Secure element application and corresponding secure element for triggering mobile devices to perform a preferred network selection procedure

Abstract

The present invention relates to a secure element application for triggering a mobile device to perform a preferred network selection procedure for attaching to a private network after an optimal period of time and after at least a specific location change, the optimal period of time and the specific location change being learned from a continuous learning phase, during which the application performs the following steps: - Attach to a private network: - a known method for triggering a mobile device to perform a network selection, or - triggering the mobile device to perform network selection in a known location where a private network is expected to be available; or - Varying the time and frequency at which the mobile device is triggered to make a network selection after a given location change. including an order to try the case through successive trials, The optimal time period is derived from a previously recorded time period measured from a location change event to a successful switch to the private network.

Description

Technical Field

[0001] The present invention relates to telecommunications, and more precisely to secure elements such as UICC (Universal Integrated Circuit Card), such as SIM cards, USIM, eUICC (embedded UICC) or iUICC (integrated UICC). These secure elements cooperate with terminals (user equipment, hereinafter also referred to as UE), such as smartphones, mobile phones, PDAs or IoT devices. These terminals are hereinafter also referred to as devices.

Background Art

[0002] The secure element has the purpose of establishing and securing communication between the terminal and the telecommunications network.

[0003] There are two main types of telecommunications networks, public and private: The public telecommunications network is also called a PLMN (Public Land Mobile Network). The PLMN can be a Home PLMN (HPLMN) or a Visited PLMN (VPLMN).

[0004] For example, in schools, airports, large factories, energy public facilities, manufacturing sites, ports or offshore platforms, private cellular networks are installed, and access to these networks can be restricted to a given identified user of the terminals covered by these private cellular networks.

[0005] The objective of this invention is to switch from a globally accessible public network to a locally accessible private network. This switch is initiated by the secure element as soon as it detects the presence of the private network (the terminals with which the secure element interacts are covered by the private network). This switch is implemented, in particular, to minimize communication costs, address security risks, and narrow the attack surface.

[0006] The term "private network" as used herein is also known as a non-public network (or NPN in the 3GPP specification), or a mobile private network.

[0007] 3GPP does not allow a UE to search for a higher-priority PLMN when it is camped on an HPLMN or EHPLMN (Equivalent Home PLMN).

[0008] Existing solutions for switching from a globally accessible public network to a local private network rely on one of the following: - Setting a global public network as the destination PLMN (VPLMN) presents related issues, such as those related to user experience or user preferences regarding restricting data access during roaming and having mobile devices set up by default to prevent data usage while roaming; - Use the SIM toolkit location status event sent to the USIM by the ME; however, such a method is not optimal when the private network is not immediately available after such an event; - Support from the ME for a specific method to force the USIM to perform a network selection procedure.

[0009] However, these existing methods impact either battery usage or user experience, for example, by causing long delays before switching to a private network or frequent UE reinitialization periods (making them unavailable for normal use). [Overview of the project] [Problems that the invention aims to solve]

[0010] Therefore, the technical problem that the present invention intends to solve is: how to automatically force the ME to switch from a globally accessible public network to a local private network as quickly as possible without modifying the ME and minimizing the impact on battery consumption and usability. [Means for solving the problem]

[0011] this is: - A step to select a method supported by the ME in order to guide the ME to perform network selection. - A step to determine the earliest time when ME should perform the network selection procedure. - Steps to minimize service interruptions caused by the switching procedure in order to optimize battery consumption. This is achieved by a method that includes [a specific method].

[0012] More precisely, the present invention proposes a secure element application for triggering a mobile device to perform a preferred network selection procedure for attaching to a private network after an optimal time period and at least after a specific location change, the optimal time period and specific location change being learned from a continuous learning phase, during which the application: - Attach to a private network: - Known methods for triggering a mobile device to make a network selection, or - Triggering a mobile device to select a network in a known location where a private network is expected to be available, or - Modify the timing and frequency at which mobile devices are triggered to make a network selection after a given location change. Includes instructions for trial through a series of trials, The optimal time period is derived from previously recorded time periods measured from the location change event until the successful switch to the private network.

[0013] Preferably, a known method for triggering a preferred network selection procedure is: - Modify the IMSI of the secure element, or - Clear the LOCI file of the secure element, or - Perform a roaming refresh steering, or - Instructing a mobile device to switch to idle mode via AT commands, or - Any combination of the above methods Includes.

[0014] For advantages, known locations are: - Pre-configured in the secure element application, or - Configured in the secure element application by a remote server, or - Learned through a series of trials in various locations, or - Based on network rejection events That is the case.

[0015] Preferably, changing how often a mobile device is triggered to make a network selection involves reducing the frequency to achieve the best trade-off between battery consumption and the time spent switching to a private network.

[0016] The present invention also relates to a secure element including such an application.

[0017] The present invention represents a UE (consisting of an ME part and an SE / USIM part) moving from a first HPLMN cell to a second HPLMN cell, and will be better understood by reading the following description of a preferred embodiment of the present invention with reference to FIG. 1 in which at least a part of a private network is within the coverage of the second HPLMN cell.

Brief Description of the Drawings

[0018] [Figure 1] It is a diagram representing a public network 10 (here an HPLMN, although this may also be, for example, a VPLMN or EHPLMN).

Embodiments for Carrying Out the Invention

[0019] The HPLMN 10 includes two HPLMN cells, a first HPLMN cell A referred to as 11 and a second HPLMN cell B referred to as 12. The coverage of the HPLMN cell 12 covers at least a part of the coverage of a private 5G (or 4G / 3G / 2G, WLAN or generally a wireless network) private network 13.

[0020] The UE 14 composed of an ME and a secure element SE (e.g., USIM) is in a first step covered by the cell 11 and is moving in the direction of the cells 12 and 13.

[0021] In a second step, the UE 14 reaches the edge of the cells A and B of the HPLMN and camps on the cell B.

[0022] The UE 14 requested by the SE to send an event regarding a location change sends a location change event to the SE.

[0023] The SE determines (for example, by searching a locally stored lookup table) that the private 5G network 13 is close to cell B (i.e., there is a high probability that the ME's attempt at network selection on the private network will succeed), but it does not yet know exactly when to switch to the private 5G network.

[0024] When UE14 enters the coverage of cell B 12, SE starts timer T and attempts to switch to and connect to the private 5G network 13 at varying frequencies.

[0025] When the private 5G network is reached (i.e., when UE14 enters the wireless coverage of the private network), in the third step, the SE's attempt to switch to private 5G is successful, and UE14 is able to connect to the private 5G network 13.

[0026] USIM then stops timer T, halts the switching trial process, and saves the value T0 of timer T. Thus, the value T0 corresponds to the duration (time period) required for ME to reach cell B13 from the moment it entered cell B12.

[0027] Next, when SE receives an event from ME indicating that ME has reached cell B12, SE starts timer T again (after reset), waits for timer T to reach the value T0* derived from T0, and then begins the switching attempt process.

[0028] When UE14 is able to connect to the private 5G network 13, UE14 stops the switching procedure and saves a new value T1 for timer T.

[0029] Next, when UE14 reaches cell B, SE derives a new value T1* from T0, T0*, and T1.

[0030] After several epochs of the above learning process, the SE can estimate one or more values ​​for timer T, and thus the private 5G network may be accessible with a higher probability than the set fixed threshold.

[0031] As a result, the SE can reduce the time and frequency of the switching trial procedure, thereby reducing the battery consumption of the UE caused by the switching trial.

[0032] Regarding learning algorithms, two alternative forms are possible: one based on a neural network model (e.g., LSTM - Long Short-Term Memory), and a second, more traditional algorithm based on a triplet of programmatically tuned parameters.

[0033] These algorithms are described below.

[0034] Neural network (e.g., LSTM)-based algorithms can be used to improve the prediction of the Time of Arrival (ToA).

[0035] ToA is the time period between the moment the target cell (i.e., a cell with a high probability of being under private network coverage) is reached and the moment the selection of the MPN (Mobile Private Network) is successful.

[0036] Radio cells are identified by their cell identifiers, which are broadcast over the radio interface by radio-based stations. Target cells are identified, in particular, by their target cell-ID. Applets can learn about the target cell's target cell-ID, for example, through pre-configuration by a home PLMN or private network operator, or after a trial-and-error phase.

[0037] This invention uses reinforcement learning algorithms to improve machine learning models (referred to as MODELs) (for example, based on one LSTM or multiple layers of an LSTM neural network).

[0038] The input to MODEL includes a series of Cell-IDs received by the applet from successive location update events until a location update event includes this target Cell-ID, indicating that the target Cell-ID has been reached.

[0039] The output of MODEL is ToA.

[0040] Reinforcement learning phase: - The MODEL is provided with a sequential list of Cell-IDs as input. - The model generates a predicted time of arrival (pToA) for each input. - The error between the following two points (Err = pToA - gtToA) is calculated. ○ Predicted arrival time (pToA) and ○ Ground truth time of arrival (gtToA) measured by the applet. - The calculated error (Err) is backpropagated within the training model to tune the model's internal parameters until the model can provide an aggregated Err (e.g., the mean of the square root sum of squared errors for a validation batch of data) that falls below a specified threshold.

[0041] The training application records a series of Cell-IDs for reinforcement training of the model.

[0042] The series of Cell-IDs recorded by the training application are the last N Cell-IDs before reaching the target cell, and include the target Cell-ID.

[0043] The training application also records the corresponding ground truth gtToA associated with a set of Cell-IDs.

[0044] A triplet-based learning algorithm that minimizes the time and number of selection attempts required to successfully select a private network when the target cell is reached: - The applet performs a first attempt to select a private network after a long time period (e.g., 10 minutes) of value TL. - If a private network could not be selected during this first trial: ○ The applet will perform subsequent trials after a shorter time period (e.g., 1 minute) than the value TS. ○ Until the MVP is selected - When the Time Limit (TL) expires, if a private network could have been selected during the first trial, the TL is reduced by Nd times the TS (new_TL = old_TL - Nd × TS), and this new value is used when the next device coming from the public mobile network reaches the target cell; - If a private network cannot be selected when the TL expires over a long period of time, and is successfully selected after the TS has expired N times, ○ If N is greater than Nd, TL increases by (N-Nd) times TS (new_TL = old_TL + (N-Nd) × TS). ○ If N is less than or equal to Nd, TL is not changed. It may be possible.

[0045] Therefore, Nd is used to reduce the time limit over long periods in order to have the opportunity to select a private network at an earlier point when the target cell is reached.

[0046] The triplet value (TL,TS,Nd) can be associated with each or a group of target cells in the target cell list.

[0047] Therefore, the present invention proposes using a secure element application to trigger a mobile device to perform a preferred network selection procedure for attaching to a private network after an optimal time period and at least after a specific location change, the optimal time period and specific location change being learned from a continuous learning phase (the application is running persistently), during which the application: - Attach to a private network: - Known methods for triggering a mobile device to make a network selection, or - Triggering a mobile device to select a network in a known location where a private network is expected to be available, or - Modify the timing and frequency at which mobile devices are triggered to make a network selection after a given location change. Includes instructions for trial through a series of trials, The optimal time period is derived from previously recorded time periods measured from the location change event until the successful switch to the private network.

[0048] The optimal time period is derived from previously recorded time periods measured from a location change event where the private network is expected to be available until a successful switch from HPLMN to the private network is achieved.

[0049] Known methods for triggering a preferred network selection procedure include, but are not limited to, the following: - Correcting the USIM's IMSI (this is the slowest method), - Clear the USIM's LOCI file. - Perform a roaming refresh steering. - Instructing ME to switch to idle mode via AT commands, - and any combination of the above methods Includes.

[0050] The target location known by the application is, for example: - Pre-configured in UICC applications, - Or, configured in the UICC application by a remote server, - Or, learned through a series of trials in various locations. - Or, based on a network rejection event It is one of the following:

[0051] Changing how often a mobile device is triggered to make a network selection depends on reducing the frequency to achieve the best trade-off between battery consumption and the time spent switching to a private network.

[0052] The present invention also relates to secure elements that include such applications.

[0053] The main advantages of this invention are: - Seamless transition: No user intervention required; - Minimize switching delays when the private network becomes available; - Allows the device to spend a maximum amount of time on a private network; - Minimize battery consumption by optimizing timing for location changes (at the edge of the public network where the private network is expected to be). Includes.

[0054] There are two use cases for this invention: - Use Case 1: In a combination of existing solutions, the SE learns when to switch to a private network. - Use Case 2: In "standalone mode," the SE is configured (by the MNO) using the Cell B identity where the private network is located, and then learning begins when the terminal enters Cell B coverage.

[0055] In summary, the present invention relies on actual trials to determine what features are supported by actual terminals in order to select the most time-efficient procedure for performing a network selection procedure, which may change over time depending on the environment, such as changes in wireless coverage and obstacles. The present invention also relies on actual trials to predict when to switch based on the location and trajectory of a device, where previous methods rely only on changes in location. Finally, parameters learned for one or more devices presenting the same usage profile may be shared with other devices presenting the same usage profile.

Claims

1. A secure element application in the secure element of a mobile device for triggering the mobile device to perform a preferred network selection procedure for attaching to a private network after an optimal time period and after at least a specific location change, wherein the optimal time period and the specific location change are learned from a continuous learning phase, and the application performs the following actions during the continuous learning phase: - Attaching to the aforementioned private network, - Known methods for triggering the mobile device to perform a network selection, or - Triggering the mobile device to make a network selection at a known location where the private network is expected to be available, or - Modify the time and frequency at which the mobile device is triggered to make a network selection after a given location change. Includes instructions to the secure element for testing through a series of trials, The aforementioned optimal time period is derived from a previously recorded time period measured from the location change event until the successful switch to the private network in the secure element application.

2. The known method for triggering a preferred network selection procedure is - Modify the IMSI of the secure element, or - Clear the LOCI file of the secure element, or - Perform a roaming refresh steering, or - Instructing the mobile device to switch to idle mode via AT command, or - Any combination of the above methods The secure element application according to claim 1, including the following:

3. The aforementioned known location - Pre-configured in the aforementioned secure element application, or - Configured in the secure element application by a remote server, or - Learned through a series of trials in various locations, or - A secure element application according to claim 1 or 2, based on a network rejection event.

4. The secure element application according to claim 1 or 2, wherein changing the frequency at which the mobile device is triggered to make a network selection depends on reducing the frequency to obtain the best trade-off between battery consumption and the time it takes to switch to the private network.

5. A secure element comprising the application described in claim 1 or 2.

Images