Power satellite communication-oriented dynamic multi-network switching method and related device

Abstract

The invention belongs to the technical field of communication, and discloses a dynamic multi-network switching method for power satellite communication and a related device. The dynamic multi-network switching method comprises the following steps: predicting and acquiring a cellular link performance prediction result in a future preset time period based on an acquired network state joint feature vector; splicing the network state joint feature vector, the power service scene label and the cellular link performance prediction result to form a network fusion feature vector, taking the network fusion feature vector as input, and performing action prediction by using the trained double-depth Q network to obtain a prediction action set; based on the predicted action set, illegal action elimination and overcharge action shielding are carried out, effective actions are obtained, corresponding configuration files are activated, and network connection is carried out. According to the technical scheme disclosed by the invention, flexible switching between the ground network and the satellite network can be realized, and meanwhile, the switching packet loss rate can be reduced.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and specifically relates to a dynamic multi-network switching method and related apparatus for power satellite communication. Background Technology

[0002] Traditional communication methods rely on terrestrial base stations, which suffer from insufficient terrestrial network coverage in remote areas, oceans, and emergency disaster relief environments. Currently, satellite-based Internet of Things (IoT) is rapidly filling these terrestrial network blind spots, but its development remains constrained by insufficient terminal intelligence and bottlenecks in cross-network integration. With the growth of IoT and global communication demands, the convergence of terrestrial and satellite networks has become a crucial direction for communication technology development.

[0003] In the field of power applications, current IoT cards generally use physical SIM (Subscriber Identity Module) cards, which cannot dynamically switch operators. This results in equipment in remote areas (such as power transmission line monitoring devices) requiring manual card replacement to maintain connectivity, with fault repair cycles lasting up to several weeks. In addition, satellite terminals are pre-installed with a single operator's SIM card, making it impossible to select the optimal network as needed, and the cost is high. Furthermore, while general eSIM (Embedded Subscriber Identity Module) supports remote configuration, it suffers from three major drawbacks: poor adaptability to long latency in high-orbit satellites (profile download failure rate >30%), seamless service switching (decision based solely on signal strength, resulting in a 12% data loss rate for power relay protection), and weak security architecture (lack of device-level identity binding). Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic multi-network handover method and related apparatus for power satellite communication, thereby solving one or more of the aforementioned technical problems. The technical solution disclosed in this invention enables flexible handover between terrestrial and satellite networks while reducing handover packet loss rate.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a dynamic multi-network switching method for power satellite communication, comprising the following steps: Obtain the joint feature vector of network state; wherein, the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; Based on the network state joint feature vector, the cellular link performance prediction results for a future preset time period are obtained; The network state joint feature vector, the power business scenario label and the cellular link performance prediction result are concatenated to form a network fusion feature vector, which is then used as input. A trained dual-deep Q network is used to predict actions to obtain a set of predicted actions. Based on the predicted action set, illegal actions are eliminated and excessive actions are masked to obtain valid actions; Based on the effective action, the corresponding configuration file is activated and a network connection is established.

[0006] A further improvement of the technical solution of the present invention is that, in the step of predicting the cellular link performance prediction result for a future preset time period based on the network state joint feature vector, a trained long short-term memory network model is used for prediction.

[0007] A further improvement to the technical solution of the present invention is that the training steps of the dual deep Q-network include: Obtain a training sample set; wherein each sample in the training sample set includes: historical network fusion feature vector sample data and sample labels in the power communication scenario; Construct a dual-network structure including an evaluation network and a target network, and initialize the parameters; Based on the training sample set, a reward function is adopted, combined with hard constraint masks and tariff filtering. The dual network structure is optimized through experience replay and gradient descent. The reward function uses user experience quality (QoE) as a positive incentive and link tariff and violation penalty as negative constraints. During forward inference, a mask is dynamically generated based on the label to eliminate illegal actions. Before the reward is calculated, over-fee actions are blocked based on tariff rules. The target network parameters are periodically synchronized until convergence, resulting in a well-trained dual-depth Q-network.

[0008] A further improvement of the technical solution of the present invention is that the action space of the dual-depth Q network includes: maintaining the cell, cell bit rate reduction, satellite switching, and satellite bit rate reduction.

[0009] A further improvement to the technical solution of the present invention lies in the step of activating the corresponding configuration file and establishing a network connection based on the aforementioned effective action. Based on the aforementioned effective action, the corresponding configuration file is activated using the eSIM chip, and a network connection is established. The eSIM chip contains a dedicated power satellite configuration file and a terrestrial network configuration file. Each configuration file includes a certificate for verifying the operator's identity and secure communication between the device and the network, network access information, and an encryption key for encrypted communication and user authentication. The dedicated power satellite configuration file also has a blank key slot for subsequent key negotiation and updates.

[0010] A further improvement of the technical solution of the present invention is that the parameter template of the power-specific satellite configuration file includes the following fields: APN configuration, heartbeat interval, data fragment size, compression algorithm, differential update flag, retransmission strategy, frequency band priority, emergency reporting strategy, and satellite resource reservation.

[0011] A second aspect of the present invention provides a dynamic multi-network switching system for power satellite communication, comprising: The feature acquisition module is used to acquire a joint feature vector of network state; wherein, the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; The performance prediction module is used to predict the cellular link performance for a future preset time period based on the network state joint feature vector. The action prediction module is used to concatenate the network state joint feature vector, the power service scenario label and the cellular link performance prediction result to form a network fusion feature vector and use it as input. It then uses a trained dual-deep Q network to perform action prediction and obtain a predicted action set. The action constraint module is used to remove illegal actions and shield excessive actions based on the predicted action set to obtain valid actions; The configuration connection module is used to activate the corresponding configuration file and establish a network connection based on the effective action.

[0012] A further improvement of the technical solution of the present invention is that, in the performance prediction module, in the step of predicting the cellular link performance prediction result for a future preset time period based on the network state joint feature vector, a trained long short-term memory network model is used for prediction.

[0013] A further improvement to the technical solution of the present invention is that, in the action prediction module, the training steps of the dual-depth Q-network include: Obtain a training sample set; wherein each sample in the training sample set includes: historical network fusion feature vector sample data and sample labels in the power communication scenario; Construct a dual-network structure including an evaluation network and a target network, and initialize the parameters; Based on the training sample set, a reward function is adopted, combined with hard constraint masks and tariff filtering. The dual network structure is optimized through experience replay and gradient descent. The reward function uses user experience quality (QoE) as a positive incentive and link tariff and violation penalty as negative constraints. During forward inference, a mask is dynamically generated based on the label to eliminate illegal actions. Before the reward is calculated, over-fee actions are blocked based on tariff rules. The target network parameters are periodically synchronized until convergence, resulting in a well-trained dual-depth Q-network.

[0014] A further improvement of the technical solution of the present invention is that the action space of the dual-depth Q network includes: maintaining the cell, cell bit rate reduction, satellite switching, and satellite bit rate reduction.

[0015] A further improvement of the technical solution of the present invention lies in that, in the configuration connection module, during the step of activating the corresponding configuration file and establishing a network connection based on the effective action, Based on the aforementioned effective action, the corresponding configuration file is activated using the eSIM chip, and a network connection is established. The eSIM chip contains a dedicated power satellite configuration file and a terrestrial network configuration file. Each configuration file includes a certificate for verifying the operator's identity and secure communication between the device and the network, network access information, and an encryption key for encrypted communication and user authentication. The dedicated power satellite configuration file also has a blank key slot for subsequent key negotiation and updates.

[0016] A further improvement of the technical solution of the present invention is that the parameter template of the power-specific satellite configuration file includes the following fields: APN configuration, heartbeat interval, data fragment size, compression algorithm, differential update flag, retransmission strategy, frequency band priority, emergency reporting strategy, and satellite resource reservation.

[0017] In a third aspect, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a dynamic multi-network switching method for power satellite communication as described in any one of the first aspects of the present invention.

[0018] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the dynamic multi-network switching method for power satellite communication as described in any one of the first aspects of the present invention.

[0019] In a fifth aspect, the present invention provides a computer program product comprising computer instructions which, when executed by a processor, implement the steps of the dynamic multi-network switching method for power satellite communication as described in any one of the first aspects of the present invention.

[0020] Compared with the prior art, the present invention has the following beneficial effects: The dynamic multi-network switching method disclosed in this invention can intelligently switch between satellite and terrestrial networks based on real-time network status and service requirements, optimizing communication efficiency and cost. Specifically, addressing the technical challenge of remote areas where devices cannot dynamically switch operators due to physical SIM cards and the long manual SIM card replacement cycle, this invention obtains a joint feature vector of network status including cellular, satellite, environmental, and tariff characteristics, predicts cellular link performance, concatenates these features into a network fusion feature vector, and uses dual-depth Q-network prediction to dynamically select a suitable network, eliminating the need for manual SIM card replacement. Addressing the technical challenge of satellite terminals being unable to select the optimal network due to a pre-installed single SIM card and high tariffs, this invention comprehensively evaluates the network using multiple features, selecting the optimal network as needed, balancing performance and cost. Based on real-time network status and power service weight labels, it automatically selects the optimal network and pre-switch connections, reducing handover packet loss rate.

[0021] In the preferred embodiment of this invention, the eSIM chip can manage multiple profiles, supporting the storage and dynamic loading of power-specific satellite profiles and terrestrial network profiles. It also features over-the-air (OTA) writing functionality, enabling remote operator switching and tariff policy matching. Addressing the technical problems of general-purpose eSIMs, such as poor adaptability to long latency with high-orbit satellites, data loss due to seamless service switching, and weak security architecture, this invention optimizes the profile download strategy based on satellite network characteristics, improving the success rate. Furthermore, it introduces power service scenario tags to ensure reliable transmission of critical service data. Through these improvements, this invention achieves flexible switching between terrestrial and satellite networks. During switching, it can select the optimal path based on actual conditions, reducing data loss and lowering the packet loss rate, thus providing reliable protection for power communication.

[0022] In a preferred embodiment of the present invention, a profile parameter template structure for a dedicated power satellite is disclosed. This design is optimized for satellite communication characteristics, power data features, and terminal power consumption. Specifically, data fragmentation optimization and a dedicated compression algorithm reduce traffic, improve efficiency, and lower costs; ultra-low power heartbeat intervals extend terminal battery life; an emergency alarm data reporting mechanism is configured, and reserved bandwidth ensures priority transmission of fault data; frequency band priority switching ensures stable communication; differential update and retransmission strategies improve communication reliability; and dynamic loading of multi-carrier profiles and flexible carrier switching via eSIM technology reduce management costs. These innovative methods significantly improve the performance of power communication systems and have broad application prospects and commercial value. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a dynamic multi-network switching method for power satellite communication in an embodiment of the present invention. Figure 2 This is a schematic diagram of the hardware architecture of a dual-mode communication terminal in a specific embodiment of the present invention; Figure 3 This is a schematic diagram of the eSIM multi-profile isolation management architecture in an embodiment of the present invention; Figure 4 This is a schematic diagram of the activation process of the power-dedicated satellite profile in an embodiment of the present invention; Figure 5 This is a schematic diagram of the feature extraction and strategy decision-making process in an embodiment of the present invention; Figure 6 This is a schematic diagram of a dynamic multi-network switching system for power satellite communication in an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention; obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0026] Based on the technical solutions disclosed in the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this invention. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0027] The terminology of this invention is explained as follows: 3GPP (3rd Generation Partnership Project) is an international communications standards development organization responsible for developing and maintaining global communications standards.

[0028] 5G NTN: 5G Non-Terrestrial Network, refers to the combination of 5G technology and satellite communication to extend the coverage of 5G networks.

[0029] APN: Access Point Name, which identifies the gateway through which a terminal accesses the network.

[0030] EID: Electronic Identity Document, a unique identifier for the eSIM chip used to identify the eSIM hardware.

[0031] eSIM: Embedded SIM is a SIM card integrated into a device that supports remote configuration and management without the need for physical replacement.

[0032] eSIM COS: eSIM Card Operating System, responsible for managing the operation and configuration of the eSIM chip.

[0033] IoT: Internet of Things (IoT) refers to connecting various devices via the Internet to enable data interaction and intelligent control between devices.

[0034] IMEI: International Mobile Equipment Identity.

[0035] NTN: Non-Terrestrial Network, refers to a network that uses non-terrestrial communication methods such as satellites to extend the coverage of terrestrial networks.

[0036] OTA: Over-the-Air.

[0037] Profile: A configuration file containing network access parameters from the carrier, security authentication information, etc., used for devices to connect to a specific network.

[0038] High-orbit satellites (GEO): Geostationary Earth Orbit satellites operate in geostationary orbit and have a relatively long communication delay.

[0039] Please see Figure 1 The present invention provides a dynamic multi-network handover method for power satellite communication, comprising the following steps: Step 1: Obtain the joint feature vector of network state; wherein the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; Step 2: Based on the network state joint feature vector, predict the cellular link performance prediction results for a future preset time period; Step 3: The network state joint feature vector, the power service scenario label and the cellular link performance prediction result are concatenated to form a network fusion feature vector and used as input. The trained dual deep Q network is used to predict actions to obtain a predicted action set. Step 4: Based on the predicted action set, perform illegal action elimination and excessive action masking to obtain valid actions; Step 5: Based on the effective action, activate the corresponding configuration file and establish a network connection.

[0040] In the technical solution disclosed in the embodiments of the present invention, effective actions are obtained through predictive analysis, dual-depth Q network action prediction and multiple screening, and the configuration file is selected to connect to the network accordingly. This solves the existing technical problems such as IoT card replacement and satellite terminal network selection, realizes flexible switching between ground and satellite networks, and reduces the switching packet loss rate. Specifically, in the technical solution of this invention, multiple network status data such as cellular and satellite are first acquired to predict future network status. A trained dual-deep Q network is then used in conjunction with service scenario labels to predict actions. Valid actions are formed through hard constraints of "service and scenario" and tariff filtering. Based on these valid actions, a configuration file is selected to connect to the network. This solution breaks the limitation that physical SIM cards cannot dynamically switch operators, enabling devices to automatically select suitable networks based on network status and service needs without manual SIM card replacement. This solves the problem of long-term fault repair caused by the inability to dynamically switch operators in remote areas. Furthermore, this invention also enables satellite terminals to dynamically select the optimal network based on real-time network status and service needs, rather than being limited to a pre-set single operator network. Simultaneously, during the formation of valid actions, tariff filtering filters out excessively expensive actions, helping to reduce tariff costs and solving the problems of inability to select networks on demand and high tariffs caused by pre-set single operator SIM cards in satellite terminals.

[0041] In the specific exemplary technical solution of the present invention, in the step of activating the corresponding configuration file and establishing a network connection based on effective action, both the power dedicated satellite profile and the terrestrial network profile are pre-written into the eSIM chip. Each configuration file includes key information such as a certificate for verifying the operator's identity and secure communication between the device and the network, network access information, and encryption keys for encrypted communication and user authentication.

[0042] In this technical solution of the present invention, the currently active profile can be dynamically switched via commands from the power terminal remote management platform, thereby connecting to different operator networks. This mechanism not only improves the flexibility of the device but also enhances the user's adaptability in different network environments. The terrestrial network profile can be used for regular data transmission, while the power-dedicated satellite profile can be automatically activated when there is no terrestrial network, ensuring reliable backhaul of power monitoring data. The intelligent network switching mechanism can flexibly adapt to the complex network environment and communication needs under different power business scenarios. In a specific exemplary optional technical solution, the satellite-ground collaborative intelligent switching mechanism can dynamically select the optimal communication path based on network signal strength, service priority, and energy consumption optimization strategies; for example, low-power cellular networks are used preferentially in regular scenarios, while satellite links are forcibly switched in emergency scenarios. In addition, the eSIM multi-profile design supports remote configuration updates to adapt to the power business needs of different regions. This solution further solves the coverage blind spots and reliability bottlenecks of traditional terminals in complex environments and can be widely used in fields such as smart grids, new energy base station monitoring, and disaster emergency response, significantly improving the flexibility and disaster resistance of power communication systems.

[0043] Please see Figure 2 and Figure 3 In this invention, an innovative dual-mode communication terminal is designed to address technical challenges in power satellite communication systems, such as coverage blind spots, unstable links, and handover delays. This terminal integrates an eSIM chip module, a cellular module, and a satellite module. Specifically, the eSIM chip module supports multi-carrier profile storage and dynamic loading, covering both dedicated power satellite profiles and terrestrial network profiles. The cellular module supports 4G / 5G protocol stacks and employs low-power technology, enabling real-time acquisition of status data from power transmission and transformation equipment. The satellite module carries a high-orbit satellite communication chip, automatically activating when no terrestrial network is available, ensuring reliable transmission of power monitoring data. Furthermore, the satellite-ground collaborative communication system features an intelligent network switching mechanism, flexibly adapting to complex network environments and communication needs in various power business scenarios.

[0044] Technically, the eSIM chip in an eSIM chip module is an embedded integrated circuit used to store and manage user identity and configuration information for mobile communication networks. Compared to traditional removable SIM cards, the eSIM chip hardware is integrated onto the communication device's circuit board, offering higher security, flexibility, and tamper resistance. The eSIM chip's multi-carrier profile storage mechanism provides mobile communication devices with high flexibility and security. The core function of the eSIM chip is to authenticate and connect the device to the mobile network by storing user identity information and encryption keys. The eSIM chip uses non-volatile memory to store user identity information and network profiles, ensuring that this information is not lost even after the device is powered off. The eSIM chip has a built-in security module that uses multiple encryption algorithms to encrypt the stored information, ensuring the security of user identity information. The eSIM chip supports remote downloading and updating of profiles via OTA technology, allowing users to switch carriers without replacing the physical SIM card. The eSIM chip supports storing profiles from multiple carriers, enabling seamless switching between multiple carrier networks. Each profile contains critical information, including certificates for verifying carrier identity and secure communication between the device and the network, network access information, and encryption keys for encrypted communication and user authentication. The eSIM chip allocates an independent storage area for each operator's configuration file. Multi-operator configuration files are stored in separate storage areas, ensuring the isolation of information from different operators. Each configuration file is protected by an encryption algorithm and has access permissions set, allowing only authorized devices and operators to access and update it. In this embodiment of the invention, the currently active configuration file is dynamically switched via commands from the power terminal remote management platform, thereby connecting to different operator networks. This mechanism not only improves the flexibility of the device but also enhances the user's adaptability to different network environments.

[0045] Technically, the dedicated power satellite profile is a key technology for achieving seamless switching between satellite and terrestrial networks. Considering the high latency and intermittent connectivity of satellite communication, the service characteristics of power satellite communication data, and the power consumption of power communication terminals, this invention's embodiment designs parameters for the dedicated power satellite profile. Specifically, the dedicated power satellite profile parameters are a communication configuration template customized for the service characteristics of the power industry. It includes key elements such as a dedicated APN, data fragmentation, ultra-low power heartbeat interval, a power-specific compression algorithm, and a fault-priority transmission strategy. Its main functions include: reducing the data upload traffic of power terminals and improving communication efficiency by adjusting data fragment size, using dedicated compression algorithms, and differential update technology; improving communication reliability through retransmission strategies and frequency band priority switching; configuring an emergency alarm data reporting strategy to ensure priority transmission of fault data; and reserving satellite resource bandwidth for direct transmission of critical services to ensure efficient power grid emergency response. The dedicated power satellite profile template design can meet the special needs of the power industry in complex communication environments and ensure the stable operation of the power communication system. Exemplary results of the dedicated power satellite profile template design are shown in Table 1. Power satellite communication terminals achieve seamless switching between terrestrial and satellite networks through a hardware-integrated eSIM chip module. The eSIM chip has the ability to store multiple operator configuration profiles, including profiles from both terrestrial operators and satellite internet service providers.

[0046] Table 1. Profile Template for Power Utility Satellites

[0047] During the eSIM chip manufacturing phase, the terrestrial network profile is first written into the chip to ensure normal communication capabilities of the terminal in conventional terrestrial network environments. Simultaneously, a dedicated power satellite profile, encrypted using national cryptographic algorithms, is pre-installed in the eSIM chip's independent secure storage area. This encryption method provides strong security for the dedicated power satellite profile, preventing information from being tampered with or stolen during storage and transmission. The pre-installed dedicated power satellite profile includes the operator's CA certificate, a crucial credential used to verify the operator's identity and ensure secure communication between the device and the satellite network. Furthermore, this profile has a blank key slot, providing space for subsequent key negotiation and updates to adapt to different security requirements and communication scenarios. The pre-installed dedicated power satellite profile is initially inactive and requires an activation process to be enabled, further enhancing the system's security and controllability.

[0048] Please see Figure 4 The dedicated power satellite profile is bound to the hardware fingerprint of the communication terminal. The hardware fingerprint includes the eSIM chip identifier (EID), the communication terminal's IMEI number, and the device serial number, ensuring that each profile can only be used on a specific hardware terminal. This binding mechanism enhances the security and tamper-proof capabilities of the profile through the uniqueness of hardware characteristics. When the terminal first connects to the network via the terrestrial network, it sends an activation request to the satellite gateway, which includes the communication terminal's hardware fingerprint. The hardware fingerprint, as the terminal's unique identifier, ensures the legitimacy and security of the activation request. After receiving the activation request, the satellite gateway first verifies the terminal's EID and hardware fingerprint in a whitelist database. The whitelist database, managed by the satellite operator, contains the EID and hardware fingerprint information of all authorized terminals. Whitelist verification effectively prevents unauthorized terminals from accessing the network. If the terminal passes verification, the gateway issues a differential activation command to the terminal. The differential activation command is only 1KB in size and includes a session key and a tariff policy. The session key is used to encrypt communication between the terminal and the satellite network, ensuring data transmission security; the tariff policy defines the terminal's usage rights and billing rules within the satellite network, ensuring the reasonable allocation and use of network resources. After receiving the differential activation command, the terminal activates the power satellite profile using the session key. During activation, the terminal writes the session key into the blank key slot of the profile, completing key negotiation. Once activation is complete, the power satellite profile's status changes from inactive to active, and the terminal can then use the satellite network service normally.

[0049] Based on eSIM multi-profile management, it meets communication needs in different scenarios through flexible profile configuration and switching. Utilizing eSIM embedded technology, cellular (4G / 5G) and satellite (BeiDou / high-orbit satellite) modes can be seamlessly switched via commands. When satellite is activated, the power supply to the cellular module is automatically cut off, and network registration can be completed without restarting the device.

[0050] The dedicated power satellite profile is optimized for long latency (GEO satellite RTT > 500ms), employing pre-acknowledgment protocols and data fragmentation technology to improve communication efficiency and reliability. It also supports over-the-air (OTA) SIM card writing, enabling remote switching of operators from cellular network operators to satellite operators, thus achieving communication coverage in areas without cellular coverage. A tariff strategy matching function allows the system to dynamically load low-cost profiles based on time and traffic demands, such as using off-peak data packages at night to reduce communication costs. The dual-mode communication module integrates satellite and cellular communication modules to adapt to different communication environments. In emergency rescue scenarios, when there is no terrestrial signal, the system automatically switches to high-orbit satellites to transmit compressed casualty location and vital sign data. After network recovery, it switches to the public network via eSIM to upload high-resolution images in batches, ensuring communication continuity and data integrity under extreme conditions.

[0051] Further explanation of the template in Table 1 above is as follows: In power satellite communication systems, optimizing APN configuration and heartbeat intervals are key measures to improve communication efficiency and terminal endurance. Dedicated APNs achieve network isolation, preventing non-power devices from occupying satellite resources and ensuring the privacy and security of power grid data. Dedicated APN configuration not only guarantees the independence of power communication but also effectively avoids public network interference, improving communication stability and reliability. While the heartbeat interval for general satellite communication profiles is typically set to 300 seconds, for power satellite communication terminals requiring ultra-low power mode, the heartbeat interval can be extended to 600 seconds to enhance terminal endurance. During periods without service data transmission, the system automatically extends the heartbeat interval to 1800 seconds, significantly reducing terminal power consumption. The technical solution of this invention, by setting a low-power mode, reduces the terminal's energy consumption in standby mode, thereby extending the endurance of battery-powered devices, making it more suitable for monitoring power transmission lines in remote areas or where frequent maintenance is difficult.

[0052] In power satellite communication, to ensure optimized, efficient, and reliable data transmission, a profile data transmission optimization parameter was specifically designed, which is then finely controlled through a traffic management module. This module integrates data compression technology, effectively reducing the amount of data transmitted and lowering communication costs; it employs a differential update mechanism, transmitting only the changed portions of the data to further improve transmission efficiency; and it implements a priority scheduling strategy to ensure that critical data is transmitted first, guaranteeing the real-time performance and reliability of the power system. This comprehensively enhances the data processing capabilities and transmission performance of power satellite communication, meeting the high data transmission requirements of the power industry.

[0053] Traditional general-purpose profiles typically use 1KB fragment sizes; however, this setting does not fully meet the characteristics of power service data packets. To optimize data transmission efficiency, the data fragment size needs to be adjusted, especially for power telemetry data, whose data packets are usually small (e.g., critical information such as current and voltage values ​​are typically less than 200 bytes). Therefore, adjusting the data fragment size to 512 bytes more accurately adapts to the actual needs of power telemetry data. This adjustment not only avoids the padding waste caused by 1KB fragments but also significantly reduces fragmentation in satellite links, thereby improving overall transmission efficiency. Compared to the common 1KB fragment size, this optimization scheme can reduce redundant data by approximately 50%, ensuring more efficient and economical transmission of power data over satellite links, further enhancing the performance and reliability of power satellite communication systems.

[0054] To improve data compression efficiency in power satellite communication, an optimized dictionary and a customized compression algorithm were designed for floating-point power data. Meter data compression employs a customized LZMA-Power algorithm, which optimizes the dictionary based on 1000 sets of measured substation data. This algorithm focuses on the characteristics of floating-point data such as current and voltage, significantly improving the compression ratio compared to general algorithms. For non-critical data, compression algorithms are used to effectively reduce data transmission volume, thereby reducing bandwidth usage, improving communication efficiency, and simultaneously reducing the burden on the satellite link.

[0055] In power satellite communication, differential update flag technology is optimized for power acquisition scenarios. The differential update mechanism only transmits changed data; for example, a full transmission is performed only once a month, with only the data changes transmitted at other times, significantly reducing bandwidth usage. Furthermore, to address the long latency characteristics of satellite communication, pre-caching and data prefetching techniques are employed, prioritizing the transmission of small data packets. Cache relay technology is used, allowing edge nodes to temporarily store data and upload it in batches after network recovery, avoiding frequent satellite link starts and stops. High-latency optimization measures include pre-caching instructions, embedding binary templates of commonly used instructions (such as location requests and status reports) in the profile, allowing terminals to directly fill in parameters and reduce interaction rounds. The differential heartbeat mechanism only transmits changed status bits; for example, only 1 byte of data is sent when the battery level drops from 78% to 77%.

[0056] In power satellite communication systems, retransmission strategies are crucial mechanisms for ensuring data integrity and communication reliability. When data packets are lost, the system initiates a retransmission mechanism to quickly recover data lost due to temporary interruptions. This mechanism effectively addresses common issues in satellite communication such as signal interference and link instability. To further enhance system robustness, a delayed retransmission strategy is designed. This avoids resource waste and inefficiency caused by frequent retransmissions during periods of persistent interference. Explained, after detecting data packet loss, the communication system first attempts a rapid retransmission to restore data transmission as quickly as possible. If multiple retransmissions fail consecutively, the system determines that persistent interference may exist and automatically switches to a delayed retransmission mode, waiting for the interference to subside before attempting retransmission. This flexible retransmission strategy in this embodiment of the invention not only improves the success rate of data transmission but also optimizes the utilization efficiency of the satellite link, ensuring the stable operation of the power communication system.

[0057] In power satellite communication systems, frequency band priority configuration is a crucial strategy for optimizing communication performance. The L-band (1-2 GHz) offers significant advantages in power communication scenarios due to its excellent resistance to 50 Hz power frequency interference. When the electromagnetic noise around a substation exceeds 25 dBm, the system automatically switches to the L-band to ensure the stability of the communication link and the reliability of data transmission. This automatic switching mechanism effectively addresses communication challenges in complex electromagnetic environments such as substations, reducing the impact of interference on communication quality. By prioritizing the L-band, power satellite communication systems can maintain efficient communication capabilities in high-noise environments, providing reliable communication support for critical business operations in the power industry. Ensuring efficient and continuous data transmission in high-noise environments enhances the adaptability of power satellite communication systems in complex electromagnetic environments.

[0058] In power satellite communication systems, the emergency reporting strategy prioritizes the transmission of fault data to ensure rapid identification and response to power grid faults. When a fault is detected, the system immediately marks the fault data as the highest priority and transmits it rapidly to the monitoring center via the satellite link. To address temporary interruptions and persistent interference in the communication link, the emergency reporting strategy combines a rapid recovery mechanism with a delayed retransmission mechanism: rapid restoration of transmission during temporary interruptions and delayed retransmission during persistent interference to avoid resource waste.

[0059] The system is also equipped with an emergency alarm data reporting mechanism and reserves satellite bandwidth for direct transmission of critical services, further ensuring the efficient transmission of fault data. These measures collectively ensure that the power grid satellite communication system can efficiently and reliably transmit fault data in complex electromagnetic environments, enhancing the power grid's emergency response capabilities. This strategy not only improves the stability of the communication link but also provides dedicated bandwidth for critical services by reserving satellite resources.

[0060] Please see Figure 5 Existing monitoring terminals mostly rely on single cellular or satellite links, making it impossible to adjust the network in real time according to service scenarios, signal quality, and cost. In extreme conditions such as remote mountainous areas and natural disasters, link interruptions or cost overruns are prone to occur, and a unified decision-making framework is lacking. A cellular-satellite fusion terminal for power transmission line monitoring scenarios achieves millisecond-level dynamic network switching and cost optimization through an 82-dimensional FSV (Fused State Vector) fusion state vector. Specific exemplary technical solutions include: real-time monitoring of signal strength, latency, bandwidth, and cost policies for feature extraction; intelligent selection of the optimal network connection based on service models and machine learning algorithms; prioritizing satellite to ensure critical data transmission in emergency situations, while switching to the NB-IoT network in low-traffic scenarios, thereby optimizing communication efficiency and cost.

[0061] In the specific exemplary technical solution of this invention, a total of 47-dimensional joint feature vectors are constructed, including four major categories: cellular network features, satellite network features, environmental features, and tariff features. Real-time data acquisition and feature fusion of heterogeneous networks are achieved by integrating dual-channel radio frequency probes for cellular and satellite networks on the terminal side.

[0062] Specifically, in the cellular mode, the terminal performs a full-beam scan of the primary serving cell and the synchronization signal blocks of the two strongest neighboring cells at a period of 100ms, and extracts 15-dimensional cellular network features in real time: high-precision synchronization signal block scanning is used to collect multiple index data of the cellular network such as RSRP, RSSI, and SINR at 100ms intervals, the three-dimensional distance between the base station and the site is output using the Haversine formula, the PRB occupancy rate is decoded from MIB and SIB1 as the base station load, the multipath RMS delay spread is estimated by the second moment of the power delay spectrum, and the Doppler spread is estimated by the time domain autocorrelation slope. All features are linearly normalized to form a 15-dimensional feature vector of the cellular network.

[0063] Specifically, in satellite mode, the terminal locks onto the top three satellites within the visible range every 100ms and sequentially calculates seven satellite link characteristics. The specific method is as follows: the elevation angle is calculated in real time using three-dimensional coordinates; the Doppler frequency shift is obtained using carrier phase difference; the satellite-to-ground round-trip time is obtained by correcting for the speed of light and superimposing the device delay; based on the real-time rainfall rate, the Ka-band rain attenuation index is estimated using the ITU-RP.838 model; the link load is statistically analyzed by parsing the DVB-S2 frame header; the ephemeris age is given by the ephemeris timestamp difference; and the link margin is calculated by the carrier-to-noise ratio. All the characteristics of the three satellites are normalized to form a 15-dimensional feature vector of the high-orbit satellite network.

[0064] Furthermore, environmental features and cost-based features are introduced and jointly normalized with network performance features to construct a 47-dimensional joint feature vector (as shown in Table 2) that includes signal quality, network topology, environmental interference, and dynamic tariffs. This vector, along with network performance features, is input into the hierarchical decision model to provide millisecond-level decision support for intelligent network handover.

[0065] Table 2. 47-dimensional joint eigenvectors

[0066] Based on the above, the network handover strategy modeling of this invention adopts a three-level hierarchical decision architecture, specifically as follows: The first layer is a random forest scene classifier: based on the 47-dimensional JFV (Joint Feature Vector) environment subset, it can distinguish scenes such as mountainous areas, plains, cities, deserts, and islands in real time. When it receives five types of business instructions, such as line inspection, fault repair, load monitoring, emergency command, or routine meter reading, it immediately sets the corresponding one-hot bit to 1 and the rest to 0, and dynamically splices them to the 48th-52nd dimensions of the fused state vector.

[0067] The second layer, the performance predictor, takes a 47-dimensional JFV with a 30s sliding window as input and outputs a three-dimensional prediction sequence of RSRP, SINR and RTT for the cellular link in the next 10s through a dual-layer LSTM. The third layer, the strategy optimizer, sends the 82-dimensional FSV (47-dimensional JFV in real time, 5-dimensional service labels and 30-dimensional prediction results, as shown in Table 3) into the scene-aware dual-depth Q network at once. Under the dual constraints of QoE and tariff limit, it selects the action of "maintaining cellular, reducing cellular bit rate, switching satellite or reducing satellite bit rate" in real time, and immediately bypasses to the forced satellite lock mode when the emergency command arrives.

[0068] Table 3. 82-dimensional Feature State Vector (FSV)

[0069] The terminal completes 82-dimensional FSV vector assembly and policy reasoning once per second, achieving millisecond-level dynamic network switching.

[0070] In a further exemplary optional technical solution of the present invention, the dynamic network handover strategy decision engine is further explained as follows: Regarding the triggering conditions: The terminal performs policy evaluation every 1 second: each time, it first reads the service tag and scenario tag to look up the table to obtain the default policy, and immediately re-evaluates when events such as service changes, emergency commands, link interruptions or tariff overruns occur; if it receives a "fault repair" or "emergency command" command, it directly sets the star lock flag and bypasses the dual-deep Q network.

[0071] Regarding decision input: Under normal circumstances, the terminal will stitch together the real-time 47-dimensional JFV, 5-dimensional service label and 30-dimensional LSTM prediction into an 82-dimensional FSV, and input it into a dual-depth Q network with scenario-service dual-layer constraints.

[0072] Regarding the policy network: Service and scenario labels constitute a hard constraint mask. The reward function uses QoE as the primary positive incentive and link cost and violation penalties as negative constraints, and introduces configurable weight coefficients to achieve flexible optimization driven by services. Only when a candidate action falls into the valid set allowed by the hard constraint mask jointly defined by service and scenario labels can the action enter the reward calculation; otherwise, it is assigned a negative infinite reward to forcibly eliminate it. Any action that violates the policy table is assigned a negative infinite reward, thereby ensuring that a legal and optimal network switching decision is output within milliseconds.

[0073] Regarding the motion space: there are 4 types of motions, as shown in Table 4.

[0074] Table 4. Motion Space

[0075] Regarding the hard constraint table: The hard constraint table built into the terminal directly maps "business tags and scenario tags" to default actions and emergency actions: Once fault repair or emergency command is triggered, regardless of the scenario, action 2 (satellite lock) is immediately forced; Line inspection in mountainous areas defaults to action 2 (satellite lock), and satellite lock is maintained after cellular disconnection, while in plains or cities the default action is 0, and satellite is switched after cellular disconnection.

[0076] Regarding real-time tariff filtering: Real-time tariff filtering takes effect before the policy is executed; the terminal first calculates the satellite link unit price multiplied by the emergency discount coefficient. If the result is higher than the user-set limit, the "switching satellite" and "satellite bitrate reduction" actions are immediately blocked. Similarly, if the cellular unit price exceeds the limit, the "keeping cellular" and "cellular bitrate reduction" actions are blocked. If both cellular and satellite networks exceed the limit, the bitrate reduction action is selected first. If the bitrate cannot be lowered below the limit after reduction, the terminal immediately triggers a local audible and visual alarm and reports the cost over-limit event to ensure uninterrupted service and controllable costs.

[0077] In a specific exemplary technical solution of this invention, the closed-loop process includes: First, the terminal collects 47-dimensional JFV data with a 1-second closed-loop cycle, parses the scene and service tags, drives LSTM to output cellular link predictions for the next 10 seconds, splices them into an 82-dimensional FSV, and sends it to the policy network to generate the original action probability π(a|s). Then, the system uses "service × scene" hard constraints to eliminate illegal actions, and then uses real-time tariff filtering to block overcharge options, forming the final set of valid actions. Finally, the terminal immediately executes the action, updates the radio frequency channel, link status, and tariff count, and re-enters the loop in the next 1 second or when triggered by events such as service changes, emergency commands, link interruptions, or tariff overruns, achieving millisecond-level dynamic network switching. The seamless switching technology is based on the 3GPP NTN standard, pre-establishes satellite + 4G / 5G dual links, ensures switching latency of less than 50ms, and achieves seamless communication connectivity. The dynamic network switching strategy of the power satellite communication system aims to intelligently switch between the satellite network and the terrestrial network according to the real-time network status and service requirements to ensure the efficiency and economy of communication.

[0078] The principle behind the advanced nature of the technical solution in this invention is explained below: Traditional multi-mode terminals, which combine physical SIM cards, offer users diverse communication options by integrating multiple communication modules (including cellular, Wi-Fi, and satellite communication modules) within the terminal. However, this approach relies on physical SIM cards rather than eSIM technology, meaning operator switching can only be done manually or semi-automatically, typically requiring users to pre-configure priority rules. This solution was common in early vehicle-mounted emergency communication devices and field monitoring terminals. Its advantage lies in eliminating the need for an eSIM chip, thus reducing hardware costs, and its mature technology ensures good compatibility with existing equipment. However, this solution also has significant limitations, such as poor flexibility, inability to dynamically switch operators, complex management requiring manual replacement of the physical SIM card, and relatively weak satellite communication support, often only supporting a single satellite standard. Therefore, this solution is more suitable for applications with limited budgets and relatively fixed network environments, such as fixed weather stations, where the demand for communication flexibility and multi-operator support is lower. In addition, pure satellite communication terminal solutions that rely on a single satellite network, such as Tiantong, Iridium, or Starlink, have built-in satellite modems to achieve communication functions. These terminals are widely used in scenarios such as ocean-going vessels, polar scientific research, and military communications because they have significant advantages, such as full coverage without reliance on terrestrial networks and strong disaster resistance, thus ensuring high reliability. However, pure satellite communication terminals also have some disadvantages, including high satellite data charges, large latency, and bulky terminal equipment, which usually has high power consumption and large size. 5G NTN (Non-Terrestrial Network) technology connects 5G base stations directly to satellites, enabling terminals to communicate using standard 5G modules. Typical applications include satellite-connected phones as defined by 3GPP R17+, such as the satellite communication function of the Huawei Mate 60 Pro. 5G NTN provides users with a seamless network switching experience, supports high-bandwidth applications like video transmission, and, as a core direction for future 6G development, has high standardization potential. However, in the early stages of deployment, global coverage is not yet complete, and terminal devices need to be equipped with dedicated radio frequency chips to support satellite communication functions, increasing hardware costs and complexity. Furthermore, the pricing structure for 5G NTN is yet to be determined, but it is expected to be higher than traditional cellular networks.

[0079] Unlike traditional solutions, the technical solution of this invention addresses the limitations of traditional single-profile devices switching between terrestrial and satellite networks, as well as the need for manual SIM card replacement for power equipment in areas without signal. It proposes a dynamic multi-profile loading mechanism with "terrestrial network priority and satellite emergency backup." This mechanism introduces power-specific profile parameter templates, meeting the power industry's security isolation requirements while optimizing for low-power scenarios. Furthermore, this invention includes profile management methods, achieving seamless switching across different network environments through intelligent traffic optimization strategies and switching processes, thus improving the flexibility, security, and economy of power communication systems. This invention not only achieves "service-aware" network switching in the power sector, reducing overall communication costs, but can also be extended to industries such as oil and gas and water conservancy, forming a standardized multi-network converged communication architecture.

[0080] Existing general-purpose eSIM profiles are not optimized for the characteristics of power data, resulting in high data consumption, high failure rates in profile downloads under high-latency satellite communication environments, and a lack of industry-specific transmission protocols. Traditional network switching solutions rely on manual configuration or simple signal strength switching, which cannot adapt to the priority requirements of power services. In view of the above problems, the new solution in this invention utilizes an eSIM chip to store multi-operator profiles (including satellite / terrestrial networks), dynamically switching based on signal strength, pricing strategies, and service priorities to improve network coverage. Hardware fingerprint binding of profiles prevents unauthorized terminal access; a fragmented transmission protocol optimized for high-orbit satellites improves profile download success rates. Satellite-terrestrial collaborative switching algorithms and power-specific compression algorithms reduce device power consumption and pricing costs, achieving optimization in reliability, energy efficiency, and response speed. At transmission line monitoring points without terrestrial base stations, eSIM-loaded satellite profiles enable data backhaul, solving the coverage problem of traditional SIM cards and significantly improving maintenance efficiency. In scenarios such as earthquakes and floods, eSIM-supported satellite-terrestrial converged terminals can automatically switch to the satellite network, quickly restoring emergency communications and ensuring command and control for power facility repairs. Power companies can leverage the multi-profile feature of eSIM to automatically select the optimal operator based on pricing strategies, thereby reducing communication costs.

[0081] Specifically, in the field of power satellite applications, traditional technical solutions have exposed serious problems: First, energy efficiency is deteriorating; the standard heartbeat mechanism (60 seconds / time) reduces the battery life of power transmission and transformation monitoring terminals to 1.8 years, far below the industry threshold of 5 years. Second, data traffic costs are out of control; due to the lack of compression and encryption mechanisms, the average monthly satellite data traffic fee exceeds 800,000 yuan. Third, service assurance is lacking; the switching strategy is not included in the priority of power services, resulting in insufficient reliability of critical control command transmission. These pain points restrict the in-depth application of satellite communication in high-value scenarios such as power emergency repair and wide-area equipment monitoring. This invention aims to solve these problems by using eSIM technology to achieve multi-operator configuration file storage, power-specific satellite profile management, and dynamic multi-network switching, providing a more efficient and flexible communication solution to meet communication needs in special environments.

[0082] This invention achieves seamless switching between terrestrial networks (4G / 5G / NB-IoT) and satellite networks (such as Tiantong and Beidou) through eSIM technology, ensuring stable communication for power monitoring terminals even without terrestrial signals. In extreme environments such as remote areas in the northwest, power grid equipment can automatically switch to satellite links, increasing the communication success rate from 60% in traditional solutions to nearly 100%, with a switching latency of less than 200ms, significantly improving the real-time monitoring capabilities of the power system. Coverage is comprehensive, with no communication dead zones; reliability is ensured through multi-link redundancy and cache relay, guaranteeing 99.9% availability. Traditional power equipment requires manual SIM card replacement or on-site network configuration, resulting in high maintenance costs and slow response times. This invention supports remote OTA management, allowing maintenance personnel to update the communication configuration of thousands of devices in batches via satellite or terrestrial networks. Furthermore, eSIM supports multi-carrier profile storage, automatically selecting the optimal connection based on tariff policies or network quality, reducing communication costs. Cost-effectiveness: Dynamic pricing strategies reduce operating costs by 30%-60%; Flexibility: eSIM remote switching supports multi-carrier switching; Power equipment is often deployed in harsh environments such as high-altitude, high-temperature, and strong electromagnetic interference, where traditional SIM cards are prone to failure due to physical damage or signal attenuation. This invention's industrial-grade eSIM design can still operate stably in scenarios such as ultra-high-voltage substations and offshore wind power platforms, reducing the failure rate by 90% compared to traditional solutions. Simultaneously, the anti-interference communication protocol ensures reliable data transmission even in environments with strong electromagnetic interference, making it suitable for critical equipment such as smart meters and fault indicators.

[0083] The power industry has extremely high requirements for communication security, and traditional SIM cards are at risk of being tampered with or cloned. This invention employs pre-built profile management, customized EID, and hardware-level encryption to prevent unauthorized devices from accessing the power grid. Furthermore, the eSIM's over-the-air writing function complies with the GSMA SGP.32 standard and supports domestic cryptographic algorithms, ensuring the confidentiality and integrity of power data during satellite transmission. Compared to the high cost of dedicated satellite communication terminals, this solution reduces hardware modification costs by more than 50% by reusing eSIM chips from existing power equipment, supporting scenarios such as ultra-high voltage power transmission and distributed photovoltaic grid connection, and promoting space-ground converged communication as a new infrastructure standard for smart grids.

[0084] In summary, in the field of power satellite communication, the technical solution of the embodiments of the present invention achieves the following three improvements: Satellite module upgrade: Deploy a high-orbit converged communication system, adopting the Tiantong high-orbit architecture to achieve full-scenario coverage; Power-specific satellite profile design: Through customized data fragmentation and compression algorithms and ultra-low power heartbeat mechanisms, the efficiency of power data transmission is optimized, traffic consumption is reduced, terminal battery life is improved, and key alarm data is prioritized for transmission. It supports dynamic loading and differential updates of multi-operator profiles, filling the gap in the standardization of satellite profiles in the power industry. Dynamic multi-network handover control: Integrating eSIM multi-profile management and intelligent decision engine, it automatically selects the optimal network and pre-switches the connection based on real-time network status (signal strength, latency) and power service weight (such as protection device data priority), reducing handover packet loss rate.

[0085] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not disclosed in the apparatus embodiments, please refer to the embodiments of the method of the present invention.

[0086] Please see Figure 6 In this embodiment of the invention, a dynamic multi-network switching system for power satellite communication is provided, comprising: The feature acquisition module is used to acquire a joint feature vector of network state; wherein, the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; The performance prediction module is used to predict the cellular link performance for a future preset time period based on the network state joint feature vector. The action prediction module is used to concatenate the network state joint feature vector, the power service scenario label and the cellular link performance prediction result to form a network fusion feature vector and use it as input. It then uses a trained dual-deep Q network to perform action prediction and obtain a predicted action set. The action constraint module is used to remove illegal actions and shield excessive actions based on the predicted action set to obtain valid actions; The configuration connection module is used to activate the corresponding configuration file and establish a network connection based on the effective action.

[0087] In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used to execute operations of a dynamic multi-network switching method for power satellite communication.

[0088] In one embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the operating system of the terminal. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM (Random Access Memory) or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the dynamic multi-network handover method for power satellite communication in the above embodiments.

[0089] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0090] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A dynamic multi-network handover method for power satellite communication, characterized in that, Includes the following steps: Obtain the joint feature vector of network state; wherein, the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; Based on the network state joint feature vector, the cellular link performance prediction results for a future preset time period are obtained; The network state joint feature vector, the power business scenario label and the cellular link performance prediction result are concatenated to form a network fusion feature vector, which is then used as input. A trained dual-deep Q network is used to predict actions to obtain a set of predicted actions. Based on the predicted action set, illegal actions are eliminated and excessive actions are masked to obtain valid actions; Based on the effective action, the corresponding configuration file is activated and a network connection is established.

2. The dynamic multi-network switching method for power satellite communication according to claim 1, characterized in that, In the step of predicting the cellular link performance prediction results for a future preset time period based on the network state joint feature vector, a trained long short-term memory network model is used for prediction.

3. The dynamic multi-network switching method for power satellite communication according to claim 1, characterized in that, The training steps of the dual deep Q network include: Obtain a training sample set; wherein each sample in the training sample set includes: historical network fusion feature vector sample data and sample labels in the power communication scenario; Construct a dual-network structure including an evaluation network and a target network, and initialize the parameters; Based on the training sample set, a reward function is adopted, combined with hard constraint masks and tariff filtering. The dual network structure is optimized through experience replay and gradient descent. The reward function uses user experience quality (QoE) as a positive incentive and link tariff and violation penalty as negative constraints. During forward inference, a mask is dynamically generated based on the label to eliminate illegal actions. Before the reward is calculated, over-fee actions are blocked based on tariff rules. The target network parameters are periodically synchronized until convergence, resulting in a well-trained dual-depth Q-network.

4. A dynamic multi-network switching method for power satellite communication according to claim 3, characterized in that, The action space of the dual-depth Q network includes: maintaining cellular, cellular bitrate reduction, satellite switching, and satellite bitrate reduction.

5. A dynamic multi-network switching method for power satellite communication according to claim 1, characterized in that, Based on the aforementioned effective action, the steps of activating the corresponding configuration file and establishing a network connection are as follows: Based on the aforementioned effective action, the corresponding configuration file is activated using the eSIM chip, and a network connection is established. The eSIM chip contains a dedicated power satellite configuration file and a terrestrial network configuration file. Each configuration file includes a certificate for verifying the operator's identity and secure communication between the device and the network, network access information, and an encryption key for encrypted communication and user authentication. The dedicated power satellite configuration file also has a blank key slot for subsequent key negotiation and updates.

6. A dynamic multi-network switching method for power satellite communication according to claim 5, characterized in that, The parameter template of the power satellite configuration file includes the following fields: APN configuration, heartbeat interval, data fragment size, compression algorithm, differential update flag, retransmission strategy, frequency band priority, emergency reporting strategy, and satellite resource reservation.

7. A dynamic multi-network switching system for power satellite communication, characterized in that, include: The feature acquisition module is used to acquire a joint feature vector of network state; wherein, the joint feature vector of network state includes: cellular network features, satellite network features, environmental features, and tariff features; The performance prediction module is used to predict the cellular link performance for a future preset time period based on the network state joint feature vector. The action prediction module is used to concatenate the network state joint feature vector, the power service scenario label and the cellular link performance prediction result to form a network fusion feature vector and use it as input. It then uses a trained dual-deep Q network to perform action prediction and obtain a predicted action set. The action constraint module is used to remove illegal actions and shield excessive actions based on the predicted action set to obtain valid actions; The configuration connection module is used to activate the corresponding configuration file and establish a network connection based on the effective action.

8. A dynamic multi-network switching system for power satellite communication according to claim 7, characterized in that, In the performance prediction module, the step of predicting the cellular link performance prediction result for a future preset time period based on the network state joint feature vector is performed using a trained long short-term memory network model.

9. A dynamic multi-network switching system for power satellite communication according to claim 7, characterized in that, In the action prediction module, the training steps of the dual-depth Q-network include: Obtain a training sample set; wherein each sample in the training sample set includes: historical network fusion feature vector sample data and sample labels in the power communication scenario; Construct a dual-network structure including an evaluation network and a target network, and initialize the parameters; Based on the training sample set, a reward function is adopted, combined with hard constraint masks and tariff filtering. The dual network structure is optimized through experience replay and gradient descent. The reward function uses user experience quality (QoE) as a positive incentive and link tariff and violation penalty as negative constraints. During forward inference, a mask is dynamically generated based on the label to eliminate illegal actions. Before the reward is calculated, over-fee actions are blocked based on tariff rules. The target network parameters are periodically synchronized until convergence, resulting in a well-trained dual-depth Q-network.

10. A dynamic multi-network switching system for power satellite communication according to claim 9, characterized in that, The action space of the dual-depth Q network includes: maintaining cellular, cellular bitrate reduction, satellite switching, and satellite bitrate reduction.

11. A dynamic multi-network switching system for power satellite communication according to claim 7, characterized in that, In the configuration connection module, the step of activating the corresponding configuration file and establishing a network connection based on the effective action is as follows: Based on the aforementioned effective action, the corresponding configuration file is activated using the eSIM chip, and a network connection is established. The eSIM chip contains a dedicated power satellite configuration file and a terrestrial network configuration file. Each configuration file includes a certificate for verifying the operator's identity and secure communication between the device and the network, network access information, and an encryption key for encrypted communication and user authentication. The dedicated power satellite configuration file also has a blank key slot for subsequent key negotiation and updates.

12. A dynamic multi-network switching system for power satellite communication according to claim 11, characterized in that, The parameter template of the power satellite configuration file includes the following fields: APN configuration, heartbeat interval, data fragment size, compression algorithm, differential update flag, retransmission strategy, frequency band priority, emergency reporting strategy, and satellite resource reservation.

13. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the dynamic multi-network switching method for power satellite communication as described in any one of claims 1 to 6.

14. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the dynamic multi-network switching method for power satellite communication as described in any one of claims 1 to 6.

15. A computer program product, characterized in that, It includes computer instructions that, when executed by a processor, implement the steps of the dynamic multi-network switching method for power satellite communications as described in any one of claims 1 to 6.

Images