Method for predicting a variation in quality of service in a cellular telecommunications network, prediction device and corresponding computer program.

Abstract

Method for predicting a variation in quality of service in a cellular telecommunications network, prediction device and corresponding computer program. The invention relates to a method for predicting a variation in quality of service (QoS) in a cellular telecommunications network comprising at least one base station to which user equipment is connected.Such a process includes: - the collection (DUE) of at least one location information (LI) including an identifier of said base station, - the selection (S_MOD), from a set of key performance indicators (KPIs) of said network, of at least one indicator of interest representative of said quality of service, - the prediction (C_MOD) by machine learning of a future value of said indicator of interest from at least one time trajectory of said indicator of interest, said time trajectory being representative of past values ​​of said indicator of interest, collected by the prediction device, - the transmission (ADAPT), within said network, of an adjustment message for at least one parameter of said network based on said predicted future value. Abbreviated figure: Figure 4.

Description

Title of the invention: Method for predicting a variation in quality of service in a cellular telecommunications network, prediction device and corresponding computer program. Scope of the invention

[0001] The field of the invention is that of cellular radiocommunications. In particular, the proposed solution applies especially, but not exclusively, in the context of LTE / 4G (Long Term Evolution) or 5G NR (New Radio) mobile networks.

[0002] More particularly, the invention relates to the use of a machine learning model for the prediction of quality of service (QoS) in a cellular telecommunications network, also called a mobile network. Previous Art

[0003] In the past, cellular telecommunications networks, or mobile networks, were designed to provide a basic level of service, such as voice and text communication. Over the last few decades, technological advances in user equipment (e.g., smartphones, tablets, laptops, etc.), as well as evolving user expectations, have led to a change in how users interact with their devices and the Internet.

[0004] Indeed, users now expect mobile networks to provide a diverse range of services, including broadband Internet, video streaming and real-time gaming.

[0005] One consequence of this evolution is the adaptation of mobile networks. In particular, to meet new user expectations, mobile network operators must be able to differentiate the mobile network configuration according to specific services and user needs. This means that mobile network operators must be able to adapt network performance based on the type of service provided and the user's requirements. For example, a mobile network may need to prioritize a video streaming service over voice communication or provide faster speeds for online gaming applications.

[0006] To enable mobile networks to adapt, new technologies have emerged such as: real-time analytics capabilities based on machine learning automatic and new radio communication technologies such as 5G. These new technologies have enabled mobile networks to provide new capabilities (e.g., data storage or analysis) and new services to users (e.g., broadband internet).

[0007] One of the main drivers of this transformation of mobile networks is the ability to store and analyze large volumes of data in real time using machine learning. Indeed, machine learning models allow mobile networks to detect patterns and anomalies in network traffic and adjust their parameters accordingly. In other words, by analyzing network data at the time of anomalies observed in the past, an artificial intelligence model can deduce the factors that cause these anomalies and therefore predict them in the future. Furthermore, the ability to differentiate mobile network parameters according to the specific needs of the service and users has become an essential requirement for mobile network operators.

[0008] Furthermore, the emergence of new capabilities (such as edge computing) linked to the deployment of 5G networks should bring significant improvements to mobile networks. For example, 5G networks are designed to provide ultra-reliable, low-latency communication (URLLC), which enables new use cases such as autonomous vehicles, remote surgery, and real-time monitoring of critical infrastructure (e.g., automated factories).

[0009] Furthermore, the implementation of 5G networks also leads to the development of network slicing, which allows mobile network operators to create multiple virtual mobile networks on a single physical mobile network infrastructure. Each mobile network slice can be customized with specific characteristics and service levels to meet the needs of different users and use cases.

[0010] Thus, the implementation of 5G networks opens up new opportunities for businesses and individuals to develop and use a variety of innovative applications. However, the most demanding use cases for 5G networks will require Quality of Service (QoS) control to ensure that they meet the performance levels required for each user and each service.

[0011] Quality of service control encompasses several critical parameters, including latency, jitter, throughput, and radio signal quality metrics such as RSRP (Reference Signal Received Power) and RSSI (Received Signal Strength Indicator (SSI). These parameters represent network performance and therefore determine the level of service quality that can be provided on the network. As such, it is essential that they be carefully managed and monitored.

[0012] The implementation of "Network Slicing" technology can enable quality of service control in 5G networks by providing different QoS levels for different network slices. Thus, network operators can ensure that the most demanding use cases receive the required level of service while optimizing the use of radio resources.

[0013] However, one of the significant challenges facing mobile network operators in providing QoS control for 5G networks is the constantly changing radio conditions. As users move through different areas (or cells) of the mobile network, they experience different radio conditions. Mobile network operators must then adjust network parameters to provide the required level of service.

[0014] However, it is extremely difficult to predict future radio conditions accurately, and mobile network operators may not be able to guarantee Service Level Agreements (SLAs) without a significant loss of radio resources.

[0015] Thus, despite the latest advances in radio communication technology and machine learning, it is still not possible today to individually adapt the characteristics of the mobile network based on a prediction of future quality of service. This means that mobile networks can still experience outages or performance issues that can impact the user experience. Although machine learning and real-time capabilities have helped mobile networks become more efficient and effective, there is still work to be done to ensure that mobile networks can anticipate and adapt to user needs in real time.

[0016] There is therefore a need to improve techniques for adapting to a change in QoS in a cellular telecommunications network context in order to ensure the best possible QoS for users. In particular, it is important to enable mobile networks to adapt upstream of a QoS variation in order to guarantee the best possible QoS for the user. Description of the invention

[0017] The present technique offers a solution to overcome certain drawbacks of the prior art. In one aspect, the present technique relates to a method for predicting a variation in the quality of service in a network cellular telecommunications comprising at least one base station to which at least one user device is connected. According to the general principle of the proposed technique, said method is implemented by a prediction device and it comprises

[0018] - a collection of at least one location information including an identifier of said at least one base station to which said at least one user device is connected,

[0019] - a selection, from a set of key performance indicators of said network, of at least one key performance indicator of interest representative of said quality of service provided by said at least one base station audited on at least one user device,

[0020] - a machine learning prediction of a future value of said at least one key performance indicator of interest from at least one time trajectory of said at least one key performance indicator of interest, said at least one time trajectory being representative of past values ​​of said at least one key performance indicator of interest, collected by said prediction device,

[0021] - a transmission, within said cellular telecommunications network, of a message adjusting at least one parameter of said cellular telecommunication network according to said predicted future value.

[0022] Thus, the invention proposes a completely new and inventive approach to predicting the variation in Quality of Service (QoS) within an operator's cellular telecommunications network (or mobile network). More specifically, the invention proposes to predict the QoS variation within the mobile network solely on the basis of metric values ​​collected by the various nodes of the operator's mobile network and on the location of a user device at the cellular level. In other words, there is no need to use geolocation information for user devices, but only to know their position in the network, for example, by knowing the base station to which the user device is connected. Indeed, the location of the user device is based on identifying the base station to which the user device is connected.In other words, the method according to the invention is based on the cellular-level localization of the user equipment and therefore on the detection of cell changes (for example, the operator's network knows if the user equipment is present in a given cell because it communicates with the cell's base station, and to which cell this user equipment moves when it connects to another base station). It is thus possible to mitigate QoS variations when the user equipment moves within the network.

[0023] The metrics collected are representative of indicators (for example: flow rate, latency, congestion...) which influence network performance and therefore directly or indirectly QoS. In particular, among these indicators, also called key performance indicators or KPIs ("Key Performance Indicator" in English), we find: KPIs not related to the service used (for example the SINR ("Signal to Interference plus Noise Ratio" in English), or related to the service used (for example: pixelation for an online video service).

[0024] The ability to predict the performance of a mobile network for each user device is an important development in the field of network optimization. Traditionally, mobile network performance could only be measured in real time or after the fact, which complicated the optimization of mobile network behavior to provide the best possible user experience. However, with the emergence of machine learning and predictive analytics, it is now possible to develop models capable of predicting the performance of a mobile network for each user.

[0025] In order to make a personalized prediction, one or more key performance indicators of interest, or KPIs, representative of the QoS provided by the telecommunications network to the user equipment, are selected from a set of network KPIs. The past values ​​of this or these KPIs of interest are collected by the prediction device and then represented as time trajectories. These time trajectories are therefore representative of the evolution of the KPIs of interest over time and space.

[0026] From these time trajectories, it is possible to predict, depending on the location at the cellular level of the user equipment, the future values ​​of the selected KPI(s) of interest and therefore the variation of the QoS of the end user.

[0027] Based on these future values ​​of the KPI(s) of interest, it is therefore possible to transmit a message to the operator to adapt the parameters of the telecommunications network (for example: increase the priority of the user equipment, reduce the quality of the video, etc.) to guarantee the best quality of service.

[0028] By using this model, network operators can offer a higher level of quality by predicting QoS degradation and taking steps to prevent it. For example, if network performance degradation is predicted, the network operator can adjust network parameters to ensure that the user continues to receive the expected level of service. This proactive approach can significantly improve the user experience, leading to greater customer satisfaction and loyalty. Furthermore, network operators can create offers tailored to the specific needs and usage patterns of each user. By predicting network performance, network operators can better understand the We understand the needs of each user and offer services that meet those needs. For example, a user who uses many jitter- and latency-sensitive applications may require a different level of service than a user who primarily uses voice services.

[0029] Thus, the ability to predict the performance of a mobile network for each user is an important development that offers numerous advantages to both mobile network operators and users. By using machine learning, mobile network operators can offer users a higher level of quality, create offers tailored to each user's needs, and improve user satisfaction and loyalty.

[0030] In a particular embodiment, said method further includes determining a user profile based on said at least one location information and at least one service nature information.

[0031] In this way, for each user, it is possible to determine a user profile based on, on the one hand, user location information (e.g., frequently moving from one cell to another, moving quickly, etc.) and, on the other hand, service information (e.g., network usage for online games, voice calls, online video, etc.). This information thus defines a user "behavior," characteristic of their habits in terms of movement and mobile network usage. By taking a user profile into account, it is therefore possible to predict the variation of QoS centered on the user.

[0032] Indeed, when there is good mobile network coverage, one of the factors that can cause QoS to decrease is the fact that the user's equipment moves within the mobile network. Thus, QoS variation is closely linked to the user's movement habits. These habits are characterized in particular by their location at the cell level over time (for example: presence in a cell every day at the same time, daily movements between the same cells, etc.).

[0033] QoS also depends on the type of service used by the user. Indeed, each service does not require the same level of network performance, and the user may also have signed a contract that guarantees a certain level of performance for a given service, for example.

[0034] Thus, travel habits, as well as service usage patterns, characterize a user profile. It is therefore possible to adapt the prediction of QoS variation according to the user profile.

[0035] In a particular embodiment, said selection of said at least one key performance indicator of interest takes into account said user profile and past values ​​of said network key performance indicators, collected by said device prediction.

[0036] Advantageously, in order to refine the prediction to the user profile, it is thus possible to select, from among all the key network performance indicators representative of QoS variation, one or more KPIs of interest. These KPIs of interest are adapted to the user profile, that is to say, they depend on the mobility of the user equipment and the type of service used.

[0037] In a particular embodiment, said method further comprises a generation of at least one prediction model associating said at least one user profile with said at least one time trajectory of said at least one key performance indicator of interest.

[0038] In this way, it is possible to predict the variation of personalized QoS, that is, QoS adapted to the user profile. Thus, for each user profile, there is a prediction model that takes into account key performance indicators of interest that depend on the user profile and, in particular, the service used. It is therefore possible to make a personalized, user-centered prediction and thus make the necessary network adaptations so that the QoS is tailored to the user's needs.

[0039] In a particular embodiment, said prediction implements an artificial intelligence module configured to implement said at least one prediction model.

[0040] In this way, through the use of such an artificial intelligence module, techniques such as, for example, deep learning techniques based on multilayer neural networks can be implemented to perform a prediction task from a large volume of diverse input data. More specifically, the network data or information already known and stored in memory in the nodes of the operator's mobile network is collected by the prediction device to train a machine learning algorithm that uses past data to predict future values ​​of key performance indicator(s) of interest representative of QoS based on the service used by the user equipment and its movements within the network (i.e., based on the user profile).Such a machine learning algorithm makes it possible to provide accurate predictions of future radio conditions and therefore of future QoS, thus allowing the network operator to adjust network parameters accordingly.

[0041] According to another aspect, the proposed technique also relates to a device for predicting a variation in quality of service in a cellular telecommunications network comprising at least one base station to which at least one user device is connected, said prediction device being configured to:

[0042] - collect at least one location information including an identifier of said at least one base station to which said at least one user device is connected,

[0043] - select, from a set of key performance indicators of said, at least a key performance indicator of interest representative of said quality of service provided by said at least one base station audited on at least one user device,

[0044] - predict by machine learning a future value of said at least one in key performance indicator of interest from at least one time trajectory of said at least one key performance indicator of interest, said at least one time trajectory being representative of past values ​​of said at least one key performance indicator of interest, collected by said prediction device,

[0045] - transmit within said cellular telecommunications network a message adjusting at least one parameter of said cellular telecommunications network according to said predicted future value.

[0046] In a particular embodiment, said device is further configured to determine a user profile based on said at least one location information and at least one service nature information.

[0047] In a particular embodiment, said selection of said at least one key performance indicator of interest takes into account said user profile and past values ​​of said key performance indicators of the network, collected by said prediction device.

[0048] In a particular embodiment, said prediction device implements a machine learning module configured to generate and implement at least one prediction model associating said at least one user profile with said at least one time trajectory of said at least one key performance indicator of interest.

[0049] The invention also relates to a computer program product comprising program code instructions for implementing a prediction method as described above, when executed by a processor.

[0050] The invention also relates to a computer-readable recording medium on which is recorded a computer program comprising program code instructions for executing the steps of the method for predicting a variation in quality of service in a cellular telecommunication network according to the invention as described above, when the program is executed by a processor.

[0051] Such a recording medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, for example a CD-ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a mobile storage (memory card) or a hard drive or an SSD.

[0052] On the other hand, such a recording medium can be a transmissible medium such as an electrical or optical signal, which can be transmitted via an electrical or optical cable, by radio, or by other means, so that the computer program it contains is executable remotely. The program according to the invention can, in particular, be uploaded to a network, for example, the Internet.

[0053] Alternatively, the recording medium may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute the steps or to be used in the execution of the method of predicting a variation in quality of service in a aforementioned cellular telecommunication network.

[0054] According to one embodiment, the present technique is implemented using software and / or hardware components. In this context, the term "module" or "device" may refer in this document to a software component, a hardware component, or a set of hardware and software components.

[0055] A software component corresponds to one or more computer programs, one or more subroutines of a program, or more generally to any element of a program or software capable of implementing a function or a set of functions, as described below for the module concerned. Such a software component is executed by a data processor of a physical entity (terminal, server, router, etc.) and is capable of accessing the hardware resources of that physical entity (memory, storage media, communication buses, electronic input / output cards, user interfaces, etc.). Hereafter, the term "resources" refers to any set of hardware and / or software elements supporting a function or service, whether they are individual or combined.

[0056] Similarly, a hardware component corresponds to any element of a hardware assembly capable of implementing a function or a set of functions, as described below for the module concerned. It may be a programmable hardware component or one with an integrated processor for executing software, for example an integrated circuit, a smart card, a memory card, an electronic card for executing firmware, etc.

[0057] Each component of the system described above naturally implements its own software modules.

[0058] The different embodiments mentioned above can be combined with each other for the implementation of the present technique.

[0059] The aforementioned prediction device, computer program and corresponding recording medium have at least the same advantages as those conferred by the method of predicting a variation in quality of service according to the present invention. Brief description of the figures

[0060] Other objects, features and advantages of the invention will become more apparent upon reading the following description, given by way of simple illustrative, and not limiting, example, in relation to the figures, among which:

[0061] [Fig-1]: [Fig.1] illustrates schematically a radio communication network cellular to which the prediction process can be applied according to different embodiments of the invention;

[0062] [Fig.2]: [Fig.2] represents in flowchart form the different stages of the learning phase of the prediction process according to an embodiment of the invention;

[0063] [Fig.3]: [Fig.3] represents an example of a heat map showing the evolution in time and space of the values ​​of a key performance indicator, or KPI;

[0064] [Fig.4]: [Fig.4] represents in flowchart form the different stages of the prediction method according to an embodiment of the invention;

[0065] [Fig.5]: [Fig.5] represents, in the form of a flowchart, an example of an application of the method for predicting a variation in quality of service according to an embodiment of the invention in a 2G, 3G or 4G type network;

[0066] [Fig.6]: [Fig.6] represents, in the form of a flowchart, an example of an application of the method for predicting a variation in quality of service according to an embodiment of the invention in a 5G type network;

[0067] [Fig.7]: [Fig.7] schematically illustrates an example of the architecture of a device for predicting a variation in quality of service in a cellular telecommunications network, according to an embodiment of the invention. Descriptions of the methods of implementation

[0068] The general principle of the invention is based on the spatio-temporal prediction of the variation in quality of service (or QoS) related to the movements of a user device within a cellular telecommunications network, or mobile network.

[0069] Indeed, when there is good mobile network coverage, one of the factors that can cause QoS to decrease is the fact that user equipment moves within the network and changes zones, or cells. In other words, when user equipment moves within the mobile network, it can pass from one zone to another and experience different radio conditions that will influence the quality of service.

[0070] Advantageously, this method for predicting QoS variation does not take into account the geolocation of the user equipment. Indeed, for a cellular telecommunications network operator, it is very complicated, or even sometimes impossible (particularly for reasons of confidentiality, for example), to retrieve the Geolocation data from a user device. Indeed, this geolocation data is not explicitly stored in the mobile network. In other words, the user device's geographic position may be known if GPS (or "Global Positioning System" or "Satellite Geopositioning") is enabled, but by default this information is not transmitted to the mobile network.

[0071] Thus, the prediction of QoS variation according to the invention is based solely on the use of data already known from the various nodes of the operator's mobile network (for example: type of service used by the user equipment, information on metrics, or KPIs, representative of network performance, location of the user equipment at the cellular level, etc.). This prediction of QoS variation then allows the operator to proactively adapt the mobile network parameters to compensate for this variation and guarantee the user the QoS best suited to their needs.

[0072] We first present, in connection with [Fig.1], a cellular radiocommunication network 1, or mobile network, composed of a network of relay antennas (or base stations) Ni to N; (i being an integer greater than zero), each covering a portion of territory delimited Ci to Ci (i being an integer greater than zero), commonly called a cell (represented schematically in hexagonal form on [Fig.1]), and routing communications in the form of radio waves to and from user equipment located in the corresponding cell.

[0073] To access the services offered by the network operator (voice or mobile data), a user device must therefore be located within the coverage area of ​​an N-cell base station. This base station has a limited range and only covers a small area around it, called a cell. To cover as much territory as possible and ensure that user terminals always have access to the services offered, operators deploy thousands of G-cells, each equipped with Ni-antennas, ensuring that their coverage areas overlap in order to provide as complete a network coverage as possible.

[0074] It should be noted that other infrastructures (hardware and / or software), well known, constitute this mobile network (such as a base station controller, etc.) thus forming the different nodes of the mobile network (not shown).

[0075] The mobile network comprises an access network and a core network. The access network is the part of the mobile network that connects end-user equipment (such as smartphones or laptops) to radio base stations and other network infrastructure. Access network data includes information such as signal strength, throughput, latency, and packet loss, which are important factors in determining network performance for individual users. The core network also plays a crucial role in determining the Network performance is affected, especially for data-intensive services such as video streaming or online gaming. Core network data includes information such as routing, service type, congestion, and load balancing, which can significantly impact network performance for individual users.

[0076] Consequently, to obtain accurate predictions, it is important to consider both access network and core network data. Indeed, by analyzing both access network and core network data, a more complete understanding of network performance can be gained, leading to a better user experience.

[0077] Thus, advantageously, the QoS variation prediction method according to the invention takes advantage of the collection and use of mobile network usage data already stored in memory in one or more mobile network nodes.

[0078] More specifically, user data or mobile network usage data refers to all data or information known as "network data" that is known to one or more network nodes. This set of network information includes both access network and core network information. This information includes, in particular: - "Location" information representing the location over time and at the cellular level of one or more user devices (for example, the presence of a user device in cell Ci, then movement to cell C2). It is the base station of the cell in which the user device is located that detects the presence of the user device when it connects to the base station. Location information includes, in particular, an identifier of the base station to which the user device is connected. It is therefore understood that by location information, we mean any information that allows the position of the user device within the network to be known at any given time; - information described as "service nature" representative of the nature of the service(s) used by a user's equipment (for example: online gaming, video, voice...), - Network metrics, or key performance indicators (also called network KPIs), which quantify network performance (e.g., throughput, latency, SINR, etc.) and therefore directly or indirectly influence QoS. These network KPIs include: • Key performance indicator(s) or KPIs known as "low-level" indicators, i.e., those not related to the service used and which will therefore influence indirectly affects QoS. These KPIs, known as "low-level" KPIs, are representative, for example, of radio signal quality, congestion conditions, etc. In other words, they do not allow for direct prediction of the impact on the service. Examples include the signal-to-interference ratio, or SINR. • High-level key performance indicators (KPIs), meaning service-dependent KPIs. In other words, for a given service (e.g., online video), one or more KPIs can be selected that represent the QoS of that service (e.g., video buffering, video freezing, pixelation). In this case, it is necessary to know in real time which service is being used by the user's device. This can be achieved either by directly knowing the nature of the service from the mobile network or by deducing it from known network data such as the bandwidth scheme, IP address, etc. It can also be predicted based on a history of all the services used by a user's device over a given period. Indeed, with a sufficient volume of data over time, a machine learning model can deduce which application a user might use over a specific period.For example, a user who watches an online video service every night, a music app on their way to work,... .

[0079] The method for predicting QoS variation in a cellular telecommunications network according to the invention comprises two main phases: a phase of learning one or more QoS variation prediction model(s), and a phase of implementing this / these prediction model(s). The various steps of the prediction method according to the invention are implemented by a machine learning (ML) module of a DISP prediction device. The DISP prediction device will be described in relation to [Fig. 7].

[0080] The machine learning module (ML) uses machine learning algorithms to predict future quality of service (QoS) from a single prediction model ("low-level" prediction) or from a combination of several prediction models ("high-level" prediction), thereby enabling the operator to adjust network parameters accordingly. This allows operators to improve network performance, enhance the user experience, and provide a more personalized and responsive service.

[0081] A flowchart representing the different stages of the learning phase of the models for predicting the variation of is now presented in connection with [Fig. 2]. QoS. This learning phase aims to train the machine learning (ML) module, implemented by the DISP prediction system, to identify different user profiles based on a set of network information, and to associate at least one prediction model with each profile. More specifically, for each user profile, one or more prediction models are associated based on previously selected metrics or KPIs of interest (low-level or high-level KPIs). The KPIs of interest are selected based on the service used by the user's equipment and the user's profile.

[0082] To this end, during the learning phase, during a step El, mobile network usage data, stored in memory in one or more mobile network nodes, is collected by the DISP prediction device. More specifically, for a group of several user devices, the DISP prediction device collects over a period P (for example, 24 hours) a set of information from an access network (ACC) and a core network (CO), such as: a set of location information for the user devices at the cellular level, a set of KPIs representative of network performance (low-level KPIs or high-level KPIs) during period P, and information on the nature of services used during period P.

[0083] User equipment does not need to send this network information because it is already known to one or more network nodes. For example, when a user device uses an online video service, the mobile network already knows the video quality, as well as other information such as bandwidth, latency, congestion, etc.

[0084] From this network information, the ML learning module creates a HIS history of network usage.

[0085] The network usage history (HIS) includes, for each user device in the group, a set of time trajectories representing the evolution, during period P, of a set of KPI values. In other words, this HIS history includes all possible values ​​of the metrics, or KPIs, collected during period P. These values ​​will therefore depend on: the radio conditions encountered by the user when moving through the network and the service(s) used by the user device. These time trajectories may or may not be anonymized depending on the user's consent.

[0086] From these temporal trajectories, the machine learning module (ML) also generates, in step 11, temporal maps representing the evolution of the KPIs over time and space. In particular, a heat map representation is used. An example of a heat map representing the evolution of the values ​​over time and space is shown in relation to [Fig. 3]. of a KPI. This heatmap (CTH_KPI_C_tps) represents the KPI values ​​(for example: a flow rate) VAL_KPI as a function of time and the attachment cell Cell_att, that is, the "cellular" location of the user equipment. A heatmap therefore allows us to represent both: - the evolution over time of the values ​​of a KPI, for a given cell C, (Val_KPI_Ci_tps), - at each past, present, or future moment, the variation of KPI values ​​across a set of cells, - the evolution over time of the KPI values ​​for all cells (Val_KPI_C_tps).

[0087] A heat map is therefore a graphical representation of data where the values ​​of this data are represented as colors (or shades of gray, in the example of [Fig. 3]). In the context of network optimization, a heat map can be used to visualize the various KPIs and services on a geographic map, allowing network operators to identify high- or low-performance areas of the network. The heat map can also help network operators optimize network performance by identifying areas where improvements are needed. For example, if many users are experiencing poor network performance in a particular area, network operators can use the heat map to identify the root cause of the problem and take steps to improve network performance in that area.

[0088] During step E2, also using the aforementioned network information, the machine learning module (ML) creates one or more user profiles (PUs). These user profiles (PUs) group together users who share the same profile, i.e., the same "behavior," in order to generate predictions for each type of user behavior, or profile. Behaviors are defined by: user location information (MOBs) (e.g., users who move frequently from one cell to another, quickly, etc.) and by service nature information (SERs) (e.g., network usage for online games, voice calls, online video, etc.). The different user profiles are therefore characteristic of users' habits. For example, there might be a group of users who take the same daily route using the same online music application.This group therefore has the same location information ([Fig. 1]: movement from cell C1 to cell C2 then cell C3) and service type (online music application). The prediction system therefore collects, for this group of users, network KPI measurements (latency, throughput, etc.) which will vary throughout the journey depending on radio conditions.

[0089] In a step E3, the temporal fineness FT, or granularity, used for the Implementation of MOD usage models. The temporal granularity of FT depends on: the nature of the KPIs considered (for example, some KPI values ​​may be obtained after 15 minutes, others after an hour), the service used, and the mobility of the user's equipment (for example, the user's equipment remains on a cell for 4 hours, then moves to another cell). Generally, "low-level" KPIs are reported in "real time." For other KPIs, particularly "high-level" KPIs, a certain processing time is required (for example, a few minutes or several hours). This processing time depends on the KPI and a predefined threshold. For example, for a KPI related to mobility, the processing time may be set to several hours to verify whether a user is passing through a cell or not.

[0090] The temporal precision (FT) thus allows us to define the prediction horizon (for example, we make a prediction for the next 15 minutes or 1 hour, etc.). In other words, the temporal precision (FT) depends on the scenario being considered: the mobility of the user equipment and the KPI(s) taken into account for the QoS evaluation.

[0091] In a step E4, the machine learning module ML of the prediction device DISP is therefore trained to associate with each user profile PU, time trajectories representative of the evolution, during the period P, of the values ​​of KPIs of interest selected from the set of network KPIs.

[0092] In order to select the most relevant KPIs for the user profile (PU) and thus associate them with appropriate time trajectories, the heatmaps generated in the EL step are used. Indeed, generating heatmaps can help identify the most relevant KPI / service pairs for each user profile (PU) by visualizing network performance data. In the previous example of the group of users with a daily commute and the use of the same online music application, a KPI of interest is, for example, bandwidth.

[0093] By displaying different KPIs and, possibly, services on the map, the heat map can help network operators identify areas with high or low network performance, and identify the specific KPIs that are most relevant to the user in each location and for each service.

[0094] Once the most relevant KPIs have been identified, the prediction model(s), centered on a user profile (PU), can use past values ​​of the KPIs of interest to predict future network performance for each user, based on their past usage patterns and other factors (e.g., the current state of the network, the user's location, etc.). In other words, based on past time trajectories of KPIs of interest and the user profile (PU), it is possible to predict the future values ​​of these KPIs of interest and thus, if necessary, adapt the network parameters to guarantee the ideal QoS for the user.

[0095] Thus, several MOD prediction models are created in step E4. These MOD prediction models depend on the user profile (PU), the selected KPIs of interest, and possibly the service used at the time of the prediction. In other words, depending on the user profile (PU), there are "low-level" prediction models and "high-level" prediction models. Indeed, each MOD prediction model is associated with its own list of KPIs of interest to be taken into account as input to the learning algorithm, as well as its own temporal precision. In other words, the same KPIs of interest are not always used for a new prediction. Prediction models are thus defined as follows: - so-called "low-level" KPIs, for which only "low-level" KPIs of interest unrelated to the service used during the implementation of the prediction process are taken into account. - so-called "high-level" KPIs for which we take into account, alternatively or in combination, KPIs of interest so-called "high-level" dependent on the service (for example: for online video, we take into account KPIs specific to this type of application).

[0096] Each prediction model MOD is then stored S_MOD (see [Fig.4]) in a memory of the prediction device DISP.

[0097] In addition, each model is also associated with a learning rate and / or an interference rate.

[0098] More specifically, the learning rate is a parameter used during the training of a machine learning model. It determines how quickly the model adjusts its weights and parameters based on the errors it makes when predicting outcomes. A high learning rate can allow the model to converge more quickly to a solution, but also carries the risk of not taking into account certain local optima. A low learning rate can slow down the training of the machine learning model, but it generally allows for a better exploration of the solution space. The choice of the learning rate is generally made based on the complexity of the model, the size of the dataset, and the nature of the problem to be solved.It is common to adjust the learning rate over time (for example, by gradually decreasing it) to improve the model's stability and accuracy. The interference rate, or regularization rate, is a measure of a model's ability to generalize training data to new data. Interference occurs when the model fits the training data too closely, which can lead to poor performance on test or real-world data. To avoid this behavior, regularization techniques are used, such as L1 / L2 regularization, stalling, etc. The interference rate... The extent of these regularization techniques is controlled. A higher interference rate reduces the model's ability to fit too closely to the training data, but can also reduce its ability to capture complex patterns and impair its performance on the training data. A lower interference rate allows the model to fit the training data better, but can also lead to overfitting and poor generalization.

[0099] It should be noted that the optimal choice of learning rate and interference rate may vary depending on the specific model, dataset, problem to be solved, and other factors. According to a particular feature of the proposed technique, cross-validation is performed to identify the best parameters (learning rate and / or interference rate) for each model used in predicting a given KPI. The various steps of the process for predicting QoS variation according to one embodiment of the invention are now presented in relation to [Fig. 4] in the form of a flowchart.

[0100] After the end of the learning phase, the ML learning module of the DISP prediction device is therefore able to implement a so-called "low level" prediction model, and / or a so-called "high level" prediction model.

[0101] Thus, in a phase of implementing the prediction models, the prediction device DISP collects DUE a new set of network information for a new user device or for a new group of user devices.

[0102] Based on this network information, the machine learning module ML then searches R_PU for a similar user profile from the set of user profiles PU created and stored in memory S_PU during the training phase. Once a similar user profile D_PU is determined, the learning module selects S_MOD, based on this profile, at least one prediction model associating the user profile PU with the time trajectories of the KPIs of interest selected according to the user profile and the service used.

[0103] The learning module then applies C_M0D the selected prediction model(s) taking into account as input the values ​​of the KPIs of interest (among the network information collected during the DUE collection) of the new user equipment or group of user equipment, and thus predicts a future value of at least one key performance indicator of interest from at least one time trajectory of the selected key performance indicator(s).

[0104] It is then possible to determine whether the predicted QoS is considered adequate (QoS_ok = 0) or not (QoS_ok = N). If the predicted QoS is considered adequate (QoS_ok = 0), the prediction process is stopped. Otherwise, if the predicted QoS is not considered sufficient, the process is stopped. (QoS_ok = N), the network operator can then modify (ADAP) the mobile network parameters to ensure that the end user has a QoS adapted to their needs.

[0105] To this end, the prediction device transmits a network parameter adjustment message to the operator. For example, the prediction device sends an adjustment message to the OAM (Operations, Administration and Maintenance).

[0106] This enables network operators to take proactive measures to optimize network performance and avoid service quality degradation, thereby providing a better user experience and increasing customer satisfaction.

[0107] Examples of QoS variation prediction and mobile network adaptation based on QoS variation prediction models according to an embodiment of the invention are now presented in connection with Figures 5 and 6.

[0108] In the example shown in connection with [Fig. 5], the prediction model is a so-called "low-level" model that only takes into account a KPI of interest unrelated to the service, representative of radio conditions, for example the SINR. In this example, the prediction model is applied to a 2G, 3G or 4G mobile network.

[0109] Over a period P' defining the temporal granularity of the prediction (for example, a few minutes), the machine learning module (ML) collects a set of network information associated with a user's network usage. It should be noted that the finer the granularity, the more accurate the prediction. From the location information and the type of service(s) used over the period P', the machine learning module determines the profile of this new user and identifies, among the user profiles (PUs) already stored in memory during the training phase, the one that is similar to the new user.

[0110] Depending on the user profile determined, and based on a so-called "low-level" prediction model, the learning module selects a KPI of interest, the SINR for example.

[0111] The machine learning module then identifies, based on the selected KPI of interest, in this case SINR, a suitable prediction model from among the prediction models stored in memory. The machine learning module then implements the identified prediction model to predict (DEG KPI), from a heat map, the future values ​​of the KPI of interest considered, here SINR.

[0112] If a degradation of the SINR, and therefore a loss of throughput, is predicted based on the movements of the user equipment, then the machine learning module (ML) checks (Ex_SER) which service is being used and whether it requires high throughput, for example. If this is not the case (N), then the prediction process is terminated (STOP). In other words, the degradation of the SINR has little or no impact on the QoS for this user.

[0113] If, on the other hand, the service used requires high bandwidth (O), for example, online video, then if this bandwidth loss (i.e., the degradation of the SINR) is significant enough to cause video degradation and therefore QoS degradation, then the operator can modify the network parameters. For example, the operator can reduce the video resolution or instruct the network to increase the client's priority level (Adap_Res).

[0114] If no SINR degradation is predicted (N), then it is possible to switch to a so-called "high-level" prediction model by taking into account other KPIs (KPI_ADD) as well. In particular, for a "high-level" model, it is possible to select one or more KPIs associated with the service used during period P' and then predict the future QoS (Pred_QoS).

[0115] To do this, the machine learning module determines, based on service nature information, which service is being used (for example, an online video service). From the user profile and the nature of the service, the machine learning module selects one or more KPIs of interest related to the service used (for example: video freezing, pixelation, etc.).

[0116] Based on the user profile and the selected KPIs of interest, the machine learning (ML) module selects the appropriate prediction model. To this end, the ML module determines (QoS_ok?) whether the QoS is satisfactory or not. According to this prediction model, if the QoS is satisfactory (O), the prediction process is terminated (STOP). Conversely, if the QoS is not satisfactory (N) and the KPI(s) degrade significantly, then the operator can modify the network parameters and reduce the video resolution or instruct the network to increase the client's priority level (Adap_Res), for example.

[0117] In connection with [Fig.6] we present an example of QoS variation prediction also based on a so-called "low-level" prediction model taking into account only one KPI of interest not related to the service, representative of radio conditions, for example the SINR, but this time applied in a 5G type network.

[0118] The steps are similar to those described in relation to [Fig. 5], except that the specific characteristics of the 5G network are taken into account. Thus, if a degradation of the SINR, and therefore a loss of throughput, is predicted based on the movements of the user equipment, then the machine learning module checks whether the network slice or the user's SLA is satisfactory (Ex_SLA). If so (O), then the prediction process is terminated (STOP). Otherwise (N), the operator can reduce the video resolution or instruct the network to increase the customer's priority level (Adap_Res).

[0119] If no degradation of the SINR is predicted (N) then the same steps are implemented as in connection with [Fig.5].

[0120] Thus, the user-centered prediction model(s) described above for predicting variations in quality of service can be applied to various types of networks, including 4G and 5G networks. However, the benefits may be greater for 5G networks due to the capabilities of the technology.

[0121] Indeed, even if the network slicing function is not required, it can improve the effectiveness of the prediction model. Network slicing allows the creation of multiple virtual networks with different characteristics, such as different levels of quality of service. With network slicing, the prediction method according to the invention can predict network performance for each user in different slices, enabling network operators to offer tailored services that meet the needs of each user.

[0122] However, even without network segmentation, the prediction method according to the invention can still be useful for optimizing network performance and providing a better user experience. By predicting network performance for each user, network operators can take proactive measures to prevent degradation and improve the overall quality of service.

[0123] In order to illustrate more precisely the principle of the invention, [Fig.7] schematically presents the architecture of a DISP prediction device, according to one embodiment of the invention.

[0124] The DISP prediction device includes a machine learning module ML.

[0125] The prediction device DISP further comprises a random access memory (RAM), a processing unit (CPU) equipped, for example, with a processor, and controlled by a computer program stored in read-only memory (ROM or a hard drive). At initialization, the code instructions of the computer program are, for example, loaded into the RAM before being executed by the processor of the CPU.

[0126] The DISP prediction device further includes a MEM memory for storing data from the communication network 1, such as the type of service used, KPIs and metrics representative of network performance, mobility information of one or more user devices...

[0127] The DISP prediction device also includes a COM communication module for receiving data from mobile network nodes and transmitting a network parameter adjustment message.

[0128] Fig. 7 illustrates only one particular way, among several possible ways, of The DISP prediction device is designed to perform the steps of the process for predicting a variation in quality of service in a communication network detailed above, in relation to Figures 2 to 6 in its various embodiments. These steps can be performed either on a reprogrammable computing machine (a PC, a DSP processor, or a microcontroller) executing a program comprising a sequence of instructions, or on a dedicated computing machine (for example, a set of logic gates such as an FPGA or an ASIC, or any other hardware module).

[0129] In the case where the prediction device DISP is made with a reprogrammable computing machine, the corresponding program (i.e. the sequence of instructions) may be stored in a removable storage medium (such as for example an SD card, a USB key, a CD-ROM or a DVD-ROM) or not, this storage medium being readable partially or totally by a computer or a processor.

Claims

Claims

1. Method for predicting a variation of a quality of service (QoS) in a cellular telecommunications network comprising at least one base station to which at least one user equipment is connected, characterized in that said method is implemented by a prediction device and comprises: - a collection (DUE) of at least one location information (IL) comprising an identifier of said at least one base station to which said at least one user equipment is connected, - a selection (S_MOD), from a set of key performance indicators (KPIs) of said network, of at least one key performance indicator of interest representative of said quality of service provided by said at least one base station to said at least one user equipment,- a prediction (C_MOD) by automatic learning of a future value of said at least one key performance indicator of interest from at least one time trajectory of said at least one key performance indicator of interest, said at least one time trajectory being representative of past values of said at least one key performance indicator of interest, collected by said prediction device, - a transmission (ADAPT), within said cellular telecommunications network, of a message for adjusting at least one parameter of said cellular telecommunications network as a function of said predicted future value.,

2. Prediction method according to claim 1, characterized in that it further comprises: - a determination of a user profile as a function of said at least one location information and at least one service nature information.

3. Prediction method according to claim 2, characterized in that said selection of said at least one key performance indicator of interest takes into account said user profile and past values of said key network performance indicators, collected by said prediction device.

4. Prediction method according to claim 3, characterized in that it further comprises a generation of at least one prediction model associating said at least one user profile with said at least one tra- temporal target of said at least one key performance indicator of interest.

5. Prediction method according to claim 4, characterized in that said prediction implements an artificial intelligence module configured to implement said at least one prediction model.

6. Device for predicting a variation in a quality of service in a cellular telecommunications network comprising at least one base station to which at least one user equipment is connected, characterized in that said prediction device is configured to: - collect at least one location information comprising an identifier of said at least one base station to which said at least one user equipment is connected, - select, from a set of key performance indicators of said network, at least one key performance indicator of interest representative of said quality of service provided by said at least one base station to said at least one user equipment, - predict by machine learning a future value of said at least one key performance indicator of interest from at least one time trajectory of said at least one key performance indicator of interest,said at least one time trajectory being representative of past values of said at least one key performance indicator of interest, collected by said prediction device, - transmitting within said cellular telecommunications network a message for adjusting at least one parameter of said cellular telecommunications network as a function of said predicted future value.,

7. Prediction device according to claim 6, characterized in that it is further configured to determine a user profile based on said at least one location information and at least one service nature information.

8. Prediction device according to claim 7, characterized in that said selection of said at least one key performance indicator of interest takes into account said user profile and past values of said key network performance indicators, collected by said prediction device.

9. Prediction device according to claim 8, characterized in that it implements a machine learning module configured to generate and implement at least one prediction model associating said at least one user profile with said at least one trajectory. temporal of said at least one key performance indicator of interest.

10. A computer program product comprising program code instructions for implementing a method according to any one of claims 1 to 5, when executed by a processor.

11. A computer-readable recording medium comprising program code instructions which, when executed by a processor, cause the processor to carry out a method according to any one of claims 1 to 5.

Images