Internet of Things equipment management method, computing platform and system

Abstract

The invention relates to the technical field of Internet of Things equipment, in particular to an Internet of Things equipment management method, computing platform and system, and the method comprises the steps: obtaining the state data of the Internet of Things equipment, carrying out the health scoring of the equipment periodically through a neural network model, and detecting the potential fault risk; constructing a task allocation strategy based on the equipment health score, and dynamically transferring a task to health equipment when the equipment has a potential fault risk; detecting the communication state of the potential fault risk equipment; potential fault risk equipment in a normal communication state is degraded into temporary communication equipment, and the health equipment is assisted to complete communication; and optimizing network topology and communication load by using temporary communication equipment based on a dynamic topology optimization algorithm driven by a graph model. According to the method, the reliability, the task completion rate and the network communication performance of the Internet of Things equipment can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of Internet of Things device technology, specifically to an Internet of Things device management method, computing platform, and system. Background Art

[0002] IoT devices typically need to operate continuously for extended periods, undertaking critical tasks such as data collection, information transmission, and control execution. Existing IoT device management technologies face the following challenges in practical applications: After prolonged operation, IoT devices experience performance degradation and increased failure risk. Existing technologies lack effective mechanisms for managing and utilizing these devices at risk of failure. Current management approaches typically involve continuing to operate devices at full capacity until they fail completely, or simply deactivating them. This approach fails to effectively prevent device damage, wastes device resources, and reduces the efficiency and reliability of the entire IoT system.

[0003] In summary, how to effectively detect device status, assess potential failure risks, and improve the overall operating efficiency of the IoT system through reasonable task allocation strategies and communication optimization mechanisms in IoT device management has become an urgent problem to be solved.

[0004] In view of this, the present application proposes an IoT device management method, computing platform, and system. Summary of the Invention

[0005] To achieve the above objectives, this application provides an IoT device management method, computing platform, and system. The specific technical solutions are as follows:

[0006] IoT device management methods, including:

[0007] Obtain IoT device status data, periodically perform health scoring on IoT devices through neural network models, and detect potential failure risks of IoT devices;

[0008] Based on the health scores of IoT devices, a task allocation strategy is constructed, task priorities are calculated, and tasks are allocated according to the task priority order and device health scores. When a potential failure risk is detected in an IoT device, the task of the potential failure risk device is transferred to a healthy device.

[0009] Detect the network communication status of devices with potential failure risks, including communication status indicators such as signal strength, bandwidth utilization, latency, and packet loss rate;

[0010] Degrade potential fault risk devices with normal network communication status to temporary communication devices, and assist healthy devices in communication through the temporary communication devices;

[0011] When an IoT device is downgraded to a temporary communication device, the communication topology is optimized using a dynamic topology optimization algorithm driven by a graph model, and the communication load of the IoT device is optimized using temporary communication nodes.

[0012] Preferably, the IoT device status data is acquired in real time by deploying multiple types of sensors on each IoT device, wherein the IoT device status data includes the temperature value, vibration frequency, operating power, cumulative operating time, operating noise decibel value, and device load rate of the IoT device;

[0013] The working cycle is used as the data collection cycle. The state data of the IoT device is sampled multiple times in each working cycle and the average value is taken to construct a state vector.

[0014] Preferably, the neural network model adopts a dual-output feedforward neural network architecture, the input layer of the neural network model includes 6 neurons, corresponding to the six-dimensional state characteristics of the IoT device state data, and the output layer of the neural network model is divided into two branches: a health score branch and a fault risk branch;

[0015] The health score branch outputs a health score through a Sigmoid activation function, and the fault risk branch outputs a potential fault risk probability through a Sigmoid activation function;

[0016] The neural network model is trained using a historical annotated dataset and a combined loss function, which is the sum of the binary cross entropy loss function of the health score and the binary cross entropy loss of the fault risk.

[0017] Preferably, a multi-dimensional priority evaluation mechanism is constructed based on the urgency, importance and resource requirements of the task, and for each task to be assigned, a priority calculation function is defined to calculate the priority of the task to be assigned;

[0018] Based on the health scores of IoT devices, all IoT devices are sorted from high to low by health scores to form an ordered device list; all tasks to be assigned are sorted from high to low by priority to form an ordered task list;

[0019] Starting from the highest priority task, the pending tasks are assigned to the IoT devices with the highest health scores.

[0020] Preferably, when it is detected that the potential failure risk probability of an IoT device exceeds a preset threshold, it is determined to be a potential failure risk device, and all tasks currently carried by the potential failure risk device are sequentially assigned to healthy devices.

[0021] Preferably, detecting the signal strength of the potential failure risk device, wherein the signal strength is determined by sending a probe data packet to adjacent communication nodes and calculating an average received signal strength;

[0022] Detecting bandwidth utilization of devices with potential failure risks, where the bandwidth utilization is determined by monitoring data transmission of IoT devices and calculating the ratio of the average bandwidth actually used to the theoretical maximum bandwidth capacity;

[0023] Evaluate the communication delay performance of potential failure risk devices through end-to-end round-trip time measurement, where the communication delay performance is determined by sending ICMP echo request packets to key nodes in the network and calculating the average network delay;

[0024] The packet loss rate of the device with potential failure risk is detected, where the packet loss rate is determined by sending a sequence of test data packets and counting the number of successfully received test data packets.

[0025] Preferably, a downgrade operation is performed on the potential failure risk device with a normal network communication status, and the potential failure risk device with a normal network communication status is downgraded to a temporary communication device;

[0026] Allocating independent buffer space for the temporary communication device, and deploying multiple queues in the temporary communication device for data reception, the multiple queues including a high priority queue, a medium priority queue, and a low priority queue;

[0027] Conversion and adaptation of multiple IoT communication protocols are performed on the temporary communication device; and active congestion control and traffic shaping mechanisms are implemented through the temporary communication device. The active congestion control is achieved by monitoring queue occupancy and adjusting the receiving rate, and the traffic shaping mechanism adopts a token bucket algorithm.

[0028] Preferably, the current IoT device network is abstracted into a directed weighted graph structure, wherein the vertex set includes all healthy devices and the temporary communication device, the edge set represents the communication links between the devices, and a comprehensive weight value is assigned to each edge;

[0029] Adopting a topology reconstruction strategy based on the shortest path, when the temporary communication device joins the network, recalculating the optimal communication path between all healthy device pairs;

[0030] Performing traffic scheduling based on the communication capacity and current load status of each network node, and utilizing the temporary communication equipment to share the forwarding pressure of overloaded network nodes;

[0031] Establish a periodic topology evaluation and adjustment mechanism, align the topology update cycle with the IoT device status data collection cycle, and dynamically adjust routing strategies based on performance data.

[0032] An IoT device management system, used in the IoT device management method, comprises: an IoT device evaluation module, a task allocation module, a communication detection module, a temporary communication module, and a communication topology optimization module;

[0033] The IoT device evaluation module is used to obtain IoT device status data, periodically perform health scoring on IoT devices through a neural network model, and detect potential failure risks of IoT devices;

[0034] The task allocation module constructs a task allocation strategy based on the health scores of IoT devices, calculates task priorities, and allocates tasks according to the task priorities and device health scores. When a potential failure risk is detected in an IoT device, the task of the potential failure risk device is transferred to a healthy device.

[0035] The communication detection module detects the network communication status of the potential failure risk device, including communication status indicators such as signal strength, bandwidth utilization, delay and packet loss rate;

[0036] The temporary communication module downgrades a potential fault risk device with normal network communication status to a temporary communication device, and assists healthy devices in communicating through the temporary communication device;

[0037] The communication topology optimization module optimizes the communication topology using a dynamic topology optimization algorithm driven by a graph model when an IoT device is downgraded to a temporary communication device, and optimizes the communication load of the IoT device using temporary communication nodes.

[0038] An IoT device management computing platform, comprising:

[0039] A processor and a memory; the memory stores execution instructions;

[0040] When the execution instruction is executed by the processor, the computing platform executes the Internet of Things device management method.

[0041] Beneficial effects of this application: This application realizes intelligent assessment of equipment health status through a neural network model, which can detect potential faults in advance, avoid sudden equipment failures, and improve system operation stability and reliability.

[0042] This application uses an intelligent task allocation mechanism based on health scores to ensure that critical tasks are performed by high-reliability devices, while implementing fault-preventive task migration to minimize the impact of failures on the business, while reducing the load on the equipment in the future time period to avoid further damage to the equipment.

[0043] This application comprehensively evaluates the communication capabilities of fault-risk devices to avoid misjudgments caused by single-dimensional judgments and provide an accurate decision-making basis for device function degradation and resource reuse.

[0044] This application makes full use of some normally functioning equipment resources and converts them into communication auxiliary nodes, which not only avoids resource waste but also enhances the communication redundancy and reliability of the overall network.

[0045] This application dynamically optimizes the network topology, intelligently adjusts communication paths and load distribution, and ensures efficient network communication performance and overall system stability when device status changes.

[0046] This approach builds an adaptive, highly reliable IoT device management system through a collaborative mechanism of health assessment, intelligent task allocation, device function degradation, and dynamic topology optimization. This approach enables proactive fault prevention, maximized resource utilization, optimized task allocation, and dynamic network adjustments, significantly improving the overall operational efficiency, reliability, and resource utilization of IoT systems, extending device lifespan, and reducing operation and maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Flowchart of the IoT device management method provided for this application;

[0048] Figure 2 Flowchart of equipment status data collection and health prediction provided for this application;

[0049] Figure 3 Flowchart of the task allocation and migration mechanism provided for this application;

[0050] Figure 4 Communication status detection and evaluation flow chart provided for this application;

[0051] Figure 5 A flow chart of temporary communication equipment deployment and optimization provided for this application;

[0052] Figure 6 The network topology optimization and load balancing flow chart provided for this application;

[0053] Figure 7 This is the structural diagram of the Internet of Things device management system provided for this application. DETAILED DESCRIPTION

[0054] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the drawings in the specification.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application can also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without violating the connotation of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present application. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0057] Example 1:

[0058] Reference Figure 1 , which is the first embodiment of this application, provides an Internet of Things device management method.

[0059] Step 1: Obtain IoT device status data, periodically perform health scoring on IoT devices through a neural network model, and detect potential failure risks of IoT devices. Figure 2 , which is the flow chart of equipment status data collection and health prediction in this step.

[0060] Through the multiple types of sensors deployed on each IoT device, the IoT device status data is obtained in real time, including the temperature value, vibration frequency, working power, cumulative operating time, working noise decibel value and equipment load rate of the core components of the IoT device.

[0061] The working cycle is used as the data collection cycle. The state data of the IoT device is sampled multiple times in each working cycle and the average value is taken to construct the state vector ,in, Indicates the The device in The state vector of a working cycle, Indicates the The device in The average temperature of a working cycle, Indicates the The device in The average vibration frequency of a working cycle, Indicates the The device in The average working power of a working cycle, Indicates the The device in The cumulative running time of a working cycle, Indicates the The device in The average working noise decibel value of a working cycle, Indicates the The device in The average load rate of a working cycle. The use of multiple sampling and averaging can reduce the impact of occasional interference on the quality of collected data and improve the accuracy of status assessment.

[0062] A standardization method is used to normalize the collected data to eliminate the impact of different dimensions on model training. For each feature, the mean and standard deviation of the feature are calculated for all devices, and then the feature value of each device is standardized. The standardization process makes all feature values ​​​​distributed in the same numerical range, avoiding certain features with larger values ​​​​dominating the model calculation process, and ensuring that the neural network model can evenly learn the impact of each feature on the health status of the equipment.

[0063] The neural network model uses a dual-output feedforward neural network architecture. The input layer contains six neurons corresponding to six-dimensional state features, and the output layer is divided into two branches: a health score branch and a fault risk branch. The health score branch includes one neuron that uses a Sigmoid activation function to output a health score between 0 and 1. The fault risk branch includes one neuron that also uses a Sigmoid activation function to output a potential fault risk probability between 0 and 1. This dual-output feedforward neural network structure enables the neural network model to simultaneously learn two related but distinct objectives: the current health status of the equipment and future fault trends, thereby improving the model's practicality.

[0064] The neural network model is trained using a historically annotated dataset and a combined loss function: the sum of the binary cross-entropy loss for the health score and the binary cross-entropy loss for the fault risk. Parameters are updated using the Adam optimizer. This combined loss function balances the importance of the two prediction tasks, enabling the neural network model to simultaneously optimize the accuracy of both health score and fault risk prediction.

[0065] The historically labeled dataset is the basis for training the neural network model. For example, a historically labeled dataset is constructed to collect the status data of all IoT devices during operation over a period of time, including temperature, vibration frequency, operating power, cumulative operating time, operating noise decibel value, and device load rate, to form a multi-dimensional state vector. Each data record is manually assessed by a health status label and a fault risk label, where the health status label is represented by a value between 0 and 1. A health status label of 0 indicates that the health status is a damaged state, and a health status label of 1 indicates that the device is in a healthy state as new from the factory. The fault risk label is also represented by a value between 0 and 1. A fault risk label of 0 indicates that the potential failure risk probability of the device in the next working cycle is 0, that is, there is no failure risk at all. A fault risk label of 1 indicates that the potential failure risk probability of the device in the next working cycle is 1, that is, there is a definite damage risk. The health status label and the fault risk label both have values ​​of two decimal places between 0 and 1. Both labels are manually set based on the device's historical maintenance records or preset rules (such as overheating or long-term operation).

[0066] During IoT device operation, various indicators are sampled multiple times within each set data collection cycle. After the final data collection within the cycle, the average data is calculated and processed. A neural network model is used to predict the health status and failure risk of the next cycle. Based on these predictions, the operating status of the IoT device for the next cycle is adjusted at the end of the current operating cycle. For example, a device with a potential failure risk can be restored to a healthy state, or a healthy device can be adjusted to a potential failure risk. This periodic monitoring mechanism continuously identifies the health status and failure risk of IoT devices and dynamically adjusts the status of IoT devices.

[0067] This step achieves simultaneous assessment and prediction of the health status and fault risk of IoT devices through periodic multi-dimensional state collection and standardized processing, combined with a dual-output neural network model.

[0068] Step 2: Based on the health score of the IoT device, build a task allocation strategy, calculate the task priority, and allocate tasks according to the task priority order and the device health score. When a potential failure risk is detected in an IoT device, transfer the task of the potential failure risk device to the healthy device. See Figure 3 , which is the flow chart of the task allocation and migration mechanism of this step.

[0069] Build a multi-dimensional priority evaluation mechanism based on the urgency, importance and resource requirements of the task; for each task to be assigned in the next cycle , for example, define its priority calculation function ,in Indicates tasks to be assigned The urgency coefficient ranges from 0 to 10 and is determined by the difference between the task deadline and the current time; Indicates tasks to be assigned The importance coefficient ranges from 0 to 10 and is preset according to the impact of the task; Indicates tasks to be assigned The resource requirement coefficient, ranging from 1 to 10, reflects the amount of computing and communication resources required to execute the task; 、 、 is the weight coefficient, which is determined by the analytic hierarchy process.

[0070] The design of taking the inverse of the resource demand coefficient gives tasks with smaller resource requirements higher priority, which helps to improve the task completion rate; the multi-dimensional priority calculation method comprehensively considers the timeliness, importance and execution resource requirements of the task, ensuring that critical tasks are given priority while taking into account resource utilization efficiency.

[0071] Establish a dynamic task allocation strategy based on the health score of IoT devices. Based on the device health score of the next work cycle predicted by the current work cycle, sort all IoT devices from high to low according to the health score to form an ordered device list. ,in Indicates the sorted devices, satisfying , After sorting Health score of each device; sort all tasks to be assigned from high to low priority to form an ordered task list ,in Indicates the sorted Tasks to meet , Indicates the sorted The priority of each task; starting from the highest priority task, the tasks to be assigned are assigned to the IoT device with the highest health score currently available. If it is greater than the number of devices n, then after each device is assigned a task, continue to follow the ordered device list. Allocate remaining tasks. Assign tasks of corresponding priority based on health scores to avoid overloading IoT devices with low health scores, which helps to extend the overall service life of IoT devices.

[0072] A task transfer mechanism is constructed to predict the potential failure risk probability of the device in the next working cycle based on the current working cycle. When it is detected that the potential failure risk probability of an IoT device exceeds the preset threshold, the corresponding IoT device will be identified as a potential failure risk device; the task migration process will be started for the potential failure risk device, and all tasks currently carried by all potential failure risk devices will be assigned to healthy devices in turn.

[0073] Continue to collect status data of all IoT devices in the current work cycle, and output health scores and potential failure risk probabilities through the neural network model for task allocation in the new work cycle.

[0074] This step achieves the optimized scheduling of work tasks in the Internet of Things by constructing an intelligent task allocation strategy based on health scores and a task transfer mechanism for fault prevention; the task allocation strategy based on health scores can fully guarantee the normal implementation of tasks, while the task transfer mechanism constructed according to the probability of potential failure risks can effectively avoid task interruptions caused by long-term high-load operation of Internet of Things devices; the combined application of these two mechanisms can significantly improve the reliability and task completion rate of Internet of Things devices during operation, providing strong guarantees for the long-term stable operation of Internet of Things devices.

[0075] Step 3: Check the network communication status of the potential fault risk device, including communication status indicators such as signal strength, bandwidth utilization, latency, and packet loss rate. Figure 4 , which is the communication status detection and evaluation flow chart of this step.

[0076] Evaluate the communication quality of all identified potential failure risk devices and establish a communication monitoring queue ,in Indicates the A potential failure risk device that needs to detect the network communication status, Indicates the total number of devices with potential failure risk, Each device with a potential fault risk to be tested is assigned a dedicated communication test time slot to avoid network congestion caused by simultaneous communication testing of multiple devices. A time-division multiplexing test scheduling mechanism ensures that each device's network communication status test results truly reflect its actual communication capabilities, providing an accurate basis for subsequent device role transition decisions.

[0077] Detect the signal strength of the potential fault risk device, obtain the received signal strength indicator value through the communication device in the IoT device, and , during the test period, to its adjacent Each communication node sends a detection data packet, records the signal strength indicator value returned by each adjacent communication node, and calculates the average received signal strength ,in Indicates a device with a potential risk of failure With the The received signal strength indicator value when the adjacent communication nodes communicate. Set the signal strength threshold ,when When the signal strength of the potential fault risk device meets the communication requirements.

[0078] Detect bandwidth utilization of devices with potential failure risks, monitor data transmission of devices with potential failure risks, and define bandwidth utilization ,in Indicates a device with a potential risk of failure During the monitoring period The average bandwidth actually used within Indicates the theoretical maximum bandwidth capacity of a device with potential failure risk. The actual average bandwidth used is calculated by continuously collecting the data sending and receiving rates of the device during the monitoring period. ,in and Respectively The data sending and receiving rates at the moment.

[0079] Setting bandwidth utilization thresholds ,when When the potential fault risk device is determined to have sufficient bandwidth margin to perform the communication task; the bandwidth evaluation method based on actual usage accurately reflects the communication load status of the device, ensuring that the selected temporary communication device has sufficient transmission capacity.

[0080] The communication delay performance of potential failure risk devices is evaluated by end-to-end round-trip time measurement. Perform a standard ping test, sending ICMP echo request packets to key nodes in the network and recording the round-trip time for each test node , calculate the average network delay ,in Indicates the test node number, Indicates the total number of test nodes.

[0081] To improve measurement accuracy, each test node performs multiple ping tests and takes the average after removing the maximum and minimum values. ,when When the network delay of the potential fault risk device is determined to meet the real-time communication requirements, the multi-node, multiple-test approach can effectively reduce the interference of network jitter on delay evaluation and ensure the accuracy of delay measurement results.

[0082] Detect the packet loss rate of potential fault risk devices, evaluate the link reliability by sending test data packet sequences, and , continuously sending to the preset test server Test data packets with sequence numbers, counting the number of test data packets successfully received , calculate the packet loss rate .

[0083] For example, the test data packet size can be set to 1KB and the sending interval to 10ms to simulate the data transmission mode of the actual application scenario. ,when When the communication link reliability of the potential fault risk device is determined to meet the communication requirements, the packet loss rate detection method based on large sample statistics can accurately assess the link quality and avoid the influence of occasional packet loss on the judgment result.

[0084] By integrating the test results of the four communication status indicators, when all four network communication status indicators meet the requirements, the network communication status of the potential fault risk device is determined to be normal and can be converted into a temporary communication device. The constructed network communication status evaluation mechanism can ensure that only devices with fully qualified communication performance can undertake data transfer tasks, ensuring high-quality network services.

[0085] This step comprehensively checks the network communication status of potential risk devices, including precise measurement and evaluation of four key metrics: signal strength, bandwidth utilization, latency, and packet loss rate. This process selects potential risk devices with normal communication capabilities as candidates for temporary communication nodes. This multi-dimensional communication quality assessment ensures that the selected potential risk devices possess reliable data transmission capabilities and fully utilizes their remaining communication value, maximizing device resource utilization and improving resource efficiency and network resilience across the IoT.

[0086] Step 4: Degrade the potential fault risk device with normal network communication status to a temporary communication device, and assist the healthy device to communicate through the temporary communication device; see Figure 5 , which is the temporary communication equipment deployment and optimization flow chart of this step.

[0087] Perform downgrade operations on potential fault risk devices with normal network communication status and establish a temporary communication device set ,in Indicates the nodes that have been downgraded to temporary communication devices, Indicates the total number of temporary communication devices. Update the device status flag, set its task processing function to sleep, release computing resources and reduce energy consumption, and activate the communication forwarding function; allocate independent cache space for each temporary communication device and set the cache capacity Used to temporarily store data to be forwarded. By clearly switching functions, temporary communication devices can be focused on communication relay tasks, preventing computing loads from interfering with communication performance.

[0088] Deploy multiple queues in the temporary communication device for data reception. When the healthy device needs to forward data through the temporary communication device, the temporary communication device First, the destination address and priority tag in the packet header are checked.

[0089] Exemplarily, a three-level priority queue structure is established, including: a high priority queue Used to store emergency control instructions and alarm information, medium priority queue Used to store regular business data, low priority queue Used to store logs and statistics. Each queue uses a first-in-first-out policy, but the high-priority queue enjoys scheduling priority. When a data packet is received, it is marked according to the priority of the data packet. Add it to the corresponding queue, for example, if Then join ,like Then join Otherwise, add The establishment of a hierarchical queue mechanism ensures the priority transmission of critical data and can enhance the differentiation capabilities of communication services.

[0090] Temporary communication equipment uses intelligent forwarding scheduling algorithm to process cached data and define the forwarding time slot length At the beginning of each time slot, the forwarding scheduler checks the three queues in order of priority. If the high priority queue is not empty, it will be extracted first. The data packets in are forwarded; if Empty but If the queue is not empty, medium-priority data is processed; low-priority data is processed only when both the high-priority and medium-priority queues are empty. For each packet to be forwarded, the temporary communication device performs an integrity check. If the check passes, the packet's hop count and timestamp are updated before it is sent to the destination device via the wireless interface. Priority-based intelligent scheduling ensures the timely and reliable transmission of important data.

[0091] Perform conversion and adaptation of multiple IoT communication protocols on temporary communication devices. When the source device and the target device use different communication protocols, the temporary communication device performs the protocol conversion function. Maintain the protocol mapping table ,in Indicates the source protocol type. Indicates the target protocol type, Indicates the corresponding conversion function. For example, when receiving data based on the CoAP protocol and needs to be forwarded to a device using the MQTT protocol, the temporary communication device extracts the payload of the CoAP message and repackages it into the MQTT format, including adding a topic field, a QoS level, and a reservation flag.

[0092] The communication protocol conversion process maintains the integrity of the data content and only modifies the protocol-related control fields. Performing communication protocol conversion and adaptation on temporary communication devices enables heterogeneous IoT devices to interconnect through temporary communication nodes, reducing the computational overhead of IoT devices performing tasks.

[0093] Active congestion control and traffic shaping mechanisms are implemented through temporary communication devices. When a healthy device associated with the temporary communication device receives downlink traffic, the temporary communication device monitors the real-time transmission rate of the downlink traffic. and queue occupancy , when the queue occupancy is detected Greater than the queue occupancy threshold hour, , start active congestion control and reduce the transmission rate to ,in is the adjustment coefficient.

[0094] Use the token bucket algorithm to shape the downstream traffic and set the token generation rate to , the barrel capacity is Each token is used to set the number of bytes allowed to be sent. Only packets with enough tokens can be forwarded, while packets exceeding the limit are temporarily queued. The active congestion control and traffic shaping mechanisms built into the network prevent network congestion and ensure smooth data transmission.

[0095] This step effectively enhances the communication capabilities of healthy devices by converting potentially faulty devices with normal communication functions into data relay nodes. Temporary communication devices provide flexible and reliable communication assistance services for healthy devices through multi-level priority queues, intelligent forwarding scheduling, protocol adaptation and conversion, active congestion control, and traffic shaping mechanisms. This resource reuse approach not only extends the service lifecycle of potentially faulty devices but also reduces their workload, significantly improving the communication capacity and service quality of the entire IoT, providing an innovative solution for building an efficient and resilient IoT communication architecture.

[0096] Step 5: When an IoT device is downgraded to a temporary communication device, the communication topology is optimized using a dynamic topology optimization algorithm driven by a graph model. The communication load of the IoT device is optimized using temporary communication nodes. Figure 6 , which is the network topology optimization and load balancing flow chart of this step.

[0097] Abstract the current IoT device network into a directed weighted graph structure. Construct a graph , where the vertex set Contains all health equipment and temporary communication equipment, edge set Indicates the reachable communication links between devices.

[0098] Each edge is assigned a comprehensive weight, which takes into account the link's signal strength, transmission delay, and current load. A smaller weight indicates better link quality. After converting the IoT's actual communication network into a graph structure, sophisticated graph algorithms are used for topology analysis and optimization, providing a mathematical foundation for subsequent path planning and load balancing.

[0099] A topology reconstruction strategy based on the shortest path is adopted. When a temporary communication device joins the network, the optimal communication path between all healthy device pairs is recalculated. For device pairs that originally require multi-hop transmission or have poor communication quality, it is evaluated whether a better relay path can be established through the temporary communication device.

[0100] For example, the optimization process follows the following principles: Prioritizing connections between pairs of devices with the longest communication distances, focusing on communication links with more than three hops, and ensuring that each healthy device has at least two independent paths to other devices. By introducing temporary communication nodes as relays, communication paths that would otherwise require detours can be significantly shortened, reducing end-to-end transmission latency and improving data transmission efficiency.

[0101] Traffic scheduling is performed based on the communication capacity and current load status of each network node. Real-time communication load statistics are collected for each network node, including metrics such as data forwarding volume, queue length, and processing delay. When certain network nodes are detected to be overloaded, some data flows are rerouted to less loaded paths. In particular, temporary communication equipment is used to offload the forwarding pressure from overloaded network nodes. Load distribution adheres to the principle of balancing to avoid situations where individual network nodes are overloaded while others are idle. For large data flows that can be transmitted in parallel, these flows are split into multiple sub-flows and transmitted simultaneously over different paths, fully utilizing the additional bandwidth resources provided by temporary communication equipment.

[0102] Establish a periodic topology assessment and adjustment mechanism, set the topology update cycle to be consistent with the IoT device data collection cycle, and collect network performance data in each cycle, including the transmission success rate, average latency, and throughput of each link.

[0103] Routing policies are dynamically adjusted based on performance data, and alternatives are promptly identified for paths experiencing performance degradation. When the communication quality of a temporary communication device changes or a new temporary communication device joins, the topology's synchronized routing table is updated. This update process uses incremental adjustments, modifying only the affected routing entries to avoid network-wide routing oscillations.

[0104] The effectiveness of topology optimization is verified by comparing network performance metrics before and after optimization. Key performance indicators evaluated include: the reduction in average network path length, the degree of reduction in communication latency, the improvement in load balancing, and the increase in network throughput. If the overall performance after optimization is significantly improved (an improvement greater than 10%), the optimization strategy is considered effective and the current topology is maintained. If the improvement is not significant or performance degrades, the system automatically reverts to the pre-optimization configuration. The parameter configuration and performance data for each optimization are recorded to form an optimization experience library, providing a reference for subsequent rapid optimization in similar scenarios.

[0105] This step leverages the relay role of temporary communication devices within the network by building a dynamic topology optimization mechanism driven by a graph model, optimizing communication paths and balancing load distribution. The dynamic topology optimization algorithm can adjust communication routes in real time based on network state changes, ensuring that data transmission always takes the optimal path.

[0106] By rationally utilizing temporary communication nodes, not only does this alleviate the communication pressure on existing health devices, but it also improves the communication capacity and service quality of the entire network. The adaptive topology optimization method in this step provides flexible and efficient communication guarantees for the Internet of Things, significantly improving network performance and enhancing overall operational efficiency and reliability.

[0107] Example 2:

[0108] Reference Figure 7 , which is the second embodiment of this application, provides an Internet of Things device management system.

[0109] The system includes: an Internet of Things device evaluation module, a task allocation module, a communication detection module, a temporary communication module and a communication topology optimization module.

[0110] The IoT device evaluation module is used to obtain IoT device status data, periodically perform health scoring on IoT devices through a neural network model, and detect potential failure risks of IoT devices.

[0111] The task allocation module constructs a task allocation strategy based on the health score of the IoT device, calculates the task priority, and allocates tasks according to the task priority order and the device health score. When a potential failure risk is detected in an IoT device, the task of the potential failure risk device is transferred to a healthy device.

[0112] The communication detection module detects the network communication status of the potential failure risk device, including communication status indicators such as signal strength, bandwidth utilization, delay and packet loss rate.

[0113] The temporary communication module downgrades a potential fault risk device with normal network communication status to a temporary communication device, and assists healthy devices in communicating through the temporary communication device.

[0114] The communication topology optimization module optimizes the communication topology using a dynamic topology optimization algorithm driven by a graph model when an IoT device is downgraded to a temporary communication device, and optimizes the communication load of the IoT device using temporary communication nodes.

[0115] In the embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interface, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0116] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also change, modify, replace and modify the above-mentioned embodiments without departing from the purpose of this application and the scope of protection of the claims, which are all within the protection of this application.

Claims

1. The method for managing IoT devices is characterized in that: include: Obtain IoT device status data, periodically perform health scoring on IoT devices through neural network models, and detect potential failure risks of IoT devices; Based on the health scores of IoT devices, a task allocation strategy is constructed, task priorities are calculated, and tasks are allocated according to the task priority order and device health scores. When a potential failure risk is detected in an IoT device, the task of the potential failure risk device is transferred to a healthy device. Detect the network communication status of devices with potential failure risks, including communication status indicators such as signal strength, bandwidth utilization, latency, and packet loss rate; Degrade potential fault risk devices with normal network communication status to temporary communication devices, and assist healthy devices in communication through the temporary communication devices; When an IoT device is downgraded to a temporary communication device, the communication topology is optimized using a dynamic topology optimization algorithm driven by a graph model, and the communication load of the IoT device is optimized using temporary communication nodes.

2. The method for managing IoT devices according to claim 1, wherein: include: The IoT device status data is acquired in real time through multiple types of sensors deployed on each IoT device. The IoT device status data includes the device's temperature, vibration frequency, operating power, cumulative operating time, operating noise decibel value, and device load rate. The working cycle is used as the data collection cycle. The state data of the IoT device is sampled multiple times in each working cycle and the average value is taken to construct a state vector.

3. The method for managing IoT devices according to claim 2, wherein: The neural network model adopts a dual-output feedforward neural network architecture. The input layer of the neural network model contains 6 neurons, corresponding to the six-dimensional state characteristics of the IoT device status data. The output layer of the neural network model is divided into two branches: a health score branch and a fault risk branch. The health score branch outputs a health score through a Sigmoid activation function, and the fault risk branch outputs a potential fault risk probability through a Sigmoid activation function; The neural network model is trained using a historical annotated dataset and a combined loss function, which is the sum of the binary cross entropy loss function of the health score and the binary cross entropy loss of the fault risk.

4. The method for managing IoT devices according to claim 3, wherein: Build a multi-dimensional priority evaluation mechanism based on the urgency, importance, and resource requirements of the task. For each task to be assigned, define a priority calculation function to calculate the priority of the task to be assigned. Based on the health scores of IoT devices, all IoT devices are sorted from high to low by health scores to form an ordered device list; all tasks to be assigned are sorted from high to low by priority to form an ordered task list; Starting from the highest priority task, the pending tasks are assigned to the IoT devices with the highest health scores.

5. The method for managing IoT devices according to claim 4, wherein: When it is detected that the potential failure risk probability of an IoT device exceeds a preset threshold, it is determined to be a potential failure risk device, and all tasks currently carried by the potential failure risk device are assigned to healthy devices in sequence.

6. The method for managing IoT devices according to claim 5, wherein: Detecting the signal strength of a device with a potential fault risk, wherein the signal strength is determined by sending a probe data packet to adjacent communication nodes and calculating an average received signal strength; Detecting bandwidth utilization of devices with potential failure risks, where the bandwidth utilization is determined by monitoring data transmission of IoT devices and calculating the ratio of the average bandwidth actually used to the theoretical maximum bandwidth capacity; Evaluate the communication delay performance of potential failure risk devices through end-to-end round-trip time measurement, where the communication delay performance is determined by sending ICMP echo request packets to key nodes in the network and calculating the average network delay; The packet loss rate of the device with potential failure risk is detected, where the packet loss rate is determined by sending a sequence of test data packets and counting the number of successfully received test data packets.

7. The method for managing IoT devices according to claim 6, wherein: Perform a downgrade operation on a potential fault risk device with normal network communication status, and downgrade the potential fault risk device with normal network communication status to a temporary communication device; Allocating independent buffer space for the temporary communication device, and deploying multiple queues in the temporary communication device for data reception, the multiple queues including a high priority queue, a medium priority queue, and a low priority queue; Conversion and adaptation of multiple IoT communication protocols are performed on the temporary communication device; and active congestion control and traffic shaping mechanisms are implemented through the temporary communication device. The active congestion control is achieved by monitoring queue occupancy and adjusting the receiving rate, and the traffic shaping mechanism adopts a token bucket algorithm.

8. The method for managing IoT devices according to claim 7, wherein: The current IoT device network is abstracted into a directed weighted graph structure, where the vertex set includes all healthy devices and the temporary communication device, the edge set represents the communication links between devices, and a comprehensive weight value is assigned to each edge; Adopting a topology reconstruction strategy based on the shortest path, when the temporary communication device joins the network, recalculating the optimal communication path between all healthy devices; Performing traffic scheduling based on the communication capacity and current load status of each network node, and utilizing the temporary communication equipment to share the forwarding pressure of overloaded network nodes; Establish a periodic topology evaluation and adjustment mechanism, align the topology update cycle with the IoT device status data collection cycle, and dynamically adjust routing strategies based on performance data.

9. An Internet of Things device management system, which is used to implement the Internet of Things device management method according to any one of claims 1 to 8, characterized in that: include: IoT device evaluation module, task allocation module, communication detection module, temporary communication module, and communication topology optimization module; The IoT device evaluation module is used to obtain IoT device status data, periodically perform health scoring on IoT devices through a neural network model, and detect potential failure risks of IoT devices; The task allocation module constructs a task allocation strategy based on the health scores of IoT devices, calculates task priorities, and allocates tasks according to the task priorities and device health scores. When a potential failure risk is detected in an IoT device, the task of the potential failure risk device is transferred to a healthy device. The communication detection module detects the network communication status of the potential failure risk device, including communication status indicators such as signal strength, bandwidth utilization, delay and packet loss rate; The temporary communication module downgrades a potential fault risk device with normal network communication status to a temporary communication device, and assists healthy devices in communicating through the temporary communication device; The communication topology optimization module optimizes the communication topology using a dynamic topology optimization algorithm driven by a graph model when an IoT device is downgraded to a temporary communication device, and optimizes the communication load of the IoT device using temporary communication nodes.

10. IoT device management computing platform, characterized by: include: processor; a memory storing execution instructions; When the execution instruction is executed by the processor, the computing platform executes the Internet of Things device management method according to any one of claims 1 to 8.

Images