Cloud-native virtualized ground station system based on full software-defined radio, working method and medium

Abstract

The application provides a cloud-native virtualized ground station system based on a full software-defined radio, a working method and a medium. An edge digital front end receives a radio frequency signal transmitted by an over-the-top satellite and performs radio frequency digitization sampling processing to generate a digital radio frequency stream; an elastic computing scheduler determines a real-time signal-to-noise ratio in the digital radio frequency stream, and performs associated calculation of a decoding algorithm complexity and a hardware resource mapping relationship based on the real-time signal-to-noise ratio to determine a target decoding mode, and meanwhile, quota calculation of a computing resource scale is performed in combination with a satellite transit schedule, and accordingly, a software modem container matched with the target decoding mode is dynamically pulled up in a cloud container pool; the software modem container generates a bit stream composed of restored user service data based on the digital radio frequency stream in response to the target decoding mode. The application realizes deep decoupling of hardware and protocols, improves resource utilization and reduces deployment costs.

Description

Technical Field

[0001] This application relates to the field of satellite ground station infrastructure, and more specifically, to a cloud-native virtualized ground station system, operating method, and medium based on fully software-defined radio. Background Technology

[0002] With the rapid expansion of commercial spaceflight, the construction and maintenance costs of ground stations, as the core infrastructure for satellite downlink data reception and command uplink, have become a key factor restricting the industry's development. Traditional ground stations typically require the deployment of a large amount of dedicated hardware at the antenna site to meet the processing needs of different satellite frequency bands and complex communication protocols.

[0003] In existing ground station construction schemes, hardware modems based on dedicated application-specific integrated modules (ASICs) or field-programmable gate arrays (FPGAs) are typically used. This scheme first constructs an equipment room near the antenna tower and configures hardware decoding equipment that is strongly bound to the target satellite protocol; then, the weak received signal is transmitted to the equipment room via analog radio frequency cables for down-conversion and demodulation processing; finally, the bitstream is restored by fixed hardware logic and delivered to the backend network.

[0004] However, this existing hardware-driven approach has significant technical drawbacks. Because the hardware logic and communication protocols are highly locked, expensive physical equipment must be replaced when a ground station needs to support a newly launched satellite or upgrade the communication standard, resulting in extremely poor system upgrade flexibility. Furthermore, satellite overhead transit times are typically highly intermittent; when a satellite is not overhead, these expensive dedicated hardware resources are idle, leading to a significant waste of computing resources and extremely high capital expenditures (CAPEX). Summary of the Invention

[0005] To at least alleviate the aforementioned technical problems, this application provides a cloud-native virtualized ground station system, operating method, and medium based on fully software-defined radio, in order to at least alleviate the aforementioned technical problems.

[0006] A cloud-native virtualized ground station system based on fully software-defined radio includes: an edge digital front-end, an elastic computing scheduler, and a software modem container; the edge digital front-end is used to receive radio frequency signals transmitted by overhead satellites and perform radio frequency digital sampling processing to generate a digital radio frequency stream; the elastic computing scheduler is used to determine the real-time signal-to-noise ratio in the digital radio frequency stream, and perform correlation calculations on the mapping relationship between decoding algorithm complexity and hardware resources based on the real-time signal-to-noise ratio to determine the target decoding mode, while combining the satellite overpass schedule to perform quota calculations on the scale of computing resources, and dynamically launch software modem containers matching the target decoding mode in the cloud container pool accordingly; the software modem container is used to generate a bit stream composed of restored user service data based on the digital radio frequency stream in response to the target decoding mode.

[0007] Optionally, the analog-to-digital conversion module is used to sample the radio frequency signal to generate an original in-phase quadrature sample sequence, and the timing interface module is used to inject a time stamp with a set precision into the original in-phase quadrature sample sequence; the original in-phase quadrature sample sequence after injecting the time stamp with the set precision is segmented and encapsulated according to the service transmission protocol specification to generate a digital radio frequency stream carried by the Internet protocol and containing the time stamp information with the set precision.

[0008] Optionally, the real-time signal-to-noise ratio is compared with a preset performance threshold to establish the correlation logic between the number of decoding iterations and the weight of computing resources, and a decoding strategy control code is generated based on the correlation logic; based on the decoding strategy control code, one of the preset low-power decoding mode and enhanced decoding mode is selected and established as the target decoding mode.

[0009] Optionally, the preset performance thresholds include a first threshold and a second threshold. The determination logic for the target decoding mode is as follows: if the real-time signal-to-noise ratio is higher than the first threshold, the elastic computing scheduler establishes the target decoding mode as a low-power decoding mode based on the decoding strategy control code, and drives the software modem container to perform decoding processing with a low number of iterations; if the real-time signal-to-noise ratio is lower than the second threshold, the elastic computing scheduler establishes the target decoding mode as an enhanced decoding mode based on the decoding strategy control code, and drives the software modem container to perform decoding processing with a high number of iterations.

[0010] Optionally, based on the established target decoding mode, the corresponding number of virtual central processing unit cores and the demand for graphics processing unit computing power are calculated to generate a resource quota instruction set; through the cloud container orchestration interface module, the cloud container pool is expanded or shrunk according to the resource quota instruction set to instantiate a software modem container corresponding to the resource quota instruction set in the cloud container pool, and use it as a physical instance to support the scale of computing resources.

[0011] Optionally, error-correcting code block features in the digital radio frequency stream are identified, and in response to the target decoding mode determined by the elastic computing scheduler, the hardware acceleration unit is invoked to perform parallel decoding processing on the error-correcting code block features to generate the restored bit stream.

[0012] Optionally, satellite orbit parameters are acquired by monitoring, and the satellite transit schedule is updated in real time based on the satellite orbit parameters to extract the communication frequency points and demodulation parameters corresponding to the visible satellites; the communication frequency points and demodulation parameters are encapsulated into container startup metadata, and the container startup metadata is injected into the initialization sequence of the software modem container to drive the software modem container to perform environment configuration according to the initialization sequence.

[0013] Optionally, the Internet Protocol header in the digital radio frequency stream is parsed to identify the stream source identification information associated with the target over-the-head satellite; using the container network interface, logical redirection processing is performed on the container network interface based on the stream source identification information to direct the digital radio frequency stream to a software modem container instantiated in response to the stream source identification information.

[0014] Optionally, monitor the attenuation trend of the real-time signal-to-noise ratio and identify physical characteristics where the real-time signal-to-noise ratio is lower than the low-power shutdown threshold; in response to the physical characteristics, issue a resource reclamation command through the cloud container orchestration interface module to release the associated computing resources from the software modem container and return them to the cloud container pool.

[0015] Optionally, based on the extracted satellite identifier, a target protocol image version corresponding to the satellite identifier is matched from the protocol image library; the configuration metadata corresponding to the target protocol image version is injected into the software modem container to drive the software modem container to dynamically instantiate and upgrade in response to the configuration metadata to perform demodulation logic.

[0016] A cloud-native virtualized ground station operating method based on fully software-defined radio, comprising: The edge digitization front end receives radio frequency signals transmitted by overhead satellites and performs radio frequency digitization sampling processing to generate a digital radio frequency stream; The elastic computing scheduler determines the real-time signal-to-noise ratio in the digital radio frequency stream and performs correlation calculations on the mapping relationship between decoding algorithm complexity and hardware resources based on the real-time signal-to-noise ratio to determine the target decoding mode. At the same time, it performs quota calculations on the scale of computing resources in conjunction with the satellite transit schedule, and dynamically launches software modulation and demodulation containers that match the target decoding mode in the cloud container pool accordingly. The software modem container responds to the target decoding mode by generating a bitstream consisting of the restored user service data based on the digital radio frequency stream.

[0017] A computer storage medium having computer-executable instructions stored thereon, the computer-executable instructions being executed to perform the described method of operation.

[0018] The cloud-native virtualized ground station solution based on fully software-defined radio proposed in this application has the following technical advantages: First, by performing RF digital sampling processing and generating a digital RF stream through an edge digital front-end, this approach alleviates the "hardware lock-in" problem inherent in traditional solutions that rely on dedicated ASIC / FPGA hardware modems. In traditional solutions, the communication protocol is directly embedded in the physical module, necessitating a complete system replacement when changing satellite protocols. This application, however, converts the RF signal into a universal digital RF stream at the antenna front-end, allowing the subsequent demodulation logic to be separated from the physical device. This approach enables the software modem container to process digital signals purely in software, achieving the technical effect of supporting new satellite protocols simply by updating the container image without modifying the hardware. This significantly reduces the difficulty of system upgrades and deployment costs.

[0019] Secondly, by using an elastic computing scheduler to calculate the quota of computing resources in conjunction with the satellite transit schedule and dynamically launching software modem containers, the problem of severe idle hardware resources during non-satellite transit periods in traditional solutions is at least alleviated. Traditional hardware ground stations typically adopt a "one satellite, one machine" exclusive mode, where hardware costs are amortized and cannot be reused regardless of signal input. The elastic computing scheduler in this application can anchor the satellite transit time window as accurately as possible, requesting and instantiating computing resources from the cloud container pool only when the actual business occurs. This "on-demand instantiation" mechanism allows multiple virtual ground stations to share the same base station computing resources, significantly improving the overall resource utilization rate of ground station infrastructure compared to traditional solutions, and achieving lower operating expenses (OPEX) and construction costs.

[0020] Furthermore, by determining the real-time signal-to-noise ratio (SNR) through an elastic computing scheduler and performing correlation calculations between decoding algorithm complexity and hardware resource mapping to determine the target decoding mode, this approach alleviates the problem of balancing decoding reliability and computational efficiency in traditional schemes where fixed processing logic is insufficient. Satellite channels are affected by weather and orbital distance, resulting in significant fluctuations in signal quality. Traditional hardware logic is typically designed with fixed parameters based on worst-case scenarios, leading to computational overflow when the signal is good. This application can dynamically adjust the target decoding mode in response to changes in the real-time SNR. For example, a lower-complexity algorithm is used to reduce CPU usage when the SNR is high, while a higher-complexity enhanced decoding mode is automatically mapped and GPU acceleration is invoked when the SNR is low (e.g., in rain-attenuated environments). This dynamic resource adaptation strategy based on channel quality feedback achieves optimal allocation of computational overhead while ensuring the reliability of data bitstream restoration, exhibiting higher environmental adaptability compared to traditional schemes.

[0021] Finally, the software modem container responds to the target decoding mode and generates the restored bitstream based on the digital radio frequency stream. This closed-loop processing flow ensures end-to-end consistency of business data processing. By running software-defined radio logic on a general-purpose computing platform, this application not only eliminates the dependence on specialized hardware but also leverages cloud-native dynamic scaling capabilities to support the parallel processing needs of large-scale constellations under peak concurrency. Attached Figure Description

[0022] Figure 1 This application provides an embodiment of a cloud-native virtualized ground station system based on fully software-defined radio. Detailed Implementation

[0023] like Figure 1 The image shows an embodiment of a cloud-native virtualized ground station system based on fully software-defined radio, comprising: an edge digital front-end, an elastic computing scheduler, and a software modem container; the edge digital front-end is used to receive radio frequency signals transmitted by overhead satellites and perform radio frequency digital sampling processing to generate a digital radio frequency stream; the elastic computing scheduler is used to determine the real-time signal-to-noise ratio in the digital radio frequency stream, and perform correlation calculations on the mapping relationship between decoding algorithm complexity and hardware resources based on the real-time signal-to-noise ratio to determine the target decoding mode, and simultaneously perform quota calculations on the scale of computing resources in conjunction with the satellite overpass schedule, thereby dynamically launching software modem containers matching the target decoding mode in a cloud container pool; the software modem container is used to generate a bit stream composed of restored user service data based on the digital radio frequency stream in response to the target decoding mode.

[0024] Optionally, the radio frequency signal is sampled using an analog-to-digital conversion module to generate an original in-phase quadrature sample sequence, and a time synchronization interface module is used to inject a timestamp of a set precision into the original in-phase quadrature sample sequence; the original in-phase quadrature sample sequence after injecting the timestamp of the set precision is segmented and encapsulated according to the service transmission protocol specification to generate the digital radio frequency stream carried by the Internet Protocol and containing the timestamp information of the set precision.

[0025] Preferably, the edge digital front-end uses its built-in high-sensitivity analog signal receiving module to perform energy capture on electromagnetic waves from the airspace to identify the physical distribution characteristics of the transmission frequency of the overriding satellite. In the specific technical implementation of this identification, the carrier power spectral density entering the antenna front-end is monitored in real time, and the transient phase shift component of the radio frequency signal in the analog channel is captured. This establishes the response boundary of the physical space signal at the hardware interface, providing the raw analog radio frequency payload for subsequent conversion from continuous analog waveforms to discrete digital symbols. By identifying these subtle electromagnetic variations, the physical starting point of the signal processing chain is established, ensuring that subsequent processing actions are precisely anchored to the satellite energy carrier with payload semantics.

[0026] Preferably, the analog-to-digital conversion module acquires the identified analog RF payload and drives its built-in high-speed quantization operator to perform waveform discretization sampling processing on the voltage envelope, thereby generating the original in-phase orthogonal sample sequence discretely distributed along the time axis. During the generation process, complex sampling mapping is performed on the signal, converting the analog waveform into in-phase and orthogonal components describing the vector phase, and generating digital sample points reflecting the transient characteristics of the waveform according to a preset sampling frequency. Since the signal strength and frequency of the satellite dynamically evolve over time during its transit, these generated sample sequences achieve a digital twin representation of the satellite's original physical waveform. This processing step constitutes the basic data template for the software-defined demodulation logic, laying a digital feature foundation for subsequent accurate channel feature estimation in a virtualized environment.

[0027] Preferably, the timing interface module accesses an external precision clock source to extract a reference clock signal synchronized with the Global Navigation Satellite System (GNSS) to monitor a globally consistent time reference and generate a highly reliable time-base synchronization payload accordingly. In this step, the time-base synchronization payload is phase-aligned with the generation cycle of the original in-phase orthogonal sample sequence, and a time protocol is used to inject a timestamp of a set precision reflecting its instantaneous physical location at the moment of acquisition into each sample unit. This achieves the assignment of absolute time coordinates to discrete digital sampling points. These generated sample data with time characteristics at least alleviate the problem of dynamic changes in propagation delay caused by the high-speed motion of the satellite, ensuring that the backend demodulation logic can reconstruct the original service payload with definite spatiotemporal characteristics.

[0028] Preferably, the edge digital front-end acquires the sample payload injected with time-domain markers and drives its built-in signaling transmission engine to perform physical layer push to the backbone fiber optic network for continuous stream output processing. During the continuous stream output process, sampling data blocks carrying the set precision timestamp information are distributed to the cloud network in real time according to a preset sampling clock step. This action achieves a substantial physical migration of the observation payload from the edge hardware side to cloud computing resources, at least mitigating the risk of broadband signal overflow caused by limited local storage space of the front-end device. This transmission output processing ensures that the cloud computing instance can acquire real-time, continuous waveform samples, providing stable data stream support for subsequent large-scale parallel decoding.

[0029] Preferably, the edge digital front-end aggregates the real-time sample streams from the above-mentioned execution flow output process, and performs mapping and encapsulation processing for wide area network transmission protocols using configured protocol encapsulation operators, thereby ultimately forming the digital radio frequency stream characterizing the dynamic features of the original waveform of the overhead satellite. In the specific technical implementation, according to the service transmission protocol specification, the original in-phase orthogonal sample sequence is encapsulated into the standardized Internet Protocol message payload field, and a digital radio frequency stream with satellite stream source identifier and sequence number is generated. The formed digital radio frequency stream achieves deep decoupling between the radio frequency signal and the underlying specific physical link, enabling the analog waveform to be transformed into a general information stream that can be freely routed in the switching network. This processing step realizes the final transformation of the ground station from a hardware-bound architecture to a software-defined architecture, providing a standardized input interface for subsequent virtual demodulation execution in the cloud container pool.

[0030] Preferably, the network environment between the "edge digital front-end" and the "cloud container pool" is designed to support RoCE (RDMA over Converged Ethernet) or high-bandwidth leased lines, so as to support the real-time distribution of sampled data blocks carrying the specified precision timestamp information to the cloud network as bit streams. The specific value of the precision timestamp can be determined according to the engineering scenario and is not limited to a single value.

[0031] In summary, in this application, the identification of radio frequency signals provides the original physical signal source for the entire processing chain; the generated original in-phase quadrature sample sequence transforms the physical excitation into digital features that can be recognized by computational logic, directly determining the sampling accuracy of subsequent processing actions. Next, by executing continuous stream outputs, a rigorous logical sequence is established for the data stream using the set precision timestamps. Finally, the generated digital radio frequency stream supports the scheduler's assessment of channel quality.

[0032] Optionally, the real-time signal-to-noise ratio is compared with a preset performance threshold to establish the correlation logic between the number of decoding iterations and the weight of computing resources, and a decoding strategy control code is generated based on the correlation logic; based on the decoding strategy control code, one of the preset low-power decoding mode and enhanced decoding mode is selected and established as the target decoding mode.

[0033] Preferably, the elastic computing scheduler performs real-time analysis on the data stream flowing into the cloud region to observe and capture the signal quality characteristics of the digital radio frequency stream. In specific implementation, the elastic computing scheduler uses its built-in energy detection operator to extract the carrier useful signal power and in-band noise floor power from the digital radio frequency stream, and performs an algebraic ratio calculation between the two. This establishes the transient perception of the physical link channel quality by the cloud virtualization environment, enabling the acquisition of the real-time signal-to-noise ratio characterizing the effects of distance variations, rain attenuation, or Doppler effects on overlying satellites. By acquiring this raw observation metric, the initial stimulus source for the adaptive resource scheduling logic is established, ensuring that subsequent processing actions can accurately respond to the dynamic evolution of the satellite carrier environment.

[0034] Preferably, the elastic computing scheduler acquires the observed real-time signal-to-noise ratio (SNR) and compares it with preset performance threshold execution values ​​loaded in memory for discrimination and interval positioning. In this step, multiple preset threshold intervals reflecting different decoding convergence speeds are retrieved. By comparing the coordinate positions of the real-time SNR within each threshold interval, the corresponding logic relating the number of decoding iterations to the computing power resource weights is generated. Since the signal jitter generated by the overpass satellite during flight is discrete on the time axis, this application maps the macroscopic physical channel quality to the microscopic algorithm complexity expectation, establishing a closed-loop mapping between computing power allocation and data restoration quality, thereby providing a logical template for generating control sequences with scheduling semantics.

[0035] Preferably, the elastic computing scheduler drives its built-in conversion coding module to perform digital mapping on the generated correlation logic to generate decoding strategy weight values ​​discretely distributed over time. In the specific technical implementation, based on the current sampling rate, the correlation logic is transformed into logic code blocks with specific widths. These code blocks serve as discrete scales describing the decoding strength requirements, recording the quantitative requirements of the system's computing resource usage within each discrete time slice. This step represents a fundamental leap from abstract mapping logic to deterministic digital control parameters. These generated weight sequences not only record the channel's changing trajectory but also provide data index support for subsequent resource quota allocation within the cloud container pool.

[0036] Preferably, the elastic computing scheduler drives the internally configured instruction transmission interface to perform signaling encapsulation on the generated discrete weight sequence to further generate decoding strategy control codes and perform continuous stream output processing. During the execution of continuous stream output, the instruction code stream reflecting the decoding complexity requirements is pushed in real-time to the control plane of the software modem container according to a preset synchronization frequency. This action realizes the physical migration of the control payload from the resource scheduling side to the algorithm execution side, at least alleviating the problem of untimely parameter updates caused by the high dynamic motion of the satellite. This transmission output ensures that the virtual demodulation operator can capture micro-jumps in signal quality in real time, providing a dynamic input stream for achieving millisecond-level adaptive switching of decoding modes.

[0037] Preferably, the elastic computing scheduler aggregates the output decoding strategy control codes and uses the configured pattern matching model to perform the final scheme selection and establishment, thereby ultimately forming the target decoding mode that characterizes the complexity of the demodulation logic. In the specific technical implementation of this formation, in response to the mode selection identifier in the control code, it matches the corresponding mode from the preset low-power decoding mode and enhanced decoding mode. This formation process not only locks the algorithm branch within the current time slice but also synchronously establishes the corresponding computing power quota through the container orchestration interface. The formed target decoding mode achieves deep alignment between the physical characteristics of the satellite-to-ground link and the distribution of cloud computing resources, at least alleviating the technical defects of the high degree of lock-in between demodulation logic and hardware capabilities under the traditional ground station architecture.

[0038] In summary, this application demonstrates that the signal-to-noise ratio (SNR) capture action provides a physical-level feedback source for the entire processing chain; the generated discrete decoding strategy weight values ​​quantify complex analog interference into digital computational basis. Next, by executing continuous stream output actions, the scheduling instructions are synchronized to the execution unit in real time using the decoding strategy control code. Finally, the generated target decoding mode guides the software modem container's invocation of the graphics processing acceleration operator. This application compensates for the non-stationarity of the physical channel through the flexibility of software-defined radio, ensuring that the ground station system possesses high computational resource reuse efficiency and decoding stability.

[0039] Optionally, the preset performance threshold includes a first threshold and a second threshold, and the determination logic of the target decoding mode is as follows: if the real-time signal-to-noise ratio is higher than the first threshold, the elastic computing scheduler establishes the target decoding mode as the low-power decoding mode based on the decoding strategy control code, and drives the software modem container to perform decoding processing with a low number of iterations; if the real-time signal-to-noise ratio is lower than the second threshold, the elastic computing scheduler establishes the target decoding mode as the enhanced decoding mode based on the decoding strategy control code, and drives the software modem container to perform decoding processing with a high number of iterations.

[0040] Preferably, the elastic computing scheduler utilizes its configured signal analysis interface to analyze the digital radio frequency (RF) stream payload flowing into the cloud data inlet in real time, thereby capturing the physical link status characteristics. In the specific implementation of this capture, the elastic computing scheduler uses its built-in power spectrum detection operator to extract the carrier useful signal power component and the in-band noise floor power component from the digital RF stream, and performs an algebraic ratio calculation between the two. This establishes the cloud virtualization environment's ability to perceive the transient quality of the satellite-to-ground link, enabling the real-time acquisition of the real-time signal-to-noise ratio characterizing the impact of atmospheric attenuation or orbital variations on overlying satellites. By capturing this raw observation indicator, the physical feedback source of the adaptive scheduling process is established, ensuring that subsequent processing actions can respond as accurately as possible to the dynamic evolution of signal quality.

[0041] Preferably, the elastic computing scheduler acquires the captured real-time signal-to-noise ratio (SNR) and performs multi-level quantitative discrimination and interval positioning processing on it and preset performance thresholds (including the first threshold and the second threshold) loaded in memory to generate a decision logic sequence discretely distributed along the time axis. During the generation process, preset threshold values ​​reflecting different decoding convergence speeds are retrieved, and the probability distribution of the real-time SNR falling into high or low SNR intervals under different time slices is compared to generate a series of discrete sample values ​​reflecting the decoding complexity requirements. Since the signal jitter generated by the overpass satellite during flight exhibits discreteness along the time axis, these generated decision sequences map macroscopic channel advantages and disadvantages to microscopic algorithm expectations, providing a quantitative logical reference for the subsequent construction of the decoding strategy control code.

[0042] Preferably, the elastic computing scheduler, based on the generated decision logic sequence, drives its internally configured control logic generation engine to execute targeted digital encoding mapping, thereby generating the decoding strategy control code representing the mode switching trend. In the specific technical implementation of the generation, based on the current sampling rhythm, the required number of decoding iterations and the allocation weight of hardware computing resources are encapsulated into an instruction sequence with an 8-bit or 16-bit width, and marked as the decoding strategy control code. Therefore, this application realizes the solidification of abstract decision conclusions into protocol payloads that can be efficiently transmitted and parsed by cloud-native networks. The generated decoding strategy control code not only carries the preliminary screening conclusion for the target decoding mode, but also provides a definite digital instruction foundation for the subsequent precise alignment of computing resources within the cloud container pool.

[0043] Preferably, the elastic computing scheduler drives its built-in command transmission interface to perform signaling stream encapsulation on the generated decoding strategy control code, and accordingly controls the elastic computing scheduler to perform continuous stream output processing. During the execution of continuous stream output, the decoding strategy control code, containing mode selection identification information, is pushed in real-time to the control plane of the software modem container at a preset synchronization frequency (e.g., once every 10 milliseconds). This action realizes the physical migration of the control payload from the resource scheduling side to the demodulation execution side, at least alleviating the problem of control parameter update lag caused by the high dynamic motion of the satellite. This transmission output ensures that the underlying algorithm operators can capture the latest intentions of the scheduling layer in real time, achieving synchronization between the physical channel state and the cloud demodulation parameters.

[0044] Preferably, the elastic computing scheduler aggregates the decision results from the aforementioned streaming output process and uses the configured resource mapping operator to perform the final scheme establishment and mode locking processing, thereby ultimately forming the target decoding mode that characterizes the complexity of the demodulation logic. In the specific technical implementation, if the real-time signal-to-noise ratio is higher than the first threshold in response to the energy-saving flag in the decoding strategy control code, the target decoding mode is established as the low-power decoding mode; conversely, if the real-time signal-to-noise ratio is detected to be lower than the second threshold, it is established as the enhanced decoding mode. The formed target decoding mode achieves deep alignment between the physical characteristics of the satellite-to-ground link and the distribution of cloud computing resources, at least alleviating the technical defect of the high degree of lock between decoding logic and specific hardware capabilities under the traditional ground station architecture.

[0045] Specifically, in one scenario, while increasing the number of iterations, the scheduler will preferably also consider whether it is necessary to increase the computing resource quota (expand vCPU / GPU) to offset the increase in computing load and further ensure real-time performance.

[0046] In summary, this application firstly provides a physical-level excitation signal for the entire processing chain by capturing the real-time signal-to-noise ratio; the generated decision sequence and the decoding strategy control code then quantify the physical excitation into digital scheduling instructions. Next, the reliable transfer of the instruction set between heterogeneous computing nodes is ensured through a network interface by executing continuous stream output. Finally, the formed target decoding mode guides the software modem container to call graphics processing acceleration operators on demand. This application compensates for the shortcomings of space channel non-stationarity through the extreme flexibility of software-defined radio, enabling the ground station to maximize the reuse efficiency of computing resources while ensuring the quality of data bitstream restoration.

[0047] Optionally, based on the established target decoding mode, the corresponding number of virtual central processing unit cores and the demand for graphics processing unit computing power are calculated to generate a resource quota instruction set; through the cloud container orchestration interface module, the cloud container pool is scaled up or down according to the resource quota instruction set to instantiate the software modem container corresponding to the resource quota instruction set in the cloud container pool, and use it as a physical instance to carry the computing resource scale.

[0048] Preferably, the elastic computing scheduler drives its built-in task status monitoring operator to perform parsing actions on the target decoding mode established through the mapping relationship between decoding algorithm complexity and hardware resources, thereby achieving the most accurate capture of computing resource requirements. In the specific technical implementation of this capture, the computational density level of the satellite downlink signal in the current processing queue is identified, and the preset decoding iteration step size and parallelism requirements in the target decoding mode are captured and sensed. To this end, a quantitative interface is implemented to establish the logical algorithm requirements and physical computing capabilities, enabling the cloud environment to perceive in real time the expected consumption of the hardware base by the satellite under different channel conditions. By capturing these discrete mode parameters, a definite business logic input is provided for the subsequent dynamic calculation of resource weights, ensuring that subsequent processing actions can respond as accurately as possible to sudden computing power demands caused by changes in satellite-to-ground link quality.

[0049] Preferably, the elastic computing scheduler acquires the captured target decoding mode and drives its configured computing power topology mapping model to perform correlation calculation processing for hardware quotas, thereby generating a sequence of virtual hardware quota indicators discretely distributed over time. During the generation process, a pre-stored database of decoding algorithm complexity and hardware resource mapping relationships is retrieved. Based on the performance benchmark of the current target decoding mode, the number of virtual central processing unit cores (e.g., 4 to 16 cores) and the graphics processing unit computing power unit requirement required to maintain real-time signal decoding within the current observation period are calculated. Since the service load generated by satellite transit has significant time-domain non-stationary characteristics, these generated quota indicators achieve a digital twin representation of the physical computing power consumption trajectory. This step constitutes the basic data template for dynamic resource optimization, laying a digital characteristic foundation for subsequent fine-grained computing power allocation in a distributed environment.

[0050] Preferably, the elastic computing scheduler aggregates the generated virtual CPU core count and GPU computing unit demand, driving its internal instruction orchestration engine to perform mapping and encapsulation processing for cloud-native scheduling protocols, thereby ultimately forming the resource quota instruction set representing the computing power scheduling intent. In the specific technical implementation, the quantified hardware requirements, memory bandwidth placeholders, and storage volume mount declarations are logically aggregated to generate the resource quota instruction set conforming to the orchestration framework syntax specification. Therefore, this application transforms macro-level business computing power requirements into micro-level system execution commands, at least alleviating the problem of unallocated computing capacity due to the fixed deployment of traditional ground station hardware. The resulting resource quota instruction set provides standardized logical credentials for subsequent elastic resource delivery across nodes and data centers.

[0051] Preferably, the elastic computing scheduler drives its configured high-speed signaling synchronization channel to perform execution-plane-oriented transmission processing on the generated resource quota instruction set, and accordingly controls the system to execute continuous streaming output. During the specific process of executing continuous streaming output, the built-in cloud container orchestration interface module is used to push the encapsulated computing power demand payload to the control kernel of the computing cluster in real time in the form of an instruction stream. This action realizes a substantial physical migration of resource scheduling strategy from static configuration to dynamic flow. By executing this transmission output processing, the underlying container engine can capture the quota update intent from the scheduling side in real time, providing stable dynamic input stream support for achieving millisecond-level alignment between the ground station demodulation capability and the satellite overpass time.

[0052] Preferably, the cloud container orchestration interface module obtains the continuously output resource quota instruction set and drives the corresponding cloud container pool to perform deep scaling operations, thereby ultimately forming a baseline container instance cluster representing the dynamic computing power profile of the ground station. In the specific technical implementation, according to the hardware scale requirements specified in the instruction set, computing slices are instantaneously allocated in the general cloud server cluster, and the software modem container carrying a specific protocol image is instantaneously generated. This formation process not only locks the physical entity carrying the current computing task, but also realizes the deployment of ground station resources through the rapid startup characteristics of container technology. The formed instance cluster realizes deep tracking of computing resource scale and real-time business load, at least alleviating the problem of idle ground station hardware resources caused by intermittent satellite overhead.

[0053] Optionally, error-correcting code block features in the digital radio frequency stream are identified, and in response to the target decoding mode determined by the elastic computing scheduler, the hardware acceleration unit is invoked to perform parallel decoding processing on the error-correcting code block features to generate the restored bitstream. The hardware acceleration unit may be, for example, a GPU.

[0054] Preferably, the software modem container utilizes its built-in frame synchronization monitoring operator to analyze the digital radio frequency stream payload transmitted back via the high-speed network in real time, thereby identifying the logical boundaries of the service load. In the specific technical implementation of this identification, decapsulation processing is performed on the incoming Internet Protocol (IP) packets, and preset code block start characteristics and error correction prefix identifiers are captured and sensed in the digital radio frequency stream. This establishes the perception of the timing characteristics of the digital waveform by the cloud virtualization environment, enabling accurate anchoring of the data payload location to be processed. By identifying these microscopic protocol characteristics, the original logical trigger source is obtained for the subsequent essential transformation from the waveform domain to the probabilistic domain, ensuring that subsequent processing actions can respond as accurately as possible to the coding system requirements of different satellite carriers.

[0055] Preferably, the software modem container acquires the identified error correction code block features and drives its built-in soft-decision quantization engine to perform probability mapping processing on the signal voltage vector, thereby generating the discrete probability feature sequence that is discretely distributed along the time axis. During the generation process, distance metric calculations are performed on in-phase orthogonal samples in the digital radio frequency stream, converting the digitally quantized amplitude values ​​into log-likelihood ratio values ​​describing the bit logic polarity. Since the satellite channel exhibits continuous dynamic fluctuations along the time axis, these generated discrete probability feature sequences realize a digital discrete representation of the physical signal evolution trajectory, constituting the core input carrier of the decoding iterative algorithm.

[0056] Preferably, the software modem container, in response to the target decoding mode issued by the elastic computing scheduler, drives its built-in computing task orchestration engine to perform logical redirection processing for the acceleration driver interface. In the specific implementation of this processing, if the target decoding mode is determined to be an enhanced decoding mode, the high-performance computing channel is automatically activated and a signaling association with the hardware acceleration unit is established, thereby controlling the system to execute continuous streaming output. During the execution of continuous streaming output, according to a preset decoding pipeline rhythm, the generated discrete probability feature sequence is sent to the video memory space in real time through a non-uniform memory access path. This action achieves a substantial physical migration of the computing load from general-purpose processing units to specific hardware acceleration units, at least alleviating the performance bottleneck caused by the surge in computing power demands due to high-order satellite modulation.

[0057] Preferably, the hardware acceleration unit aggregates the received discrete probability feature sequences and drives its internally configured thousands of parallel computing cores to perform iterative deduction of the error correction and verification matrix. In the specific technical implementation, the hardware acceleration unit is invoked to perform parallel decoding processing on the error correction code block features, and graphics processing operators are used to perform real-time correction of signal distortion features in the discrete probability feature sequence. Because graphics processors possess a naturally large-scale thread concurrency architecture, this processing step achieves synchronous throughput of massive probability samples, enabling complex error correction codes to converge within millisecond-level time windows. The technical essence of this action is to trade computational power for processing time, ensuring that the virtualized ground station has processing real-time performance similar to that of a hardware ground station.

[0058] Preferably, the software modem container acquires the iterative results after the above correction processing, drives its configured hard decision discriminant operator to perform the final logical restoration, thereby ultimately forming the bitstream representing the restored user service data. In the specific technical implementation, based on the ultimate nature of the probability sequence, the error-corrected feature payload is restored to standard binary code, and logical reassembly of the original user data frame is performed based on the service transmission protocol. The formed bitstream realizes a morphological closed loop from discrete analog sampling points to a logical codestream with defined service semantics. This processing step marks the final completion of the cloud demodulation process, enabling the satellite downlink payload to be transparently delivered to the subsequent service application layer for processing.

[0059] In summary, in this application, the identification action provides a logical anchor point for the entire processing chain; the generated discrete probability feature sequence quantizes the physical signal into the input payload of the digital algorithm, directly determining the computational accuracy of subsequent processing actions. Next, by executing continuous stream output actions, the large-scale throughput capability of the dedicated acceleration unit compensates for the latency bottleneck when the software processes high-bit-rate signals. Finally, the formed bitstream closes the data delivery loop of the ground station. This application, through the extreme elasticity of computing resource pooling, maximizes the reuse efficiency of computing resources at the ground station while ensuring the quality of high-dynamic signal reconstruction.

[0060] Optionally, satellite orbit parameters are acquired by monitoring, and the satellite transit schedule is updated in real time according to the satellite orbit parameters to extract the communication frequency points and demodulation parameters corresponding to the visible satellites; the communication frequency points and demodulation parameters are encapsulated into container startup metadata, and the container startup metadata is injected into the initialization sequence of the software modem container to drive the software modem container to perform environment configuration according to the initialization sequence.

[0061] Preferably, the elastic computational scheduler, through its built-in orbit tracking interface, retrieves external ephemeris databases or two lines of orbital element data in real time to perform attitude identification actions for its target satellite. In the specific technical implementation of identification, it captures core physical quantities such as perigee altitude, orbital inclination, and mean perigee angle entering the monitoring window, and identifies the microscopic changes in these physical quantities caused by perturbation factors. To this end, it establishes the spatiotemporal perception boundary of the cloud-based virtualization system for spatial physical entities, thereby obtaining the original physical characteristic payloads for the subsequent essential mapping from the orbit evolution model to the ground station mission timing. By identifying these dynamic satellite orbital parameters, the initial trigger source of the entire automated scheduling process is established, ensuring that subsequent processing actions can respond as accurately as possible to the real-time displacement of the overriding satellite in physical space.

[0062] Preferably, the elastic computing scheduler acquires the identified satellite orbit parameters and drives its internally configured spatiotemporal topology evolution operator to perform predictive processing for visibility windows, thereby generating a satellite transit schedule discretely distributed along the time axis. In the specific generation process, the geographical coordinates of the ground station antenna are combined to calculate the discrete time nodes for the satellite's entry into, reaching its highest point, and leaving the antenna beam range, and the corresponding communication frequency points and demodulation parameters (e.g., code rate, modulation scheme, and error correction coding type) are extracted synchronously based on the satellite's identification. Since the satellite exhibits cyclical motion in its orbit, this series of time windows and their associated frequency domain characteristics achieve a digital discrete representation of the ground station's workload along the time axis. This processing step constitutes the basic data template for the pre-allocation of virtualized resources at the ground station, laying a digital feature foundation for subsequent precise container lifecycle management on heterogeneous computing units.

[0063] Preferably, the elastic computing scheduler aggregates the generated satellite transit schedule, communication frequency, and demodulation parameters, and drives the internally configured configuration orchestration module to perform syntax mapping processing for the cloud-native object model, thereby forming the container startup metadata with structured descriptive semantics. In the specific implementation, complex orbital mechanics conclusions and physical layer protocol parameters are encapsulated into configuration files or sets of environment variables that meet the requirements of the container runtime environment. This achieves the assignment of configuration semantics that can be recognized by the software system to discrete physical channel characteristics. The generated container startup metadata at least alleviates the problem of dynamic parameter configuration caused by the intermittent nature of satellite transits, supporting backend container instances to perform real-time environmental adaptation based on defined digital credentials, thereby achieving logical decoupling between business logic and underlying parameters.

[0064] Preferably, the elastic computing scheduler drives its built-in high-speed signaling transmission engine to perform push processing on the metadata of the generated containers to the computing nodes, and controls the system to perform continuous streaming output accordingly. During the execution of continuous streaming output, according to a preset synchronization step size, the encapsulated configuration payload is injected in real-time into the initialization interface of the software modem container through the cloud management channel. This action achieves a substantial physical migration of scheduling intent from the macro-decision plane to the micro-execution plane, at least alleviating the task switching delay problem caused by the high dynamic motion of satellites. This transmission output processing ensures that newly launched computing instances can capture the latest link configuration intent in real time, providing stable dynamic parameter stream support for the rapid instantiation of demodulation capabilities within the cloud container pool.

[0065] Preferably, the software modem container acquires the container startup metadata received during the aforementioned stream output process, drives its internal loading logic to perform construction processing for the demodulation algorithm environment, thereby ultimately forming a baseline initialization environment configuration representing the demodulation function's ready state. In the specific technical implementation, the communication frequency points in the metadata are injected into the frequency control unit of the software-defined radio, and the demodulation parameters are mapped to the computing kernel of the hardware acceleration unit, thereby establishing the initialization sequence of the software modem container. This formation process not only locks the protocol algorithm stack within the current time slice, but also achieves zero-latency reconfiguration of ground station hardware resources across different satellite missions through a software-defined approach. The formed baseline environment configuration achieves deep alignment between the container's processing capabilities and satellite signal characteristics, at least alleviating the technical shortcomings of rigid demodulation logic and difficulty in upgrading under the traditional hardware ground station architecture.

[0066] In summary, this application demonstrates that the identification of orbital parameters provides physical-level incentives for the entire scheduling chain; the generated discrete time schedule and parameter sequence quantify these physical incentives into digital data. Subsequently, continuous streaming output utilizes configuration metadata to ensure the reliable deployment of the scheduling strategy across heterogeneous computing nodes in the cloud. Finally, the resulting baseline initialization environment configuration verifies the adaptive performance of the entire system. This application compensates for the intermittent nature of satellite overhead transits through the flexibility of software-defined radio, achieving efficient pooling and low-cost evolution of ground station infrastructure.

[0067] Optionally, the Internet Protocol header in the digital radio frequency stream is parsed to identify the stream source identification information associated with the target over-the-head satellite; using the container network interface, a logical redirection process is performed on the container network interface according to the stream source identification information to direct the digital radio frequency stream to the software modem container instantiated in response to the stream source identification information.

[0068] Preferably, the elastic computing scheduler uses its built-in message monitoring interface to perform real-time traffic characteristic detection on the bit streams flowing into the cloud region, thereby identifying the distribution pattern of digital radio frequency (RF) streams at the network layer. In the specific technical implementation of this identification, the message payload carrying the RF stream is extracted in real time, and the source address field, destination address field, and protocol type identifier in the Internet Protocol (IP) header are captured and sensed. This establishes physical routing awareness of the inbound RF payload within the cloud virtualization environment, thus obtaining the original signaling payload for subsequent essential mapping from network packets to logical satellite streams. By identifying these microscopic network protocol characteristics, the initial trigger source of the traffic distribution process is established, ensuring that subsequent processing actions are anchored as accurately as possible to valid satellite service sessions.

[0069] Preferably, the elastic computing scheduler acquires the identified Internet Protocol header and drives its internally configured deep packet inspection engine to perform metadata extraction processing on the extended fields, thereby generating the flow source identification information discretely distributed along the time axis. During the generation process, differential operations are performed on the virtual switching interface identifier encapsulated in the header, and unique samples associated with the current overriding satellite are identified. These samples serve as discrete scales describing the physical source of the satellite signal, recording the attribution attribute of the radio frequency code stream within each discrete time slice. Since the traffic generated by satellite overpasses has time-varying pulse characteristics, the generated flow source identification information achieves a digital discrete representation of the traffic source in a multi-satellite concurrent environment, providing a definite digital index support for subsequent fine-grained path redirection in the virtualized network.

[0070] Preferably, the elastic computing scheduler aggregates the generated flow source identification information and drives the internally configured routing mapping operator to perform logical addressing processing for the container pool topology. In this step, a pre-stored satellite identity-container instance mapping table is retrieved, the virtual medium access control address of the target container is determined based on the current identification sample, and a logical redirection instruction representing the traffic destination is generated accordingly. This achieves the assignment of forwarding semantics that can be recognized by software-defined networks to discrete flow source characteristics. The generated redirection instruction at least alleviates the traffic aliasing problem caused by multiple satellites sharing fiber optic backhaul links, supporting the backend forwarding plane to perform real-time payload separation based on determined digital credentials.

[0071] Preferably, the container network interface obtains the generated logical redirection processing instruction, drives its internal virtual switching matrix to perform routing conversion processing for the digital radio frequency stream, and controls the system to perform continuous stream output accordingly. During the execution of continuous stream output, the digital radio frequency stream is directionally transmitted to the instantiated software modem container using a high-speed backplane channel according to a preset message forwarding rhythm. This action achieves a substantial physical migration of the radio frequency payload from the general access plane to the dedicated processing plane. This transmission output processing ensures that the cloud demodulation instance can obtain real-time, continuous waveform samples, providing stable data flow support for the dynamic instantiation of signal demodulation functions.

[0072] Preferably, the software modem container receives the redirected payload during the aforementioned stream output process and uses its built-in memory mapping operator to perform reassembly and buffering processing on the incoming packets, thereby ultimately forming a reference redirected service stream payload representing the demodulation-ready state. In the specific technical implementation of this formation, the received digital samples are stripped of their network layer encapsulation and logically sorted in the video memory space according to their sequence numbers to reconstruct a digital radio frequency sequence with definite spatiotemporal characteristics. This formation process not only locks the computational input source within the current service slice but also achieves the independence of computing resources between different satellite missions through logical isolation. The formed reference service stream payload achieves deep integration between physical radio frequency data and the demodulation algorithm kernel, at least alleviating the technical deficiency of the inability to route multiple signals in parallel under the traditional hardware ground station architecture.

[0073] Preferably, the identification, generation, output, and formation processes, which are collaboratively executed by the elastic computing scheduler, container network interface, and software modem container, have deep technical roles and functional support relationships. The header identification action provides network-level incentives for the entire distribution chain; the generated discrete identification information quantifies abstract connection attributes into digital routing basis. Then, by executing continuous stream output, virtual switching technology ensures reliable delivery of massive amounts of radio frequency data between heterogeneous computing nodes. Finally, the formed reference redirection service stream payload reverses the conversion loop from Internet Protocol to demodulation logic. This application compensates for the shortcomings of fixed-wire analog hardware through the routing flexibility of cloud computing, achieving efficient distribution and on-demand carrying of satellite ground station infrastructure.

[0074] Optionally, the decay trend of the real-time signal-to-noise ratio is monitored, and physical characteristics of the real-time signal-to-noise ratio being lower than the low-power shutdown threshold are identified. In response to the physical characteristics, a resource reclamation command is issued through the cloud container orchestration interface module to release the associated computing resource scale from the software modem container and return it to the cloud container pool.

[0075] Preferably, the elastic computing scheduler utilizes its configured channel quality analysis interface to perform trend monitoring on the digital radio frequency stream accessed from the cloud data inlet, thereby sensing the signal energy attenuation pattern. In the specific technical implementation of this sensing, the useful signal power spectral density in the radio frequency payload is extracted in real time, and the rate of decrease in the real-time signal-to-noise ratio within the continuous sampling window is sensed. This establishes the cloud virtualization system's awareness of the physical boundary of the impending termination of the satellite transit mission, thus obtaining the original physical characteristic payload for the subsequent essential switch from the mission-carrying state to the resource recovery state. By sensing these micro-indicators reflecting link quality degradation, the initial trigger source of the automated shutdown process is established, ensuring that subsequent processing actions can respond as accurately as possible to the signal drop process caused by the satellite leaving the antenna's field of view.

[0076] Preferably, the elastic computing scheduler acquires the sensed attenuation index and drives its internally configured logic discrimination engine to perform quantitative location processing for the low-power shutdown threshold, thereby generating a decision logic sequence discretely distributed along the time axis. During the generation process, the physical depth of the real-time signal-to-noise ratio falling into the invalid communication interval in each discrete time slice is compared to identify physical characteristics below the shutdown threshold, and discrete sample values ​​reflecting the shutdown probability are generated. Since the signal quality changes of the satellite during orbital operation have time-domain discreteness, these generated decision logic sequences map the macroscopic energy dissipation process to the microscopic system shutdown expectation. This processing step constitutes the basic data template for dynamic resource optimization, laying a digital decision-making foundation for the subsequent, as accurate as possible release of computing resources in a distributed environment.

[0077] Preferably, the elastic computing scheduler aggregates the generated decision logic sequence and drives its built-in orchestration instruction generation operator to perform syntax mapping processing for the cloud-native object model, thereby ultimately forming the resource reclamation instruction representing the intention to deregister resources. In the specific technical implementation, the identified software modem container identifier to be released, the core quota to be returned, and the mounted virtual storage space handle are logically aggregated and encapsulated into a structured code block conforming to container cluster management specifications. This assigns an administrative semantic that can be executed by the automated management system to the discrete physical decision results. The generated resource reclamation instruction at least alleviates the static computing power lock-in problem caused by the intermittency of satellite transits, supporting the backend container platform to perform immediate resource occupancy release based on a defined digital certificate.

[0078] Preferably, the elastic computing scheduler drives the internally configured high-speed signaling transmission interface to perform real-time push to the execution plane for the resource reclamation instructions formed above, and controls the system to execute continuous streaming output accordingly. During the specific process of executing continuous streaming output, the cloud container orchestration interface module utilizes the encapsulated resource deregistration declaration to be injected into the control kernel of the computing cluster in real-time as an instruction stream. This action achieves a substantial physical migration of resource scheduling strategies from the decision plane to the physical carrying plane, at least alleviating the problem of wasted computing resources caused by intervention lag. This transmission output processing ensures that the container orchestration engine can capture quota reclamation intentions from the scheduling side in real time, providing stable dynamic input stream support for aligning demodulation capabilities with the satellite dynamic window.

[0079] Preferably, the cloud container orchestration interface module receives the instruction payload during the aforementioned stream output process, drives the corresponding cloud container pool to perform a deep recycling operation, and ultimately forms benchmark idle computing power resource data representing the dynamic computing power balance of the ground station. In the specific technical implementation of this formation, according to the mapping entries in the instruction set, a lifecycle deregistration operation for the software modem container is executed in the general server cluster, and the corresponding computing resource scale is redirected from the active running queue to the globally available resource pool. This formation process not only unlocks the hardware quota of the current task slice but also achieves rapid flow of computing power resources through a cloud-native approach. The formed benchmark idle computing power resource data achieves deep tracking of hardware supply and business load, at least alleviating the technical deficiency of traditional hardware ground station architectures where computing power cannot be time-sharing multiplexed.

[0080] In summary, this application provides a physical-level stimulus signal for the entire recycling chain by sensing the signal-to-noise ratio trend; the generated discrete decision sequence quantifies the physical stimulus into digital criteria. Next, by executing continuous streaming output, a standardized cloud container orchestration interface module ensures the reliable delivery of the recycling strategy across heterogeneous computing nodes. Finally, the resulting baseline idle computing resource data inversely closes the dynamic pooling loop of computing resources. This application compensates for the spatial non-stationary characteristics of satellite signals through the extreme elasticity of cloud-native computing, achieving high computing power reuse efficiency and low operating costs for the satellite ground station system.

[0081] Optionally, based on the extracted satellite identifier, a target protocol image version corresponding to the satellite identifier is matched from the protocol image library; the configuration metadata corresponding to the target protocol image version is injected into the software modem container to drive the software modem container to dynamically instantiate and upgrade in response to the configuration metadata to perform demodulation logic.

[0082] Preferably, the elastic computing scheduler utilizes its built-in feature extraction operator to parse the asymmetric signaling fields in the digital radio frequency stream transmitted back via optical fiber in real time, thereby identifying the identity attributes of the target satellite. In the specific technical implementation of identification, the ephemeris index encapsulated in the message header is decoded, and the physical carrier source corresponding to the digital radio frequency stream is identified, thus obtaining the extracted satellite identifier. This achieves the most accurate possible determination of the space service type by the cloud virtualization system, obtaining the original task source identifier for the subsequent essential mapping from hardware resource reuse to multi-protocol adaptation. Identifying this core identifier establishes the initial incentive source for the entire protocol automated upgrade process, ensuring that subsequent processing actions are anchored as accurately as possible to the correct target satellite communication system.

[0083] Preferably, the elastic computing scheduler acquires the identified and extracted satellite identifiers, drives its internally configured protocol topology library addressing logic to perform association matching for demodulation, thereby generating a protocol version mapping sequence discretely distributed over time. During the generation process, the protocol image library index loaded in memory is retrieved, and the binary executable logic corresponding to the physical attributes of the satellite identifier is searched within discrete time slices, ultimately locking the corresponding target protocol image version. Since ground stations need to carry different tasks at different times, this application maps macroscopic satellite task switching to microscopic container software version iterations, establishing a digital logical benchmark for the flexible definition of the protocol stack, providing a definite protocol template for the subsequent construction of the demodulation environment.

[0084] Preferably, the elastic computing scheduler aggregates the locked target protocol image versions and their associated carrier characteristics, driving the internally configured structured description operators to perform syntax transformation processing on the container runtime parameters, thereby ultimately forming the configuration metadata with standardized descriptive semantics. In the specific technical implementation, complex physical layer error correction coding rules, modulation order, and sampling bit width are encapsulated into configuration code blocks conforming to microservice scheduling specifications. To this end, dynamic runtime instructions with time-domain characteristics are assigned to static protocol images. The generated configuration metadata at least alleviates the incompatibility problem of processing logic caused by differences in different constellation protocols, supporting backend container instances to perform real-time environment adaptation based on determined digital credentials, thereby achieving logical decoupling between business logic and underlying parameters.

[0085] Preferably, the elastic computing scheduler drives its built-in high-speed signaling delivery pipeline to perform real-time push to the execution plane for the generated configuration metadata, and controls the system to perform continuous streaming output processing accordingly. During the specific process of executing continuous streaming output, the configuration metadata, including the mirror pull-up position and initialization parameter payload, is delivered to the control plane of the software modem container in real time according to a preset task switching step size. This action realizes the substantial physical migration of protocol upgrade instructions from the decision center to the computing nodes, at least alleviating the problem of long service interruption times caused by maintaining ground station mirrors. This real-time delivery output processing ensures that the cloud computing unit can capture the latest protocol evolution intentions in real time, providing stable dynamic parameter stream support for the rapid instantiation of demodulation capabilities within the cloud container pool.

[0086] Preferably, in response to the configuration metadata received during the aforementioned stream output process, the software modem container drives its internal dynamic library loading operator to perform logical reconstruction processing for the demodulation kernel, thereby ultimately forming a baseline instantiation upgrade result representing the ready state of the demodulation function. In the specific technical implementation, according to the instructions in the metadata, the signal processing algorithm stream corresponding to the target protocol image version is instantaneously loaded in the container runtime environment, and the dynamic instantiation upgrade action of the demodulation logic is executed. This formation process not only locks the decoding pipeline within the current mission slice but also achieves a physical leap from hardwired modules to microservice images for ground station functionality through software-defined radio. The formed baseline instantiation upgrade result achieves deep alignment of demodulation capabilities with real-time satellite standards, at least alleviating the technical deficiency of traditional ground stations in supporting real-time protocol expansion.

[0087] In summary, in this application, the identification action provides the task trigger source for the entire upgrade chain; the generated protocol sequence and the configuration metadata quantify the satellite standard into digital container management instructions. Next, by executing continuous stream output actions, a standardized communication interface ensures the reliable distribution of image parameters between distributed computing nodes. Finally, the resulting benchmark instantiation upgrade result supports the accurate reconstruction of the bitstream. This application compensates for the rigidity of the hardware ground station architecture through the flexibility of software-defined radio, achieving a high level of computing power reuse efficiency for the satellite ground system.

[0088] This application also provides a cloud-native virtualized ground station operating method based on fully software-defined radio, which includes: The edge digitization front end receives radio frequency signals transmitted by overhead satellites and performs radio frequency digitization sampling processing to generate a digital radio frequency stream; The elastic computing scheduler determines the real-time signal-to-noise ratio in the digital radio frequency stream and performs correlation calculations on the mapping relationship between decoding algorithm complexity and hardware resources based on the real-time signal-to-noise ratio to determine the target decoding mode. At the same time, it performs quota calculations on the scale of computing resources in conjunction with the satellite transit schedule, and dynamically launches software modulation and demodulation containers that match the target decoding mode in the cloud container pool accordingly. The software modem container responds to the target decoding mode by generating a bitstream consisting of the restored user service data based on the digital radio frequency stream.

[0089] Optionally, the edge digitization front-end is internally configured with an analog-to-digital conversion module and a timing interface module. The step of performing radio frequency digital sampling processing to generate a digital radio frequency stream specifically includes: The analog-to-digital conversion module is used to sample the radio frequency signal to generate an original in-phase quadrature sample sequence, and the timing interface module is used to inject a timestamp of a set precision into the original in-phase quadrature sample sequence. According to the service transmission protocol specification, the original in-phase orthogonal sample sequence after being injected with the set precision timestamp is segmented and encapsulated to generate the digital radio frequency stream carried by the Internet Protocol and containing the set precision timestamp information.

[0090] Optionally, the step of determining the target decoding mode by performing the correlation calculation between the decoding algorithm complexity and the hardware resource mapping relationship based on the real-time signal-to-noise ratio specifically includes: The real-time signal-to-noise ratio is compared with a preset performance threshold to determine the correlation logic between the number of decoding iterations and the weight of computing resources, and a decoding strategy control code is generated based on the correlation logic. Based on the decoding strategy control code, one of the preset low-power decoding mode and enhanced decoding mode is selected and established as the target decoding mode.

[0091] Optionally, the software modulation and demodulation container internally operates a hardware acceleration unit, and the step of generating a bitstream composed of the restored user service data based on the digital radio frequency stream specifically includes: The error correction code block features in the digital radio frequency stream are identified, and in response to the target decoding mode determined by the elastic computing scheduler, the hardware acceleration unit is invoked to perform parallel decoding processing on the error correction code block features to generate the restored bit stream.

[0092] Optionally, the method further includes: The elastic computing scheduler acquires satellite orbit parameters by monitoring them and updates the satellite transit schedule in real time based on the satellite orbit parameters in order to extract the communication frequency points and demodulation parameters corresponding to the visible satellites. The elastic computing scheduler encapsulates the communication frequency and the demodulation parameters into container startup metadata, and injects the container startup metadata into the initialization sequence of the software modem container to drive the software modem container to perform environment configuration according to the initialization sequence.

[0093] Optionally, the elastic computing scheduler is internally configured with a container network interface, and the method further includes the elastic computing scheduler performing traffic distribution processing for the digital radio frequency stream, with the following specific steps: The Internet Protocol header in the digital radio frequency stream is parsed to identify the stream source identification information associated with the target overhead satellite; Using the container network interface, a logical redirection process is performed on the container network interface based on the stream source identification information to direct the digital radio frequency stream to the software modem container instantiated in response to the stream source identification information.

[0094] Optionally, the elastic computing scheduler is preset with a low-power shutdown threshold and configured with a cloud container orchestration interface module. The method further includes the elastic computing scheduler performing dynamic release processing for computing resources, with the specific steps as follows: Monitor the attenuation trend of the real-time signal-to-noise ratio and identify the physical characteristics of the real-time signal-to-noise ratio being lower than the low-power shutdown threshold; In response to the physical characteristics, a resource reclamation command is issued through the cloud container orchestration interface module to release the associated computing resource scale from the software modem container and return it to the cloud container pool.

[0095] Optionally, the method further includes the elastic computing scheduler dynamically upgrading the ground station communication protocol by accessing a protocol mirror library, with the specific steps as follows: Based on the extracted satellite identifier, a target protocol image version corresponding to the satellite identifier is matched from the protocol image library; The configuration metadata corresponding to the target protocol image version is injected into the software modem container to drive the software modem container to dynamically instantiate and upgrade in response to the configuration metadata to perform demodulation logic.

[0096] This application also provides a computer storage medium storing computer-executable instructions thereon, which are executed to implement the described working method.

[0097] The above description is merely an exemplary embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A cloud-native virtualized ground station system based on fully software-defined radio, characterized in that, include: Edge digital front-end, elastic computing scheduler, and software modem container; The edge digital front end is used to receive radio frequency signals transmitted by overpass satellites and perform radio frequency digital sampling processing to generate a digital radio frequency stream; the elastic computing scheduler is used to determine the real-time signal-to-noise ratio in the digital radio frequency stream, and perform correlation calculation of the decoding algorithm complexity and hardware resource mapping relationship based on the real-time signal-to-noise ratio to determine the target decoding mode, and at the same time perform quota calculation of computing resource scale in combination with the satellite overpass schedule, and dynamically launch software modulation and demodulation containers that match the target decoding mode in the cloud container pool accordingly; The software modulation and demodulation container is used to generate a bit stream consisting of restored user service data based on the digital radio frequency stream in response to the target decoding mode.

2. The cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The edge digital front-end is internally configured with an analog-to-digital conversion module and a timing interface module. The process by which the edge digital front-end performs radio frequency digital sampling processing to generate a digital radio frequency stream is as follows: the analog-to-digital conversion module is used to sample the radio frequency signal to generate an original in-phase quadrature sample sequence, and the timing interface module is used to inject a timestamp of a set precision into the original in-phase quadrature sample sequence. According to the service transmission protocol specification, the original in-phase orthogonal sample sequence after being injected with the set precision timestamp is segmented and encapsulated to generate the digital radio frequency stream carried by the Internet Protocol and containing the set precision timestamp information.

3. The cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, When the elastic computing scheduler performs the correlation calculation between the decoding algorithm complexity and the hardware resource mapping relationship to determine the target decoding mode, it performs the following operations: it distinguishes the real-time signal-to-noise ratio with the preset performance threshold execution value to establish the correlation logic between the number of decoding iterations and the computing power resource weight, and generates decoding strategy control code based on the correlation logic. Based on the decoding strategy control code, one of the preset low-power decoding mode and enhanced decoding mode is selected and established as the target decoding mode.

4. A cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The software modulation and demodulation container runs a hardware acceleration unit. When the software modulation and demodulation container generates a bit stream based on the digital radio frequency stream, it performs the following operations: identifies the error correction code block features in the digital radio frequency stream, and in response to the target decoding mode determined by the elastic computing scheduler, calls the hardware acceleration unit to perform parallel decoding processing on the error correction code block features to generate the restored bit stream.

5. A cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The elastic computing scheduler is also used to perform the following operations: by monitoring and acquiring satellite orbit parameters, and updating the satellite transit schedule in real time according to the satellite orbit parameters, to extract the communication frequency points and demodulation parameters corresponding to the visible satellites; encapsulating the communication frequency points and demodulation parameters into container startup metadata, and injecting the container startup metadata into the initialization sequence of the software modem container, so as to drive the software modem container to perform environment configuration according to the initialization sequence.

6. A cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The elastic computing scheduler is internally configured with a container network interface. When performing traffic distribution processing for the digital radio frequency stream, it performs the following operations: parsing the Internet Protocol header in the digital radio frequency stream to identify the stream source identification information associated with the target over-the-head satellite; and using the container network interface, performing logical redirection processing for the container network interface based on the stream source identification information to direct the digital radio frequency stream to the software modem container instantiated in response to the stream source identification information.

7. A cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The elastic computing scheduler is preset with a low-power shutdown threshold and configured with a cloud container orchestration interface module to perform dynamic release processing for computing resources. Specifically, it includes the following operations: monitoring the attenuation trend of the real-time signal-to-noise ratio and identifying physical characteristics where the real-time signal-to-noise ratio is lower than the low-power shutdown threshold; in response to the physical characteristics, issuing a resource reclamation command through the cloud container orchestration interface module to release the associated computing resource scale from the software modem container and return it to the cloud container pool.

8. A cloud-native virtualized ground station system based on fully software-defined radio according to claim 1, characterized in that, The elastic computing scheduler is also used to dynamically upgrade the ground station communication protocol by accessing the protocol image library. Specifically, it includes the following operations: matching the target protocol image version corresponding to the satellite identifier from the protocol image library according to the extracted satellite identifier; injecting the configuration metadata corresponding to the target protocol image version into the software modem container to drive the software modem container to dynamically instantiate and upgrade the demodulation logic in response to the configuration metadata.

9. A cloud-native virtualized ground station operating method based on fully software-defined radio, characterized in that, include: The edge digitization front end receives radio frequency signals transmitted by overhead satellites and performs radio frequency digitization sampling processing to generate a digital radio frequency stream; The elastic computing scheduler determines the real-time signal-to-noise ratio in the digital radio frequency stream and performs correlation calculations on the mapping relationship between decoding algorithm complexity and hardware resources based on the real-time signal-to-noise ratio to determine the target decoding mode. At the same time, it performs quota calculations on the scale of computing resources in conjunction with the satellite transit schedule, and dynamically launches software modulation and demodulation containers that match the target decoding mode in the cloud container pool accordingly. The software modem container responds to the target decoding mode by generating a bitstream consisting of the restored user service data based on the digital radio frequency stream.

10. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The edge digitization front-end is internally configured with an analog-to-digital conversion module and a timing interface module. The step of performing radio frequency digital sampling processing to generate a digital radio frequency stream specifically includes: The analog-to-digital conversion module is used to sample the radio frequency signal to generate an original in-phase quadrature sample sequence, and the timing interface module is used to inject a timestamp of a set precision into the original in-phase quadrature sample sequence. According to the service transmission protocol specification, the original in-phase orthogonal sample sequence after being injected with the set precision timestamp is segmented and encapsulated to generate the digital radio frequency stream carried by the Internet Protocol and containing the set precision timestamp information.

11. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The step of determining the target decoding mode by performing a correlation calculation between the decoding algorithm complexity and hardware resource mapping based on the real-time signal-to-noise ratio specifically includes: The real-time signal-to-noise ratio is compared with a preset performance threshold to determine the correlation logic between the number of decoding iterations and the weight of computing resources, and a decoding strategy control code is generated based on the correlation logic. Based on the decoding strategy control code, one of the preset low-power decoding mode and enhanced decoding mode is selected and established as the target decoding mode.

12. The cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The software modulation and demodulation container internally operates a hardware acceleration unit. The step of generating a bitstream composed of the restored user service data based on the digital radio frequency stream specifically includes: The error correction code block features in the digital radio frequency stream are identified, and in response to the target decoding mode determined by the elastic computing scheduler, the hardware acceleration unit is invoked to perform parallel decoding processing on the error correction code block features to generate the restored bit stream.

13. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The method further includes: The elastic computing scheduler acquires satellite orbit parameters by monitoring them and updates the satellite transit schedule in real time based on the satellite orbit parameters in order to extract the communication frequency points and demodulation parameters corresponding to the visible satellites. The elastic computing scheduler encapsulates the communication frequency and the demodulation parameters into container startup metadata, and injects the container startup metadata into the initialization sequence of the software modem container to drive the software modem container to perform environment configuration according to the initialization sequence.

14. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The elastic computing scheduler is internally configured with a container network interface. The method further includes the elastic computing scheduler performing traffic distribution processing for the digital radio frequency stream, with the following specific steps: The Internet Protocol header in the digital radio frequency stream is parsed to identify the stream source identification information associated with the target overhead satellite; Using the container network interface, a logical redirection process is performed on the container network interface based on the stream source identification information to direct the digital radio frequency stream to the software modem container instantiated in response to the stream source identification information.

15. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The elastic computing scheduler is preset with a low-power shutdown threshold and configured with a cloud container orchestration interface module. The method also includes the elastic computing scheduler performing dynamic release processing for computing resources, with the specific steps as follows: Monitor the attenuation trend of the real-time signal-to-noise ratio and identify the physical characteristics of the real-time signal-to-noise ratio being lower than the low-power shutdown threshold; In response to the physical characteristics, a resource reclamation command is issued through the cloud container orchestration interface module to release the associated computing resource scale from the software modem container and return it to the cloud container pool.

16. A cloud-native virtualized ground station operating method based on fully software-defined radio according to claim 9, characterized in that, The method also includes the elastic computing scheduler dynamically upgrading the ground station communication protocol by accessing the protocol mirror library, with the specific steps as follows: Based on the extracted satellite identifier, a target protocol image version corresponding to the satellite identifier is matched from the protocol image library; The configuration metadata corresponding to the target protocol image version is injected into the software modem container to drive the software modem container to dynamically instantiate and upgrade in response to the configuration metadata to perform demodulation logic.

17. A computer storage medium, characterized in that, It stores computer-executable instructions that are executed to implement the working method as described in claims 9-16.

Images