Broadband / narrowband adaptive switching power internet of things communication system and control method

Abstract

The invention relates to the technical field of power internet of things, in particular to a broadband / narrowband self-adaptive switching power internet of things communication system and a control method, which intelligently identify the priority and quality of service (QoS) requirements of different services by collecting network states, terminal states and service flow data in real time. And an optimal communication mode switching strategy is dynamically generated based on a pre-trained machine learning model, seamless switching of broadband and narrowband communication links is realized, and the decision accuracy is continuously improved through a feedback optimization mechanism. According to the method, the problems that a traditional electric power Internet of Things communication mode is fixed and single and cannot intelligently adapt to services and network states are effectively solved, the system communication reliability is remarkably improved, the transmission time delay and energy consumption are reduced, and high-adaptability and high-reliability intelligent communication guarantee in a complex electric power scene is achieved.

Description

Technical Field

[0001] This invention relates to the field of power Internet of Things (IoT) technology, and in particular to a power IoT communication system and control method with broadband / narrowband adaptive switching. Background Technology

[0002] With the rapid development of the power Internet of Things (IoT), various terminal devices are widely used in complex environments such as substations, distribution rooms, and transmission lines, placing higher demands on the reliability, real-time performance, and adaptability of communication systems. Currently, power IoT communication utilizes multiple communication methods, including 5G, 4G, LoRa, and Wi-Fi, each suitable for both broadband high-speed and narrowband low-power scenarios. However, existing systems often use a fixed communication method, unable to dynamically switch according to the field environment, business needs, or network status. This leads to decreased communication quality or even outages in scenarios with high interference, weak signals, or sudden high bandwidth demands, severely impacting the real-time performance and reliability of power monitoring and control.

[0003] To overcome the aforementioned problems, some existing technologies have attempted to introduce multi-mode communication modules, supporting terminals to switch between multiple networks manually or under simple rules. For example, some power IoT gateways support 5G and LoRa dual-mode operation, and can trigger switching based on signal strength thresholds. These methods improve communication flexibility to some extent and reduce the risks associated with single-network failures.

[0004] However, these existing technologies still have obvious shortcomings: First, the switching strategies are simple and crude, based only on signal strength or network availability, and fail to comprehensively consider multi-dimensional factors such as service type, power consumption requirements, and network load; second, they lack intelligent prediction and adaptive learning capabilities, and cannot predict changes in network status in advance, resulting in switching delays or frequent ping-pong switching; third, they are not deeply integrated with the power business system, and cannot dynamically adjust communication strategies according to service priorities (such as alarms, control, and monitoring), making it difficult to meet the stringent requirements of the power system for high reliability and low latency communication. Summary of the Invention

[0005] The purpose of this invention is to provide a power Internet of Things (IoT) communication system and control method with broadband / narrowband adaptive switching, which solves the problems of low communication reliability, large latency and high energy consumption caused by the fixed and single communication mode of the existing power IoT, which cannot dynamically and intelligently switch according to business needs and network status.

[0006] To achieve the above objectives, the present invention provides a power Internet of Things (IoT) communication control method with broadband / narrowband adaptive switching, comprising the following steps:

[0007] Real-time acquisition of multi-dimensional status data, including network status data, terminal status data, and service flow data;

[0008] Identify the priority and service quality requirements of the current business based on business flow data;

[0009] Multidimensional state data, current business priorities, and service quality requirements are input into a pre-trained intelligent decision-making model to generate communication mode switching strategies.

[0010] Perform seamless switching of communication links according to the communication mode switching strategy;

[0011] Record the communication performance data after this switchover and feed the performance data back to the intelligent decision-making model for model optimization.

[0012] This includes real-time acquisition of multi-dimensional status data, which includes network status data, terminal status data, and service flow data.

[0013] Network status data includes signal strength, network latency, packet loss rate, and network load for currently available communication modes;

[0014] Terminal status data includes remaining battery power and computing load;

[0015] Business flow data includes the business type and data volume of the data to be transmitted.

[0016] Among them, identifying the priority and service quality requirements of the current business based on business flow data, in the following aspects:

[0017] The service types include alarm control commands, video stream data, instrument reading data, and environmental sensor data.

[0018] The specific steps for identifying the priority and service quality requirements of the current business based on business flow data include:

[0019] Parse the protocol headers and payloads of the business flow data to identify its business type;

[0020] Based on the pre-defined business type-policy mapping table, query the default priority and service quality requirement parameters corresponding to the business type;

[0021] If the business type is instrument readings or environmental sensor data, further analysis is conducted to determine whether the data change rate or value exceeds a preset threshold, and the priority is dynamically increased accordingly.

[0022] In this process, multidimensional state data, the priority of the current business, and service quality requirements are input into a pre-trained intelligent decision-making model to generate a communication mode switching strategy.

[0023] The intelligent decision-making model is a machine learning model trained based on historical handover performance data, used to output the optimal communication mode selection decision;

[0024] Communication mode switching strategies include deciding whether to use broadband communication mode, narrowband communication mode, or maintain the current mode.

[0025] In this process, multidimensional state data, the priority of the current business, and service quality requirements are input into a pre-trained intelligent decision-making model to generate a communication mode switching strategy.

[0026] When an alarm control command with the highest priority for the current service is identified, broadband communication mode is selected first.

[0027] When network status data indicates that broadband communication mode is unavailable or the quality is below a preset threshold, and the current service can tolerate the delay, narrowband communication mode is selected.

[0028] By combining historical network status data to predict trends, the handover process can be triggered in advance before the network quality deteriorates to the threshold requiring handover.

[0029] Among them, seamless switching of communication links is performed according to the communication mode switching strategy.

[0030] For connection-oriented session services, maintain session state continuity during handover;

[0031] For non-session services, data is cached before switching and resent or resumed after successful switching.

[0032] A power Internet of Things communication system with broadband / narrowband adaptive switching includes a data acquisition module, a service identification module, an intelligent decision-making module, a switching execution module, and a feedback optimization module;

[0033] The data acquisition module is used to collect multi-dimensional status data in real time, including network status data, terminal status data, and service flow data.

[0034] The business identification module, connected to the data acquisition module, is used to identify the priority and service quality requirements of the current business based on business flow data.

[0035] The intelligent decision-making module, connected to the data acquisition module and the service identification module, is used to input multi-dimensional status data, the priority of the current service, and service quality requirements into the pre-trained intelligent decision-making model to generate communication mode switching strategies.

[0036] The switching execution module, connected to the intelligent decision-making module, is used to perform seamless switching of communication links based on the communication mode switching strategy.

[0037] The feedback optimization module, connected to the switching execution module and the intelligent decision-making module, is used to record the communication performance data after this switching and feed the performance data back to the intelligent decision-making model for model optimization.

[0038] This invention discloses a broadband / narrowband adaptive switching communication system and control method for the power Internet of Things (IoT). By real-time collection of network status, terminal status, and service flow data, it intelligently identifies the priority and Quality of Service (QoS) requirements of different services and dynamically generates the optimal communication mode switching strategy based on a pre-trained machine learning model. This achieves seamless switching between broadband and narrowband communication links, while continuously improving decision accuracy through a feedback optimization mechanism. This method effectively solves the problems of traditional power IoT communication methods being fixed and singular, unable to intelligently adapt to service and network conditions. It significantly improves system communication reliability, reduces transmission latency and energy consumption, and achieves highly adaptable and reliable intelligent communication assurance in complex power scenarios. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0040] Figure 1 This is a flowchart of the steps of the power Internet of Things communication control method with broadband / narrowband adaptive switching according to the first embodiment of the present invention.

[0041] Figure 2 This is a schematic diagram of a power Internet of Things communication system with broadband / narrowband adaptive switching according to the second embodiment of the present invention.

[0042] In the diagram: 201 - Data Acquisition Module, 202 - Business Identification Module, 203 - Intelligent Decision Module, 204 - Switching Execution Module, 205 - Feedback Optimization Module. Detailed Implementation

[0043] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention.

[0044] The first embodiment of this application is as follows:

[0045] Please see Figure 1 ,in, Figure 1 This is a flowchart of the steps of the power Internet of Things communication control method with broadband / narrowband adaptive switching according to the first embodiment of the present invention.

[0046] This invention provides a power Internet of Things (IoT) communication control method with broadband / narrowband adaptive switching, comprising the following steps:

[0047] S101: Real-time acquisition of multi-dimensional status data, including network status data, terminal status data, and service flow data;

[0048] Specifically, the data acquisition module 201, deployed on the edge computing gateway or smart terminal, is responsible for this task. Its core mission is to continuously and in parallel collect real-time system operating status data from multiple dimensions. The data acquisition module 201 acquires three types of key data periodically or triggered by events by calling the application programming interfaces and driver interfaces at the device's operating system level and by listening to data streams. Network status data is primarily collected by querying the built-in status registers of the communication modules and actively sending probe messages. For example, for 5G / 4G modules, it queries their received signal strength indicator and reference signal received power using standard AT command sets; for Wi-Fi modules, it obtains their link quality and signal strength using wireless extension tools; and for LoRa modules, it reads the signal-to-noise ratio and received signal strength value of their most recently received data packets. Simultaneously, it assesses network latency by periodically sending lightweight Ping messages to reliable nodes in the network and calculating their round-trip time, calculates packet loss rate by counting the number of probe messages sent and received over a period of time, and obtains the current uplink and downlink bandwidth utilization, i.e., network load, by reading network interface traffic statistics.

[0049] The collection of terminal status data focuses on the device's own resource status. The data acquisition module 201 obtains this information by reading the device's hardware sensors and system status information. For battery-powered terminals, the current battery voltage and remaining power percentage are read through the power management chip; for AC-powered devices, this parameter can be marked as sufficient. The computing load is determined by querying the operating system kernel to obtain the current utilization rate of the central processing unit and the available memory capacity, thereby comprehensively judging whether the device's current processing capacity is sufficient.

[0050] The collection of service flow data is achieved through deep packet inspection (DPI) technology or application layer protocol parsing. The data acquisition module 201 listens to and analyzes network data packets flowing through the device. By parsing the protocol header information of the data packets, such as source / destination IP addresses, port numbers, protocol types, and specific identification fields in the payload content, the service type is identified. For example, it determines whether it is a Modbus-TCP based instrument reading, an RTSP-based video stream, or a 104 protocol-based control command. At the same time, the amount of data to be transmitted per unit time is counted to provide a basis for subsequent judgment on whether a high-bandwidth link is required. All collected raw data, after being timestamped and initially formatted, is sent to a shared data buffer for real-time use by the subsequent service identification module 202 and intelligent decision-making module 203, thereby providing comprehensive and accurate data support for the entire adaptive handover process.

[0051] S102: Identify the priority and service quality requirements of the current business based on business flow data;

[0052] Specifically, based on the successful acquisition of service flow data, the service identification module 202 performs deep parsing and feature extraction on the data packets to accurately determine the type of the service to be transmitted and map its corresponding priority and quality of service requirements, providing key input for subsequent intelligent decision-making. The service identification module 202 first parses the service flow data sent to the shared data buffer in S101. By analyzing the protocol header information of the network data packets, such as source and destination IP addresses, port numbers, and transport layer protocol types, it performs preliminary service classification. For example, it identifies traffic destined for port 554 as likely belonging to RTSP video streams, while data packets connected to a specific server port and using the Modbus-TCP protocol are likely meter reading data. Based on this, the module further performs deep packet inspection on the payload content of the data packets, accurately identifying the service type by matching preset feature codes or regular expressions. For example, it identifies H.264 or H.265 encoding formats in video streams and parses function codes in power protocol messages to distinguish between alarm control commands, telemetry data, and remote signaling data.

[0053] Based on the service type-policy mapping table pre-installed in the system memory, the service identification module 202 assigns default priority and quality of service parameters to the identified service types. This mapping table is a predefined key data structure, the contents of which are usually configured by the system administrator according to the actual needs and policies of the power business. For example, the mapping table explicitly specifies that alarm control commands and trip signals have the highest priority, with service quality requirements of extremely low latency and high reliability; real-time video stream data has the second highest priority, requiring high bandwidth and low latency; while periodic meter reading data and environmental sensor data (such as temperature and humidity) are set to normal priority by default, with a certain tolerance for latency, but requiring high data integrity.

[0054] Specifically, for identified instrument readings or environmental sensor data, the service identification module 202 does not simply assign them a default ordinary priority, but instead initiates a dynamic analysis sub-process. This sub-process calls the historical database or cache to calculate the rate of change of recently reported data from the data source, or determines whether its current value exceeds a preset safety threshold. For example, if the temperature reading of a certain line suddenly rises sharply, or its absolute value exceeds the warning value, even if it is ordinary telemetry data, the service identification module 202 will dynamically and temporarily significantly increase its priority and service quality level, treating it as a quasi-alarm signal, thereby ensuring that critical status changes can be transmitted to the monitoring master station with priority and reliability. After completing all identification and judgment, the service identification module 202 encapsulates the current service type, the final determined priority, and the specific service quality requirement parameters into a structured decision result, which is then output to the downstream intelligent decision module 203 for use.

[0055] S103: Input multidimensional state data, current business priority, and service quality requirements into a pre-trained intelligent decision-making model to generate a communication mode switching strategy;

[0056] Specifically, after the service identification module 202 outputs the judgment result of the current service, the intelligent decision-making module 203 is immediately activated. Its core task is to integrate the multi-dimensional status data from the data acquisition module 201 with the priority and service quality requirements from the service identification module 202, and calculate the optimal communication mode switching strategy through a pre-trained machine learning model. This module first standardizes and vectorizes all input data. For example, it normalizes network status data such as signal strength, latency, and packet loss rate to a unified numerical range, and transforms parameters in service priority and service quality requirements (such as maximum tolerable latency and minimum required bandwidth) into numerical features. These features are then combined with terminal status data (such as remaining battery power and CPU load) to form a comprehensive feature vector, which serves as the input to the decision-making model.

[0057] This intelligent decision-making model is a machine learning model trained offline and deployed online. Its training data comes from historical switching performance data accumulated during long-term system operation. This data records the switching decisions made under different combinations of state features, and the resulting communication performance (e.g., whether the switching was successful, whether the latency met requirements, and energy consumption). The model employs algorithms such as deep Q-learning or random forests to learn the complex mapping relationships between states, actions, and rewards. Once deployed at the edge gateway, the model can be used for online inference. Its output is a probability distribution of communication mode selection decisions or a direct optimal policy suggestion, such as whether to continue using the current broadband mode, switch to narrowband mode, or maintain the current mode.

[0058] When generating specific strategies, the model strictly adheres to the constraints of business requirements and system status. For example, when input features indicate that the current service is identified as a high-priority alarm control command, the model will almost inevitably decide to prioritize broadband communication mode to ensure extremely low latency and ultra-high reliability transmission of the command. Conversely, when network status data indicates that broadband communication mode is unavailable or its signal quality is below a preset availability threshold, and the service identification module 202 determines that the current service is non-urgent and can tolerate a certain delay (such as periodic environmental sensor data), the model will decide to switch to narrowband communication mode to save energy and maintain the connection. In addition, the model also has a certain predictive capability. By analyzing historical network status data sequences, it infers the short-term trend of network quality changes, thereby triggering the switching process to a more stable mode in advance when the network quality has not deteriorated to the switching threshold but shows a significant downward trend, achieving proactive link maintenance and avoiding service interruption. Finally, the intelligent decision module 203 encapsulates the decision result output by the model into a clear and executable communication mode switching strategy command and sends it to the switching execution module 204.

[0059] S104: Perform seamless switching of communication links according to the communication mode switching strategy;

[0060] Specifically, upon receiving the communication mode switching strategy instruction from the intelligent decision-making module 203, the switching execution module 204 immediately starts. Its core task is to execute the link switching operation accurately, efficiently, and smoothly, minimizing the impact on upper-layer business applications and achieving a seamless switching experience. This module first parses the strategy instruction, clarifies the target communication mode, and notifies the corresponding communication interface driver to prepare, such as waking up a dormant narrowband module or reconnecting a broadband module to optimize link parameters. To ensure uninterrupted service, the switching execution module 204 adopts differentiated switching strategies based on the characteristics of the currently transmitted service data.

[0061] For connection-oriented services that require maintaining session state, such as TCP-based remote control sessions or video surveillance sessions, the handover execution module 204 works in conjunction with the operating system protocol stack. Utilizing Mobile IP or specific proxy technologies, it pre-establishes a backup transmission path through the target network link before the handover, and rapidly migrates the service flow to the new link during the handover. During this process, a session persistence mechanism maintains the continuity of connection state and sequence numbers, making the brief network fluctuations virtually imperceptible to upper-layer applications. For connectionless or non-session-oriented services that can tolerate brief interruptions, such as periodic sensor data reporting, the data packets to be sent are temporarily stored in a handover buffer created locally on the device before the handover. Once the handover execution module 204 confirms that the target link has been successfully established and is stable, it reads the data from the buffer and retransmits or resumes the data transmission to the platform through a new communication interface, effectively avoiding the risk of data loss during link instability.

[0062] The entire handover process is closely monitored by the handover execution module 204, which determines the success of the handover by real-time confirmation of key indicators such as the signal quality of the target link and gateway reachability. Once it is confirmed that the new link is ready and the service flow has been smoothly transitioned, the module will notify the original communication interface to enter standby or disconnect state, thereby completing the entire seamless handover process, and will notify the feedback optimization module 205 of the handover completion event and result for recording. This refined handover control mechanism ensures that power Internet of Things services can obtain continuous and reliable communication service guarantees in various complex network environments.

[0063] S105: Record the communication performance data after this switchover and feed the performance data back to the intelligent decision-making model for model optimization.

[0064] Specifically, after the handover execution module 204 completes the communication link handover operation, the feedback optimization module 205 is immediately activated. This module is responsible for comprehensively recording detailed communication performance data before and after the handover event and constructing training samples for model optimization. This module collects and correlates data from multiple dimensions, including the timestamp of the handover, network status indicators before and after the handover (such as signal strength, latency, and packet loss rate comparison), the type of handover strategy used, the total time consumed during the handover process, the stability performance of the new link after the handover (such as the transmission success rate over a continuous period), and the impact of this handover on terminal energy consumption. In particular, the module also records the type of service served by this handover and its final quality perception results, such as whether high-priority alarm commands were successfully delivered within the required time limit, and whether stuttering or image quality degradation occurred after the video stream handover.

[0065] All this raw performance data, after being anonymized and standardized, is packaged into a complete sample data package and uploaded via a secure transmission channel to a model training server deployed at the edge or in the cloud. The machine learning platform on the training server merges these new samples with the massive amount of historically accumulated sample data to form an incremental training dataset. Subsequently, the platform initiates the model's retraining process periodically or on a triggered basis, using algorithms such as reinforcement learning to adjust the weights of neural network nodes or the decision paths of policy trees in the intelligent decision-making model based on the new sample data. The core goal of optimization is to enable the model to more accurately evaluate the long-term returns of different switching strategies under specific state characteristics, thereby learning to make better decisions in complex scenarios, such as better balancing latency and energy consumption, or predicting network quality deterioration earlier and initiating a switch in advance.

[0066] After training is complete, the generated new model version undergoes rigorous testing and verification to ensure improved performance and stability. It is then distributed to edge computing gateways or smart terminals across the entire network via a remote upgrade mechanism, replacing the old decision-making model and enabling continuous evolution of decision-making capabilities. This entire feedback optimization process forms a closed-loop system from data acquisition and model training to deployment and updates. This allows the entire communication system to continuously learn from practical operational experience, gradually adapting to more diverse power business scenarios and network environments, ultimately achieving a spiral increase in overall communication efficiency and reliability.

[0067] It effectively solves the problems of fixed and single communication methods in traditional power Internet of Things, which cannot intelligently adapt to business and network conditions. It significantly improves the reliability of system communication, reduces transmission latency and energy consumption, and achieves highly adaptable and reliable intelligent communication guarantee in complex power scenarios.

[0068] The second embodiment of this application is as follows:

[0069] Based on the first embodiment, please refer to Figure 2 ,in, Figure 2 This is a schematic diagram of a power Internet of Things communication system with broadband / narrowband adaptive switching according to the second embodiment of the present invention.

[0070] The broadband / narrowband adaptive switching power Internet of Things communication system of this embodiment includes a data acquisition module 201, a service identification module 202, an intelligent decision-making module 203, a switching execution module 204, and a feedback optimization module 205.

[0071] Data acquisition module 201 is used to collect multi-dimensional status data in real time, including network status data, terminal status data and service flow data;

[0072] The service identification module 202, connected to the data acquisition module 201, is used to identify the priority and service quality requirements of the current service based on service flow data.

[0073] The intelligent decision-making module 203, connected to the data acquisition module 201 and the service identification module 202, is used to input multi-dimensional status data, the priority of the current service, and service quality requirements into the pre-trained intelligent decision-making model to generate a communication mode switching strategy.

[0074] The switching execution module 204, connected to the intelligent decision module 203, is used to perform seamless switching of communication links according to the communication mode switching strategy.

[0075] The feedback optimization module 205 is connected to the switching execution module 204 and the intelligent decision-making module 203. It is used to record the communication performance data after this switching and feed the performance data back to the intelligent decision-making model for model optimization.

[0076] Using a broadband / narrowband adaptive switching power IoT communication system according to this embodiment, through the coordinated operation of the data acquisition module 201, service identification module 202, intelligent decision-making module 203, switching execution module 204, and feedback optimization module 205, the system sequentially achieves real-time acquisition of multi-dimensional status data, intelligent identification of service priorities and quality of service requirements, generation of switching strategies based on machine learning models, seamless switching of communication links, and performance data feedback optimization, forming a closed-loop control. It can dynamically and intelligently select the optimal communication mode, effectively solving the problems of low reliability, high latency, and high energy consumption caused by fixed and single communication methods in traditional power IoT systems, and significantly improving the adaptive capability and service assurance level of communication in complex power environments.

[0077] The above-disclosed embodiments are merely one or more preferred embodiments of this application and should not be construed as limiting the scope of this application. Those skilled in the art can understand that all or part of the processes for implementing the above embodiments and equivalent changes made in accordance with the claims of this application still fall within the scope of this application.

Claims

1. A power Internet of Things communication control method with broadband / narrowband adaptive switching, characterized in that, Includes the following steps: Real-time acquisition of multi-dimensional status data, including network status data, terminal status data, and service flow data; Identify the priority and service quality requirements of the current business based on business flow data; Multidimensional state data, current business priorities, and service quality requirements are input into a pre-trained intelligent decision-making model to generate communication mode switching strategies. Perform seamless switching of communication links according to the communication mode switching strategy; Record the communication performance data after this switchover and feed the performance data back to the intelligent decision-making model for model optimization.

2. The power IoT communication control method with broadband / narrowband adaptive switching as described in claim 1, characterized in that, Real-time acquisition of multi-dimensional status data, including network status data, terminal status data, and service flow data. Network status data includes signal strength, network latency, packet loss rate, and network load for currently available communication modes; Terminal status data includes remaining battery power and computing load; Business flow data includes the business type and data volume of the data to be transmitted.

3. The power IoT communication control method with broadband / narrowband adaptive switching as described in claim 2, characterized in that, Based on business flow data, the priority and service quality requirements of the current business are identified. The service types include alarm control commands, video stream data, instrument reading data, and environmental sensor data.

4. The power IoT communication control method with broadband / narrowband adaptive switching as described in claim 3, characterized in that, Identifying the priority and service quality requirements of the current business based on business flow data involves the following steps: Parse the protocol headers and payloads of the business flow data to identify its business type; Based on the pre-defined business type-policy mapping table, query the default priority and service quality requirement parameters corresponding to the business type; If the business type is instrument readings or environmental sensor data, further analysis is conducted to determine whether the data change rate or value exceeds a preset threshold, and the priority is dynamically increased accordingly.

5. The power IoT communication control method with broadband / narrowband adaptive switching as described in claim 4, characterized in that, Multidimensional state data, current business priorities, and service quality requirements are input into a pre-trained intelligent decision-making model to generate a communication mode switching strategy. The intelligent decision-making model is a machine learning model trained based on historical handover performance data, used to output the optimal communication mode selection decision; Communication mode switching strategies include deciding whether to use broadband communication mode, narrowband communication mode, or maintain the current mode.

6. The power Internet of Things communication control method with broadband / narrowband adaptive switching as described in claim 5, characterized in that, Multidimensional state data, current business priorities, and service quality requirements are input into a pre-trained intelligent decision-making model to generate a communication mode switching strategy. When an alarm control command with the highest priority for the current service is identified, broadband communication mode is selected first. When network status data indicates that broadband communication mode is unavailable or the quality is below a preset threshold, and the current service can tolerate the delay, narrowband communication mode is selected. By combining historical network status data to predict trends, the handover process can be triggered in advance before the network quality deteriorates to the threshold requiring handover.

7. The power IoT communication control method with broadband / narrowband adaptive switching as described in claim 6, characterized in that, According to the communication mode switching strategy, seamless switching of communication links is performed. For connection-oriented session services, maintain session state continuity during handover; For non-session services, data is cached before switching and resent or resumed after successful switching.

8. A power IoT communication system with broadband / narrowband adaptive switching, used in the power IoT communication control method with broadband / narrowband adaptive switching as described in claim 1, characterized in that, It includes a data acquisition module, a business identification module, an intelligent decision-making module, a switching execution module, and a feedback optimization module; The data acquisition module is used to collect multi-dimensional status data in real time, including network status data, terminal status data, and service flow data. The business identification module, connected to the data acquisition module, is used to identify the priority and service quality requirements of the current business based on business flow data. The intelligent decision-making module, connected to the data acquisition module and the service identification module, is used to input multi-dimensional status data, the priority of the current service, and service quality requirements into the pre-trained intelligent decision-making model to generate communication mode switching strategies. The switching execution module, connected to the intelligent decision-making module, is used to perform seamless switching of communication links based on the communication mode switching strategy. The feedback optimization module, connected to the switching execution module and the intelligent decision-making module, is used to record the communication performance data after this switching and feed the performance data back to the intelligent decision-making model for model optimization.

Images