. Attached Figure Description
[0035] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.
[0036] Figure 1 A system block diagram of a satellite tracking method based on communication carrier detection and neural network error correction;
[0037] Figure 2 Diagram of a BP neural network model;
[0038] Figure 3 This is a flowchart of the satellite tracking process. Detailed Implementation
[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.
[0040] like Figure 1 The diagram shows a system block diagram of a satellite tracking method based on communication carrier detection and neural network error correction. The satellite tracking flowchart is shown below. Figure 3 As shown.
[0041] The implementation steps are as follows:
[0042] Step 1: After the phased array mobile communication antenna completes the initial positioning and carrier attitude measurement through integrated navigation, it calculates the azimuth and elevation angles of the satellite beam based on the carrier coordinates by transforming the coordinates according to the location of the communication satellite.
[0043] Let the attitude measurement information provided by the inertial navigation system be (θ, ρ, ψ) (where θ represents the heading angle, ρ represents the pitch angle, and ψ represents the roll angle), then the transformation matrix from the geographical coordinate system where the satellite is located to the carrier coordinate system is:
[0044]
[0045] Let the coordinates of the satellite's location in the carrier coordinate system be obtained after transformation using a transformation matrix. Then the azimuth angle α0 and elevation angle β0 of the star beam based on the phased array plane can be obtained as follows:
[0046]
[0047]
[0048] Step 2: The phased array antenna forms a satellite beam based on the calculated azimuth and elevation angles to complete the initial satellite alignment and network entry. It continuously updates the direction of the satellite beam in the carrier coordinate system by periodically outputting the measured heading, elevation, and roll angles of the integrated navigation system, and performs low-precision satellite program tracking.
[0049] Step 3: The communication carrier signal received by the antenna is split into two paths using a 1-to-2 power divider. One path is sent to the carrier signal energy detection circuit, and the other output is connected to the modem for satellite communication.
[0050] Step 4: Design an RF filter circuit to constrain the detection signal within the bandwidth of the communication carrier (specifically related to the selected communication satellite). Use an RF logarithmic detector circuit to detect the carrier energy and convert it into a level value. In conjunction with the beam scanning optimization process, perform multiple multi-channel high-speed AD samplings of the level value and calculate the average value to obtain the final sampled value.
[0051] Step 5: The integrated navigation updates the position and attitude measurements every 20 milliseconds. Within the 20-millisecond interval, it performs 5 small-step beam scan optimization carrier energy detections, which means dividing the 360-degree circle into 5 equal parts and measuring once every 72 degrees. The beam azimuth and elevation angles of the point with the strongest signal energy are marked as α1 and β1.
[0052] Step 6: [The sentence is incomplete and requires more context to be translated accurately.] Figure 1 Switches S1 and S3 are closed, and switch S2 is opened to enter the neural network learning and training mode. Since the main sources of attitude measurement error are the gyroscopes and accelerometers (accelerometers) of the integrated navigation system, the azimuth and elevation angles of the satellite beam based on the phased array plane, obtained after coordinate transformation of the integrated navigation measurement vehicle's attitude values, and the gyroscope angular velocity and accelerometer acceleration measurements output by the inertial navigation system are used as inputs to the neural network. The azimuth and elevation angles of the optimal communication beam, detected based on the communication carrier energy, are used as the outputs of the neural network for online training of the three-layer BP neural network. Before completing the online training of the neural network, the system calculates the azimuth and elevation angles of the satellite beam from the integrated navigation system's output attitude values for lower-precision satellite tracking.
[0053] Artificial neural networks learn through self-training, without needing to determine the mathematical mapping relationship between input and output, and obtain the output value closest to the desired value given an input value. A BP neural network is a multi-layer feedforward network that uses gradient descent for gradient search and is trained by backpropagation of error to minimize the root mean square error between the network's output value and the desired output value. A BP neural network includes forward propagation of the signal and backward propagation of the error. Forward propagation calculates the error; the input signal passes through hidden layers and undergoes a nonlinear transformation to generate the output signal. If the output signal does not match the desired output value, the backward propagation process begins. Backpropagation propagates the output error back through the hidden layers to the input layer layer by layer, distributing the error among the units and adjusting the weights of each unit. To determine the network parameters, such as weights and thresholds, corresponding to the minimum error, it is necessary to adjust the connection strength and thresholds of each node, allowing the error to decrease along the gradient direction, and to repeatedly train the network. After training, inputting similar samples into the neural network yields the output information with the minimum error after nonlinear transformation.
[0054] The input to the BP neural network is the coordinate transformation of the inertial navigation system's attitude values, which are then used to obtain the azimuth angle α and pitch angle β of the star-aligned beam based on the phased array plane, as well as the 3-axis gyro angular velocity and accelerometer acceleration measurements (v) from the inertial navigation system output. x v y v z a x a y a z The output is the azimuth angle α1 and elevation angle β1 of the satellite beam for communication carrier detection optimization. After the BP neural network is trained, it outputs the corrected azimuth angle α1 of the satellite beam. AI Pitch angle β AI .
[0055] The BP neural network design structure diagram is as follows: Figure 2 As shown, the structure has 8 nodes in the input layer, 9 nodes in the hidden layer, and 2 nodes in the output layer. The learning rate η in the gradient descent method is set to 0.3, the momentum factor α is set to 0.05, and the initial values of the weighting coefficients are randomly generated within the range of (-0.5, 0.5).
[0056] (1) Forward propagation
[0057] The input signal is:
[0058]
[0059] The inputs and outputs of the hidden layer are as follows:
[0060]
[0061]
[0062] The input to the output layer is:
[0063]
[0064] The output of the output layer is:
[0065]
[0066]
[0067] (2) Error Backpropagation Algorithm
[0068] The connection weights of a backpropagation (BP) neural network are adjusted using the gradient descent algorithm. In the BP neural network, the selected performance metric function is...
[0069]
[0070] The formula for adjusting the connection weights from the hidden layer to the output layer is:
[0071]
[0072]
[0073] Where η is the learning efficiency; α is the momentum factor.
[0074] The formula for adjusting the connection weights from the input layer to the hidden layer is:
[0075]
[0076]
[0077] Step 7: When the optimal beam position obtained from 5 small-step carrier energy detections within a 20-millisecond interval deviates from the satellite beam position calculated by the neural network within 0.3 degrees, the BP neural network model is deemed to have reached a practical level, and training is stopped. Subsequently, the azimuth and elevation angles of the satellite beam based on the phased array plane, obtained by coordinate transformation of the combined inertial navigation measurement carrier attitude values, and the angular velocity and acceleration measurements output by the inertial navigation system are input into the BP neural network model to directly obtain the corrected optimal azimuth and elevation angles of the satellite beam. This achieves the suppression of combined navigation errors, completes accurate satellite tracking, and ensures that the antenna beam is always at a high gain position, enabling the system to conduct higher bandwidth communication for extended periods.
[0078] Step 8: After the BP neural network calculates the azimuth and elevation angles of the satellite beam for a certain period of time, return to the carrier signal energy measurement mode to find the optimal communication beam position. If the deviation between the optimal beam position obtained by the BP neural network and the carrier signal energy measurement reaches 0.8 degrees, return to Step 6 to enter the BP neural network training mode. When the deviation of the training result is within 0.3 degrees, return to the error correction mode based on the BP neural network to enter the precise tracking mode.
[0079] Step 9: After multiple iterations of training, the error suppression effect of the BP neural network becomes increasingly better, eventually reaching a point where it completely stops using the BP neural network to correct errors. At this point, the loop is closed. Figure 1 Switch S2 in the middle opens switches S1 and S3, outputting the azimuth angle α of the BP neural network error correction for the star-aligned beam. AI Pitch angle β AI Used for precise tracking of star beams.
[0080] Step 10: Each time the system outputs the corrected azimuth and elevation angles of the satellite beam, it compares the deviation between the neural network-calculated correction value and the integrated navigation measurement output value. When the deviation reaches half the angle of the 3dB beam of the phased array antenna, it returns to Step 6 for online neural network training mode correction. A relatively loose constraint is applied to the final output of the satellite beam azimuth and elevation angles to eliminate some abnormal situations.
[0081] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the scope of the technology disclosed in the present invention, and such modifications or substitutions should all be covered within the scope of protection of the present invention.
Claims
1. A high-precision satellite tracking method while on the move, characterized in that... The steps are as follows: The azimuth and elevation angles of the satellite tracking beam, obtained by transforming the attitude values of the integrated navigation measurement vehicle into coordinates, and the acceleration measurements from the inertial navigation system output are used as inputs to the neural network. The azimuth and elevation angles of the optimal communication beam, detected by the communication carrier energy, are used as the outputs of the neural network for online training. Before completing the online training of the neural network, the azimuth and elevation angles of the satellite tracking beam are calculated from the attitude values of the measurement vehicle output by the integrated navigation system for low-precision satellite tracking. When the deviation between the optimal beam position obtained by carrier energy detection and the satellite beam position calculated by the neural network is within the threshold range, the neural network model is deemed to have reached a practical level and training is stopped. Subsequently, the azimuth and elevation angles of the satellite beam based on the phased array plane and the acceleration measurement values output by the inertial navigation system are obtained by transforming the attitude values of the integrated navigation measurement vehicle into coordinates and inputting them into the neural network model to directly obtain the corrected optimal azimuth and elevation angles of the satellite beam, thereby suppressing integrated navigation errors and completing accurate satellite tracking. The step of detecting the azimuth and elevation angles of the optimal communication beam based on the communication carrier energy includes the following steps: The communication carrier signal received by the antenna is split into two paths by a 1-to-2 power divider, one path of which is sent to the carrier signal energy detection circuit, and the other path is connected to the modem for satellite communication. Design an RF filter circuit to constrain the detection signal within the bandwidth of the communication carrier. Use an RF logarithmic detector circuit to detect the carrier energy and convert it into a level value. Combined with a beam scanning optimization process, the level value is sampled multiple times by a multi-channel high-speed AD converter and the final sampled value is obtained after averaging. Carrier energy detection is performed n times with small step size beam scans within a fixed interval. The beam azimuth and elevation angles of the point with the strongest signal energy are marked as follows: and ; After the neural network calculates the azimuth and elevation angles of the satellite beam for a certain period of time, it returns to the carrier signal energy measurement mode to find the optimal communication beam position. When the deviation between the optimal beam position obtained by the neural network and the carrier signal energy measurement reaches 0.8 degrees, it re-enters the neural network training mode. When the deviation of the training result is within the threshold range, it returns to the error correction mode based on the neural network and enters the precise tracking mode. Each time the corrected azimuth and elevation angles of the satellite beam are output, the deviation between the neural network-calculated correction value and the combined navigation measurement calculation output value is compared. When the deviation reaches 1 / 2 angle of the 3dB beam of the phased array antenna, the neural network online training mode correction is performed again.
2. The high-precision satellite tracking method on the move according to claim 1, characterized in that... The process of performing low-precision star tracking includes the following steps: After the phased array mobile communication antenna completes initial positioning and carrier attitude measurement through integrated navigation, it calculates the azimuth and elevation angles of the satellite beam based on the carrier coordinates through coordinate transformation according to the location of the communication satellite. The phased array antenna forms a satellite beam based on the calculated azimuth and elevation angles of the satellite beam, completing the initial satellite alignment and network integration. It continuously updates the direction of the satellite beam in the carrier coordinate system by periodically outputting the measured heading, elevation, and roll angles of the integrated navigation system, and performs satellite tracking with lower precision.
3. The high-precision satellite tracking method on the move according to claim 2, characterized in that... The calculation, based on the carrier coordinates and the azimuth and elevation angles of the star beam, specifically involves: Let the attitude measurement information provided by the inertial navigation system be (θ, ρ, ψ), where θ represents the heading angle, ρ represents the pitch angle, and ψ represents the roll angle. Then the transformation matrix from the geographical coordinate system where the satellite is located to the carrier coordinate system is: Let the coordinates of the satellite's location in the carrier coordinate system be obtained after transformation using a transformation matrix. Then the azimuth angle of the star beam based on the phased array plane can be obtained. Pitch angle for: 。 4. The high-precision satellite tracking method on the move according to claim 1, characterized in that... The neural network is a BP neural network, which consists of three layers: an input layer, a hidden layer, and an output layer; the input layer has 8 nodes, the hidden layer has 9 nodes, and the output layer has 2 nodes.
5. A computer system, characterized in that... include: One or more processors, a computer-readable storage medium, for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the method of any one of claims 1-4.
6. A computer-readable storage medium, characterized in that... The device stores computer-executable instructions, which, when executed, are used to implement the method described in any one of claims 1-4.
Citation Information
Patent Citations
Satellite tracking method and communication-in-moving equipment
CN115149994A
Advanced navigation satellite system positioning method and system using delayed precise information
EP3293549A1
