Low-power-consumption client terminal equipment (CPE) integrated with RedCap technology and Internet of Things application system

Abstract

The invention relates to the technical field of communication, and discloses low-power-consumption client terminal equipment (CPE) integrated with RedCap technology, which comprises a RedCap communication module, a multi-service access unit, an edge computing unit, an intelligent power consumption management unit, a power management unit, a security module, an environment sensing module and a network selection circuit. According to the low-power-consumption client terminal equipment (CPE) integrated with the RedCap technology and the Internet of Things application system, the RedCap technology and an intelligent power consumption management architecture are deeply integrated, so that the overall power consumption and the operation cost of the CPE are remarkably reduced; and a service awareness-dynamic scheduling-adaptive wake-up cooperation mechanism is adopted, and parameters such as an eDRX period and PSM sleep duration are dynamically adjusted and optimized in combination with a machine learning optimization model, so that the equipment realizes deep energy saving in an inactive state, the cruising ability is greatly improved, and the maintenance-free operation requirements of more than 5 years in fields, industries and other scenes can be met.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of communication, in particular to a low-power customer terminal equipment (CPE) and an Internet of Things application system fusing RedCap technology. BACKGROUND

[0002] In the current field of communication technology, 5G customer terminal equipment (CPE) plays an important role, which can convert mobile communication signals into local area network signals to provide network connection for various devices and is widely used in home broadband, enterprise access and industrial Internet of Things scenarios.

[0003] However, the 5G customer terminal equipment (CPE) in the prior art is designed for high-bandwidth and low-latency scenarios, and has problems such as high power consumption, high cost, short endurance, etc., and it is difficult to meet the needs of large-scale Internet of Things applications for low power consumption, long life and wide coverage. Although the RedCap technology is introduced by 3GPP to reduce the complexity of the terminal, the current RedCap CPE still follows the traditional communication architecture and lacks a system-level power consumption optimization mechanism, especially in terms of business dynamic perception, deep energy saving in non-active state and edge intelligence cooperation. Therefore, a low-power customer terminal equipment (CPE) and an Internet of Things application system fusing RedCap technology are proposed to solve the above-mentioned problems. SUMMARY

[0004] (I) Technical problems to be solved In view of the deficiencies of the prior art, the application provides a low-power customer terminal equipment (CPE) and an Internet of Things application system fusing RedCap technology, which has the advantages of significantly reduced power consumption, greatly extended endurance, support for intelligent access of multiple services, edge AI reasoning capability and seamless switching of heterogeneous networks, etc. The technical problems of high power consumption, high cost and short endurance of traditional 5G CPE, and the lack of system-level power consumption optimization, weak service perception capability, insufficient edge intelligence and easy interruption of network switching of existing RedCap terminals are solved.

[0005] (II) Technical solutions In order to achieve the above-mentioned purposes of "significantly reduced power consumption, greatly extended endurance, support for intelligent access of multiple services", etc., the application provides the following technical solutions: a low-power customer terminal equipment (CPE) fusing RedCap technology, comprising a RedCap communication module, a multi-service access unit, an edge computing unit, an intelligent power consumption management unit, a power management unit, a security module, an environment perception module and a network selection circuit, wherein, The RedCap communication module is electrically connected with the intelligent power consumption management unit and the network selection circuit, and supports eDRX mode and PSM mode. The multi-service access unit comprises an Ethernet interface, a Wi-Fi 6 module and a Bluetooth 5.2 module, and is electrically connected with the edge computing unit input end; The edge computing unit comprises a data compression circuit, a protocol conversion circuit, a local decision circuit and an AI inference engine, the data compression circuit output end is electrically connected with the protocol conversion circuit input end, the protocol conversion circuit output end is electrically connected with the local decision circuit input end, and the AI inference engine is bidirectionally connected with the local decision circuit. The intelligent power management unit comprises a service awareness engine, a dynamic power consumption scheduler, an adaptive wake-up sub-module and a machine learning optimization model, the service awareness engine output end is electrically connected with the dynamic power consumption scheduler input end, the dynamic power consumption scheduler output end is electrically connected with the RedCap communication module control end, the machine learning optimization model is embedded in the dynamic power consumption scheduler, and a parameter adjustment strategy set trained based on historical operation data is stored; The power management unit is electrically connected with the intelligent power management unit and the RedCap communication module respectively; The security module integrates a hardware encryption chip, and is in communication connection with the edge computing unit and the RedCap communication module; The environment awareness module comprises a temperature and humidity sensor, a signal strength detector, a position locator, an air quality monitor and a noise level detector, and each sensor output end is electrically connected with the intelligent power management unit and the network selection circuit; The network selection circuit comprises a signal comparator, a preset threshold storage and a dual-mode switching controller, the signal comparator first input end is electrically connected with the environment awareness module, the second input end is electrically connected with the preset threshold storage, the output end is electrically connected with the dual-mode switching controller input end, the dual-mode switching controller is electrically connected with the RedCap communication module, the Wi-Fi 6 module and the Bluetooth 5.2 module respectively, and a redundant link channel is configured between the dual-mode switching controller and the RedCap communication module.

[0006] Preferably, the service awareness engine comprises an interface identification circuit, the interface identification circuit is provided with at least three different types of physical interfaces, which are an RJ45 interface, a USB Type-C interface and a dedicated radio frequency interface, corresponding to a periodic data source, an event trigger source and a video stream source.

[0007] Preferably, the dynamic power consumption scheduler comprises a register group, the register group stores eDRX cycle parameters, PSM sleep duration parameters, transmission power parameters and data transmission rate parameters, and the register group write port is electrically connected with the edge computing unit and the machine learning optimization model.

[0008] Preferably, the adaptive wake-up sub-module comprises a Bluetooth Low Energy chip and a Zigbee chip, and the wake-up signal output ends of the Bluetooth Low Energy chip and the Zigbee chip are electrically connected to the wake-up signal input end of the RedCap communication module through an OR gate circuit.

[0009] Preferably, the AI inference engine is a lightweight neural network processor loaded with an image classification model and a time series anomaly detection model trained by transfer learning, and the input end of the AI inference engine is electrically connected to the video stream source, and the output end is electrically connected to the local decision circuit.

[0010] An Internet of Things application system integrating RedCap technology comprises a plurality of low-power customer premises equipment (CPE), a 5G core network, an Internet of Things cloud platform, a central power consumption management server and an edge node controller, wherein, The 5G core network comprises a RedCap network slice, and the bandwidth of the RedCap network slice is configured as 20MHz, matching the RedCap communication module; The Internet of Things cloud platform comprises a data storage server, a data analysis server, a decision support server and a security gateway; The central power consumption management server comprises a network congestion detector, a service load counter, a battery state receiver and a policy generation circuit, the output ends of the network congestion detector, the service load counter and the battery state receiver are electrically connected to the input end of the policy generation circuit, and the output end of the policy generation circuit is in communication connection with the intelligent power consumption management unit of the low-power CPE device; The edge node controller comprises a data forwarding circuit and a task allocation circuit, the input end of the data forwarding circuit is in communication connection with the edge computing unit of the low-power CPE device, the output end is in communication connection with the Internet of Things cloud platform, and the output end of the task allocation circuit is in communication connection with the edge computing unit; The security gateway establishes an end-to-end encryption tunnel with the security module of the low-power CPE device, and adopts a national encryption SM9 algorithm or an AES-256 encryption protocol.

[0011] Preferably, the machine learning optimization model is provided with a global training instance in the central power consumption management server, the global training instance performs incremental training by collecting historical running logs of a plurality of low-power CPE devices, and periodically issues an updated parameter adjustment strategy set to each device.

[0012] Preferably, the task allocation circuit of the edge node controller issues an edge side inference task instruction to the AI inference engine, and the inference task instruction comprises target detection, behavior recognition or device fault prediction.

[0013] Preferably, the air quality data and noise level data collected by the environment perception module are uploaded to the data analysis server of the Internet of Things cloud platform after being processed by the edge computing unit, for generating a regional environmental quality heat map.

[0014] (Three) beneficial effects Compared with the prior art, the present application provides a low-power customer terminal equipment (CPE) and Internet of Things application system integrating RedCap technology, which has the following beneficial effects: 1. The low-power customer terminal equipment (CPE) and Internet of Things application system integrating RedCap technology significantly reduces the overall power consumption and operating cost of the customer terminal equipment (CPE) by deeply integrating RedCap technology and intelligent power consumption management architecture; and adopts a "service awareness-dynamic scheduling-self-adaptive wake-up" collaborative mechanism to dynamically optimize parameters such as eDRX cycle and PSM sleep duration in combination with a machine learning optimization model, so that the device realizes deep energy saving in an inactive state, the endurance is greatly improved, and the maintenance-free operation demand for more than 5 years in the field, industrial and other scenes can be met; at the same time, the AI inference engine completes image recognition, anomaly detection and other computing tasks on the edge side, effectively reducing invalid data upload, further reducing communication frequency and energy consumption, and improving the overall energy efficiency of the system.

[0015] 2. The low-power customer terminal equipment (CPE) and Internet of Things application system integrating RedCap technology enhances the intelligence and environmental adaptability of the CPE, supports multi-source environmental perception and seamless switching of heterogeneous networks, improves connection reliability and deployment flexibility; through the real-time monitoring of temperature, humidity, signal strength, air quality and other parameters by the environment perception module, in combination with the network selection circuit to realize the automatic switching of the optimal link of 5G, Wi-Fi and Bluetooth, the business continuity is ensured; the system-level collaborative design enables the central power consumption management server to globally optimize the power consumption strategy of multiple devices, realizes the energy efficiency closed-loop management from the terminal to the cloud, solves the technical bottleneck of single function and lack of intelligent collaboration of existing RedCap terminals, and provides a high-reliability, low-power and intelligent terminal solution for large-scale Internet of Things applications. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The internal architecture diagram of the low-power customer terminal equipment (CPE) integrating RedCap technology of the present application; Figure 2 The collaborative architecture diagram of the Internet of Things application system integrating RedCap technology of the present application; Figure 3 The intelligent power consumption optimization key process diagram of the present application; Figure 4This is a flowchart illustrating the operation method of the low-power client terminal device and IoT system integrating RedCap technology according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1 Please see Figure 1 , Figure 2 and Figure 3 As shown, a low-power customer premises equipment (CPE) integrating RedCap technology includes a RedCap communication module, a multi-service access unit, an edge computing unit, an intelligent power management unit, a power management unit, a security module, an environmental awareness module, and a network selection circuit, wherein; The RedCap communication module is one of the core components, supporting eDRX and PSM modes. It can dynamically adjust its working state according to service requirements, thereby achieving significant energy-saving effects. This module is electrically connected to the intelligent power management unit and network selection circuit, ensuring that the device can enter a deep sleep mode when inactive, and quickly wake up when network demand occurs.

[0019] The multi-service access unit includes an Ethernet interface, a Wi-Fi 6 module, and a Bluetooth 5.2 module, which can flexibly connect to various local IoT devices. These interfaces are electrically connected to the input of the edge computing unit, enabling data to be initially processed locally, reducing the amount of data uploaded to the cloud and further reducing power consumption.

[0020] The edge computing unit includes a data compression circuit, a protocol conversion circuit, a local decision circuit, and an AI inference engine. The data compression circuit compresses the raw data and then passes it to the protocol conversion circuit, which converts it into a format suitable for transmission and passes it to the local decision circuit. The AI ​​inference engine is responsible for performing complex tasks such as image classification and anomaly detection, and communicates bidirectionally with the local decision circuit to improve the intelligence level of the device.

[0021] The intelligent power management unit consists of a service awareness engine, a dynamic power scheduler, an adaptive wake-up submodule, and a machine learning optimization model. The service awareness engine identifies different types of service traffic and transmits the information to the dynamic power scheduler. The dynamic power scheduler adjusts the device's operating parameters, such as eDRX cycle and PSM sleep duration, according to current service requirements. The machine learning optimization model generates an optimal parameter adjustment strategy set based on historical operating data to further improve the device's energy efficiency.

[0022] The power management unit is electrically connected to the intelligent power management unit and the RedCap communication module to ensure stable power supply to the device in various operating modes and to optimize energy consumption through the intelligent power management unit.

[0023] The security module integrates a hardware encryption chip to provide end-to-end security protection; it communicates with the edge computing unit and the RedCap communication module to ensure data security during transmission.

[0024] The environmental sensing module includes temperature and humidity sensors, signal strength detectors, location locators, air quality monitors, and noise level detectors. These sensors monitor external environmental conditions in real time and transmit the data to the intelligent power management unit and network selection circuit for dynamic adjustment and optimization.

[0025] The network selection circuit includes a signal comparator, a preset threshold memory, and a dual-mode switching controller. The signal comparator compares the data collected by the environmental perception module with the preset threshold and transmits the result to the dual-mode switching controller. The dual-mode switching controller selects the best network (such as 5G, Wi-Fi, or Bluetooth) based on the comparison result and ensures seamless connection during the switching process through redundant link channels.

[0026] like Figure 2 As shown, an IoT application system architecture integrating RedCap technology includes the following: The 5G core network includes RedCap network slices with a bandwidth configuration of 20MHz, which are matched with the RedCap communication module to ensure efficient communication of devices under wide-area coverage.

[0027] The IoT cloud platform includes a data storage server, a data analysis server, a decision support server, and a security gateway. The data storage server stores the data uploaded from various CPE devices, the data analysis server analyzes this data, the decision support server generates corresponding control commands, and the security gateway uses the national cryptographic SM9 algorithm or the AES-256 encryption protocol to ensure the security of data transmission.

[0028] The central power management server includes a network congestion detector, a service load counter, a battery status receiver, and a policy generation circuit. These components work together to collect historical operating logs from multiple CPE devices, perform incremental training through machine learning models, and periodically send updated parameter adjustment policy sets to each device to achieve global optimization.

[0029] The edge node controller includes a data forwarding circuit and a task allocation circuit. The data forwarding circuit forwards the data received from the CPE device to the IoT cloud platform, while the task allocation circuit issues inference task instructions to the AI ​​inference engine on the edge side, such as object detection, behavior recognition, or device fault prediction.

[0030] Example 2 In this embodiment, the machine learning optimization model is trained based on historical operating data to generate a set of parameter adjustment strategies, and its output is an optimal power consumption parameter vector. It can be represented as: ; in: : Optimal parameter vector, representing eDRX period, PSM sleep duration, transmit power and data transmission rate respectively; : Business traffic feature vector, including data packet arrival rate, average packet length, and business type identifier (periodic = 1, event triggered = 2, video stream = 3). : Environmental state vector, including signal strength (RSRP), temperature and humidity, air quality index (AQI), and noise level; : Current remaining battery power (%); : Current timestamp (used to identify peak / off-peak business periods); : A nonlinear mapping function trained through supervised learning, which can be a lightweight neural network or a gradient boosting tree (GBDT) model.

[0031] This function minimizes the objective function. Conduct training: ; in: Total energy consumption per unit time; Average data upload latency; Network connection reliability (0~1); Weighting coefficients are dynamically configured by the central power management server according to system policies.

[0032] In this embodiment, the image classification model loaded by the AI ​​inference engine adopts a lightweight convolutional neural network (such as MobileNetV2), and its forward propagation process can be represented as follows: ; in: Input image tensor (e.g.) ); : No. Layer weight matrix and bias vector; Network layers; The classification output probability vector is used for object detection or behavior recognition. In this embodiment, an LSTM network model is used for time series anomaly detection: in: : No. Sensor data at any given time (such as temperature, humidity, and signal strength); Hidden state and cell state; Anomaly score, with a value close to 1 indicating anomaly; : Sigmoid activation function.

[0033] In this embodiment, the network selection logic of the dual-mode switching controller can be expressed as: ; in: :network The signal strength (RSRP or RSSI); :network The load rate (provided by the network congestion detector); Using the network The power consumption cost; Weighting coefficients are stored in a preset threshold memory. The selected optimal network.

[0034] During the handover process, the redundant link channel maintains dual-mode activation time. satisfy: ; in To test the switching latency and ensure uninterrupted service.

[0035] In this embodiment, the global training instances of the central power management server are incrementally updated using a federated learning framework: ; in: : No. Round global model parameters; Learning rate; : No. Local loss function reported by each CPE device; Number of devices participating in the training.

[0036] Updated The new set of parameter adjustment strategies was distributed to each device.

[0037] In this embodiment, as Figure 4 As shown, a method for operating a low-power client terminal device and IoT system integrating RedCap technology includes the following steps: Step 1: Multi-source data access and business type identification Data access: Periodic, event-triggered, and video stream data can be accessed via Ethernet interface, Wi-Fi 6 module, and Bluetooth 5.2 module.

[0038] Business type identification: The business awareness engine identifies the data source type and passes the information to the dynamic power scheduler.

[0039] Step Two: Environmental Perception and Network Status Monitoring Collect environmental data: Temperature and humidity sensors, signal strength detectors, etc. collect environmental parameters.

[0040] Network status monitoring: The signal comparator compares the current signal strength with a preset threshold to determine whether to switch networks.

[0041] Step 3: Edge Data Processing and AI Inference Data compression and protocol conversion: The edge computing unit compresses and converts the original data according to the protocol.

[0042] Perform AI inference: The AI ​​inference engine processes video stream data, performs object detection or anomaly detection, and generates local decision instructions.

[0043] Step 4: Intelligent power consumption scheduling and communication parameter optimization Dynamic power consumption scheduling: Dynamically adjust communication parameters such as eDRX cycle and PSM sleep duration based on service type, environmental status and battery level.

[0044] Optimization parameters are issued: The machine learning optimization model generates the optimal parameters and writes them into the register group to control the RedCap communication module.

[0045] Step 5: Adaptive wake-up and seamless network switching Wake-up mechanism: The low-power Bluetooth chip or Zigbee chip receives an external wake-up signal, triggering the RedCap communication module to wake up.

[0046] Network switching: The dual-mode switching controller seamlessly switches between 5G, Wi-Fi, and Bluetooth to ensure uninterrupted data transmission.

[0047] Step Six: System-level Collaborative Optimization and Cloud Interaction Data Upload and Analysis: The processed data is uploaded to the IoT cloud platform to generate a regional environmental quality heat map.

[0048] Global optimization and policy updates: The central power management server collects historical operation logs from multiple CPEs, performs incremental training, and periodically sends updated parameter adjustment policy sets to each device.

[0049] Step Seven: Secure Communication Assurance Data encryption: The security module encrypts the transmitted data, establishes an end-to-end encrypted tunnel, and adopts the national cryptographic standard SM9 or AES-256 protocol.

[0050] Example 3 Scenario Example 1. Smart City Environmental Monitoring In a smart city project, multiple low-power CPE devices are deployed in different areas of the city to monitor environmental parameters such as air quality, noise levels, and traffic flow. These devices are designed to provide real-time data support to help city managers make more informed decisions.

[0051] Implementation steps: Multi-source data access and processing: Environmental sensing module: Temperature and humidity sensor, signal strength detector, location locator, air quality monitor and noise level detector collect environmental data in real time.

[0052] Edge computing unit: The data compression circuit compresses the raw data, the protocol conversion circuit converts it into a format suitable for transmission, the AI ​​inference engine performs image classification or anomaly detection tasks, and the local decision circuit generates preliminary analysis results.

[0053] Network selection and switching: Network selection circuit: The signal comparator compares the data from the environmental sensing module with a preset threshold. The dual-mode switching controller seamlessly switches between 5G, Wi-Fi, or Bluetooth based on the comparison result, ensuring the stability and efficiency of data transmission.

[0054] Cloud-based interaction and analytics: IoT Cloud Platform: The processed data is uploaded to the data storage server of the IoT cloud platform through the data forwarding circuit. The data analysis server performs in-depth analysis on the data and generates regional environmental quality heat maps and traffic flow prediction reports.

[0055] Intelligent power management: Intelligent power management unit: The service awareness engine identifies the service type, the dynamic power scheduler adjusts parameters such as eDRX cycle and PSM sleep duration according to the current service requirements, and the machine learning optimization model generates the optimal parameter adjustment strategy set based on historical operating data to extend battery life.

[0056] Results: City managers can monitor the environmental quality of each area in real time and formulate targeted governance measures; local preliminary processing of data reduces the amount of data uploaded to the cloud, lowering communication costs and power consumption.

[0057] 2. Devices in an inactive state for extended periods in the Industrial Internet of Things In industrial IoT applications, devices are often inactive for long periods of time, but need to upload data at specific times. This embodiment demonstrates how to significantly extend battery life through the dynamic power scheduler and machine learning optimization model of the intelligent power management unit.

[0058] Implementation steps: Business awareness and data collection: Business Awareness Engine: Identifies different types of business traffic, such as periodic data reporting and event-triggered alarms, and transmits the information to the dynamic power consumption scheduler.

[0059] Environmental sensing module: Real-time monitoring of environmental parameters such as temperature, humidity, and signal strength around the device, providing a reference for dynamic power consumption scheduling.

[0060] Intelligent power consumption scheduling: Dynamic power scheduler: Based on business needs and environmental conditions, it dynamically adjusts parameters such as eDRX cycle and PSM sleep duration to achieve the best energy-saving effect.

[0061] Adaptive wake-up submodule: When there is an external wake-up signal (such as a Bluetooth Low Energy chip or a Zigbee chip), the RedCap communication module is triggered to wake up from PSM mode to ensure timely data upload.

[0062] Machine learning optimization: Machine learning optimization model: Incremental training is performed based on historical operation logs to generate an optimal set of parameter adjustment strategies, which are then regularly updated and distributed to each device to further improve the energy efficiency of the devices.

[0063] Central power management server: Works collaboratively to collect historical operation logs from multiple CPE devices, performs incremental training through global training instances, and distributes updated parameter adjustment strategy sets to each device.

[0064] Secure communication guarantee: Security module: Integrates a hardware encryption chip to ensure data security during transmission, and establishes an end-to-end encrypted tunnel using the national cryptographic standard SM9 or AES-256 encryption protocol.

[0065] Results: The device can maintain deep sleep for extended periods of inactivity, significantly extending battery life; when data needs to be uploaded, it can quickly wake up and complete data transmission, ensuring uninterrupted business operations; through machine learning optimization models, the device can predict the optimal wake-up time and transmission frequency based on historical data, further improving energy efficiency.

[0066] In summary, this low-power customer premises equipment (CPE) and IoT application system integrating RedCap technology achieves low-power, high-reliability, and adaptive IoT access by integrating edge computing, intelligent power management, and multi-mode communication, supporting efficient deployment and long-term stable operation in smart city and industrial scenarios.

[0067] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0068] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A low-power customer premises equipment (CPE) integrating RedCap technology, characterized in that: It includes a RedCap communication module, a multi-service access unit, an edge computing unit, an intelligent power management unit, a power management unit, a security module, an environmental awareness module, and a network selection circuit. The RedCap communication module is electrically connected to the intelligent power management unit and the network selection circuit, and supports eDRX mode and PSM mode. The multi-service access unit includes an Ethernet interface, a Wi-Fi 6 module, and a Bluetooth 5.2 module, which are electrically connected to the input of the edge computing unit. The edge computing unit includes a data compression circuit, a protocol conversion circuit, a local decision circuit, and an AI inference engine. The output of the data compression circuit is electrically connected to the input of the protocol conversion circuit, the output of the protocol conversion circuit is electrically connected to the input of the local decision circuit, and the AI ​​inference engine is bidirectionally connected to the local decision circuit. The intelligent power management unit includes a service awareness engine, a dynamic power scheduler, an adaptive wake-up submodule, and a machine learning optimization model. The output of the service awareness engine is electrically connected to the input of the dynamic power scheduler, and the output of the dynamic power scheduler is electrically connected to the control terminal of the RedCap communication module. The machine learning optimization model is embedded in the dynamic power scheduler and stores a set of parameter adjustment strategies trained based on historical running data. The power management unit is electrically connected to the intelligent power consumption management unit and the RedCap communication module, respectively; The security module integrates a hardware encryption chip and is communicatively connected to the edge computing unit and the RedCap communication module. The environmental sensing module includes a temperature and humidity sensor, a signal strength detector, a location locator, an air quality monitor, and a noise level detector. The output terminals of each sensor are electrically connected to the intelligent power consumption management unit and the network selection circuit. The network selection circuit includes a signal comparator, a preset threshold memory, and a dual-mode switching controller. The first input terminal of the signal comparator is electrically connected to the environmental perception module, the second input terminal is electrically connected to the preset threshold memory, and the output terminal is electrically connected to the input terminal of the dual-mode switching controller. The dual-mode switching controller is electrically connected to the RedCap communication module, the Wi-Fi 6 module, and the Bluetooth 5.2 module, respectively, and a redundant link channel is configured between the dual-mode switching controller and the RedCap communication module.

2. A low-power customer premises equipment (CPE) integrating RedCap technology according to claim 1, characterized in that: The business awareness engine includes an interface identification circuit, which has at least three different types of physical interfaces: an RJ45 interface, a USB Type-C interface, and a dedicated radio frequency interface, which are connected to periodic data sources, event triggering sources, and video stream sources, respectively.

3. A low-power customer premises equipment (CPE) integrating RedCap technology according to claim 1, characterized in that: The dynamic power scheduler includes a register group that stores eDRX cycle parameters, PSM sleep duration parameters, transmit power parameters, and data transmission rate parameters. The register group write port is electrically connected to the edge computing unit and the machine learning optimization model.

4. A low-power customer premises equipment (CPE) integrating RedCap technology according to claim 1, characterized in that: The adaptive wake-up submodule includes a low-power Bluetooth chip and a Zigbee chip. The wake-up signal output terminals of the low-power Bluetooth chip and the Zigbee chip are combined through an OR gate circuit and electrically connected to the wake-up signal input terminal of the RedCap communication module.

5. A low-power customer premises equipment (CPE) integrating RedCap technology according to claim 1, characterized in that: The AI ​​inference engine is a lightweight neural network processor, loaded with an image classification model and a time series anomaly detection model trained by transfer learning. The input end of the AI ​​inference engine is electrically connected to the video stream source, and the output end is electrically connected to the local decision circuit.

6. An Internet of Things (IoT) application system integrating RedCap technology, applied to a low-power customer premises equipment (CPE) integrating RedCap technology as described in claims 1-5, characterized in that: This includes multiple low-power customer premises equipment (CPE), 5G core network, IoT cloud platform, central power management server, and edge node controller, wherein... The 5G core network includes a RedCap network slice, the RedCap network slice bandwidth is configured to be 20MHz, and it matches the RedCap communication module; The IoT cloud platform includes a data storage server, a data analysis server, a decision support server, and a security gateway; The central power management server includes a network congestion detector, a service load counter, a battery status receiver, and a policy generation circuit. The outputs of the network congestion detector, the service load counter, and the battery status receiver are all electrically connected to the input of the policy generation circuit. The output of the policy generation circuit is communicatively connected to the intelligent power management unit of the low-power CPE device. The edge node controller includes a data forwarding circuit and a task allocation circuit. The input of the data forwarding circuit is communicatively connected to the edge computing unit of the low-power CPE device, and the output is communicatively connected to the IoT cloud platform. The output of the task allocation circuit is communicatively connected to the edge computing unit. The security gateway establishes an end-to-end encrypted tunnel with the security module of the low-power CPE device, using the national cryptographic algorithm SM9 or the AES-256 encryption protocol.

7. An IoT application system integrating RedCap technology according to claim 6, characterized in that: The machine learning optimization model has a global training instance in the central power management server. The global training instance performs incremental training by collecting historical operation logs of multiple low-power CPE devices and periodically sends updated parameter adjustment strategy sets to each device.

8. An IoT application system integrating RedCap technology according to claim 6, characterized in that: The task allocation circuit of the edge node controller sends edge-side inference task instructions to the AI ​​inference engine. The inference task instructions include target detection, behavior recognition, or device fault prediction.

9. An IoT application system integrating RedCap technology according to claim 6, characterized in that: The air quality data and noise level data collected by the environmental sensing module are processed by the edge computing unit and then uploaded to the data analysis server of the Internet of Things cloud platform to generate a regional environmental quality heat map.

Images