Multi-mode cabinet intelligent management terminal for narrowband Internet of Things communication

Abstract

A multimode cabinet intelligent management terminal for narrowband Internet of Things communication comprises an environment monitoring module, a load sensing module, a protocol switching module and a bandwidth scheduling module. The environment monitoring module obtains key indexes of network signal strength, delay and packet loss rate and calculates environment data; the load sensing module collects CPU usage rate and memory usage condition load data, evaluates the load state of an Internet of Things sensor node, and generates load data and environment data. The protocol switching module dynamically evaluates network environment and equipment load changes, predicts a network environment change trend by adopting a weighted decision method, and screens an optimal protocol based on a protocol switching strategy; and the bandwidth scheduling module performs optimal allocation on the bandwidth, automatically adjusts bandwidth resources, and performs release and reallocation by monitoring the network condition in real time to obtain bandwidth resource information. The invention provides a multimode cabinet intelligent management terminal for narrowband Internet of Things communication. Accurate point prediction and reliable interval prediction and probability prediction of wind power are realized.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) technology, and more specifically to a multi-mode cabinet intelligent management terminal for narrowband IoT communication. Background Technology

[0002] With the deepening of digital transformation, the demand for intelligent management of data centers and server rooms is growing rapidly. Existing rack management technologies have achieved information-based management of data center assets, including full asset information recording, operational status monitoring, location tracking, and full lifecycle management. They also provide intuitive equipment identification through signage information display systems. In terms of operation and maintenance management, the application of work order systems allows maintenance tasks to be accurately assigned to designated racks, ensuring that designated personnel perform maintenance tasks through identity verification mechanisms, and recording task completion status, realizing an online management process from task assignment to completion. Furthermore, the application of 3D modeling technology in data center management allows the data center structure and rack layout to be presented in the form of simulated 3D models, supporting the visualization and real-time alarm of fault points in specific rack locations. It also displays the low-voltage wiring connections within the modeling environment, facilitating fault location. These technological advancements have laid an important foundation for intelligent rack management, driving the evolution of traditional racks towards intelligent, visualized, and remote management.

[0003] The "Multi-mode Cabinet Intelligent Management Terminal for Narrowband IoT Communication" refers to a multi-mode management terminal device used in narrowband IoT communication environments. This patent primarily addresses how to achieve compatibility and intelligent management of multiple communication modes in narrowband IoT application scenarios. This technology integrates multiple communication modules into a single cabinet terminal to enable switching and interoperability between different communication protocols (such as NB-IoT, LoRa, GPRS, etc.), ensuring stable operation of the device under various network conditions. This patent solves the problem of traditional devices being difficult to manage and switch effectively in multi-network environments. Through an intelligent management system, it monitors device status in real time and optimizes communication paths, ensuring efficient and reliable operation of narrowband IoT devices in various usage environments.

[0004] Traditional management terminals rely on instantaneous data for direct decision-making, triggering frequent protocol switching during network fluctuations and reducing communication stability. For example, a brief signal attenuation can lead to an incorrect switch to a low-speed protocol, impacting service transmission efficiency. Load assessment depends on a single metric, failing to comprehensively reflect device operating status and prone to load classification errors. Protocol selection lacks predictive mechanisms, adjusting reactively based on current conditions, resulting in frequent switching in rapidly changing environments and consuming system resources. Bandwidth allocation uses fixed quotas, unable to be dynamically adjusted, leading to uneven resource distribution. In multi-device environments, each device independently selects a protocol, causing conflicts and interference, affecting overall network performance.

[0005] In terms of operation and maintenance management, the application of the work order system enables the precise assignment of maintenance tasks to designated racks. An identity verification mechanism ensures that designated personnel execute maintenance tasks, and the system records task completion status, realizing an online management process from task assignment to completion. Furthermore, the application of 3D modeling technology in data center management allows the data center structure and rack layout to be presented in the form of simulated 3D models. This supports the visualization and real-time alarm of fault points at specific rack locations, and also displays the low-voltage wiring connections within the modeling environment, facilitating fault location. These technological advancements have laid an important foundation for intelligent rack management, driving the evolution of traditional racks towards intelligent, visualized, and remote management. Summary of the Invention

[0006] Purpose of the invention: The purpose of this invention is to propose a multi-mode cabinet intelligent management terminal for narrowband Internet of Things communication, which can realize accurate point prediction and reliable interval prediction and probability prediction of wind power.

[0007] A multi-mode cabinet intelligent management terminal for narrowband IoT communication includes an environmental monitoring module, a load sensing module, a protocol switching module, and a bandwidth scheduling module. The environmental monitoring module collects environmental data through built-in sensors, obtains key indicators such as network signal strength, latency, and packet loss rate, and uses the sliding window method to calculate the median and variance of the data to obtain environmental data. Based on the environmental data, the load sensing module collects CPU usage and memory usage load data, evaluates the load status of IoT sensor nodes, classifies sensor nodes based on thresholds, and generates load data and environmental data. The protocol switching module dynamically evaluates changes in network environment and device load based on the load data and environmental data, uses a weighted decision method to predict the trend of network environment changes, and selects the optimal protocol based on the protocol switching strategy. If the IoT gateway has a high load and long latency, the low-bandwidth, low-power protocol LoRa is selected; if the network quality is good, the high-bandwidth protocol NB-IoT is selected. The module also provides feedback on the protocol information after the switch and performs resource scheduling. The bandwidth scheduling module optimizes bandwidth allocation based on resource scheduling, automatically adjusts bandwidth resources, and releases and reallocates bandwidth resources by monitoring network conditions in real time, thereby obtaining bandwidth resource information.

[0008] As a further aspect of the present invention, the environmental data includes signal strength indication, transmission delay, packet loss rate, and network throughput; the load data and environmental data specifically include load classification labels, threshold comparison results, and protocol reservation identifiers; the protocol information includes radio frequency parameters, network bandwidth configuration, and protocol switching instructions; and the bandwidth resource information specifically includes resource allocation schemes, priority sorting, and a list of load devices.

[0009] As a further embodiment of the present invention, the environmental monitoring module includes an environmental data acquisition submodule, a signal evaluation submodule, and a preprocessing submodule; The environmental data acquisition submodule detects the operating status of the built-in sensors, reads the voltage and current values ​​of the sensor interfaces, collects the ambient temperature, humidity percentage, and lux value of light intensity, monitors the status of the network connection port, measures the decibel value of the network signal strength, records the data packet transmission time interval, counts the number of packet loss events, calculates the number of bytes transmitted per unit time, and encodes the environmental parameters and network performance parameters according to a unified data format to generate a standardized environmental dataset. The signal evaluation submodule calls the standardized environment dataset to extract the network signal strength in decibels, latency in milliseconds, packet loss rate percentage, and throughput in megabits. The values ​​of each indicator are checked against the boundary of the value range, and abnormal values ​​that exceed the normal range are removed. The rate of change of each indicator in the time series is calculated to obtain the key network performance indicators. Based on the key network performance indicators, the preprocessing submodule sets the sliding window time length parameter, sorts the data within the time window in ascending order of numerical value, selects the value in the middle position after sorting as the median, calculates the sum of squares of the differences between each value and the median, divides it by the total number of data to obtain the variance value, and uses the median and variance value as the statistical feature of the time period. The moving average is used to process the statistical feature value to obtain environmental data.

[0010] As a further aspect of the present invention, the load sensing module includes a device monitoring submodule, a status assessment submodule, and a classification and labeling submodule. Based on the environmental data, the device monitoring submodule reads the real-time operating parameters of the sensor nodes through the device interface, collects the CPU usage percentage, memory usage bytes, task queue pending count, and network bandwidth usage rate, subtracts the CPU usage percentage from the set CPU monitoring benchmark value to obtain the deviation, divides the memory usage bytes from the total capacity to calculate the utilization rate, and establishes a set of load parameters. The status assessment submodule calls the load parameter set to set assessment coefficients such as CPU load weight factor, memory usage weight factor, queue length weight factor, and bandwidth usage weight factor. It multiplies the CPU usage percentage by the CPU weight factor, the memory usage rate by the memory weight factor, the queue count by the queue weight factor, and the bandwidth usage rate by the bandwidth weight factor. The sum of these four products yields the node's comprehensive load index. The comprehensive index is then compared with the load assessment threshold to generate a node status rating. The classification and labeling submodule rates the node status. When the rating value is less than the load classification threshold, it maintains the current protocol parameter configuration and adjusts the bandwidth allocation ratio parameter. When the rating value is greater than or equal to the load classification threshold, it assigns a load identifier code to the node status field. It then merges the load data records and environment data records according to the timestamp to obtain the load data and environment data.

[0011] As a further aspect of the present invention, the protocol switching module includes an environmental assessment submodule, a trend prediction submodule, and a protocol selection submodule; The environmental assessment submodule analyzes the IoT gateway's CPU utilization, memory usage (in bytes), and processor temperature based on the load and environmental data. It also reads the network signal strength (in decibels), data transmission latency, and packet loss event count. It establishes a load threshold reference standard and a network quality evaluation benchmark. The module subtracts the CPU utilization from the load threshold to obtain the load difference, compares the network latency with the latency benchmark, classifies the load status based on the positive or negative load difference, classifies the network quality based on the latency comparison results, and combines the load status category code and the network quality category code to generate a network load assessment level. The trend prediction submodule calls the network load assessment level, sets the load weight factor and network weight factor parameters, multiplies the assessment level value with the corresponding weight factor, calculates the change of the weighted value at continuous time points, solves the ratio of the change to the time interval to obtain the rate of change, determines the sign and magnitude of the rate of change, and obtains the network change trend coefficient. Based on the network change trend coefficient, the protocol selection submodule sets protocol selection judgment rules and threshold boundary parameters. When the trend coefficient exceeds the upper limit threshold and the network latency coefficient exceeds the latency threshold, the LoRa protocol identifier is matched. When the trend coefficient is lower than the lower limit threshold and the network quality coefficient exceeds the quality threshold, the NB-IoT protocol identifier is matched. The protocol identifier and the switching timestamp are packaged and encapsulated, and the protocol switching command is transmitted to the target gateway device to start the device protocol parameter reconfiguration process and establish resource scheduling.

[0012] As a further aspect of the present invention, the bandwidth scheduling module includes a bandwidth load assessment submodule, a bandwidth allocation submodule, and a resource reconfiguration submodule; Based on the resource scheduling, the bandwidth load assessment submodule parses the device CPU utilization rate, memory usage in bytes, and number of network connection ports. It reads the running task count and pending queue length from the system process table, establishes a load status judgment threshold standard and a device type weight coefficient table, performs a subtraction operation between the CPU utilization rate and the load threshold, classifies the device load status level according to the positive or negative result, divides the memory usage in bytes and the total memory capacity of the device to obtain the memory usage ratio, and performs a weighted multiplication operation between the load status level code and the memory usage ratio to generate the device load status coefficient. The bandwidth allocation submodule calls the device load status coefficient, reads the device priority identifier code and bandwidth demand capacity parameter. When the load status coefficient value exceeds the load upper limit threshold, it performs a subtraction operation on the bandwidth allocation amount of the device, calculates the difference between the reduced bandwidth value and the device's minimum bandwidth demand parameter, sorts the device priority identifier codes in ascending order according to their values, and allocates the remaining bandwidth resource capacity level by level according to the sorting results to obtain the optimized bandwidth allocation amount. The resource reallocation submodule, based on the optimized bandwidth allocation, releases the bandwidth resource capacity occupied by the device when the device load status coefficient value is lower than the load lower limit threshold. It calculates the result of the division operation between the total released bandwidth capacity and the number of devices to be allocated, and reallocates the released bandwidth capacity according to the load status coefficient value. It then updates the device bandwidth configuration parameter table and network routing table data items to establish bandwidth resource information.

[0013] A multi-mode cabinet intelligent management terminal for narrowband IoT communication also includes a collaborative working module; Based on the bandwidth resource information, the collaborative working module adjusts the bandwidth allocation priority among IoT device clusters, shares signal quality data, bandwidth utilization data, and latency data through low-power communication channels between device clusters, collaboratively selects communication protocols, and dynamically compares protocol selection priorities based on a weighted decision-making mechanism in a multi-device environment according to the device cluster load status and signal quality. The weighted algorithm determines the protocol selection for multiple devices, optimizes the collaborative working between device clusters in the entire network, and reduces latency caused by protocol switching.

[0014] As a further aspect of the present invention, the device protocol selection includes protocol matching results, weight calculation values, and priority sorting tables.

[0015] As a further aspect of the present invention, the collaborative working module includes a priority scheduling submodule, a data sharing submodule, and a protocol decision submodule; Based on the bandwidth resource information, the priority scheduling submodule reads the IoT device cluster identifier code and the current bandwidth allocation value parameter from the system configuration table, obtains the load weight coefficient and communication priority level parameter of each cluster from the device management table, establishes a bandwidth allocation priority sorting benchmark table and adjustment threshold parameter standard, performs a product operation on the device cluster load weight coefficient value and priority level code to obtain a comprehensive score value, sorts the device clusters in descending order according to the comprehensive score value, adjusts the bandwidth allocation ratio value of each cluster by addition and subtraction according to the sorting result, recalculates the bandwidth quota parameter and allocation time interval parameter value of each cluster, and generates a cluster priority configuration table. The data sharing submodule calls the cluster priority configuration table to establish a low-power communication channel connection between device clusters, monitors communication parameters such as signal quality strength in decibels, bandwidth utilization percentage, and network latency in milliseconds for each cluster, classifies and stores the monitoring data according to the device cluster identification code, calculates the rate of change and fluctuation of various parameters in the time series, performs a filtering operation on key communication parameters whose rate of change exceeds the fluctuation threshold, and encapsulates the key parameter values ​​and cluster identification codes to obtain a shared communication dataset. The protocol decision submodule sets weight allocation matrix parameters based on the shared communication dataset and a weighted decision-making mechanism in a multi-device environment. It extracts the load status values ​​and signal quality level codes of the device cluster, calculates the weighted load value by multiplying the load status values ​​with the corresponding weight coefficients using a weighted algorithm, compares the performance evaluation scores of different protocols in the current network environment, determines the protocol selection sequence for each device cluster based on the evaluation scores, and combines the protocol selection sequence with the switching time parameters to reduce the latency caused by protocol switching.

[0016] A multi-mode cabinet intelligent management terminal for narrowband IoT communication also includes a sensor layer, a data acquisition and transmission layer, a cloud platform layer, and an application layer. The sensor layer includes a temperature sensor, a humidity sensor, a smoke sensor, an access control sensor, a current and voltage sensor, and an environmental monitoring module. Each sensor is used to monitor the internal temperature, humidity, smoke status, cabinet door open / closed status, and equipment power status. The environmental monitoring module collects environmental parameters such as network signal strength, latency, and packet loss rate. The data acquisition and transmission layer includes a data acquisition module, a load sensing module, a protocol switching module, and an NB-IoT communication module. The data acquisition module is responsible for collecting sensor data and processing and formatting it. The load sensing module monitors the system's CPU and memory usage and performs load assessment. The protocol switching module intelligently switches between protocols such as NB-IoT, LoRa, and GPRS according to the network environment and load status. The NB-IoT communication module transmits data to the cloud platform. The cloud platform layer includes a data storage unit, a data analysis and processing unit, an alarm management unit, a remote control unit, and a bandwidth scheduling module. The data storage unit stores rack data, the data analysis and processing unit performs big data analysis, the alarm management unit performs threshold monitoring and automatic alarms, the remote control unit supports remote operation of devices, and the bandwidth scheduling module optimizes resource allocation. The application layer includes a monitoring center, a mobile application, and a report generation module. The monitoring center provides a visual monitoring interface, the mobile application supports viewing and receiving alarms on mobile devices, and the report generation module generates operational reports periodically.

[0017] Compared with the prior art, the present invention has the following significant advantages: (1) This invention monitors the internal temperature, humidity, smoke status, cabinet door opening and closing status and equipment power status of the cabinet through the sensor layer. Combined with the data acquisition and transmission layer and other modules, it manages the full information, operation status, location tracking and service life of the data center assets in a multi-dimensional and full life cycle and displays the cabinet label information. At the same time, it is combined with the 3D module to perform 3D modeling of the data center structure and cabinet layout. The fault point of the specific cabinet location in the data center can be simulated and displayed remotely and alarmed in real time. At the same time, the weak current routing diagram can be displayed in the modeling to facilitate accurate positioning. (2) The median and variance of signal strength and delay network indicators are calculated by the sliding window method to eliminate data fluctuation interference and obtain a stable decision benchmark. The load of sensor nodes is evaluated by combining CPU utilization, memory status and queue length to achieve accurate load hierarchical management. The weighted decision method is used to predict the trend of network environment changes, perceive quality fluctuations in advance, and provide a forward-looking basis for protocol switching. The LoRa and NB-IoT protocols are intelligently selected based on load status and delay characteristics to balance power consumption and transmission requirements. The release and reallocation of execution bandwidth resources are monitored to avoid resource waste. A signal quality data sharing mechanism is established between device clusters. The protocol selection is unified through the weighted algorithm to reduce communication delay caused by protocol mismatch between devices. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a multi-mode cabinet intelligent management terminal for narrowband IoT communication according to the present invention. Figure 2 This is a flowchart of the environmental monitoring module of the present invention; Figure 3 This is a flowchart of the load sensing module of the present invention; Figure 4 This is a flowchart of the protocol switching module of the present invention; Figure 5 This is a flowchart of the wide scheduling module of the present invention; Figure 6This is a flowchart of the collaborative working module of the present invention; Figure 7 This is a flowchart of the terminal process of the present invention. Detailed Implementation

[0019] The following detailed description, in conjunction with the accompanying drawings, illustrates a specific implementation of the present invention: a multi-mode cabinet intelligent management terminal for narrowband Internet of Things communication.

[0020] Please see Figure 1 A multi-mode cabinet intelligent management terminal for narrowband IoT communication includes: an environmental monitoring module, a load sensing module, a protocol switching module, a bandwidth scheduling module, and a collaborative work module; The environmental monitoring module collects environmental data in real time through built-in sensors, preprocesses the collected data, and obtains key network performance indicators such as network signal strength, latency, packet loss rate, and throughput. Based on the environmental data obtained by the environmental monitoring module, the median and variance of each data item are calculated using the sliding window method. After the data is preprocessed, the environmental data is obtained. The load sensing module collects device load data in real time based on environmental data, and combines it with parameters such as CPU utilization, memory usage, queue length, and bandwidth usage to evaluate the load status of IoT sensor nodes. Based on the load data threshold, the sensor nodes are classified according to their load. If the load is below the set threshold, the existing protocol is retained and bandwidth optimization is performed. If the sensor node load is above the set threshold, it is marked as high load, and load data and environmental data are generated. The protocol switching module dynamically evaluates network environment and device load changes based on load data and environmental data, uses a weighted decision method to predict future network environment change trends, and selects the optimal protocol based on the protocol switching strategy. If the IoT gateway has a high load and long latency, the low-bandwidth, low-power protocol LoRa is selected; if the network quality is good, the high-bandwidth protocol NB-IoT is selected. The switched protocol information is fed back for resource scheduling. The bandwidth scheduling module optimizes bandwidth allocation based on resource scheduling. It automatically adjusts bandwidth resources according to the device load status. If a device is under high load, the bandwidth resource allocation is reduced to prioritize the bandwidth needs of high-priority devices. By monitoring the network status in real time, the module releases and redistributes bandwidth resources. Under low load conditions, the module releases bandwidth resources and redistributes them to devices with higher loads to obtain bandwidth resource information. Based on bandwidth resource information, the collaborative working module adjusts the bandwidth allocation priority among IoT device clusters. Through low-power communication channels between device clusters, it shares signal quality data, bandwidth utilization data, latency data, etc., and collaboratively selects the most suitable communication protocol. Based on a weighted decision-making mechanism in a multi-device environment, it dynamically compares the protocol selection priority according to the device cluster load status and signal quality, and determines the protocol selection for multiple devices through a weighted algorithm. This optimizes the collaborative working between device clusters in the entire network and reduces the latency caused by protocol switching. Environmental data includes signal strength indication, transmission delay, packet loss rate, and network throughput. Load data and environmental data specifically include load classification labels, threshold comparison results, and protocol reservation identifiers. Protocol information includes radio frequency parameters, network bandwidth configuration, and protocol switching instructions. Bandwidth resource information specifically includes resource allocation schemes, priority ranking, and a list of load devices. Device protocol selection includes protocol matching results, weight calculation values, and priority ranking tables. Signal strength indicator: A numerical parameter used to measure the power of received signals in wireless communication, usually quantified in decibels and milliwatts; Transmission delay: refers to the time interval required for data to be transmitted from the sending end to the receiving end, and is one of the core indicators for network performance evaluation; Packet loss rate: This represents the percentage of packets lost during network transmission out of the total number of packets transmitted, and is used to measure the reliability of network transmission. Network throughput: refers to the total amount of data that a network can transmit per unit of time, usually measured in bits or bytes per second; Wireless radio frequency parameters: These include basic configuration parameters of a radio frequency communication system, such as carrier frequency, modulation method, and transmit power. Network bandwidth configuration: refers to the data transmission capacity settings allocated to a specific communication link, which determines the upper limit of data transmission rate; Protocol switching command: A control command used to switch between different communication protocols, containing protocol identifier and switching timing information; Please see Figure 2 The environmental monitoring module includes an environmental data acquisition submodule, a signal evaluation submodule, and a preprocessing submodule. The environmental data acquisition submodule detects the operating status of the built-in sensors, reads the voltage and current values ​​of the sensor interfaces, collects the ambient temperature, humidity percentage, and lux value of light intensity, monitors the status of the network connection port, measures the decibel value of the network signal strength, records the data packet transmission time interval, counts the number of packet loss events, calculates the number of bytes transmitted per unit time, and encodes the environmental parameters and network performance parameters according to a unified data format to generate a standardized environmental dataset. Specifically, upon system startup, the system checks the status of the built-in sensors to determine if each sensor is functioning correctly. Next, it monitors the voltage and current values ​​at the sensor interfaces. The standard range for voltage and current is 5V ± 0.1V, and for current, it's 50mA ± 5mA. If these ranges are exceeded, the sensor is considered faulty. Then, it collects ambient temperature, humidity, and light intensity data. Temperature should be between -10°C and 40°C, humidity between 40% and 90%, and light intensity in lux (typically around 500 lx). Next, the system monitors the network port connection status and checks the network signal strength in decibels (dBm). If the signal strength is below -80dBm, it's considered a weak signal. The system records data packet transmission time intervals and packet loss events. If the transmission time exceeds 300ms and the packet loss rate exceeds 5%, an alarm is triggered. The system calculates the number of bytes transmitted per unit time and generates a standardized environmental dataset. The signal evaluation submodule calls the standardized environment dataset to extract the network signal strength in decibels, latency in milliseconds, packet loss rate percentage, and throughput in megabits. The values ​​of each indicator are checked against the boundary of the value range, and abnormal values ​​that exceed the normal range are removed. The rate of change of each indicator in the time series is calculated to obtain the key network performance indicators. Specifically, the system calls a standardized environment dataset to extract data such as network signal strength, latency, packet loss rate, and throughput. Signal strength should be between -90dBm and -30dBm; data outside this range is considered abnormal. Latency is generally between 10ms and 200ms, packet loss rate should typically be less than 2%, and throughput above 5Mbps is considered acceptable; data outside these ranges is marked as abnormal. Then, the system sorts the normal values ​​by time series and calculates the rate of change of each indicator over time. For example, if throughput changes from 10Mbps to 12Mbps, the rate of change is (12-10) / 10 = 20%. Finally, by analyzing these rates of change, key network performance indicators are obtained. The preprocessing submodule sets the sliding window time length parameter based on key network performance indicators, sorts the data in the time window in ascending order according to the numerical value, selects the value in the middle position after sorting as the median, calculates the sum of squares of the differences between each value and the median, divides it by the total number of data to obtain the variance value, and uses the median and variance value as the statistical feature of the time period. The moving average is used to process the statistical feature value to obtain environmental data. Specifically, based on key network performance indicators, a sliding window of 10 seconds is set. Key network performance data within the window, such as throughput and latency, are collected and sorted in ascending order. For example, if the throughput data collected within 10 seconds is 10Mbps, 12Mbps, 15Mbps, 18Mbps, and 20Mbps, the median after sorting is 15Mbps. Then, the difference between each data point and the median is calculated, and the squares of these differences are summed. Assuming the differences are -5Mbps, -3Mbps, 0, 3Mbps, and 5Mbps, the sum of squares is (25+9+0+9+25)=68. Dividing this by the total number of data points, 5, yields a variance of 13.6. Finally, a moving average is used to process the variance and median to generate the environmental data for that time period. Please see Figure 3 The load perception module includes a device monitoring submodule, a status assessment submodule, and a classification and labeling submodule. Based on environmental data, the equipment monitoring submodule reads real-time operating parameters of sensor nodes through the device interface, collects CPU usage percentage, memory usage bytes, task queue pending count, and network bandwidth usage rate, subtracts the CPU usage percentage from the set CPU monitoring benchmark value to obtain the deviation, divides the memory usage bytes by the total capacity to calculate the utilization rate, and establishes a set of load parameters. Specifically, based on environmental data, the system reads sensor data through the device interface. First, it obtains the CPU usage percentage, assuming it is 70%, and subtracts it from the set baseline value (e.g., 50%), with the deviation being 70% - 50% = 20%. Then, it obtains the number of bytes used in memory, assuming the current memory usage is 4GB and the total memory is 8GB, resulting in a memory usage rate of 4 / 8 = 50%. Next, it obtains the task queue count, assuming the current queue count is 30. Finally, it measures the network bandwidth usage rate, assuming it is 60Mbps. These values ​​(deviation, memory usage, queue count, and bandwidth usage rate) are used as parameters to form a load parameter set. The status assessment submodule calls the load parameter set to set assessment coefficients such as CPU load weight factor, memory usage weight factor, queue length weight factor, and bandwidth usage weight factor. It multiplies the CPU usage percentage by the CPU weight factor, the memory usage rate by the memory weight factor, the queue count by the queue weight factor, and the bandwidth usage rate by the bandwidth weight factor. The sum of these four products yields the node's comprehensive load index. The comprehensive index is then compared with the load assessment threshold to generate a node status rating. Specifically, the load parameter set is called, and various weight factors are set. For example, CPU weight is 0.4, memory weight is 0.3, queue length weight is 0.1, and bandwidth weight is 0.2. Then, these values ​​are multiplied by their respective weight factors: CPU utilization percentage 70% × 0.4 = 28%, memory utilization 50% × 0.3 = 15%, queue count 30 × 0.1 = 3, and bandwidth utilization 60Mbps × 0.2 = 12. The sum of these four results yields the node's overall load index: 28% + 15% + 3 + 12 = 58%. This overall index is then compared to the load assessment threshold (assumed to be 60). If the overall index is less than 60, a normal state is generated; if it is greater than or equal to 60, an overloaded state is generated, and a node status rating is produced. The classification and labeling submodule rates the node status. When the rating value is less than the load classification threshold, it maintains the current protocol parameter configuration and adjusts the bandwidth allocation ratio parameter. When the rating value is greater than or equal to the load classification threshold, it assigns a load identification code to the node status field and merges the load data record and the environment data record according to the timestamp to obtain the load data and environment data. Specifically, based on the node status rating, if the node status rating is less than the load classification threshold (assuming the threshold is 50), the current protocol configuration is maintained and the bandwidth allocation ratio is adjusted, for example, the bandwidth allocation is reduced by 10%; if the rating value is greater than or equal to the load classification threshold, a load identifier code (such as "overloaded") is added to the node status field; then, the load data (such as CPU usage, memory usage, etc.) and environmental data (such as temperature, humidity, etc.) are merged according to the timestamp to obtain the load data and environmental data; Please see Figure 4 The protocol switching module includes an environmental assessment submodule, a trend prediction submodule, and a protocol selection submodule. The environmental assessment submodule analyzes the CPU utilization, memory usage (in bytes), and processor temperature of the IoT gateway based on load and environmental data. It also reads the network signal strength (in decibels), data transmission latency, and packet loss event count. It establishes a load threshold reference standard and a network quality evaluation benchmark. The module subtracts the CPU utilization from the load threshold to obtain the load difference, compares the network latency with the latency benchmark, classifies the load status based on the positive or negative load difference, classifies the network quality based on the latency comparison results, and combines the load status category code and the network quality category code to generate a network load assessment level. Specifically, based on load and environmental data, the system acquires real-time data from the IoT gateway, including CPU utilization, memory usage (in bytes), and processor temperature. Assuming CPU utilization is 80%, memory usage is 4GB, total memory is 8GB, and processor temperature is 45°C, the system also acquires network quality data, including signal strength, latency, and packet loss events. Assuming a signal strength of -60dBm, latency of 120ms, and 3 packet loss events, the system then sets load thresholds of 70% (CPU utilization) and 50ms (latency) and calculates the load difference: 80% - 70% = 10%. If the load difference is positive, it is classified as a high-load state. The latency of 120ms is compared with the baseline latency of 50ms, and the result indicates excessive latency. Finally, the load state (high load) and network quality state (high latency) are combined and encoded to generate a network load assessment level. The trend prediction submodule calls the network load assessment level, sets the load weight factor and network weight factor parameters, multiplies the assessment level value with the corresponding weight factor, calculates the change of the weighted value at continuous time points, solves the ratio of the change to the time interval to obtain the rate of change, determines the sign and magnitude of the rate of change, and obtains the network trend coefficient. Specifically, the system retrieves load assessment level data, sets the load weight factor to 0.6, and the network quality weight factor to 0.4. Assuming that at time point 1, the assessment level is "Load Level 3," and at time point 2, the assessment level is "Load Level 4," then the load level change is 4-3=1. Assuming the time interval is 10 minutes, the change amount is 1, and the change rate is 1 / 10=0.1. Next, if the change rate is positive and the change magnitude is greater than 0.05, the network load change trend is considered "intensifying," and a network change trend coefficient of 1 is assigned. If the change rate is negative or the change magnitude is less than 0.05, the trend is considered "weakening." The sign and magnitude of the change rate are then determined to obtain the network change trend coefficient. The protocol selection submodule sets protocol selection judgment rules and threshold boundary parameters based on the network change trend coefficient. When the trend coefficient exceeds the load upper limit threshold and the network latency coefficient exceeds the latency threshold, the LoRa protocol identifier code is matched. When the trend coefficient is lower than the load lower limit threshold and the network quality coefficient exceeds the quality threshold, the NB-IoT protocol identifier code is matched. The protocol identifier code and the switching timestamp are packaged and encapsulated into data, and the protocol switching command is transmitted to the target gateway device to start the device protocol parameter reconfiguration process and establish resource scheduling. Specifically, based on the network change trend coefficient, assuming a load upper limit threshold of 0.8, a latency threshold of 150ms, a load lower limit threshold of 0.2, and a network quality threshold of -65dBm; when the network change trend coefficient is 1 and the latency exceeds 150ms, the LoRa protocol identifier is selected; when the network change trend coefficient is -0.3 and the network quality coefficient is -70dBm, the NB-IoT protocol identifier is selected; the protocol identifier is packaged together with the switching timestamp (assumed to be 2025-09-18 12:30:00), and a switching command is transmitted to the target gateway to initiate the protocol switching process, reconfigure the device protocol parameters, and establish resource scheduling; Please see Figure 5 The bandwidth scheduling module includes a bandwidth load assessment submodule, a bandwidth allocation submodule, and a resource reconfiguration submodule. The bandwidth load assessment submodule is based on resource scheduling. It parses the device's CPU utilization, memory usage (in bytes), and number of network connection ports. It reads the running task count and pending queue length from the system process table, establishes a load status judgment threshold standard and a device type weight coefficient table, performs a subtraction operation between the CPU utilization and the load threshold, classifies the device load status level according to the positive or negative result, divides the memory usage (in bytes) with the device's total memory capacity to obtain the memory usage ratio, and performs a weighted multiplication operation between the load status level code and the memory usage ratio to generate the device load status coefficient. Specifically, based on resource scheduling, the system analyzes the device's real-time CPU utilization, assuming it to be 85%. Then, a load threshold of 75% is set, and the load difference is calculated by subtraction: 85% - 75% = 10%. If the difference is positive, the device is in a high-load state. Next, the system reads the device's memory usage, assuming it's 6GB and the total memory is 8GB. The memory usage ratio is calculated by division: 6GB / 8GB = 75%. Then, a weighted product is performed based on the load status level and the memory usage ratio. If the load status is high (coded as 1), the device load status coefficient is 1 × 75% = 0.75. Finally, the device load status coefficient is obtained. The bandwidth allocation submodule calls the device load status coefficient, reads the device priority identifier code and bandwidth demand capacity parameter. When the load status coefficient value exceeds the load upper limit threshold, it performs a subtraction operation on the bandwidth allocation amount of the device, calculates the difference between the reduced bandwidth value and the device's minimum bandwidth demand parameter, sorts the device priority identifier codes in ascending order according to their values, and allocates the remaining bandwidth resource capacity level by level according to the sorting results to obtain the optimized bandwidth allocation amount. Specifically, the system calls the device load status coefficient, reads the device priority identifier code and bandwidth requirement capacity parameters. Assuming the device load status coefficient is 0.75, the load limit threshold is 0.7, and the device bandwidth requirement is 100Mbps, when the load status coefficient exceeds the load limit, the system subtracts 10Mbps from the device bandwidth, leaving 90Mbps. Then, it calculates the difference between the remaining bandwidth and the device's minimum bandwidth requirement. Assuming the minimum bandwidth requirement is 50Mbps, the difference is 90Mbps - 50Mbps = 40Mbps. Next, the system sorts the devices in ascending order of priority based on their priority identifier codes (e.g., device A has priority 1, device B has priority 2), and finally allocates the remaining 40Mbps of bandwidth to each device according to their priority: 20Mbps to device A, 20Mbps to device B, resulting in the optimized bandwidth allocation. The resource reconfiguration submodule releases the bandwidth resource capacity occupied by a device when the device load status coefficient value is lower than the load lower limit threshold, based on the optimized bandwidth allocation amount. It calculates the division result of the total released bandwidth capacity and the number of devices to be allocated, and redistributes the released bandwidth capacity according to the load status coefficient value. It also updates the device bandwidth configuration parameter table and network routing table data items to establish bandwidth resource information. Specifically, based on the optimized bandwidth allocation, assuming the device load status coefficient is lower than the load threshold (e.g., 0.3), and the device occupies 50Mbps of bandwidth, the system will release this portion of bandwidth resources. Then, the system calculates the total released bandwidth capacity by dividing it by the number of devices waiting to be allocated. Assuming there are 5 devices waiting to be allocated bandwidth, and the total released bandwidth capacity is 50Mbps, the bandwidth allocated to each device is calculated as 50Mbps / 5 = 10Mbps. Next, the system redistributes bandwidth according to the load status coefficient of each device. If device A has a load coefficient of 0.2 and device B has a load coefficient of 0.5, then device A will receive more bandwidth, such as 12Mbps for device A and 8Mbps for device B. Finally, the bandwidth configuration parameter table and network routing table are updated to establish bandwidth resource information. Please see Figure 6 The collaborative work module includes a priority scheduling submodule, a data sharing submodule, and a protocol decision submodule; The priority scheduling submodule, based on bandwidth resource information, reads the IoT device cluster identifier code and current bandwidth allocation value from the system configuration table, obtains the load weight coefficient and communication priority level parameter of each cluster from the device management table, establishes a bandwidth allocation priority sorting benchmark table and adjustment threshold parameter standard, performs a product operation on the device cluster load weight coefficient value and priority level code to obtain a comprehensive score value, sorts the device clusters in descending order according to the comprehensive score value, adjusts the bandwidth allocation ratio value of each cluster by addition and subtraction according to the sorting result, recalculates the bandwidth quota parameter and allocation time interval parameter value of each cluster, and generates a cluster priority configuration table. Specifically, based on bandwidth resource information, the device cluster identifier code and bandwidth allocation data are read from the system configuration table. For example, cluster A is allocated 100Mbps bandwidth and cluster B is allocated 150Mbps bandwidth. Then, the load weight and priority level of the clusters are obtained from the device management table. Assuming that the weight of cluster A is 0.8 and the priority is 1, and the weight of cluster B is 0.6 and the priority is 2, a comprehensive score is calculated. The score of cluster A is 0.8 × 1 = 0.8, and the score of cluster B is 0.6 × 2 = 1.2. Then, the clusters are sorted in descending order of their scores, with cluster B having a higher priority than cluster A. Based on the sorting, the bandwidth allocation ratio of cluster B is increased to 160Mbps, and cluster A is adjusted to 90Mbps. The cluster priority configuration table is then recalculated and generated. The data sharing submodule calls the cluster priority configuration table, establishes a low-power communication channel connection between device clusters, monitors communication parameters such as signal quality strength in decibels, bandwidth utilization percentage, and network latency in milliseconds for each cluster, classifies and stores the monitoring data according to the device cluster identification code, calculates the rate of change and fluctuation of various parameters in the time series, performs filtering operations on key communication parameters whose rate of change exceeds the fluctuation threshold, and encapsulates the key parameter values ​​and cluster identification codes to obtain a shared communication dataset. Specifically, the system establishes a low-power communication channel connection based on the cluster priority configuration table, and obtains the cluster signal quality strength (e.g., cluster A signal strength is -70dBm, cluster B is -65dBm), bandwidth utilization (cluster A is 70%, cluster B is 85%), and network latency (cluster A is 120ms, cluster B is 100ms). Next, the system stores these communication parameters according to the device cluster identifier code, and then calculates the rate of change and fluctuation amplitude of various parameters. For example, if the signal strength of cluster A changes from -70dBm to -65dBm within 5 seconds, the rate of change is (-65 - (-70)) / (-70) = 0.0714. If the rate of change exceeds a threshold (e.g., 0.05), the parameter is considered a critical parameter. Finally, these critical communication parameters are encapsulated and processed together with the cluster identifier to generate a shared communication dataset. The protocol decision submodule sets the weight allocation matrix parameters based on the shared communication dataset and the weighted decision mechanism in a multi-device environment. It extracts the load status values ​​and signal quality level codes of the device cluster, uses a weighted algorithm to multiply the load status values ​​with the corresponding weight coefficients to calculate the weighted load value, compares the performance evaluation scores of different protocols in the current network environment, determines the protocol selection sequence for each device cluster based on the evaluation scores, and combines the protocol selection sequence with the switching time parameters to reduce the latency caused by protocol switching. Specifically, the system extracts the device cluster load status values ​​and signal quality levels from the shared communication dataset (e.g., cluster A has a load status of 0.7 and a signal quality of 3, while cluster B has a load status of 0.5 and a signal quality of 4). It then sets a weighting matrix (e.g., load status weight is 0.6, and signal quality weight is 0.4). Next, it performs a weighted calculation on the load status and signal quality of each cluster. The weighted load value for cluster A is 0.7 × 0.6 + 3 × 0.4 = 2.22, and the weighted load value for cluster B is 0.5 × 0.6 + 4 × 0.4 = 2.1. Then, it compares the performance evaluation scores of different protocols (e.g., LoRa and NB-IoT) in the current network environment, assuming LoRa's score is 0.8 and NB-IoT's score is 0.9. Based on the scores, it selects the NB-IoT protocol and combines the protocol selection sequence with the switching time (e.g., 2025-09-18 12:30:00) to reduce latency caused by protocol switching.

[0021] A multi-mode cabinet intelligent management terminal for narrowband IoT communication also includes a sensor layer, a data acquisition and transmission layer, a cloud platform layer, and an application layer.

[0022] Please see Figure 7The sensor layer includes temperature sensors, humidity sensors, smoke sensors, access control sensors, current and voltage sensors, and an environmental monitoring module. Each sensor monitors the cabinet's internal temperature, humidity, smoke levels, door open / closed status, and equipment power status. The environmental monitoring module collects environmental parameters such as network signal strength, latency, and packet loss rate. The data acquisition and transmission layer includes a data acquisition module, a load sensing module, a protocol switching module, and an NB-IoT communication module. The data acquisition module collects and processes sensor data, while the load sensing module monitors system CPU and memory usage and performs load assessment. The protocol switching module intelligently switches between NB-IoT, LoRa, and GPRS protocols based on the network environment and load status. NB-IoT communication... The module transmits data to the cloud platform; the cloud platform layer includes a data storage unit, a data analysis and processing unit, an alarm management unit, a remote control unit, and a bandwidth scheduling module. The data storage unit stores rack data, the data analysis and processing unit performs big data analysis, the alarm management unit performs threshold monitoring and automatic alarms, the remote control unit supports remote device operation, and the bandwidth scheduling module optimizes resource allocation; the application layer includes a monitoring center, a mobile application, and a report generation module. The monitoring center provides a visual monitoring interface, the mobile application supports viewing and alarm reception on mobile devices, and the report generation module regularly generates operational reports to achieve multi-dimensional, full-lifecycle management of all information, operating status, location tracking, and service life of data center assets, as well as display of rack label information.

[0023] Furthermore, a multi-mode cabinet intelligent management terminal for narrowband IoT communication also includes a 3D module. This module performs simulated 3D modeling of the server room structure and cabinet layout, allowing for remote simulation and real-time alarm display of fault locations in specific cabinets. The model also displays a diagram of low-voltage wiring connections for accurate location. Additionally, a work order management module is included, working with the data acquisition and transmission layer to establish a database containing authorized operations for rapid matching during login. Task triggering enables online timing of maintenance tasks, achieving remote online real-time closed-loop management. System linkage via work orders ensures tasks are sent to designated cabinets and verified by designated personnel before implementation, effectively recording task completion and achieving remote online real-time closed-loop management from task issuance to completion.

[0024] Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention and not intended to limit it. Those skilled in the art should understand that modifications or equivalent substitutions can be made to the specific embodiments of the present invention, but such modifications or alterations are all within the scope of protection of the pending claims.

Claims

1. A multi-mode cabinet intelligent management terminal for narrowband Internet of Things (IoT) communication, characterized in that, The intelligent management terminal includes an environmental monitoring module, a load sensing module, a protocol switching module, and a bandwidth scheduling module; The environmental monitoring module collects environmental data through built-in sensors, obtains key indicators such as network signal strength, latency, and packet loss rate, and uses the sliding window method to calculate the median and variance of the data to obtain environmental data. Based on the environmental data, the load sensing module collects CPU usage and memory usage load data, evaluates the load status of IoT sensor nodes, classifies sensor nodes based on thresholds, and generates load data and environmental data. The protocol switching module dynamically evaluates changes in network environment and device load based on the load data and environmental data, uses a weighted decision method to predict the trend of network environment changes, and selects the optimal protocol based on the protocol switching strategy. If the IoT gateway has a high load and long latency, the low-bandwidth, low-power protocol LoRa is selected; if the network quality is good, the high-bandwidth protocol NB-IoT is selected. The module also provides feedback on the protocol information after the switch and performs resource scheduling. The bandwidth scheduling module optimizes bandwidth allocation based on resource scheduling, automatically adjusts bandwidth resources, and releases and reallocates bandwidth resources by monitoring network conditions in real time, thereby obtaining bandwidth resource information.

2. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 1, characterized in that, The environmental data includes signal strength indication, transmission delay, packet loss rate, and network throughput. The load data and environmental data specifically include load classification labels, threshold comparison results, and protocol reservation identifiers. The protocol information includes radio frequency parameters, network bandwidth configuration, and protocol switching instructions. The bandwidth resource information specifically includes resource allocation schemes, priority sorting, and a list of load devices.

3. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 2, characterized in that, The environmental monitoring module includes an environmental data acquisition submodule, a signal evaluation submodule, and a preprocessing submodule. The environmental data acquisition submodule detects the operating status of the built-in sensors, reads the voltage and current values ​​of the sensor interfaces, collects the ambient temperature, humidity percentage, and lux value of light intensity, monitors the status of the network connection port, measures the decibel value of the network signal strength, records the data packet transmission time interval, counts the number of packet loss events, calculates the number of bytes transmitted per unit time, and encodes the environmental parameters and network performance parameters according to a unified data format to generate a standardized environmental dataset. The signal evaluation submodule calls the standardized environment dataset to extract the network signal strength in decibels, latency in milliseconds, packet loss rate percentage, and throughput in megabits. The values ​​of each indicator are checked against the boundary of the value range, and abnormal values ​​that exceed the normal range are removed. The rate of change of each indicator in the time series is calculated to obtain the key network performance indicators. The preprocessing submodule sets a sliding window time length parameter based on the key network performance indicators, sorts the data within the time window in ascending order of numerical value, selects the value in the middle position after sorting as the median, calculates the sum of squares of the differences between each value and the median, divides it by the total number of data to obtain the variance value, and uses the median and variance value as statistical features for this time period. The statistical features are then processed by moving average to obtain environmental data.

4. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 3, characterized in that, The load perception module includes a device monitoring submodule, a status assessment submodule, and a classification and labeling submodule. Based on the environmental data, the device monitoring submodule reads the real-time operating parameters of the sensor nodes through the device interface, collects the CPU usage percentage, memory usage bytes, task queue pending count, and network bandwidth usage rate, subtracts the CPU usage percentage from the set CPU monitoring benchmark value to obtain the deviation, divides the memory usage bytes by the total capacity to calculate the utilization rate, and establishes a set of load parameters. The status assessment submodule calls the load parameter set to set evaluation coefficients for CPU load weight factor, memory usage weight factor, queue length weight factor, and bandwidth usage weight factor. It multiplies the CPU usage percentage by the CPU weight factor, the memory usage rate by the memory weight factor, the queue count by the queue weight factor, and the bandwidth usage rate by the bandwidth weight factor. The sum of these four products yields the node's comprehensive load index. The comprehensive index is then compared with the load assessment threshold to generate a node status rating. The classification and labeling submodule rates the node status. When the rating value is less than the load classification threshold, it maintains the current protocol parameter configuration and adjusts the bandwidth allocation ratio parameter. When the rating value is greater than or equal to the load classification threshold, it assigns a load identifier code to the node status field and merges the load data record and the environment data record according to the timestamp to obtain the load data and environment data.

5. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 4, characterized in that, The protocol switching module includes an environmental assessment submodule, a trend prediction submodule, and a protocol selection submodule. The environmental assessment submodule analyzes the IoT gateway's CPU utilization, memory usage (in bytes), and processor temperature based on the load and environmental data. It also reads the network signal strength (in decibels), data transmission latency, and packet loss event count. It establishes a load threshold reference standard and a network quality evaluation benchmark. The module subtracts the CPU utilization from the load threshold to obtain the load difference, compares the network latency with the latency benchmark, classifies the load status based on the positive or negative load difference, classifies the network quality based on the latency comparison results, and combines the load status category code and the network quality category code to generate a network load assessment level. The trend prediction submodule calls the network load assessment level, sets the load weight factor and network weight factor parameters, multiplies the assessment level value with the corresponding weight factor, calculates the change of the weighted value at continuous time points, solves the ratio of the change to the time interval to obtain the rate of change, determines the sign and magnitude of the rate of change, and obtains the network change trend coefficient. The protocol selection submodule sets protocol selection judgment rules and threshold boundary parameters based on the network change trend coefficient. When the trend coefficient exceeds the upper limit threshold and the network latency coefficient exceeds the latency threshold, the LoRa protocol identifier is matched. When the trend coefficient is lower than the lower limit threshold and the network quality coefficient exceeds the quality threshold, the NB-IoT protocol identifier is matched. The protocol identifier and the switching timestamp are packaged and encapsulated, and the protocol switching command is transmitted to the target gateway device to start the device protocol parameter reconfiguration process and establish resource scheduling.

6. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 5, characterized in that, The bandwidth scheduling module includes a bandwidth load assessment submodule, a bandwidth allocation submodule, and a resource reconfiguration submodule. The bandwidth load assessment submodule, based on the resource scheduling, parses the device CPU utilization rate, memory usage in bytes, and number of network connection ports. It reads the running task count and pending queue length from the system process table, establishes a load status judgment threshold standard and a device type weight coefficient table, performs a subtraction operation between the CPU utilization rate and the load threshold, classifies the device load status level according to the positive or negative result, divides the memory usage in bytes and the total memory capacity of the device to obtain the memory usage ratio, and performs a weighted multiplication operation between the load status level code and the memory usage ratio to generate the device load status coefficient. The bandwidth allocation submodule calls the device load status coefficient, reads the device priority identifier code and bandwidth demand capacity parameter. When the load status coefficient value exceeds the load upper limit threshold, it performs a subtraction operation on the bandwidth allocation amount of the device, calculates the difference between the reduced bandwidth value and the device minimum bandwidth demand parameter, sorts the device priority identifier codes in ascending order according to the value, and allocates the remaining bandwidth resource capacity level by level according to the sorting result to obtain the optimized bandwidth allocation amount. The resource reconfiguration submodule, based on the optimized bandwidth allocation, releases the bandwidth resource capacity occupied by the device when the device load status coefficient value is lower than the load lower limit threshold. It calculates the result of the division operation between the total released bandwidth capacity and the number of devices to be allocated, and redistributes the released bandwidth capacity according to the load status coefficient value. It then updates the device bandwidth configuration parameter table and network routing table data items to establish bandwidth resource information.

7. The multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 6, characterized in that, The intelligent management system also includes a collaborative work module; Based on the bandwidth resource information, the collaborative working module adjusts the bandwidth allocation priority among IoT device clusters, shares signal quality data, bandwidth utilization data, and latency data through low-power communication channels between device clusters, collaboratively selects communication protocols, and dynamically compares protocol selection priorities based on a weighted decision-making mechanism in a multi-device environment according to the device cluster load status and signal quality. The weighted algorithm determines the protocol selection for multiple devices, optimizes the collaborative working between device clusters in the entire network, and reduces latency caused by protocol switching.

8. A multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 7, characterized in that, The device protocol selection includes protocol matching results, weight calculation values, and priority sorting tables.

9. A multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 8, characterized in that, The collaborative work module includes a priority scheduling submodule, a data sharing submodule, and a protocol decision submodule; Based on the bandwidth resource information, the priority scheduling submodule reads the IoT device cluster identifier code and the current bandwidth allocation value parameter from the system configuration table, obtains the load weight coefficient and communication priority level parameter of each cluster from the device management table, establishes a bandwidth allocation priority sorting benchmark table and adjustment threshold parameter standard, performs a product operation on the device cluster load weight coefficient value and priority level code to obtain a comprehensive score value, sorts the device clusters in descending order according to the comprehensive score value, adjusts the bandwidth allocation ratio value of each cluster by addition and subtraction according to the sorting result, recalculates the bandwidth quota parameter and allocation time interval parameter value of each cluster, and generates a cluster priority configuration table. The data sharing submodule calls the cluster priority configuration table to establish a low-power communication channel connection between device clusters, monitors communication parameters such as signal quality strength in decibels, bandwidth utilization percentage, and network latency in milliseconds for each cluster, classifies and stores the monitoring data according to the device cluster identification code, calculates the rate of change and fluctuation of various parameters in the time series, performs a filtering operation on key communication parameters whose rate of change exceeds the fluctuation threshold, and encapsulates the key parameter values ​​and cluster identification codes to obtain a shared communication dataset. The protocol decision submodule sets weight allocation matrix parameters based on the shared communication dataset and a weighted decision-making mechanism in a multi-device environment. It extracts the load status values ​​and signal quality level codes of the device cluster, calculates the weighted load value by multiplying the load status values ​​with the corresponding weight coefficients using a weighted algorithm, compares the performance evaluation scores of different protocols in the current network environment, determines the protocol selection sequence for each device cluster based on the evaluation scores, and combines the protocol selection sequence with the switching time parameters to reduce the latency caused by protocol switching.

10. A multi-mode cabinet intelligent management terminal for narrowband IoT communication according to claim 9, characterized in that, The intelligent management terminal also includes a sensor layer, a data acquisition and transmission layer, a cloud platform layer, and an application layer; The sensor layer includes a temperature sensor, a humidity sensor, a smoke sensor, an access control sensor, a current and voltage sensor, and an environmental monitoring module. Each sensor is used to monitor the internal temperature, humidity, smoke status, cabinet door open / closed status, and equipment power status. The environmental monitoring module collects environmental parameters such as network signal strength, latency, and packet loss rate. The data acquisition and transmission layer includes a data acquisition module, a load sensing module, a protocol switching module, and an NB-IoT communication module. The data acquisition module is responsible for collecting sensor data and processing and formatting it. The load sensing module monitors the system's CPU and memory usage and performs load assessment. The protocol switching module intelligently switches between protocols such as NB-IoT, LoRa, and GPRS according to the network environment and load status. The NB-IoT communication module transmits data to the cloud platform. The cloud platform layer includes a data storage unit, a data analysis and processing unit, an alarm management unit, a remote control unit, and a bandwidth scheduling module. The data storage unit stores rack data, the data analysis and processing unit performs big data analysis, the alarm management unit performs threshold monitoring and automatic alarms, the remote control unit supports remote operation of devices, and the bandwidth scheduling module optimizes resource allocation. The application layer includes a monitoring center, a mobile application, and a report generation module. The monitoring center provides a visual monitoring interface, the mobile application supports viewing and receiving alarms on mobile devices, and the report generation module generates operational reports periodically.

Images