A method and system for terminal automatic network selection using artificial intelligence

Abstract

The present application relates to the technical field of mobile communication, and particularly relates to a method and system for realizing automatic network selection of a terminal by using artificial intelligence. After starting up or detecting network drop, a user network selection module of the terminal collects input parameters related to network selection; the user network selection module sends the collected input parameters to an artificial intelligence entity and requests to obtain a network selection strategy; the artificial intelligence entity performs artificial intelligence reasoning according to the received input parameters, generates network selection strategy data, and returns the network selection strategy data to the user network selection module; the user network selection module performs a network selection operation according to the received network selection strategy data, and successfully or unsuccessfully registers to a target network; the artificial intelligence entity records and stores result parameters of the network selection operation, which are used for subsequent artificial intelligence learning. The artificial intelligence entity can continuously train and learn based on the result of each network selection operation, constantly optimize the network selection model, and improve the accuracy of future network selection decisions.

Description

Technical Field

[0001] This invention relates to the field of mobile communication technology, and specifically to a method and system for enabling automatic network selection by a terminal. Background Technology

[0002] As the mobile communications industry has evolved into the 5G era, the internal implementations of both the network system and the terminal have become extremely complex. For multi-user, multi-access communication terminals, how to quickly and efficiently find the optimal network and perform login operations in complex and diverse network environments has always been an important research topic for the industry.

[0003] In existing technologies, some communication terminals employ artificial intelligence to learn from their own network selection data, or they obtain network selection data from multiple terminals from a network-side server to learn a network selection strategy, and then perform network selection based on this strategy. However, current 3GPP (3rd Generation Partnership Project) technical specifications do not provide specific and operable technical specifications for how multi-user, multi-access communication terminals can achieve intelligent network selection. Both terminal-side AI learning methods and network-side collaborative learning schemes face the problems of unclear technical specifications and inconsistent implementation paths in practical applications.

[0004] The aforementioned existing technologies have the following drawbacks: There is a lack of clear and operable technical specifications for methods of using artificial intelligence to learn and reason about terminal network selection data. Especially for multi-user, multi-access communication terminals, there is a multitude of information affecting network selection decisions on the terminal side, and the network selection technical specifications of different standard organizations cannot be unified, making it very difficult to achieve efficient and unified intelligent network selection. Summary of the Invention

[0005] To address the aforementioned problems, the present invention aims to provide a method for automatically selecting a network for a terminal using artificial intelligence.

[0006] Another objective of this invention is to provide a system that uses artificial intelligence to enable automatic network selection for terminals. A method for automatically selecting a network using artificial intelligence includes the following steps: Step 1: After powering on or detecting a network outage, the terminal's user network selection module collects network selection-related input parameters. Step 2: The user network selection module sends the collected input parameters to the artificial intelligence entity and requests the network selection strategy. Step 3: The artificial intelligence entity performs artificial intelligence reasoning based on the received input parameters, generates network selection strategy data, and returns the network selection strategy data to the user network selection module; Step 4: The user network selection module performs a network selection operation based on the received network selection strategy data, and registers to the target network successfully or unsuccessfully. Step 5: The AI ​​entity records and stores the result parameters of the net selection operation for subsequent AI learning.

[0007] The method for automatic network selection by a terminal using artificial intelligence as described in this invention includes network operator's fixed configuration information, user's access standard preference information, historical information of networks registered or hosted, network selection related information sent by the network side via signaling, user's terminal interface operation information, user's subscribed network tariff information, terminal's geographic location information, terminal shutdown or network loss time information, current time information, terminal configured service information, and geographic environment information collected by terminal sensors.

[0008] The method for automatic network selection of terminals using artificial intelligence as described in this invention includes the following constraints for the artificial intelligence inference: the network operator's network selection technical specifications, the industry network selection technical specifications for each access standard of the communication terminal, and the product capability settings information of the terminal equipment manufacturer.

[0009] The method for automatic network selection by a terminal using artificial intelligence as described in this invention includes network selection strategy data including: network selection order, access standard, network name or number, start search time, and indication information on whether to connect in parallel with other users.

[0010] The present invention describes a method for automatic network selection by a terminal using artificial intelligence. The network side deploys a first part of the artificial intelligence entity, and the terminal deploys a second part of the artificial intelligence entity. During the network selection operation and registration process, the terminal transmits the input parameters and constraints for artificial intelligence learning in batches to the network side. The first part of the artificial intelligence entity on the network side performs artificial intelligence training and learning, and optionally sends updated artificial intelligence model parameters to the terminal. The second part of the artificial intelligence entity on the terminal performs artificial intelligence inference based on the received input parameters.

[0011] The method for automatic network selection of a terminal using artificial intelligence as described in this invention involves deploying an artificial intelligence entity on the terminal. After the terminal completes the boot process, the artificial intelligence entity of the terminal performs local artificial intelligence training and learning based on stored artificial intelligence learning input parameters and constraints, and performs artificial intelligence inference based on the input parameters.

[0012] The method for automatic network selection of a terminal using artificial intelligence, as described in this invention, involves the artificial intelligence entity using the network selection-related input parameters of the users who have already completed network selection as the input parameters of the subsequent users, performing artificial intelligence inference for the subsequent users, and outputting the network selection strategy for the subsequent users.

[0013] The method for automatic network selection of terminals using artificial intelligence described in this invention, in the scenario of network loss recovery, the artificial intelligence entity updates and infers the network selection strategy for subsequent users based on the recorded network loss information of the current user and the previously successful network selection results.

[0014] The method for automatic network selection using artificial intelligence described in this invention further includes: when the user network selection module of the terminal registers to a new network or goes to register a network, it feeds back the relevant network selection operation results and learning parameters to the artificial intelligence entity, which then performs artificial intelligence learning and updates the network selection model based on the learning results.

[0015] A system for automatically selecting a network using artificial intelligence, and a method for automatically selecting a network using artificial intelligence, comprising: A multi-user network selection module, deployed on the terminal, is used to collect network selection-related input parameters after power-on or after detecting network loss, and send a network selection strategy request to the artificial intelligence entity; it is also used to receive and execute the network selection strategy data returned by the artificial intelligence entity to register with the target network; An artificial intelligence entity is used to receive the net selection strategy request, perform artificial intelligence inference based on the input parameters, generate and return the net selection strategy data; it is also used to record and store the result parameters of the net selection operation, which are used for subsequent artificial intelligence learning. The first part of the artificial intelligence entity is deployed on the network side, and the second part of the artificial intelligence entity is deployed on the terminal. The first part of the artificial intelligence entity on the network side receives the input parameters and constraints of the artificial intelligence learning algorithm transmitted by the terminal during network registration, performs artificial intelligence training and learning, and updates the artificial intelligence model parameters according to the learning results. The second part of the artificial intelligence entity on the terminal performs artificial intelligence inference based on the input parameters; or The artificial intelligence entity is deployed on the terminal. After the power-on process is completed, the terminal performs local artificial intelligence training and learning based on the input parameters and constraints of the stored artificial intelligence learning algorithm, and performs artificial intelligence inference based on the input parameters.

[0016] Beneficial effects: The artificial intelligence entity of the present invention can continuously train and learn based on the results of each network selection operation, continuously optimize the network selection model, and improve the accuracy of future network selection decisions. In addition, the input parameters related to network selection can be updated in a timely manner, so that subsequent update inference can dynamically adapt to changes in the network environment, helping the terminal to restore network connection faster and more stably in complex environments. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method for automatic network selection by a terminal using artificial intelligence according to the present invention; Figure 2 This is a schematic diagram of the power-on network selection process of the present invention, which uses the network as the center of artificial intelligence. Figure 3 This is a schematic diagram of the power-on network selection process of the present invention, with the terminal as the artificial intelligence center; Figure 4 This is a schematic diagram of the AI ​​network selection process for network recovery after a network outage, as described in this invention. Figure 5 This is a schematic diagram of other artificial intelligence learning processes centered on the network, as described in this invention; Figure 6 This is a schematic diagram of other artificial intelligence learning processes centered on the terminal, according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0020] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0021] This invention proposes a method for automated network selection using artificial intelligence in multi-user, multi-access terminal devices, primarily targeting scenarios of network search after power-on and network recovery after network outage.

[0022] Reference Figure 1 A method for automatically selecting a network for a terminal using artificial intelligence includes the following steps: Step 1: After powering on or detecting a network outage, the terminal's user network selection module collects network selection-related input parameters. Step 2: The user network selection module sends the collected input parameters to the artificial intelligence entity and requests the network selection strategy. Step 3: The AI ​​entity performs AI reasoning based on the received input parameters, generates network selection strategy data, and returns the network selection strategy data to the user's network selection module; Step 4: The user network selection module performs the network selection operation based on the received network selection strategy data, and registers to the target network successfully or unsuccessfully. Step 5: The AI ​​entity records and stores the result parameters of the net selection operation for subsequent AI learning.

[0023] The artificial intelligence entity of this invention can continuously train and learn based on the results of each network selection operation, constantly optimize the network selection model, improve the accuracy of future network selection decisions, and update the network selection-related input parameters in a timely manner, so that subsequent update inference can dynamically adapt to changes in the network environment, helping the terminal to restore network connection faster and more stably in complex environments.

[0024] The network selection scheme of the present invention is a method for using artificial intelligence technology on the terminal side to train, learn and reason about information data related to network selection for multiple users and multiple access standards, to obtain an efficient network selection strategy and to perform network selection according to this strategy.

[0025] The input parameters related to net selection in this invention include input parameters for artificial intelligence inference, which may include: The fixed configuration information of the network operator, for 3GPP terminals, includes the network list configuration information in the USIM card; User-defined access standard preferences, for 3GPP terminals, include preferences for access standards such as 5G NR, LTE, and WCDMA; Historical information about the network where the terminal is camped or registered. For 3GPP terminals, this includes information such as the identification of each cell and channel where the terminal is camped or registered, and signal quality. The network-side transmits network selection-related information via signaling. For 3GPP terminals, this includes downlink inter-cell handover and redirection results. User terminal interface operation information, for 3GPP terminals, includes historical information of the user manually selecting a specific network number; Information on network tariffs subscribed by users, such as the packages and tariffs offered by network operators. The terminal's geographic location information; Information on the time when the terminal was powered off or lost network access, current time information, and the terminal's configured service information; and The geographic environment information collected by the terminal sensors includes, but is not limited to, temperature, light, humidity, smoke, terminal speed, and acceleration.

[0026] The input parameters related to network selection include input parameters for artificial intelligence learning, which include input parameters for artificial intelligence inference, and at least real-time information on whether the end user successfully or unsuccessfully selects a network each time, including the time information, network name / number, access standard, etc.

[0027] The constraints of the artificial intelligence reasoning in this invention include: Network operator network selection technical specifications; Industry network selection technical specifications for various access standards of communication terminals, and for 3GPP terminals, including the technical specifications specified in 3GPP spec.23.122; Product capability settings information for terminal equipment manufacturers.

[0028] The network selection strategy data of this invention is the output data of artificial intelligence inference, including: network selection priority, access standard, network name or number, search start time, and indication information of whether it is connected in parallel with other users. Examples are shown in Table 1 below:

[0029] The present invention provides a method for automatic network selection for terminals using artificial intelligence. When multiple users belong to the same operator and network service configuration, the network selection strategy output by the artificial intelligence entity can instruct subsequent users to directly select the network currently registered by the user who has already successfully selected a network.

[0030] This invention is particularly applicable to communication terminals with multiple users and multiple access standards. Through artificial intelligence reasoning, it can reuse the successful network selection information of other users (especially users with the same operator and configuration), so that subsequent users can directly select the verified network, skipping the process of searching for the network, thereby significantly improving the overall network selection efficiency.

[0031] The first embodiment of the power-on network selection process is as follows: Figure 2 To address the issue of generally weak computing power in current commercial terminals, the AI ​​entity can be primarily deployed on the network side, with a secondary deployment on the terminal side. The AI ​​entity deployed on the network side is responsible for training and learning, while the AI ​​entity on the terminal side is responsible for inference. Includes the following steps: Step 2a: The user powers on the device through the user interface. The power-on notification sequentially notifies each functional module to start. After the AI ​​entity starts, the algorithm initialization configuration is completed.

[0032] Step 2b: After the first user network selection module is started, it collects data related to the power-on network selection, including the input parameters of the aforementioned artificial intelligence inference. It then sends this data to the artificial intelligence entity to request the first user network selection strategy.

[0033] Step 2c: The artificial intelligence entity performs artificial intelligence (AI) reasoning based on the received power-on network selection data, generates the first user network selection strategy data, and returns the network selection strategy data to the first user network selection module.

[0034] In step 2d, the first user network selection module starts to perform network selection operation for the first user based on the received first user network selection strategy data. If the operation is successful, the terminal successfully resides and registers with a certain first operator network. During the process of successfully registering with the first operator network, the terminal transmits the stored input parameters and constraints of the AI ​​learning algorithm to the network in batches. The network then uses this information to perform subsequent AI training and learning.

[0035] Step 2e: The artificial intelligence entity records and stores the network selection result parameters received from the first user's successful network selection operation upon power-on.

[0036] Step 2f: After the second user network selection module starts up, it collects data related to network selection at startup. This data serves as the input parameters for artificial intelligence inference. Then, it sends this data to the artificial intelligence entity to request the second user network selection strategy.

[0037] In step 2g, the AI ​​entity performs AI reasoning based on the received network selection data and the network selection parameter information of the first user obtained in step 2e, and generates network selection strategy data for the second user, which is then returned to the second user network selection module. In particular, when the first user and the second user belong to the same operator and have the same network service configuration, the network selection strategy output by the AI ​​entity can instruct the second user to directly select the network that the first user has already successfully selected. In this case, the second user network selection module can omit the network search process when executing step 2h.

[0038] In step 2h, the second user network selection module, based on the received network selection strategy data from the second user, begins the network selection operation for the second user and successfully selects a specific operator's network, such as the second operator's network. During the successful registration process with this network, the terminal transmits the stored input parameters and constraints of the AI ​​learning algorithm to the network in batches. The network then uses this information for subsequent artificial intelligence training and learning.

[0039] Step 2i: The AI ​​entity records and stores the result parameters of the second user's successful power-on network selection operation.

[0040] If there are other users waiting to be selected for network operation, then the process from step 2f to step 2i above can be repeated for that user.

[0041] During the network selection and registration process of this invention, the terminal transmits the input parameters and constraints of artificial intelligence learning to the network side in batches. The network side then performs artificial intelligence training and learning, and optionally sends the updated artificial intelligence model parameters back to the terminal, thereby continuously optimizing the terminal's network selection strategy and improving the accuracy and efficiency of network selection.

[0042] The second embodiment of the power-on network selection process is as follows: Figure 3 For terminals with high computing power, the AI ​​entity is deployed on the terminal side. AI training and learning are performed directly within the terminal. After the terminal boots up, the terminal-side AI entity performs local AI training and learning based on stored AI learning input parameters and constraints, and performs AI inference based on the input parameters. This includes the following steps: Step 3a: The user powers on the device through the user interface. The power-on notification sequentially notifies each functional module to start. After the artificial intelligence entity starts, it completes the initial configuration of the algorithm, including the configuration of constraints and the reading of historical learning data. Step 3b: After the first user network selection module is started, it collects the data related to network selection at startup, and then sends the data related to network selection at startup to the artificial intelligence entity to request the network selection strategy.

[0043] Step 3c: The AI ​​entity performs AI inference based on the received power-on network selection data and returns the network selection strategy data as the inference result to the first user network selection module.

[0044] Step 3d: The first user network selection module starts the network selection operation for the first user based on the received network selection strategy data and successfully (or unsuccessfully) selects the first operator's network.

[0045] Step 3e: After successfully (or unsuccessfully) registering with the first operator's network, the AI ​​entity will store the received parameters of the first user's network selection operation upon startup for subsequent AI learning.

[0046] Step 3f: After the second user network selection module is started, it collects the data related to the power-on network selection and then sends the data to the artificial intelligence entity to request the second user network selection strategy.

[0047] In step 3g, the AI ​​entity performs AI inference based on the received power-on network selection data and returns the second user network selection strategy data as the inference result to the second user network selection module. The AI ​​inference algorithm can use the first user's power-on network selection operation result parameter information obtained in step 3e to infer the second user's network selection strategy. In particular, when the first user and the second user belong to the same operator and network service configuration, the network selection strategy output by the AI ​​entity can instruct the second user to directly select the network that the first user has already successfully selected. In this way, for step 3h below, the second user network selection module can actually omit the process of searching for the network.

[0048] In step 3h, the second user network selection module starts the network selection operation for the second user based on the received second user network selection strategy data and successfully (or unsuccessfully) selects a second operator's network.

[0049] Step 3i: After successfully (or unsuccessfully) registering with the second operator's network, the AI ​​entity learns the parameters of the second user's network selection operation result received during startup for subsequent AI learning.

[0050] If there are other users waiting to be selected for network operation, then the process from step 3f to step 3i above can be repeated for that user.

[0051] After the boot process is completed, the AI ​​entity on the subsequent terminal can perform AI training and learning at any time based on batch data such as the input parameters and constraints of the stored AI learning algorithm.

[0052] In implementations centered on the terminal as the AI ​​hub, the AI ​​entity on the terminal side can independently acquire network selection strategies without real-time signaling interaction with the network side. This is particularly suitable for scenarios where the terminal cannot normally connect to the network, such as when searching for a network after powering on or recovering from a network outage. It can help the terminal restore network connectivity faster and more stably in complex environments.

[0053] An example of the AI-powered network selection process for network recovery after a network outage, referred to... Figure 4 This includes the following steps: Step 4a: The first user network selection module notifies the AI ​​entity that the network has been lost (equivalent to network selection failure).

[0054] Step 4b: The AI ​​entity performs AI inference based on the received network selection parameters (mainly the first user's network loss information) and returns the first user's network selection strategy data as the inference result to the first user's network selection module; the AI ​​entity records and stores the received network selection parameters, mainly the first user's network loss information.

[0055] Step 4c: The second user network selection module notifies the AI ​​entity that the network has been lost (equivalent to network selection failure).

[0056] In step 4d, the AI ​​entity performs AI inference based on the received network selection parameters (mainly the second user's network loss information) and returns the second user's network selection strategy data as the inference result to the second user's network selection module; the AI ​​entity records and stores the received network selection parameters, mainly the second user's network loss information.

[0057] It is important to note in this step that the input to the artificial intelligence inference algorithm may include the first user's network loss information recorded and stored in step 4b, that is, the information of the first user is used to assist in inferring the network selection strategy of the second user.

[0058] If there are other users who have lost their network and are waiting for a network to be selected, then the process from step 4c to step 4d above can be repeated for that user.

[0059] Step 4e: The first user network selection module performs network selection for the first user based on the received first user network selection strategy data and successfully selects a certain operator's network.

[0060] Step 4f: The AI ​​entity records and stores the received parameters of the first user's network selection operation result. Based on these parameters, the AI ​​entity performs further AI inference and updates the second user network selection strategy data (the inference result) to the second user network selection module. It is important to note in this step that the input to the artificial intelligence inference algorithm includes the received network selection result information of the first user, that is, the network selection strategy of the second user is inferred from the network selection result of the first user; in particular, when the first user and the second user belong to the same operator and network service configuration, the network selection strategy of the second user output by the artificial intelligence entity can be to instruct the second user to directly select the network that the first user has already successfully selected. In this way, for step 4g below, the network selection module of the second user can actually omit the process of searching for the network.

[0061] Step 4g: The second user network selection module performs network selection for the second user based on the received second user network selection strategy data and successfully selects a certain operator's network.

[0062] In step 4h, the AI ​​entity records and stores the received parameters of the second user's network selection operation.

[0063] If there are other users who have lost their network and are waiting for a network to be selected, then the process from step 4f to step 4h above can be repeated for that user.

[0064] The present invention provides a method for automatic network selection for terminals using artificial intelligence. In network loss recovery scenarios, the artificial intelligence entity updates and infers the network selection strategy for subsequent users based on the recorded network loss information of the current user and the previous successful network selection results. By using historical successful network selection results and current network loss information for update and inference, subsequent users can learn from previous successful experiences, thereby accelerating the overall recovery speed and improving the network connection stability and recovery efficiency of the terminal in complex and ever-changing network environments.

[0065] In addition to the above-described power-on and network disconnection recovery processes, this invention also provides other artificial intelligence learning processes. (See reference...) Figure 5 Other AI learning processes centered on networks include: Application Scenario A: During the process of a first user successfully registering to a new network, the terminal transmits the stored input parameters and constraints of the artificial intelligence learning algorithm to the network in batches. The network then uses this information for subsequent artificial intelligence training and learning. After artificial intelligence learning, the network can subsequently download new artificial intelligence model parameters to the terminal at any time via data or signaling connections for updating the terminal's artificial intelligence entity.

[0066] Application Scenario B: During the process of a second user successfully registering with the network, the terminal transmits the stored input parameters and constraints of the artificial intelligence learning algorithm to the network in batches. The network then uses this information for subsequent artificial intelligence training and learning. Optionally, during the registration process, the network can download new artificial intelligence model parameters to the terminal via a data connection or signaling connection for updating the terminal's artificial intelligence entity.

[0067] Reference Figure 6 Other AI learning processes centered on the terminal include: Application Scenario C: The first user's network selection module notifies the AI ​​entity that it has registered on a new network. The AI ​​entity then performs AI learning on batch data (including the network selection-related parameters received this time) such as the input parameters and constraints of the stored AI learning algorithm.

[0068] Application Scenario D: The second user's network selection module notifies the AI ​​entity that it has settled on a new network under restricted conditions. The AI ​​entity performs AI learning on batch data (including the network selection-related parameters received this time), such as the input parameters and constraints of the stored AI learning algorithm. Optionally, after AI learning, AI inference is performed, and the network selection strategy data as the inference result is returned to the second user's network selection module as the network selection strategy under restricted conditions.

[0069] The method for automatic network selection using artificial intelligence in this invention further includes: when the user network selection module of the terminal registers with a new network or deregisters with a network, it feeds back the relevant network selection operation results and learning parameters to the artificial intelligence entity, which then performs artificial intelligence learning and updates the network selection model based on the learning results. By feeding back the operation results and learning parameters when the user registers with or deregisters with a network, the artificial intelligence entity can continuously learn and update the network selection model, thereby achieving dynamic optimization of the network selection strategy and helping the terminal adapt to changes in complex network environments more accurately and stably.

[0070] This invention also improves a system for automatic network selection by a terminal using artificial intelligence, and a method for automatic network selection by a terminal using artificial intelligence, comprising: The multi-user network selection module, deployed on the terminal, is used to collect network selection-related input parameters after power-on or after detecting network loss, and send network selection policy requests to the artificial intelligence entity; it is also used to receive and execute the network selection policy data returned by the artificial intelligence entity to register with the target network. The artificial intelligence entity (AI entity) is used to receive net selection strategy requests, perform AI inference based on input parameters, generate and return net selection strategy data; it is also used to record and store the result parameters of net selection operations for subsequent AI learning. When the AI ​​entity is partially deployed on the network side, the first part of the AI ​​entity is deployed on the network side, and the second part of the AI ​​entity is deployed on the terminal. The first part of the AI ​​entity on the network side receives the input parameters and constraints of the AI ​​learning algorithm transmitted by the terminal during the network registration process, performs AI training and learning, and updates the AI ​​model parameters according to the learning results. The second part of the AI ​​entity on the terminal performs AI inference according to the input parameters. When the AI ​​entity is fully deployed on the terminal, after the terminal completes the boot process, it performs local AI training and learning based on the input parameters and constraints of the stored AI learning algorithm, and performs AI inference based on the input parameters.

[0071] This invention utilizes artificial intelligence learning and reasoning on terminal network selection data, specifying two concrete and operable implementation steps: one centered on the network and the other on the terminal. It is particularly suitable for multi-user, multi-access communication terminals, effectively coordinating the network selection process for multiple users and significantly improving overall network selection efficiency.

[0072] The artificial intelligence entity of the present invention has at least artificial intelligence reasoning or learning functions, is built on the network side or the terminal side, and can be a component of a machine learning, a program of a neural network algorithm, an artificial intelligence model of a professional field, or an AI intelligent agent.

[0073] The description and accompanying drawings provide typical embodiments of specific structures for specific implementations. Other modifications are possible based on the spirit of the invention. While the above-described invention presents preferred embodiments, these are not intended to be limiting.

[0074] For those skilled in the art, various changes and modifications will undoubtedly be apparent after reading the above description. Therefore, the appended claims should be construed as covering all changes and modifications that encompass the true intent and scope of the invention. Any and all equivalent scope and content within the scope of the claims should be considered to remain within the intent and scope of the invention.

Claims

1. A method for automatic network selection by a terminal using artificial intelligence, characterized in that, Includes the following steps: Step 1: After powering on or detecting a network outage, the terminal's user network selection module collects network selection-related input parameters. Step 2: The user network selection module sends the collected input parameters to the artificial intelligence entity and requests the network selection strategy. Step 3: The artificial intelligence entity performs artificial intelligence reasoning based on the received input parameters, generates network selection strategy data, and returns the network selection strategy data to the user network selection module; Step 4: The user network selection module performs a network selection operation based on the received network selection strategy data, and registers to the target network successfully or unsuccessfully. Step 5: The AI ​​entity records and stores the result parameters of the net selection operation for subsequent AI learning.

2. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, The input parameters related to network selection include fixed configuration information of the network operator, access standard preference information set by the user, historical information of residing or registering on various networks, network selection related information issued by the network side through signaling, user terminal interface operation information, network tariff information subscribed by the user, terminal geolocation information, terminal shutdown or network loss time information, current time information, terminal configured service information, and geographical environment information collected by terminal sensors.

3. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, The constraints of the artificial intelligence inference include: the network operator's network selection technical specifications, the industry network selection technical specifications for each access standard of the communication terminal, and the product capability settings information of the terminal equipment manufacturer.

4. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, The network selection strategy data includes: network selection order, access standard, network name or number, start search time, and indication information on whether to connect in parallel with other users.

5. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, The network side deploys the first part of the artificial intelligence entity, and the terminal deploys the second part of the artificial intelligence entity. During the network registration process of the network selection operation, the terminal transmits the input parameters and constraints of artificial intelligence learning in batches to the network side. The first part of the artificial intelligence entity on the network side performs artificial intelligence training and learning, and optionally sends the updated artificial intelligence model parameters to the terminal. The second part of the artificial intelligence entity on the terminal performs artificial intelligence inference based on the received input parameters.

6. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, The artificial intelligence entity is deployed on the terminal, and after the terminal completes the boot process, the artificial intelligence entity on the terminal performs local artificial intelligence training and learning based on the stored artificial intelligence learning input parameters and constraints, and performs artificial intelligence inference based on the input parameters.

7. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, When there are multiple users on the terminal, the artificial intelligence entity uses the network selection-related input parameters of the users who have already completed network selection as the input parameters of the subsequent users, performs artificial intelligence inference for the subsequent users, and outputs the network selection strategy for the subsequent users.

8. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, In the network loss recovery scenario, the artificial intelligence entity updates and infers the network selection strategy for subsequent users based on the recorded network loss information of the current user and the previously successful network selection results.

9. The method for automatic network selection using artificial intelligence as described in claim 1, characterized in that, Also includes: When the user network selection module of the terminal registers to a new network or goes to register a network, it feeds back the relevant network selection operation results and learning parameters to the artificial intelligence entity, which then performs artificial intelligence learning and updates the network selection model based on the learning results.

10. A system that uses artificial intelligence to achieve automatic network selection at the terminal, characterized in that, The method for using artificial intelligence to achieve automatic network selection of a terminal as described in any one of claims 1 to 9 includes: A multi-user network selection module, deployed on the terminal, is used to collect network selection-related input parameters after power-on or after detecting network loss, and send a network selection strategy request to the artificial intelligence entity; it is also used to receive and execute the network selection strategy data returned by the artificial intelligence entity to register with the target network; An artificial intelligence entity is used to receive the net selection strategy request, perform artificial intelligence inference based on the input parameters, generate and return the net selection strategy data; it is also used to record and store the result parameters of the net selection operation, which are used for subsequent artificial intelligence learning. The first part of the artificial intelligence entity is deployed on the network side, and the second part of the artificial intelligence entity is deployed on the terminal. The first part of the artificial intelligence entity on the network side receives the input parameters and constraints of the artificial intelligence learning algorithm transmitted by the terminal during network registration, performs artificial intelligence training and learning, and updates the artificial intelligence model parameters according to the learning results. The second part of the artificial intelligence entity on the terminal performs artificial intelligence inference based on the input parameters; or The artificial intelligence entity is deployed on the terminal. After the power-on process is completed, the terminal performs local artificial intelligence training and learning based on the input parameters and constraints of the stored artificial intelligence learning algorithm, and performs artificial intelligence inference based on the input parameters.

Images