Internet of Things gateway control method fusing local communication and AI computing power

Abstract

The invention discloses an Internet of Things gateway control method fusing local communication and AI computing power, and the method comprises the following steps: S1, constructing a multi-protocol adaptation layer for an Internet of Things gateway, recognizing the communication protocol type of access equipment, and building a unified data channel; relates to the technical field of Internet of Things gateway control, and provides an Internet of Things gateway multi-protocol adaptation layer based computing power resource distribution method, device and system, and the Internet of Things gateway multi-protocol adaptation layer based computing power resource distribution method and system improve device access compatibility by constructing a multi-protocol adaptation layer for an Internet of Things gateway, and generate computing power resource quantity distributed to an access device by obtaining the total residual computing power resource quantity of the Internet of Things gateway and based on QoS metadata of the access device. According to the invention, refined computing power resource allocation according to the task characteristics of the access device is realized, and the AI computing power unit is scheduled by creating the AI computing power unit, according to the QoS metadata of the access device and the computing power resource allocation strategy of the access device and through the container technology, so that the gateway of the Internet of Things can cope with different types of AI computing tasks, and the AI computing power unit scheduling efficiency is improved. And diversified processing requirements of the access equipment are met.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of Internet of Things gateway control, and in particular to a control method of an Internet of Things gateway integrating local communication and AI computing power. BACKGROUND

[0002] In the current rapid development of Internet of Things technology, as the key hub connecting the perception layer device and the network layer, the performance and function of the Internet of Things gateway play an important role in the operation of the entire Internet of Things system.

[0003] In the existing Internet of Things environment, there are various types of devices, and their communication protocols are also various. Different manufacturers often use their own unique communication protocols, such as Zigbee, Bluetooth, Wi-Fi, Modbus, and a large number of private protocols. This results in that the existing Internet of Things gateway often cannot establish an effective connection when accessing devices because it cannot identify the communication protocol of some devices, which seriously limits the expansibility of the Internet of Things system and the diversity of device access. The existing Internet of Things gateway adopts a fixed allocation model in terms of computing power resource allocation. This model does not consider the task characteristic differences of the access devices. All access devices obtain computing power resources in a fixed proportion regardless of the priority of the task, the sensitivity to delay, and the energy consumption limit. This makes some tasks with extremely high real-time requirements, such as real-time monitoring and feedback tasks in industrial control, delayed due to the inability to obtain sufficient computing power in time, affecting the normal operation of the system. Some low-power devices with strict energy consumption limits may be allocated too much computing power, causing energy waste and device heating, and also reducing the overall computing power utilization efficiency of the Internet of Things gateway. SUMMARY

[0004] To solve the technical problems in the background art, the application provides a control method of an Internet of Things gateway integrating local communication and AI computing power.

[0005] The control method of the Internet of Things gateway integrating local communication and AI computing power provided by the application comprises the following steps:

[0006] S1, a multi-protocol adaptation layer is constructed for the Internet of Things gateway to identify the communication protocol type of the access device and establish a unified data channel;

[0007] S2, QoS metadata of the access device is obtained; the QoS metadata of the access device includes task priority, maximum allowed delay, and task energy consumption limit; the real-time computing load of the Internet of Things gateway is monitored to generate the computing power resource amount of the access device;

[0008] S3, an AI computing power unit is created, and the AI computing power unit is dispatched through container technology according to the QoS metadata of the access device and the computing power resource allocation strategy of the access device.

[0009] S4, when the real-time computing load of the Internet of Things gateway is monitored to exceed a set threshold, initiating a cooperative computing request between neighboring gateways and distributing sub-tasks to neighboring Internet of Things gateways.

[0010] Preferably, in S1, a multi-protocol adaptation layer is constructed for the Internet of Things gateway, as follows:

[0011] According to the factory built-in basic library of the Internet of Things gateway, a matching database containing protocol feature codes is generated;

[0012] The protocol feature codes include preamble structure, check bit position, and interframe spacing features;

[0013] The matching database containing protocol feature codes is used to identify the communication protocol type of known access devices;

[0014] For the communication protocol type of unknown access devices, based on the matching database containing protocol feature codes, a sliding window comparison method is used to identify the communication protocol type of unknown access devices.

[0015] Preferably, for the matching database containing protocol feature codes, a multi-source mechanism is used to construct, including:

[0016] The factory built-in basic library of the Internet of Things gateway, as the core initial data source for protocol identification, pre-integrates feature code information of mainstream communication protocols in the Internet of Things field, covering Zigbee, Bluetooth, Wi-Fi, and Modbus general protocols, forming a standardized protocol fingerprint library. The factory built-in basic library of the Internet of Things gateway provides basic access capabilities out of the box for the Internet of Things gateway, ensuring fast identification and connection of mainstream devices, and solving the basic compatibility problem of fragmentation of Internet of Things device protocols;

[0017] Through edge computing nodes or cloud management platforms, the protocol feature library is regularly synchronized and updated. The edge nodes can collect communication data of unknown devices in the region in real time, extract potential protocol features through machine learning algorithms, and synchronize to the Internet of Things gateway after cloud verification. The cloud management platform integrates global device protocol data and regularly pushes new or updated protocol feature codes to the Internet of Things gateway, breaking the static limitations of traditional gateway protocol libraries, enabling the gateway to adapt to the rapid iteration of Internet of Things devices, identifying emerging market protocols in a timely manner, avoiding device access failures caused by protocol library lag, and significantly improving the long-term compatibility and technical follow-up capabilities of the system;

[0018] Support administrators to manually enter the characteristic parameters of private protocols through the local configuration interface or remote management tools, form a localized exclusive protocol library, this function gives users the flexibility to extend the gateway protocol adaptation, especially suitable for high security and customized industrial Internet of Things, smart park scenarios, solves the pain point of traditional gateway unable to access enterprise private devices;

[0019] This design database is based on the built-in basic library of the Internet of Things gateway; through edge computing nodes or cloud management platforms, regularly synchronize and update protocol feature libraries for expansion; through the support of administrators to manually enter the characteristic parameters of private protocols through the local configuration interface or remote management tools, form a localized exclusive protocol library for user customization, which contains both general protocol features of the Internet of Things gateway and continuously iterates through edge-cloud collaboration and manual intervention to form scalable protocol recognition capabilities, ensuring compatibility with unknown devices.

[0020] Preferably, in S2, the amount of computing resource of the access device is generated as follows:

[0021] The total amount of remaining computing resources of the Internet of Things gateway is obtained, and the amount of computing resources allocated to the access device is generated based on the QoS metadata of the access device, to realize the allocation of computing resources of the Internet of Things gateway.

[0022] Preferably, in S2, the amount of computing resource of the access device is generated as follows:

[0023] For an access device, let the task priority of the access device be q, the maximum allowed delay be l, the unit of l be second, for example, 0.1 second, and the task energy consumption limit be E;

[0024] Quantize the task priority q to obtain q';

[0025] Let the value of the maximum allowed delay l be l';

[0026] Let the value of the task energy consumption limit E be E', and the value range of E' be 1-100, E' = 100 representing the most relaxed energy consumption limit, and E' = 1 representing the most strict energy consumption limit;

[0027] The resource allocation function F of the access device is:

[0028] In the formula, α, β, and γ are weight coefficients;

[0029] The values of α, β, and γ are set by humans or calculated by entropy weight method;

[0030] Let the number of access devices be n, and the resource allocation function F of each access device be calculated i , i = 1, 2,..., n, to obtain the sum of the resource allocation function values of the n access devices

[0031] For a single access device i, the resource allocation ratio P of the single access device i i is:

[0032] Obtain the total amount of residual computing resources of the Internet of Things gateway, and let the total amount of residual computing resources of the Internet of Things gateway be G, then the amount of computing resources R allocated to the single access device i i is: i R = P i ·G;

[0033] Based on the QoS metadata, a resource allocation function of the access device is constructed, the demand of the task for computing power is quantified, the computing power allocation ratio of each device is dynamically adjusted according to the calculation result of the resource allocation function of the access device, the real-time, energy consumption and task priority demand are balanced, and the resource utilization efficiency and task processing reliability of the Internet of Things gateway are significantly improved.

[0034] Preferably, in S3, the creation of the AI computing unit includes:

[0035] The Internet of Things gateway is built-in with multiple AI model templates; the AI model template includes an image recognition template, a time series prediction template and an audio processing template;

[0036] According to the QoS metadata of the access device and the computing resource allocation strategy of the access device, an AI model template is selected.

[0037] Preferably, in S4, the collaborative computing request between adjacent gateways is initiated and the sub-tasks are distributed to adjacent Internet of Things gateways as follows:

[0038] Suppose the Internet of Things gateway AS whose real-time computing load exceeds the set threshold, in the Internet of Things gateway AS, the amount of computing resources of the access device is sorted in descending order, the computing task of the Internet of Things gateway AS is generated, and the computing task with the smallest amount of computing resources of the access device is set as the sub-task of the Internet of Things gateway AS;

[0039] Obtain the adjacent Internet of Things gateway of the Internet of Things gateway AS, obtain the total amount of residual computing resources of each adjacent Internet of Things gateway, and let the adjacent Internet of Things gateway with the largest total amount of residual computing resources be the Internet of Things gateway AD;

[0040] The sub-task of the Internet of Things gateway AS is distributed to the Internet of Things gateway AD.

[0041] Preferably, in S4, before distributing the sub-tasks to the adjacent Internet of Things gateway, the identity legality of the adjacent Internet of Things gateway is verified through the TEE trusted execution environment; when the sub-tasks are transmitted to the adjacent Internet of Things gateway, homomorphic encryption processing is adopted for the sub-tasks in the transmission.

[0042] The control method of the Internet of Things gateway integrating local communication and AI computing power has the following beneficial technical effects:

[0043] 1. The application improves device access compatibility by constructing a multi-protocol adaptation layer for the Internet of Things gateway. By obtaining the total amount of residual computing resources of the Internet of Things gateway, and based on the QoS metadata of the access device, the amount of computing resources allocated to the access device is generated, realizing refined computing resource allocation according to the task characteristics of the access device. Compared with the fixed allocation model, the utilization efficiency of the Internet of Things gateway computing resources is improved, and the differentiated demand for computing power of different tasks is ensured. By creating an AI computing unit, according to the QoS metadata of the access device and the computing resource allocation strategy of the access device, the AI computing unit is scheduled through container technology, which enables the Internet of Things gateway to handle different types of AI computing tasks, meet the diversified processing needs of access devices, and improve the processing capacity of the Internet of Things gateway.

[0044] 2. The application realizes load balancing between adjacent gateways through collaborative computing request between adjacent gateways. When the real-time computing load of the Internet of Things gateway is monitored to exceed the set threshold, the subtask is allocated to the adjacent Internet of Things gateway, and the allocated subtask is the computing task with the smallest amount of computing resources of the access device. This collaborative computing mechanism realizes load balancing between adjacent gateways, effectively alleviates the computing pressure of a single gateway, improves the computing performance and stability of the entire Internet of Things gateway system, and ensures that the system can still operate efficiently under high load. At the same time, before allocating the subtask, the identity legality of the adjacent Internet of Things gateway is verified through the TEE trusted execution environment, and homomorphic encryption processing is adopted during subtask transmission, which guarantees data security and privacy, and further enhances the reliability of the system.

[0045] Additional aspects and advantages of the application will be described in the following description, some of which will become apparent from the following description, or will be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 The flowchart of the control method of the Internet of Things gateway integrating local communication and AI computing power of the application. DETAILED DESCRIPTION

[0047] The embodiments of the application are described in detail below, examples of which are shown in the drawings, wherein the same or similar symbols represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below by reference to the drawings are exemplary and are only used to explain the application, and cannot be understood as a limitation on the application.

[0048] As Figure 1 shown, a control method of an Internet of Things gateway integrating local communication and AI computing power includes the following steps:

[0049] S1, a multi-protocol adaptation layer is constructed for the Internet of Things gateway, a communication protocol type of an access device is identified, and a unified data channel is established;

[0050] In an optional embodiment, in S1, a multi-protocol adaptation layer is constructed for the Internet of Things gateway, as follows:

[0051] According to a built-in basic library of the Internet of Things gateway, a matching database containing protocol characteristic codes is generated; the protocol characteristic codes include preamble structure, check bit position, and frame interval characteristics;

[0052] The matching database containing protocol characteristic codes is used to identify the communication protocol type of a known access device;

[0053] For the communication protocol type of an unknown access device, a sliding window comparison method is used to identify the communication protocol type of the unknown access device based on the matching database containing protocol characteristic codes;

[0054] The built-in basic library of the Internet of Things gateway is used as the core initial data source for protocol identification, and the characteristic code information of mainstream communication protocols in the Internet of Things field is integrated in advance, covering Zigbee, Bluetooth, Wi-Fi, and Modbus general protocols, forming a standardized protocol fingerprint library. The built-in basic library of the Internet of Things gateway provides basic access capabilities out of the box for the Internet of Things gateway, ensures fast identification and connection of mainstream devices, and solves the basic compatibility problem of fragmentation of Internet of Things device protocols;

[0055] Zigbee is an existing short-range, low-power, and low-cost wireless communication technology based on IEEE 802.15.4 standard, used for interconnection between Internet of Things devices;

[0056] Wi-Fi is an existing wireless local area network technology based on IEEE 802.11 standard, allowing devices to access the Internet or local area network through wireless means to achieve data transmission;

[0057] Modbus is an existing serial communication protocol applied in the field of industrial automation, used for communication between industrial devices such as sensors and frequency converters;

[0058] Through the edge computing node or the cloud management platform, the protocol feature library is updated regularly. The edge node can collect the communication data of unknown devices in the region in real time, extract potential protocol features through machine learning algorithms, and synchronize to the Internet of Things gateway after verification by the cloud. The cloud management platform integrates global device protocol data and regularly pushes new or updated protocol feature codes to the Internet of Things gateway, breaking the static limitations of traditional gateway protocol libraries. This allows the gateway to adapt to the rapid iteration of Internet of Things devices, identify emerging market protocols in a timely manner, avoid device access failures caused by protocol library lag, and significantly improve the long-term compatibility and technology follow-up capabilities of the system.

[0059] The administrator can manually enter the characteristic parameters of the private protocol through the local configuration interface or remote management tool, forming a localized exclusive protocol library. This function gives users the flexibility to extend the gateway protocol adaptation, especially for industrial Internet of Things and smart park scenarios that require high security and customization. It solves the pain point of traditional gateways that cannot access enterprise private devices.

[0060] This design database is based on the built-in basic library of the Internet of Things gateway. The protocol feature library is updated regularly through the edge computing node or the cloud management platform. The administrator can manually enter the characteristic parameters of the private protocol through the local configuration interface or remote management tool, forming a localized exclusive protocol library for user customization. It contains both general protocol features of the Internet of Things gateway and continuously iterates through edge-cloud collaboration and manual intervention, forming an extensible protocol recognition capability, ensuring compatibility with unknown devices, and effectively solving the access problems caused by the diversity of device protocols in the Internet of Things environment.

[0061] S2, obtain the QoS metadata of the access device; the QoS metadata of the access device includes task priority, maximum allowed delay, and task energy consumption limit; monitor the real-time computing load of the Internet of Things gateway, obtain the total amount of remaining computing resources of the Internet of Things gateway, and generate the amount of computing resources allocated to the access device;

[0062] QoS stands for Quality of Service, which is a series of service guarantee mechanisms provided in network communication or system resource allocation to meet the differentiated needs of different tasks or services.

[0063] In an optional embodiment, in S2, the total amount of remaining computing resources of the Internet of Things gateway is obtained, and based on the QoS metadata of the access device, the amount of computing resources allocated to the access device is generated, realizing the allocation of computing resources of the Internet of Things gateway;

[0064] In S2, the computing resource allocation strategy for the access device is generated as follows:

[0065] For an access device, the task priority of the access device is q, the maximum allowed delay is l, the unit of l is second, for example, 0.1 second, the task energy consumption limit is E;

[0066] The task priority q is quantized to obtain q';

[0067] The value of the maximum allowed delay l is l';

[0068] The value of the task energy consumption limit E is E', the value range of E' is 1-100, E'=100 represents the most relaxed energy consumption limit, and E'=1 represents the most strict energy consumption limit;

[0069] The resource allocation function F of the access device is:

[0070] In the formula, α, β, and γ are weight coefficients;

[0071] The values of α, β, and γ are set by human or calculated by entropy weight method;

[0072] The number of access devices is n, and the resource allocation function F of each access device is calculated i , i=1, 2,..., n, and the sum of the resource allocation function values of the n access devices is obtained

[0073] For a single access device i, the resource allocation proportion P of the single access device i is: i

[0074] The total amount of remaining computing resources of the Internet of Things gateway is obtained, and the total amount of remaining computing resources of the Internet of Things gateway is G, and the amount of computing resources R allocated to the single access device i is: i i R i =P·G;

[0075] Based on the QoS metadata, the resource allocation function of the access device is constructed, the demand of the task for computing power is quantized, the computing power allocation proportion of each device is dynamically adjusted according to the calculation result of the resource allocation function of the access device, the real-time performance, energy consumption, and task priority demand are balanced, and the resource utilization efficiency and task processing reliability of the Internet of Things gateway are significantly improved;

[0076] ​In generating the computing resource amount of the access device, according to the QoS metadata of the access device, the resource allocation function of the access device is set to obtain the computing resource allocation ratio of the access device, and then the total amount of the remaining computing resources of the Internet of Things gateway is determined to determine the amount of the computing resources allocated to the single access device. This allocation method realizes fine computing resource allocation according to the task characteristics of the device, improves the utilization efficiency of the computing resources of the Internet of Things gateway, and guarantees the differentiated demand for computing power of different tasks.

[0077] S3, creating an AI computing unit according to the QoS metadata of the access device and the computing resource allocation strategy of the access device,

[0078] According to the task priority and delay requirement, the AI computing unit is scheduled through the container technology;

[0079] In an optional embodiment, the creation of the AI computing unit includes:

[0080] The Internet of Things gateway is built-in with multiple AI model templates; the AI model template includes an image recognition template, a time series prediction template, and an audio processing template;

[0081] According to the QoS metadata of the access device and the computing resource allocation strategy of the access device, the AI model template is selected;

[0082] The AI computing unit is scheduled through the container technology, which flexibly supports the type of AI computing task, meets the diversified processing needs of the access device, and improves the intelligent processing capability and task adaptability of the gateway;

[0083] The application constructs a multi-protocol adaptation layer for the Internet of Things gateway, improves the device access compatibility, obtains the total amount of the remaining computing resources of the Internet of Things gateway, generates the amount of the computing resources allocated to the access device based on the QoS metadata of the access device, realizes fine computing resource allocation according to the task characteristics of the access device, improves the utilization efficiency of the computing resources of the Internet of Things gateway compared with the fixed allocation model, guarantees the differentiated demand for computing power of different tasks, creates an AI computing unit according to the QoS metadata of the access device and the computing resource allocation strategy of the access device, and schedules the AI computing unit through the container technology, which enables the Internet of Things gateway to cope with different types of AI computing tasks, meet the diversified processing needs of the access device, and improve the processing capability of the Internet of Things gateway;

[0084] S4, when the real-time computing load of the Internet of Things gateway exceeds a set threshold, a cooperative computing request between adjacent gateways is initiated and sub-tasks are allocated to adjacent Internet of Things gateways.

[0085] In an optional embodiment, in S4, the cooperative computing request between adjacent gateways is initiated and the sub-tasks are allocated to the adjacent Internet of Things gateways as follows:

[0086] The Internet of Things gateway whose real-time computing load exceeds the set threshold is set as an Internet of Things gateway AS, in the Internet of Things gateway AS, the computing task of the Internet of Things gateway AS is generated by sorting the computing resource amount of the access device from large to small, and the computing task with the smallest computing resource amount of the access device is set as the subtask of the Internet of Things gateway AS;

[0087] The neighboring Internet of Things gateways of the Internet of Things gateway AS are acquired, the total amount of the residual computing resource of each neighboring Internet of Things gateway is acquired, and the neighboring Internet of Things gateway with the largest total amount of the residual computing resource is set as an Internet of Things gateway AD;

[0088] The subtask of the Internet of Things gateway AS is distributed to the Internet of Things gateway AD;

[0089] In an optional embodiment, in S4, before the subtask is distributed to the neighboring Internet of Things gateway, the identity legitimacy of the neighboring Internet of Things gateway is verified by the TEE trusted execution environment; and the homomorphic encryption processing is adopted for the subtask in the transmission when the subtask is transmitted to the neighboring Internet of Things gateway.

[0090] The TEE trusted execution environment is an existing security technical solution based on the combination of hardware and software, which aims to provide a secure execution space for a computing device to protect sensitive data and critical operations from interference and attack by external malicious software and untrusted processes.

[0091] The application cooperates with the computing request between the neighboring gateways, distributes the subtask to the neighboring Internet of Things gateway when the real-time computing load of the monitoring Internet of Things gateway exceeds the set threshold, and the distributed subtask is the computing task with the smallest computing resource amount of the access device. This cooperative computing mechanism realizes the load balancing between the neighboring gateways, effectively relieves the computing pressure of the single gateway, improves the computing performance and stability of the entire Internet of Things gateway system, and ensures that the system can still operate efficiently under high load. At the same time, the identity legitimacy of the neighboring Internet of Things gateway is verified by the TEE trusted execution environment before the subtask is distributed, and the homomorphic encryption processing is adopted when the subtask is transmitted, which guarantees the data security and privacy and further enhances the reliability of the system.

[0092] Meanwhile, the contents not described in detail in the specification all belong to the prior art known by those skilled in the art.

[0093] In the embodiments provided by the present application, it should be understood that the disclosed system or method can be implemented in other ways. For example, the above-described embodiments are merely illustrative, and the division of modules is merely a logical function division. In actual implementation, there can be another division way.

[0094] The modules described as separate components may or may not be physically separate, and the components displayed as modules may or may not be physical modules, and may be located in one place or distributed to multiple network modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment.

[0095] In addition, each functional module in each embodiment of the present application can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module. The integrated module can be realized in the form of hardware or in the form of hardware plus software functional module.

[0096] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the essential characteristics of the present application.

[0097] The above is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited to this. Any skilled person in the art can make equivalent replacements or changes according to the technical solution and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A control method for an IoT gateway integrating local communication and AI computing power, characterized in that: The following steps are involved: S1. Build a multi-protocol adaptation layer for the IoT gateway to identify the communication protocol type of the access device and establish a unified data channel; S2. Obtain QoS metadata of the access device; the QoS metadata of the access device includes task priority, maximum allowed delay, and task energy consumption limit; monitor the real-time computing load of the IoT gateway and generate the computing power resource amount of the access device; S3. Create an AI computing unit and schedule it using container technology based on the QoS metadata of the access device and the computing resource allocation policy of the access device. S4. When the real-time computing load of the monitored IoT gateway exceeds the set threshold, a collaborative computing request between neighboring gateways is initiated and subtasks are assigned to the neighboring IoT gateways.

2. The control method of the Internet of Things gateway integrating local communication and AI computing power according to claim 1 is characterized in that: In S1, a multi-protocol adaptation layer is built for the IoT gateway as follows: Generate a matching database containing protocol signature codes based on the factory-built-in basic library of the IoT gateway; The protocol signature includes the preamble structure, check bit position, and frame interval characteristics; For the communication protocol type of unknown access devices, a sliding window comparison method is used to identify the communication protocol type of the unknown access device based on a matching database containing protocol feature codes.

3. The control method of the Internet of Things gateway integrating local communication and AI computing power according to claim 2 is characterized in that: The matching database containing protocol signatures is constructed using a multi-source mechanism, including: The IoT gateway has a built-in basic library when it leaves the factory; Regularly update the protocol signature library through edge computing nodes or cloud management platforms; Administrators can manually enter the characteristic parameters of private protocols through the local configuration interface or remote management tools to form a localized exclusive protocol library.

4. The control method of the Internet of Things gateway integrating local communication and AI computing power according to claim 1 is characterized in that: In S2, the computing power resources of the access device are generated as follows: Obtain the total remaining computing power resources of the IoT gateway and generate the computing power resources allocated to the access device based on the QoS metadata of the access device.

5. The control method of the Internet of Things gateway integrating local communication and AI computing power according to claim 1 is characterized in that: In S3, the creation of an AI computing unit includes: The IoT gateway has multiple built-in AI model templates; AI model templates include image recognition templates, time series prediction templates, and audio processing templates; Select an AI model template based on the QoS metadata of the access device and the computing resource allocation policy of the access device.

6. The control method of the Internet of Things gateway integrating local communication and AI computing power according to claim 1 is characterized in that: In S4, a collaborative computing request is initiated between neighboring gateways and subtasks are assigned to neighboring IoT gateways as follows: The IoT gateway whose real-time computing load exceeds the set threshold is designated as the IoT gateway AS. In the IoT gateway AS, the computing resources of the connected devices are sorted from largest to smallest, and computing tasks for the IoT gateway AS are generated. The computing task with the smallest computing resource of the connected device is set as the subtask of the IoT gateway AS. Obtain the neighboring IoT gateways of the IoT gateway AS, obtain the total amount of remaining computing power resources of each neighboring IoT gateway, and set the neighboring IoT gateway with the largest total amount of remaining computing power resources as the IoT gateway AD; Assign the subtasks of the IoT gateway AS to the IoT gateway AD.

7. The control method for an IoT gateway integrating local communication and AI computing power according to claim 1 or 6, characterized in that: In S4, before assigning the subtask to the neighboring IoT gateway, the identity legitimacy of the neighboring IoT gateway is verified through the TEE trusted execution environment; when the subtask is transmitted to the neighboring IoT gateway, homomorphic encryption is used for the subtask in transmission.

Images