Satellite signal lock losing detection method, system and device and storage medium

Abstract

The invention discloses a satellite signal lock losing detection method, system and device and a storage medium. The detection method comprises the following steps: acquiring tracking data of each channel of a satellite signal receiver; extracting first window data of a plurality of time steps in the tracking data based on a preset sliding window to generate a first feature matrix; inputting the first feature matrix into a trained preset model, extracting local mutation features in a time sequence of the first feature matrix through a convolutional neural network, and performing time sequence modeling on the local mutation features through a long-short-term memory network to output time-dependent features; obtaining a lock losing probability of each channel output by a preset model; and for each channel, determining that the channel has a lock loss risk under the condition that the lock loss probability of the channel is greater than a preset probability threshold. According to the technical scheme, the lock losing condition of each channel of the GNSS receiver can be automatically learned and adapted, a more efficient early warning mechanism is improved, and the intelligence and response capability of the system are enhanced.

Description

Technical Field

[0001] The present application relates to the field of satellite navigation signal processing technology, and in particular to a method, system, device and storage medium for detecting loss of satellite signal lock. Background Art

[0002] In actual GNSS applications, receivers often operate in complex environments, such as densely populated urban areas, shaded by trees, and on dynamic acceleration platforms. Under these conditions, GNSS signals are extremely susceptible to obstruction, multipath interference, and weak signal attenuation. This can cause the receiver's signal tracking loop to be unable to continuously and stably lock onto the target signal, ultimately resulting in "loss of lock," seriously impacting the continuity and reliability of the navigation system. Currently, mainstream receivers generally rely on traditional methods such as carrier-to-noise ratio threshold determination and loop divergence feature detection to determine signal loss of lock. However, these methods often have a response lag and can only determine when loss of lock has occurred or is imminent. They lack early warning capabilities and are susceptible to transient noise disturbances, posing the risk of misjudgment and missed detection. Summary of the Invention

[0003] The purpose of the embodiments of the present application is to provide a method, system, device and storage medium for detecting satellite signal loss, so as to solve the technical problem in the prior art that it is impossible to accurately identify satellite signal receiver channel loss.

[0004] To achieve the above objectives, the present application provides, in a first aspect, a method for detecting loss of satellite signal lock, the method comprising: Obtain tracking data for each channel of the satellite signal receiver; For the tracking data of each channel, extract the first window data of multiple time steps in the tracking data based on a preset sliding window; Generate a first feature matrix according to the first window data of all time steps corresponding to each channel; Inputting the first feature matrix corresponding to each channel into a trained preset model, the preset model including a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract local mutation features in the time series of the first feature matrix, and the long short-term memory network is used to perform time series modeling on the local mutation features to output time-dependent features; Obtain the unlocking probability of each channel output by the preset model based on the local mutation characteristics and time-dependent characteristics; For each channel, if the unlock probability of the channel is greater than a preset probability threshold, it is determined that the channel has an unlock risk.

[0005] In an embodiment of the present application, the method also includes: obtaining historical tracking data and historical unlocking status of each channel of the satellite signal receiver; extracting second window data of multiple time steps in the historical tracking data based on a preset sliding window; generating a second feature matrix based on the second window data of all time steps; training the preset model based on the second feature matrices corresponding to multiple channels and the corresponding historical unlocking status to obtain a trained preset model.

[0006] In an embodiment of the present application, the method also includes: using the TensorFlow tool to perform integer quantization processing on the preset model to obtain a processed preset model; deploying the processed preset model to an embedded platform of the Cortex-A7 architecture so that the data type output by the preset model is structured data.

[0007] In an embodiment of the present application, the TensorFlow tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, including: when the system environment of the embedded platform is no operating system or a lightweight operating system, the TensorFlow Lite for Microcontrollers tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is in the form of a constant array; when the system environment of the embedded platform is a complete operating system, the TensorFlow Lite tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is a tflite model file, and the tflite model file is loaded through the file system to call the processed preset model.

[0008] In an embodiment of the present application, the tflite model file is burned into Flash or eMMC as a firmware resource.

[0009] In an embodiment of the present application, the tracking data is at least one of a code phase error, a carrier phase error, a Doppler frequency shift, and a carrier-to-noise ratio.

[0010] A second aspect of the present application provides a satellite signal loss detection system, the system comprising: A first computing core is configured to obtain tracking data for each channel of a satellite signal receiver, extract first window data of multiple time steps in the tracking data for each channel based on a preset sliding window, generate a first feature matrix based on the window data of all time steps corresponding to each channel, write the first feature matrix corresponding to each channel into a shared memory area, set a flag bit of the shared memory area to a first value, read the unlock probability of each channel from the shared memory area when the flag bit is a second value, and determine that the channel is at risk of unlocking when the unlock probability of the channel is greater than a preset probability threshold; a second computing core, configured to periodically poll a flag bit, and when the flag bit is a first value, read a first feature matrix corresponding to each channel from the shared memory area, so as to input the first feature matrix into a trained preset model, wherein the preset model includes a convolutional neural network and a long short-term memory network, wherein the local mutation features in the time series of the first feature matrix are extracted by the convolutional neural network, the local mutation features are time-series modeled by the long short-term memory network to output time-dependent features, the lock-loss probability of each channel output by the preset model according to the local mutation features and the time-dependent features is obtained, the lock-loss probability of each channel is written into the shared memory area, and the flag bit of the shared memory area is set to the second value; The shared memory area is used to store the first characteristic matrix and the unlock probability corresponding to each channel.

[0011] In an embodiment of the present application, the first computing core is also used to determine whether to trigger a recapture process for tracking data or adjust tracking parameters based on the unlock probability; the shared memory area includes: an input buffer for storing the first characteristic matrix corresponding to each channel; an output buffer for storing the unlock probability corresponding to each channel; and a status flag area for storing flags.

[0012] A third aspect of the present application provides a satellite signal loss detection device, comprising: a memory configured to store instructions; The processor is configured to call instructions from the memory and implement the above-mentioned method for detecting loss of satellite signal lock when executing the instructions.

[0013] A fourth aspect of the present application provides a machine-readable storage medium having stored thereon instructions for causing a machine to execute the above-mentioned satellite signal loss-of-lock detection method.

[0014] Through this technical solution, the pre-configured model built using CNN-LSTM can automatically learn and adapt to the loss of lock conditions of each GNSS receiver channel, providing a more efficient early warning mechanism and enhancing the system's intelligence and responsiveness. Furthermore, by introducing a probabilistic output method, the receiver can set sensitivity thresholds based on the loss of lock probability, flexibly adjusting reacquisition timing, and improving the reliability and stability of the entire system, making it particularly suitable for navigation applications in complex environments.

[0015] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present application but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings: Figure 1 The following schematically illustrates a flow chart of a method for detecting satellite signal loss of lock according to an embodiment of the present application; Figure 2 The following schematically shows an architecture diagram of a dual-core collaborative system according to an embodiment of the present application; Figure 3 The following schematically shows a structural block diagram of a satellite signal loss detection device according to an embodiment of the present application; Figure 4 The schematic diagram shows the structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] To make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0018] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the fact that they can be implemented by ordinary technicians in this field. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0019] Figure 1 The following schematically shows a flow chart of a method for detecting satellite signal loss of lock according to an embodiment of the present application. Figure 1 As shown, an embodiment of the present application provides a method for detecting satellite signal loss of lock, which may include the following steps.

[0020] S102, obtaining tracking data of each channel of the satellite signal receiver; S104, extracting first window data of multiple time steps in the tracking data of each channel based on a preset sliding window; S106, generating a first feature matrix according to the first window data of all time steps corresponding to each channel; S108: Inputting the first feature matrix corresponding to each channel into a trained preset model, the preset model including a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract local mutation features in the time series of the first feature matrix, and the long short-term memory network is used to perform time series modeling on the local mutation features to output time-dependent features; S110, obtaining the unlocking probability of each channel output by the preset model according to the local mutation characteristics and the time-dependent characteristics; S112: For each channel, if the unlocking probability of the channel is greater than a preset probability threshold, it is determined that the channel has an unlocking risk.

[0021] GNSS (Global Navigation Satellite System) receivers achieve high-precision measurements of time, position, and velocity by tracking signals from satellites. However, in practical application scenarios, such as urban canyons, high-speed movement, severe obstruction, or strong interference, the receiver's signal tracking loop is prone to phase lock failure, divergence, or even complete loss of lock, seriously impacting the continuity and reliability of the navigation system. To overcome the technical issues of existing GNSS signal loss detection methods, such as delayed response and low prediction accuracy, tracking data from each channel of the satellite signal receiver can be acquired. A deep learning model based on a convolutional neural network and long-short-term memory network architecture can be used to identify the loss probability of each channel and, therefore, determine whether each channel is at risk of loss of lock. Specifically, the GNSS receiver periodically collects characteristic data from the tracking loop output of each channel, including at least code phase error, carrier phase error, Doppler shift, and carrier-to-noise ratio (C / N), to characterize signal quality and tracking stability. For each channel's tracking data, a first window of tracking data is extracted based on a preset sliding window. This first window of data is the data segment captured by the preset sliding window. Generally, first window data for multiple time steps can be continuously extracted to generate a first feature matrix. The first feature matrix can be two-dimensional feature data. The preset model combines a convolutional neural network (CNN) to extract local mutation features from the input first feature matrix and a long short-term memory network (LSTM) to model temporal dependencies.

[0022] In this solution, the pre-configured model built using CNN-LSTM can automatically learn and adapt to the loss of lock conditions of each GNSS receiver channel, providing a more efficient early warning mechanism and enhancing the system's intelligence and responsiveness. Furthermore, by introducing a probabilistic output method, the receiver can set sensitivity thresholds based on the loss of lock probability, flexibly adjusting reacquisition timing, and improving the reliability and stability of the entire system, making it particularly suitable for navigation applications in complex environments.

[0023] In one specific embodiment, a sliding window strategy is used to continuously extract data for N time steps (where N is typically 10-20), forming an N×4 two-dimensional feature matrix to capture the temporal trends of channel states. This matrix serves as the model input for the subsequent neural network, effectively modeling the evolution of signal anomalies.

[0024] In one specific embodiment, the CNN module includes 12 Conv1D convolutional layers, each with 1632 filters and a kernel size of 35, followed by a max pooling layer (pool size = 2) using the ReLU activation function. The LSTM module includes 12 layers, each with 16 to 32 neurons, and uses a Dropout mechanism for regularization to prevent overfitting.

[0025] In an embodiment of the present application, the method also includes: obtaining historical tracking data and historical unlocking status of each channel of the satellite signal receiver; extracting second window data of multiple time steps in the historical tracking data based on a preset sliding window; generating a second feature matrix based on the second window data of all time steps; training the preset model based on the second feature matrices corresponding to multiple channels and the corresponding historical unlocking status to obtain a trained preset model.

[0026] It can be understood that the above process is the training process of the preset model. The second window data is extracted in the same manner as the first window data, except that the data extraction object is different: the second window data extracts historical tracking data. The second feature matrix can be a two-dimensional feature matrix. Training samples are generated using a sliding window method. The preset model is trained based on the second feature matrices corresponding to multiple channels and the corresponding historical loss of lock states to obtain a trained preset model. The historical loss of lock states in the data are based on manual annotation or comparative error analysis to improve label accuracy.

[0027] In an embodiment of the present application, the method also includes: using the TensorFlow tool to perform integer quantization processing on the preset model to obtain a processed preset model; deploying the processed preset model to an embedded platform of the Cortex-A7 architecture so that the data type output by the preset model is structured data.

[0028] Specifically, model training can be performed using the TensorFlow or PyTorch frameworks, with Adam as the optimizer and Binary Cross Entropy (BCE) as the loss function. After training, the model is quantized to INT8 integers using TensorFlow Lite to adapt it for deployment on Cortex-A7 architecture embedded platforms, ensuring both operational efficiency and inference accuracy.

[0029] In an embodiment of the present application, the TensorFlow tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, including: when the system environment of the embedded platform is no operating system or a lightweight operating system, the TensorFlow Lite for Microcontrollers tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is in the form of a constant array; when the system environment of the embedded platform is a complete operating system, the TensorFlow Lite tool is used to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is a tflite model file, and the tflite model file is loaded through the file system to call the processed preset model.

[0030] Understandably, the trained deep neural network model initially uses floating-point precision (float32), resulting in a model size of approximately 12MB, making it difficult to run directly on resource-constrained embedded platforms. To address this, the model was integer-quantized (INT8) using the TensorFlow Lite tool, converting all weight parameters and intermediate activation tensors to 8-bit fixed-point format. This effectively compresses the model size to less than 200-300KB, significantly reducing memory usage and computational overhead without significantly compromising prediction performance, making it suitable for the storage and computing power requirements of ARM Cortex-A7-class platforms. Specifically, "no operating system" can refer to a bare-metal platform, while "lightweight operating system" can refer to a lightweight RTOS. Using TensorFlow Lite for Microcontrollers, the inference engine eliminates the need for dynamic memory allocation and is suitable for environments without a file system. When the embedded platform's system environment is the Linux kernel (AArch32 architecture), the standard TFLite runtime is used to load the model and perform inference via the file system.

[0031] In an embodiment of the present application, the tflite model file is burned into Flash or eMMC as a firmware resource. The model file can be burned into on-chip Flash or eMMC as a firmware resource and loaded into the memory when the system starts. The inference process includes: initializing the model → reading the input two-dimensional feature matrix → calling the inference function → outputting the probability of loss of lock. The entire process runs independently in the model inference core (such as the Cortex-A7 core), and the inference delay is controlled within 15ms, supporting data update frequencies of 100Hz and above. Through the above deployment strategy, both model prediction accuracy and system operation efficiency are taken into account, which can meet the application requirements of embedded GNSS receivers for low power consumption, high response, and strong real-time performance.

[0032] In this solution, the model, after INT8 quantization, runs on an ARM Cortex-A7 single-core processor, keeping single-inference latency within 10-15 milliseconds. The model supports feature data input at a 100Hz frequency and completes a full prediction within 10 milliseconds, meeting the real-time processing cycle of the GNSS channel. The overall end-to-end response latency, including data preprocessing, model inference, and result transmission, is kept within 20 milliseconds, meeting the fast response requirements of highly dynamic navigation systems.

[0033] The model, trained and tested on a standard GNSS lock loss sample set, has demonstrated prediction accuracy exceeding 90%. Its F1-score exceeds 0.88, balancing low false alarm rates with high detection rates. Its recall rate exceeds 90%, ensuring excellent identification of high-risk channel lock loss scenarios. The model can output a warning probability value 100 to 300 milliseconds before lock loss occurs, providing the receiver with a response window for reacquisition or parameter adjustment.

[0034] The CPU utilization of the model execution core (AI core) during inference does not exceed 60%, maintaining the multi-tasking scheduling capabilities of the embedded platform. The model itself is kept within 300KB after quantization, facilitating storage and deployment. Peak memory usage during model execution does not exceed 1MB, making it compatible with low-power SoC platforms. The average power consumption increase during model execution does not exceed 0.3 watts, meeting the power consumption requirements of portable GNSS devices.

[0035] A second aspect of the present application provides a satellite signal loss detection system, the system comprising: A first computing core is configured to obtain tracking data for each channel of a satellite signal receiver, extract first window data of multiple time steps in the tracking data for each channel based on a preset sliding window, generate a first feature matrix based on the window data of all time steps corresponding to each channel, write the first feature matrix corresponding to each channel into a shared memory area, set a flag bit of the shared memory area to a first value, read the unlock probability of each channel from the shared memory area when the flag bit is a second value, and determine that the channel is at risk of unlocking when the unlock probability of the channel is greater than a preset probability threshold; a second computing core, configured to periodically poll a flag bit, and when the flag bit is a first value, read a first feature matrix corresponding to each channel from the shared memory area, so as to input the first feature matrix into a trained preset model, wherein the preset model includes a convolutional neural network and a long short-term memory network, wherein the local mutation features in the time series of the first feature matrix are extracted by the convolutional neural network, the local mutation features are time-series modeled by the long short-term memory network to output time-dependent features, the lock-loss probability of each channel output by the preset model according to the local mutation features and the time-dependent features is obtained, the lock-loss probability of each channel is written into the shared memory area, and the flag bit of the shared memory area is set to the second value; The shared memory area is used to store the first characteristic matrix and the unlock probability corresponding to each channel.

[0036] In an embodiment of the present application, the first computing core is also used to determine whether to trigger a recapture process for tracking data or adjust tracking parameters based on the unlock probability; the shared memory area includes: an input buffer for storing the first characteristic matrix corresponding to each channel; an output buffer for storing the unlock probability corresponding to each channel; and a status flag area for storing flags.

[0037] Specifically, if Figure 2As shown in the figure, this system uses a dual-core collaborative architecture, assigning the GNSS main task and the neural network inference task to different processing cores. The main core (Core 1) is responsible for GNSS signal acquisition, tracking, and channel data management, while the AI ​​inference core (Core 2) is dedicated to running the TFLite-quantized CNN+LSTM model. The two cores communicate and interact via shared memory. The shared memory area is divided into three parts: an input buffer, an output buffer, and a status flag area. The input buffer stores feature data constructed using a sliding window (code phase error, carrier phase error, Doppler shift, and carrier-to-noise ratio); the output buffer stores model prediction results (loss of lock probability); and the status flag controls inter-core synchronization to prevent data conflicts. The system scheduling process is as follows: the main core samples, filters, and writes channel data to shared memory every 10ms, simultaneously setting the status word to "1." The inference core periodically polls the status word. Upon detecting a "1" value, it immediately loads data, executes neural network inference, writes the loss of lock probability to shared memory, and then sets the status word to "2." When the main core detects a status word value of "2," it reads the result and uses it to determine whether to trigger the signal reacquisition process. Specifically, the loss-of-lock probability value output by the model is compared with a preset threshold. If it exceeds the threshold, the channel is considered at risk of loss of lock. The receiver can initiate the reacquisition process or adjust tracking parameters in advance to avoid navigation interruption. This process is implemented based on the dual-core architecture of the ARM Cortex-A7, featuring clear task separation, simple communication mechanisms, and excellent real-time performance. The overall system completes data acquisition, inference, and control response in a 10ms cycle, achieving a high-speed and stable loss-of-lock detection closed loop.

[0038] The aforementioned system has excellent performance in predictive performance, system resource control, and embedded platform operating efficiency, and is suitable for integration into various applications such as GNSS receiver chip modules, navigation terminals, vehicle-mounted equipment, or unmanned system navigation subsystems.

[0039] Figure 1 FIG. 1 is a flow chart of a method for detecting satellite signal loss of lock in one embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0040] Figure 3 The following schematically shows a structural block diagram of a satellite signal loss detection device according to an embodiment of the present application. Figure 3 As shown, an embodiment of the present application provides a satellite signal loss detection device, which may include: a memory configured to store instructions; The processor is configured to call instructions from the memory and implement the above-mentioned method for detecting loss of satellite signal lock when executing the instructions.

[0041] Specifically, in the embodiment of the present application, the processor may be configured to: Acquire tracking data for each channel of a satellite signal receiver; for the tracking data of each channel, extract first window data of multiple time steps in the tracking data based on a preset sliding window; generate a first feature matrix based on the first window data of all time steps corresponding to each channel; input the first feature matrix corresponding to each channel into a trained preset model, the preset model including a convolutional neural network and a long short-term memory network, wherein the local mutation features in the time series of the first feature matrix are extracted through the convolutional neural network, and the local mutation features are time-series modeled through the long short-term memory network to output time-dependent features; obtain the unlocking probability of each channel output by the preset model based on the local mutation features and the time-dependent features; for each channel, determine that the channel is at risk of unlocking when the unlocking probability of the channel is greater than a preset probability threshold.

[0042] In an embodiment of the present application, the processor may further be configured to: Historical tracking data and historical unlocking status of each channel of a satellite signal receiver are obtained; second window data of multiple time steps in the historical tracking data are extracted based on a preset sliding window; a second feature matrix is ​​generated based on the second window data of all time steps; and a preset model is trained based on the second feature matrices corresponding to the multiple channels and the corresponding historical unlocking status to obtain a trained preset model.

[0043] In an embodiment of the present application, the processor may further be configured to: The TensorFlow tool is used to perform integer quantization processing on the preset model to obtain a processed preset model; the processed preset model is deployed to an embedded platform of the Cortex-A7 architecture so that the data type output by the preset model is structured data.

[0044] In an embodiment of the present application, the processor may further be configured to: Using the TensorFlow tool to perform integer quantization processing on the preset model to obtain a processed preset model includes: when the system environment of the embedded platform is no operating system or a lightweight operating system, using the TensorFlowLite for Microcontrollers tool to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is in the form of a constant array; when the system environment of the embedded platform is a complete operating system, using the TensorFlow Lite tool to perform integer quantization processing on the preset model to obtain a processed preset model, and the processed preset model is a tflite model file, and the tflite model file is loaded through the file system to call the processed preset model.

[0045] In an embodiment of the present application, the tflite model file is burned into Flash or eMMC as a firmware resource.

[0046] An embodiment of the present application further provides a GNSS signal receiver, which may include the above-mentioned satellite signal loss detection device.

[0047] An embodiment of the present application further provides a machine-readable storage medium having stored thereon instructions for causing a machine to execute the above-mentioned method for detecting loss of satellite signal lock.

[0048] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and computer program B02 in the non-volatile storage medium A04. The database of the computer device is used to store satellite signal loss detection data. The network interface A02 of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor A01, the computer program B02 implements a method for detecting satellite signal loss.

[0049] Those skilled in the art will understand that Figure 4The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0050] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0051] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0052] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0053] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0054] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0055] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0056] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology for information storage. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0057] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0058] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for detecting satellite signal loss, characterized in that: The detection method comprises: Obtain tracking data for each channel of the satellite signal receiver; For the tracking data of each channel, extracting first window data of multiple time steps in the tracking data based on a preset sliding window; Generate a first feature matrix according to the first window data of all time steps corresponding to each channel; Inputting the first feature matrix corresponding to each channel into a trained preset model, the preset model comprising a convolutional neural network and a long short-term memory network, wherein the convolutional neural network is used to extract local mutation features in the time series of the first feature matrix, and the long short-term memory network is used to perform time series modeling on the local mutation features to output time-dependent features; Obtaining the unlocking probability of each channel output by the preset model according to the local mutation feature and the time-dependent feature; For each channel, if the unlocking probability of the channel is greater than a preset probability threshold, it is determined that the channel has an unlocking risk.

2. The method for detecting satellite signal loss of lock according to claim 1, wherein: The method further comprises: Obtain historical tracking data and historical unlock status of each channel of the satellite signal receiver; Extracting second window data of multiple time steps in the historical tracking data based on the preset sliding window; Generate a second feature matrix based on the second window data of all time steps; The preset model is trained according to the second characteristic matrices corresponding to the multiple channels and the corresponding historical unlocked states to obtain a trained preset model.

3. The method for detecting satellite signal loss of lock according to claim 1, wherein: The method further comprises: Performing integer quantization processing on the preset model using the TensorFlow tool to obtain a processed preset model; The processed preset model is deployed to an embedded platform of the Cortex-A7 architecture so that the data type output by the preset model is structured data.

4. The method for detecting satellite signal loss of lock according to claim 3, wherein: The use of the TensorFlow tool to perform integer quantization processing on the preset model to obtain the processed preset model includes: In a case where the system environment of the embedded platform is no operating system or a lightweight operating system, the preset model is integer quantized using the TensorFlow Lite for Microcontrollers tool to obtain the processed preset model, where the processed preset model is in the form of a constant array; When the system environment of the embedded platform is a complete operating system, the TensorFlow Lite tool is used to perform integer quantization processing on the preset model to obtain the processed preset model, which is a tflite model file. The tflite model file is loaded through the file system to call the processed preset model.

5. The method for detecting satellite signal loss of lock according to claim 4, wherein: The tflite model file is burned into Flash or eMMC as a firmware resource.

6. The method for detecting satellite signal loss of lock according to claim 1, wherein: The tracking data is at least one of a code phase error, a carrier phase error, a Doppler frequency shift, and a carrier-to-noise ratio.

7. A satellite signal loss detection system, characterized in that: The system comprises: A first computing core is configured to obtain tracking data for each channel of a satellite signal receiver, extract first window data of multiple time steps in the tracking data for each channel based on a preset sliding window, generate a first feature matrix based on the window data of all time steps corresponding to each channel, write the first feature matrix corresponding to each channel into a shared memory area, set a flag bit of the shared memory area to a first value, read an unlock probability of each channel from the shared memory area when the flag bit is a second value, and determine that there is a risk of unlocking the channel when the unlock probability of the channel is greater than a preset probability threshold; a second computing core, configured to periodically poll the flag bit, and when the flag bit is the first value, read a first feature matrix corresponding to each channel from the shared memory area, so as to input the first feature matrix into a trained preset model, wherein the preset model includes a convolutional neural network and a long short-term memory network, wherein local mutation features in the time series of the first feature matrix are extracted through the convolutional neural network, the local mutation features are time-series modeled through the long short-term memory network to output time-dependent features, the lock-loss probability of each channel output by the preset model according to the local mutation features and the time-dependent features is obtained, the lock-loss probability of each channel is written into the shared memory area, and the flag bit of the shared memory area is set to the second value; The shared memory area is used to store the first characteristic matrix and the unlock probability corresponding to each channel.

8. The satellite signal loss detection system according to claim 7, characterized in that: The first computing core is further configured to determine whether to trigger a recapture process for the tracking data or adjust tracking parameters according to the unlock probability; the shared memory area includes: Input buffer, used to store the first feature matrix corresponding to each channel; Output buffer, used to store the unlock probability corresponding to each channel; The status flag bit area is used to store the flag bit.

9. A satellite signal loss detection device, characterized in that: include: a memory configured to store instructions; The processor is configured to call the instructions from the memory and implement the method for detecting satellite signal loss of lock according to any one of claims 1 to 6 when executing the instructions.

10. A machine-readable storage medium, characterized in that The machine-readable storage medium stores instructions, which are used to enable a machine to execute the satellite signal loss detection method according to any one of claims 1 to 6.

Images