Method for managing resources of at least one access device of a telecommunications network, device, access equipment, control equipment, system and corresponding computer programs.

Abstract

The invention relates to resource management of at least one access device of a telecommunications network, comprising:- obtaining (31) location information (IL) of a geographical area (ZC) served by said access device;- obtaining (32) demographic (ID) and topographical (IT) information relating to said geographical coverage area;- predicting (36) a set of values ​​of at least one communication data traffic indicator of terminal equipment connected to said access device for a given time period, at least from the demographic and topographical information obtained; and- deciding (37) on at least one action for managing the resources of the access device at least as a function of the set of values ​​predicted for the given time period and of technical characteristics relating to the resources of said at least one access device. FIGURE 3

Description

Title of the invention: Method for managing resources of at least one access device of a telecommunications network, device, access equipment, control equipment, system and corresponding computer programs. 1. Field of the invention

[0001] The invention lies in the field of telecommunications, and in particular in that of telecommunications networks.

[0002] In particular, the invention relates to the management of resources for the implementation of such a telecommunications network.

[0003] It applies in particular but not exclusively to a mobile telecommunications network whose architecture complies with the 3GPP standard (for “Third Generation Partnership Project”, in English), in one of its current or future versions. 2. Prior art and its drawbacks

[0004] Methods are known for planning the deployment of new radio antennas to develop a mobile telecommunications network infrastructure in emerging countries or EMEA (from the English, "Europe Middle East & Africa").

[0005] These methods generally include two phases: 1. An initial phase of research into geographic areas to be covered or strengthened. It uses public and private data to identify and characterize these areas: demographic, topographic (infrastructure, etc.), performance indicators of the current radio network, wealth indicator (an average), etc. In rural areas, satellite data is sometimes used in addition. This first phase allows, in a fairly rough manner, to identify and prioritize population areas located outside radio coverage.

[0006] This research phase relies, for example, on a software tool such as Facebook's Network Planner © tool, which allows demographic information and mobile network usage information to be retrieved using data collected from users of the Facebook © mobile application. Such a tool is especially usable for planning an update (or "upgrade" in English) of equipment at an existing site, for example a base station, the addition of a radio antenna or the replacement of a radio antenna with another of a later generation, etc. Because it relies on traffic indicators in existing radio cells, it is more suited to extending coverage or strengthening capacity of a cellular access network at the edge as well as the deployment of a new cellular access network; 2. A second phase of selection and negotiation of locations for the new site or additional radio equipment. For example, the chosen locations are a vacant lot or a building roof. This second phase is based on a radio planning tool, such as Atoll's Forsk © tool. Such a tool makes it possible to evaluate candidate locations by taking into account the topology of the terrain (relief and buildings in particular) and to determine the one that offers the best coverage / cost ratio.

[0007] A disadvantage of existing techniques is that they allow at best to qualitatively identify a need for deployment of new radio antennas or updating of an existing site, which is not sufficient for effective planning and more generally for fine management of the resources of a cellular network over time.

[0008] The invention improves the situation. 3. Presentation of the invention

[0009] The invention meets this need by proposing a method for managing resources of at least one access device of a telecommunications network, comprising: - obtaining location information of a geographical area served by said access device; - obtaining demographic and topographical information relating to the said geographical area; - the prediction of a set of values ​​of at least one indicator of communication data traffic of terminal equipment connected to said access equipment for a given time period, at least from the demographic and topographical information obtained; and - the decision of at least one action for managing the resources of the access equipment at least as a function of the set of values ​​predicted for the given time period and of technical characteristics relating to the resources of said at least one access equipment.

[0010] The invention is based on a completely new and inventive approach to the management of resources implemented or to be implemented in access equipment of a telecommunications network, which consists of recovering demographic and topographical information of the geographical area served by this equipment and quantitatively predicting the values ​​of at least one communication data traffic indicator and their variations in this geographical area for a given time period. With the invention, the data available for the supervision of an access network are enriched by the time sequence thus predicted, which is then used to manage the resources of this network more precisely and efficiently.

[0011] The invention applies both to the deployment of new access equipment in a geographical area not covered or poorly served by the current network, and to the reinforcement or updating of resources already deployed on an existing site.

[0012] According to one aspect of the invention, the access equipment is a base station of a mobile telecommunications network and the geographical area corresponds to the geographical radio coverage area of ​​a radio cell associated with the base station.

[0013] The invention is particularly interesting in the case of a cellular access network to a mobile telecommunications network, to manage all the hardware and / or software resources of a base station. Of course, the invention is not limited to this type of network and more generally concerns any access network to a telecommunications network, fixed or mobile, whatever the wireless access technology, for example Wifi, or wired, for example xDSL or fiber, used.

[0014] According to another aspect of the invention, the management action comprises a command to deploy at least one additional resource in the access equipment and belongs to a group comprising at least: - an order to deploy an additional radio antenna in the access equipment, when the access equipment is a base station of a mobile telecommunications network; - an order to replace an existing antenna of the access equipment with a later generation antenna, when the access equipment is a base station of a mobile telecommunications network; - an order to remove an existing antenna from the access equipment, when the access equipment is a base station of a mobile telecommunications network; - an order for the deployment of access equipment in the telecommunications network; - an order to move the access equipment in the telecommunications network.

[0015] The invention applies to the updating and strengthening of an already existing site. In this case, the quantitative prediction of a user data traffic indicator over time allows it to more precisely evaluate the needs in the geographical area considered, to more precisely predict the type of equipment most suitable for deployment and therefore to optimize the investment. Equipment already deployed can also be replaced, removed or even moved. It is for example possible to move a base station on board an aerial platform to to cope with a disaster or anticipate an event requiring a concentration of resources at a geographical point.

[0016] The deployment command in question can then be transmitted to a telecommunications network management equipment and / or mapped in the technical information system of this network.

[0017] According to yet another aspect of the invention, the management action comprises an adjustment of the resources allocated to the access equipment and belongs to a group comprising at least: - an adjustment of a transmission power level and / or an orientation of at least one antenna of the access equipment, when the access equipment is a base station; - an adjustment of a quantity of hardware and / or software resources allocated to the operation of the access equipment; - an adjustment of traffic management functions carried by the access equipment.

[0018] The invention also applies to the management of resources already deployed on a site of the telecommunications network and in this case, the quantitative knowledge of a data traffic indicator over a given time period is used to dynamically adjust the allocation of resources already deployed at the level of the access equipment and, in this way, improve the quality of service offered to users.

[0019] According to yet another aspect of the invention, the communication data traffic indicator belongs to a group comprising at least: - a downlink and / or uplink communication data rate; - a number of terminal equipment connected to the access equipment; - an overall volume of data traffic carried by the access equipment; - a typology of communication data carried by the access equipment.

[0020] Typology is understood to mean a set of characteristics of traffic data defined according to a type of traffic considered. For data traffic for example, it includes for example quality of service information relevant to this type of traffic, such as latency, quality of service class, throughput associated with the traffic, security level associated with the data traffic.

[0021] The advantage of such a typology indicator is that it makes it possible to decide on a management action that responds to the typology and the evolution of this typology over time. The action may consist of deploying or updating traffic management functions (such as edge computing functions, adding a class of service, optimization functions, sensitive traffic security services, depending on the predicted traffic data types.

[0022] According to another aspect of the invention, the prediction also takes into account said technical characteristics of the resources of said at least one access device.

[0023] For a base station, this concerns, for example, the height of the radio antenna(s), their power, their orientation. These technical characteristics concern resources already deployed and in operation or resources, the deployment of which is envisaged in the future. It is assumed that they are known to the device which implements the invention and have been obtained beforehand. An advantage is to supplement the demographic and cartographic data in order to improve the quality of the prediction.

[0024] According to yet another aspect of the invention, the method comprises obtaining measurement information of at least said communication data traffic indicator over the given time period and the prediction also takes into account said measurement information.

[0025] These measurements are for example reported by the access equipment itself when it is already deployed or by neighboring access equipment.

[0026] Advantageously, the measurements obtained relate to several performance indicators of communication data traffic in the geographical area concerned.

[0027] According to another aspect of the invention, the method comprises the prior learning of a prediction model of the information obtained in said set of values ​​of at least one communication data traffic indicator and in that said prediction implements an artificial intelligence module configured to use said prediction model.

[0028] One advantage of using artificial intelligence techniques such as machine learning is that it allows the execution of a very complex task which would be problematic or even impossible to achieve using conventional algorithmic means.

[0029] According to one embodiment, a neural network is used, for example of the deep learning network type or a machine learning algorithm of the decision tree type.

[0030] According to yet another aspect of the invention, the method comprises the aggregation of the information obtained for said geographical area into an input vector comprising a number, called the input number, of first components equal to a number of information items obtained, and in that said prediction comprises the transformation, from said prediction model, of the input vector into an output vector, representative of said set of values ​​of the traffic indicator, comprising a number of components, called the output number, equal to a number of time intervals in the given time period.

[0031] The information aggregated in the input vector comprises at least the demographic and topographical information collected for the geographical area concerned. Advantageously, they comprise all the information obtained and in particular the technical characteristics of the access equipment and any measurement information sent back by the access equipment when it is already deployed in the geographical area.

[0032] For example, the number of inputs is equal to 271, which corresponds to as many demographic and topographical information of different types and the number of outputs is equal to 168 = 7 days x 24 h, which corresponds to an hourly granularity over a time period of one week.

[0033] According to another aspect of the invention, said transformation comprises the discretization of the first components of the input vector and the formation of a vector, called the discretized vector, the components of which comprise the first discretized components, the translation of the discretized vector into a so-called compressed vector, representative of said set of values ​​of said traffic indicator and comprising a number of components, less than the output number, and the decoding of the compressed vector into the output vector.

[0034] Advantageously, the neural network used is of the transformer type. It is a particular neural network initially used for the processing of natural language and in particular the translation from one language to another. The invention here proposes to adapt it to the automatic conversion of demographic / topographical information into a time series of values ​​of a user data traffic indicator.

[0035] An advantage of the transformer neural network implemented by the invention is that it is configured to first produce a compressed output time series, which is then decoded so as to return an output vector having the desired dimensions.

[0036] According to yet another aspect of the invention, the method comprises, following the execution of the management action, obtaining a new set of measurements of the communication data traffic indicator over a new time period and updating the learning of the data transformation model from said new set.

[0037] An advantage of this updating of the learning of the analysis model using real measurements of the traffic indicator is to make this model evolve over time to make a prediction of a time sequence of this indicator that is increasingly fine and precise.

[0038] The invention also relates to a device for managing resources of at least one access device in a telecommunications network, said device being configured to implement: - obtaining location information for a geographic area served by said access equipment; - obtaining demographic and topographical information relating to the said geographical area; - the prediction of a set of values ​​of at least one indicator of communication data traffic of terminal equipment connected to said access equipment for a given time period, at least from the demographic and topographical information obtained; and - the decision of at least one action for managing the resources of the access equipment based on the set of values ​​predicted for the given time period.

[0039] Advantageously, said device is configured to implement the steps of the resource management method as described previously.

[0040] Advantageously, said device is integrated into control equipment configured to control at least one access equipment of said telecommunications network.

[0041] Advantageously, said control equipment is included in a resource management system in a telecommunications network, said system comprising at least one access equipment of said network.

[0042] The system, the control equipment and the resource management device have at least the same advantages as those conferred by the aforementioned resource management method.

[0043] According to an alternative embodiment, the resource management device is integrated into access equipment of a telecommunications network. The system, the access equipment and the resource management device have at least the same advantages as those conferred by the aforementioned resource management method.

[0044] The invention also relates to a computer program product comprising program code instructions for implementing the resource management method as described above, when executed by a processor.

[0045] A program may use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0046] The invention also relates to a recording medium readable by a computer on which is recorded a computer program comprising instructions for program code for executing the steps of the methods according to the invention as described above.

[0047] Such a recording medium may be any entity or device capable of storing the program. For example, the medium may comprise 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 medium (memory card) or a hard disk or an SSD.

[0048] On the other hand, such a recording medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means, so that the computer program it contains is remotely executable. The program according to the invention may in particular be downloaded over a network, for example the Internet.

[0049] Alternatively, the recording medium may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the aforementioned method.

[0050] According to an exemplary embodiment, the present technique is implemented by means of software and / or hardware components. In this regard, the term "module" may correspond in this document to a software component, a hardware component or a set of hardware and software components.

[0051] A software component corresponds to one or more computer programs, one or more sub-programs of a program, or more generally to any element of a program or software capable of implementing a function or a set of functions, according to what is described below for the module concerned. Such a software component is executed by a data processor of a physical entity (terminal, server, gateway, set-top-box, router, etc.) and is capable of accessing the hardware resources of this physical entity (memories, recording media, communication buses, electronic input / output cards, user interfaces, etc.). Subsequently, resources are understood to mean all sets of hardware and / or software elements supporting a function or a service, whether unitary or combined.

[0052] 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.

[0053] Each component of the system described above of course implements its own software modules.

[0054] The different embodiments mentioned above can be combined with each other for the implementation of the present technique. 4. Brief description of the figures

[0055] Other aims, characteristics and advantages of the invention will appear more clearly on reading the following description, given as a simple illustrative, and non-limiting, example, in relation to the figures, among which:

[0056] [Fig-1]: [Fig.l] schematically illustrates the context of an embodiment of the invention;

[0057] [Fig.2]: [Fig.2] schematically illustrates an example of architecture of a resource management system of a base station in a mobile telecommunications network implemented according to an embodiment of the invention;

[0058] [Fig.3]: [Fig.3] describes in the form of a flowchart the steps of a method of resource management of equipment providing access to a telecommunications network, according to an exemplary embodiment of the invention;

[0059] [Fig.4A]: [Fig.4A] schematically illustrates an example of location information of a geographic radio coverage area associated with a base station, obtained by a Voronoi decomposition;

[0060] [Fig.4B]: [Fig.4B] schematically illustrates an example of a map representing the lines of the borders of the radio coverage areas associated with the base stations deployed in a territory, obtained from the location information of these radio coverage areas provided by such a Voronoi decomposition;

[0061] [Fig.5]: [Fig.5] illustrates an example of a topographic map obtained from topographical information relating to a radio coverage area of ​​a base station;

[0062] [Fig.6]: [Fig.6] schematically illustrates the prediction of an output vector representative of a time sequence of values ​​of a user data traffic indicator in a radio coverage area associated with a base station, from an input vector comprising demographic and topographical information of the radio coverage area, according to a first embodiment of the invention;

[0063] [Fig.7]: [Fig.7] details the steps of said prediction, when it implements a “transformer” type neural network according to a second embodiment of the invention; and

[0064] [Fig.8]: [Fig.8] describes an example of hardware structure of a management device of resources of an access device according to the invention.

[0065] 5. Detailed description of the invention

[0066] The general principle of the invention is based on the prediction of a set of values ​​of at least one communication data traffic indicator of terminal equipment attached to the geographical area served by access equipment of a telecommunications network, for example that of the coverage area of ​​a radio cell of a base station of a mobile telecommunications network, during a given time period.

[0067] According to the invention, this prediction is carried out from demographic and topographical information collected in the geographical area concerned and aggregated to form a sequence of input information representative of the geographical coverage area concerned.

[0068] From the predicted set, at least one action for managing the resources of the access equipment is decided.

[0069] The invention thus makes it possible to estimate the needs of the terminal equipment in the geographical coverage area of ​​the access equipment, in terms of communication data traffic over a given time period and to deduce therefrom one or more resource management actions to be implemented so that the telecommunications network can satisfy these needs.

[0070] In the following, this notion of communication data traffic indicator is used to designate any information relevant to measuring a level of activity in the geographical area served by the access equipment considered.

[0071] Without loss of generality, this is for example information relating to: - a downlink communication data rate, from the access equipment to the terminal equipment; - a number of terminal equipment connected to the access equipment; - an overall volume of data traffic in the geographic area; - a level of use of hardware or software resource units of the access equipment, for example PRB (Physical Resource Blocks) frequency resource units by a base station of a mobile telecommunications network; or - a volume of data exchanged for a given type of service or class of service; or of a volume of energy data consumed by the access equipment.

[0072] We commonly speak of a key performance indicator (KPI) of communication data traffic in the geographical area served by the access equipment.

[0073] In the following, this communication data traffic will simply be referred to as user data traffic or data traffic, without loss of generality, the terminal equipment capable of connecting to the access equipment. including both user terminals, such as mobile phones or laptops for example, as well as loT (Internet of Things) connected objects, configured to communicate with each other without user intervention.

[0074] Demographic information is understood here to mean quantitative and qualitative information on characteristics of populations in a geographical area, such as age, sex, marital status, birth rate, level of education, socio-professional category, etc.

[0075] Topographical information of a geographical area means information relating to the layout and relief of this area, including for example elements of natural origin (rivers, mountains, forests, etc.) and elements of human origin (cities, infrastructures such as buildings, roads, bridges, bus stops, airports, etc.) and information relating to their surface area, their destination (commercial, residential, etc.), their economic activity, etc.

[0076] In the following, resources are used to designate both hardware resources and software resources of the access equipment.

[0077] This notion of resource encompasses the access equipment itself, other resources of the access equipment, such as a data modulation module, a module for allocating time slots for data transmission, a module for managing the quality of service associated with the different types of data transmitted (voice, loT (from the English, "Internet of Things"), video, etc.). For a base station of a cellular access network for example, these are for example particular components of this radio antenna, such as PRB resource units, modulation modules, time slot allocation modules for data transmission, management of the quality of service QoS (from the English, "Quality of Service") for the different types of data transmitted, etc.

[0078] The management action is decided on the basis of the predicted set and advantageously of information relating to the capacity of the resources already implemented or to be implemented of the access equipment and, when available, of performance measurements of the resources already deployed in the geographical area served by the access equipment. For example, the time sequence of the traffic indicator thus predicted is compared with such performance measurements, for example values ​​of communication data traffic indicators measured and returned by the access equipment via a management interface with at least one network control equipment.

[0079] This makes it possible to assess whether the resources of the access equipment are sufficient to meet the needs of the terminal equipment, currently or in the short / medium term.

[0080] For example, the evolution of the values ​​of the user data traffic indicator over the time period, due to an increase in the number of inhabitants or the construction of a residential subdivision in the area, may highlight that the resources currently deployed are insufficient to guarantee an acceptable level of quality of service at peak times.

[0081] The management action decided may then include a recommendation or prioritization of an extension of radio coverage, by the deployment of new radio equipment, such as a new base station in a geographical area located in a rural or semi-rural environment, on the edge or not of an area already covered. Conversely, it may also include a recommendation to reduce radio coverage by the withdrawal of radio resources that have become superfluous.

[0082] Alternatively, the management action may consist of a recommendation to update or reinforce the resources already implemented in a geographical area already covered by the mobile telecommunications network, for example located in an urban environment, in order to improve the quality of service or to cope with a change in the demographics or topography in this geographical area. For example, the upcoming opening of a new shopping center may cause an influx of population in certain time slots and therefore lead to increased demand on the network resources, creating a need to reinforce the existing resources. The reinforcement may be provided by the installation of a new Wi-Fi terminal at the shopping center or, if it is an extension of an existing shopping center, a Wi-Fi repeater.For a mobile telecommunications network, installing a femtocell can also be a solution to consider.

[0083] According to another example, the installation of an industrial or office zone may generate a need for a stable and high-speed Internet connection, quality mobile access, and an evolution of data traffic towards more voice or inter-object loT (Internet of Things) communications. To support such an evolution of traffic, the management action may also include a recommendation to deploy new data traffic management resources such as optimization functions, cache management, quality of service management, virtual private network or VPN (Virtual Private Network) management, implementing a secure data communications tunnel within the telecommunications network, or even edge computing functions which make it possible to optimize data processing while remaining close to the source of the data.The invention finally makes it possible to improve the management of resources already deployed at the level of access equipment. Indeed, knowledge of the variations of a user data traffic indicator on . a time period with a sufficiently fine granularity, for example one hour, makes it possible to allocate the hardware and / or software resources of the base station more precisely according to the time of day, the movements of people from their place of residence to their place of work, for example. This dynamic management is made possible by the development of self-organized networks (SON), which are configured to implement feedback loops, i.e. monitoring data is measured and fed back by the access equipment and used by control equipment to adjust the configuration of this access equipment or dynamically allocate / de-allocate virtual machines within this access equipment.The characteristics of these virtual machines in terms of CPU (Central Processing Unit) computing capacity and memory, such as RAM and Disk, are limited by the capabilities of the underlying server equipment and can be dynamically modified. Servers can therefore be switched off to save energy, or switched on to support an additional load (due, for example, to a peak in simultaneous connections) without interrupting the operation of the virtual machines.

[0084] To do this, such SON networks advantageously implement edge computing techniques previously mentioned, in particular to locate some of these functions as close as possible to the terminal equipment, for example by using distributed storage techniques.

[0085] In the following, embodiments of the invention are described in more detail in a mobile telecommunications network, the architecture of which complies with the 3GPP standard (in one of its current or future versions) and implements radio access equipment of the base station type. Of course, and as already mentioned, the invention is not limited to the examples described and applies equally well to other types of fixed or mobile telecommunications networks and to other access technologies, wireless or wired.Consequently, the access equipment which is the subject of the resource management proposed by the invention comprises, in a non-restrictive manner, a base station, a femtocell associated with this base station, a digital subscriber line access multiplexer (DSLAM) equipment for an ADSL type access network, a Wi-Fi access terminal (or "hotspot" in English) or a Wi-Fi repeater associated with this terminal, an optical link termination equipment (ONT) for a fiber access network, of the PON type (from the English, "Passive Optical Network"), etc.

[0086] In relation to [Fig.l], a simplified example of the architecture of a mobile telecommunications network RM is presented for illustrative purposes in which the invention. Such a network comprises radio network access equipment such as the base station BS. The base station BS is generally equipped with one or more radio antennas AR1, AR2, AR3. Here, antenna networks or matrices are shown as an example. These radio antennas form a radio cell of the base station, associated with a geographic radio coverage area ZC of the base station.

[0087] In the following, the term "coverage zone" refers to a geographical zone within which user terminals configured to connect to the mobile network RM can attach to the base station, i.e. they are within radio range of this base station and can communicate with it by radio.

[0088] In the example of [Fig.l], this concerns the mobile telephone UE1, the laptop UE2 equipped with a 4G / 5G key K2 and the car UE3 also equipped with a 4G / 5G key K3 or any other means of connection to the 4G / 5G network.

[0089] The mobile network RM also comprises other network equipment, such as for example a control equipment CTR configured to control the base station BS to which it is connected.

[0090] It also comprises a memory M', for example organized in a database DB', configured to store information relating to the network, such as for example information on measuring values ​​of one or more key user data traffic indicators and which can be accessed by other equipment of the network RM or even information relating to technical characteristics of the resources deployed in the base station BS.

[0091] The invention applies to any type of mobile telecommunications network, the architecture of which conforms, for example, to the 3GPP standard in one of its current or future versions.

[0092] Of course, the embodiments, particularly the radio access equipment and terminals implemented in these embodiments, are cited only as examples.

[0093] [Fig.2] schematically illustrates an example of architecture of a system S for managing the resources of a base station in a mobile telecommunications network, according to one embodiment of the invention.

[0094] According to this embodiment of the invention, the system S comprises the radio access equipment or base station BS, configured to allow user terminals to access the mobile network RM by radio. This base station BS comprises at least one radio antenna ANT, a transmission / reception module E / R allowing it to communicate with other equipment of the network RM. Advantageously, it has the hardware structure of a computer and comprises including a CPU processor and a memory M”, in which computer programs are stored.

[0095] The system S also comprises a control device, or CTR controller, configured to control several radio access devices to the mobile network RM and in particular the station BS. For example, this is a network management element ENM (from the English, “Element Network Manager”) forming part of a network management system NMS (from the English, “Network Management System”), associated with the 5G / 4G base stations, also called respectively “gNodeB” or “eNodeB”.

[0096] It is responsible for managing hardware and / or software resources of the base station BS, in particular the feedback of KPI traffic indicator measurements and the (re)configuration of the base station and the dynamic allocation of resources of the BS station. The base station BS and the CTR controller can be instantiated in the form of physical equipment dedicated or not to the respective functions of the base station BS and the CTR controller, or in the form of virtualized entities.

[0097] In this exemplary embodiment of the invention, the CTR control equipment comprises a device 100 for managing resources implemented or to be implemented in a base station of a mobile telecommunications network, configured to obtain a geographical coverage area of ​​a radio cell associated with said base station of said network, obtain demographic and / or topographical information in said coverage area, predict a time sequence of values ​​of at least one (key) indicator of communication data traffic of user terminals attached to said radio cell for a given time period, at least from the information obtained and decide to execute an action for managing the resources of the base station as a function of the predicted time sequence for the given time period.Advantageously, the management device 100 obtains the information relating to the geographical coverage area, and the demographic and topographical information of this area from the database DB, for example stored in a memory M external to the control equipment CTR. Of course, the invention is not limited to this example of implementation, several memories and several databases can be used to store the different types of information relating to the geographical radio coverage area of ​​the base station BS.

[0098] The device 100 thus implements the method for managing resources of a base station according to the invention which will be detailed below in relation to [Fig.3].

[0099] Advantageously, the CTR control equipment has the hardware structure of a computer and comprises a CPU processor, a memory M” in which computer programs are stored, for example, as well as a transmission / reception module E / R which allows it to communicate with other equipment in the RM network.

[0100] Alternatively, the device 100 may be independent of the CTR control equipment, but connected to it by any link, wired or not. For example, it may be integrated into other equipment of the mobile telecommunications network or into radio access equipment, such as the base station BS itself. It is noted that according to the 5G architecture, a gNodeB base station is composed of several elements, including a central unit CU (from the English, "Central Unit"), a distributed unit DU (from the English, "Distributed Unit"), and a remote entity or RU (from the English, "Remote Unit"). These elements are potentially hosted in separate network equipment. The device 100 according to the invention may be embedded in one of these three units, preferably in the unit CU.

[0101] We now present, in relation to [Fig. 3], in the form of a flowchart, an example of implementation of a method for managing resources of an access device, here the base station BS, of a telecommunications network, here the mobile network RM, according to the invention. In what follows, this method is implemented by the aforementioned device 100.

[0102] At 31, location information IL of the geographic radio coverage area ZC of the base station BS is obtained. This is, for example, positioning information for key points in the area ZC, for example located at the border of this area, in a World reference frame, such as that of the geodetic system which models the shape of the Earth. This information includes the latitude which is an angular value, expressing the north or south positioning of a point on Earth. From a mathematical point of view, the latitude of a point is the central angle formed by the normal at this point with the equatorial plane. It also includes the longitude, which is an angular value, expressing the east or west positioning of a point on Earth and corresponds to the central angle formed by the plane passing through this point and through the axis of rotation of the Earth with the plane of the Greenwich meridian.

[0103] For a future base station intended to be equipped with one or more radio antennas, the plans and technical characteristics of which, such as for example the transmission power, the frequency bands of the radio spectrum used for transmitting and receiving, the orientation and the height, are known, such information can in theory be estimated, for example using radio wave propagation calculation software. This estimation is all the more reliable as it is based on a realistic antenna model.

[0104] In practice, it is difficult to obtain this estimate and we know from the document by Soto et al., entitled "Automated land use identification using cell-phone records", published in June 2011 in the proceedings of the 3rd ACM international workshop on MobiArch conference (pp. 17-22), a simplified alternative solution based on a so-called Voronoi decomposition (or "Voronoi Tessellation", in English). This technique consists of tiling the plan of a territory (here for example a city, a district, a municipality) into adjacent regions, or cells, determined from distances to a discrete set of points called seeds, corresponding to the actual or planned positions of the base stations in the territory considered. Each cell, corresponding to the geographical coverage area of ​​a radio cell, encloses a single seed, corresponding to the base station associated with this radio cell, and represents, in a way, the zone of influence of the seed.

[0105] At the end of the calculation of this decomposition and, as illustrated by the table in [Fig.4A], we obtain for each geographic radio coverage zone ZC of the territory considered the following information associated with an identifier of the base station BS_ID considered: - information on the geometry of the ZC zone (here, a PLG polygon); and - the geographic coordinates (latitude, longitude) of each of the vertices of the PLG polygon.

[0106] This location information can be displayed directly on a map like that of [Fig.4B], on which the base stations are represented by a point inside each polygon.

[0107] For example, the device 100 obtains directly from the database DB the geographical coordinates of the vertices of the polygon of the zone ZC associated with the base station BS. According to a variant, it obtains the geographical coordinates of the base stations deployed on a geographical territory T including the actual or planned geographical position of the base station BS and it implements itself the Voronoi decomposition method to obtain the geographical coordinates of the zone ZC associated with the base station BS of interest. In the case where the base station BS is not yet deployed, it adds the geographical position envisaged for this new base station to those of the base stations already in place and calculates the resulting Voronoi decomposition for the entire territory T. At the end of this step 31, the device 100 therefore has the geographical coordinates of the borders of the zone ZC.

[0108] At 32, the device 100 obtains demographic ID and topographic IT information of the ZC zone, for example by querying dedicated public data sources, such as the demographic data source HDE from the company Facebook Research, accessible on the website “humanitarian data exchange” via the link https: / / data.humdata.org / organization / facebook or the cartographic data source OSM (or “Open Street Map”) accessible via the internet link https: / / www.openstreetmap.org / . An example of a MAP map constructed using the cartographic information from the OSM source is available on this last link is presented in [Fig.5]. It represents objects of interest of a given territory T, which are roads, buildings, etc. and specifies their use, their surface area, etc.

[0109] Of course, additional information may optionally be obtained at 33 from other sources, public or private, for example in relation to the socio-economic activity of the geographical area or with the technical characteristics of the radio antenna(s) to be installed on the site of the base station BS or even information relating to the mobility of network users, such as for example public transport timetables, information on the location of buses or even two-wheelers for hire (bicycles or scooters, for example).

[0110] Advantageously, information on measuring values ​​of a communication data traffic indicator is obtained at 34 for the given time period or for a previous or subsequent period. For example, they have been collected by the base station itself, when it is already deployed, or by neighboring access equipment, then sent back to the CTR control equipment or to another equipment of the network.

[0111] The various data obtained at 31, 32, 33 and 34 are stored in memory M', for example organized in the form of a database DB'.

[0112] For example, the demographic and cartographic information thus obtained is manipulated using known database management software, such as PostgreSQL software enhanced with a PostGIS extension for spatial data management.

[0113] For example, the database DB' is of type PostgreSQL and it is populated using the previous information, by constructing, within this database, at least the following three tables: - a first table TB 1 including the location information of the coverage areas resulting from the Voronoi decomposition and obtained in 30; - a second table TB2 including the IT topographic information from the OSM source; and - a third table TB3 including demographic information ID from the HDE source. For example, according to this data source, the grid of a geographical area is done with a granularity of 30 m by 30 m and a row of the table includes the geographical coordinates of a square corner of dimensions 30x30 associated with a quantity of populations in this square; - where applicable, a fourth table TB4 comprising measurement information for a data traffic indicator routed by the BS access equipment.

[0114] Once the DB' database is populated with these four tables, it is sufficient to launch a spatial aggregation query between the Voronoi cells of table 1 and the data OSM from table 2, and an intersection query between the Voronoi cells of table 1 and the population data.

[0115] At 35, the information obtained at 31 and 32 and optionally at 33 and 34, is aggregated, for example by launching a spatial aggregation query between the IL data of the first table TB1 and the IT data of the second table TB2, an intersection query between the IL data of the first table TB1 and the ID data of the third table TB3 and an intersection query between the IL data of the first table and the MKPI data of the fourth table TB4.

[0116] For example, the following query is addressed to the database DB' to count a number of independent buildings in the area ZC of the base station BS:

[0117] sql = SELECT idf_vorocells.site_id, points.buildings, count(points.way) AS feature_count FROM idf_vorocells LEFT JOIN points ON st_intersects(idf_vorocells.geom, points.way) WHERE points .building is not null GROUP BY idf_vorocells.site_id, idf_vorocells.geom, points.building;

[0118] In response, we obtain a list of identifiers of objects of interest (for example, a bridge, a building, a road, a bus stop, etc.) and a number of iterations encountered for each type in the geographical area ZC.

[0119] Once the intersection operations have been carried out, a join operation of the resulting tables is implemented, for the base station BS.

[0120] Using these different queries, a TA table is obtained comprising the demographic, topographical and traffic information relating to the same Voronoi cell and therefore to the same geographic radio coverage area ZC, which have been aggregated into 35. Advantageously, a line of the TA table is associated with the identifier BS_ID of the base station BS and comprises, for each type of object of interest, each associated with a column, a number of occurrences encountered, such as, for example, for the object “bus stop”, the number of bus stops present in the area ZC. It is noted that due to the aggregation, a type of object of interest and therefore a column of the table corresponds to the population present in the area ZC.

[0121] It is also noted that it is possible to select a subset of objects of interest in order to keep only a reasonable number p of columns in the TA table. For example, we assume that p is 271.

[0122] At 36, the device 100 uses the data of the line TA(BS_ID) of this table to predict a time sequence of values ​​of a user data traffic indicator in the geographic radio coverage area ZC.

[0123] In a first embodiment which will now be described in relation to [Fig.6], this data from the line TA(BS_ID) is provided as input to a module of automatic MPA prediction implementing an MP data analysis model obtained by machine learning.

[0124] For example, this MPA module is a neural network or any other artificial intelligence module capable of performing the same functions, based on one of the following techniques - decision tree, for example Random Forests type, - boosted decision tree (or “Gradient boosted decision trees”, in English), for example of the Catboost, XGBoost or LightGBM type; - support vector machine (or “Support Vector Machine” in English); - Multi-layer Perceptron type neural network; - etc.

[0125] Advantageously, this learning or training was implemented during a prior step 30, which made it possible to construct the MP analysis model of the MPA module from a set of training data. The MP analysis model thus constructed is used at 36 by the MPA module to carry out the prediction according to the invention.

[0126] It is noted that the training of the MP model and the analysis of demographic and topographic information using the MP model to predict a user data traffic indicator are presented here in two stages, or in two distinct phases, for simplicity. It is understood, however, that the training can be carried out several times (in particular in parallel or after the analysis) and that the analysis can be continuous.

[0127] Thus, in certain embodiments, the learning may comprise a learning phase 30, “upstream” (initial and prior to the analysis phase) to learn to analyze a set of demographic and topographical information of a coverage area and define parameters to then allow, from any set of demographic and topographical information of another radio coverage area, i.e. presented as input to the MPA module, to predict a time sequence of values ​​of the desired KPI traffic indicator.

[0128] The upstream learning phase is therefore based on a learning set or base comprising, for the different geographical areas of radio coverage of base stations in operation, sets of demographic, topographical information and possibly technical characteristics of the base station, collected for each area, and labeled using time sequences of actual measured values ​​of the KPI traffic indicator.

[0129] The learning phase may also include on-the-fly learning from demographic and topographical information and traffic value measurements collected in the geographic coverage area, in order to refine the configuration. from upstream learning. Both learnings can be carried out on the same equipment (for example on the CTR control equipment or on different equipment (for example the upstream learning can be carried out on another equipment of the mobile telecommunications network, dedicated to this task, the on-the-fly learning being carried out by the CTR control equipment or locally at the base station BS.

[0130] From the point of view of such an artificial intelligence module, the line TA(BS_ID) of the table obtained at the end of the previous step is a vector V1N of real values ​​of size p where p is a non-zero integer corresponding to the number p of columns of the table.

[0131] According to the invention, the prediction module MPA is configured to produce as output a vector VOut comprising a non-zero integer q of real variables, for example equal to 7 days x 24h, or q= 168. More precisely, in this exemplary embodiment of the invention, the output vector VOut provides a time sequence of values ​​of a KPI indicator of user data traffic in the zone ZC, comprising an averaged value of this indicator per hour for a time period PT of one week.

[0132] Such a set of values ​​is a time sequence of average values ​​of the KPI indicator for the ZC zone considered, which can be represented in the form of a CV curve in order to illustrate the variations of this indicator over time.

[0133] At 37, the output vector Vourest is used to decide on a management action to be implemented at the base station BS.

[0134] In the case where this base station is already deployed in the network, the management action decided may be an action of dynamic allocation or deallocation of resources of the base station BS, in order to adapt to the traffic variations predicted over the time period considered, for example the traffic peaks or troughs at certain times of the day. In the case of a software-defined network architecture or SDN (from the English, "Software Defined Networks"), the control equipment CTR may instantiate functions in the base station BS.

[0135] In the case where this base station is not yet deployed in the RM network, the management action may take the form of an order or recommendation or even prioritization of deployment of new equipment in physical or virtualized form. It may also recommend the shutdown of resources that have become superfluous.

[0136] It can finally recommend the addition of new management functions in the BS access equipment, such as for example quality of service management functions, optimization, location of functions at the edge (“edge computing”)•

[0137] For example, this or these recommendations may be transmitted in the form of a signaling message to equipment in the telecommunications network dedicated to planning the deployment of new resources in the network.

[0138] It is now assumed that the management action decided by the method according to the invention has been implemented. For example, an additional radio antenna has been installed on the site of the base station BS.

[0139] At 38, actual measurements of values ​​of the user data traffic indicator predicted at 36 are obtained by the device 100 for the zone ZC, which triggers at 39 an update of the MP prediction model of the MPA module. The updated MP model will be used during future implementations of the resource management method according to the invention for other base stations, which will make it possible to produce more precise and accurate predictions of user data traffic indicators.

[0140] We will now detail in relation to [Fig.7], a second embodiment of the invention, according to which the neural network implemented in the MPA module is a “transformer” type neural network. This is a convolutional neural network of a particular type, implementing an attention mechanism inspired by the functioning of the human brain. An example of the structure of such a network is for example described in the document by Vaswani et al., entitled “Attention is ail you need”, published in 2017, in the proceedings of the “Neural Information Processing Systems” conference (NIPS 2017), which was held in Long Beach, CA, USA.

[0141] Initially, transformer neural network architectures were developed specifically for natural language processing and in particular the translation of a text from one language to another language. Indeed, the attention mechanism learns to sequentially translate a sequence of words and, simultaneously, to align target words with the most relevant source words.

[0142] Subsequently, neural networks were adapted to perform other transformation tasks, such as text-to-image conversion, for example according to the technique described in the paper by Ramesh et al., entitled "Zero-shot text-to-image generation" and published in 2021, on the website / https / 'arXiv.org, preprint arXiv:2102.12092.

[0143] According to this second proposed embodiment, the MPA prediction module implements a transformer neural network specifically adapted to convert an input vector comprising demographic and topographical information into a temporal sequence of values ​​of a KPI indicator of user data traffic over a given temporal period.

[0144] For this type of prediction module, the input vector V1N is seen as a sentence of p words, for example a sentence of 271 words. The meaning of a word is given by the numerical value associated with this word, that is to say with a particular type of object of interest or demographic or topographic variable for example.

[0145] [Fig.7] details the prediction step 36 according to this second embodiment.

[0146] In 61, and to avoid having to manage an infinity of values ​​and therefore of meanings possible of each word of the sentence, these are discretized, for example by quantizing them according to a quantization technique known per se. An intermediate vector VD of p discretized components is obtained.

[0147] At 62, the vector VD is transformed or converted into a compressed output vector VC comprising n components, with n being an integer less than q. For example, n is 4.

[0148] This VC vector is a compressed representation of the final output. An advantage of such compression, described in particular in the document by Ramesh et al. already cited, for an application to the conversion of a text into an image, makes it possible to limit the complexity of the model and to reduce the calculation time.

[0149] At 63, the vector VC is decompressed using the decoder part of a discrete variational autoencoder, for example, such as that described in the document by Ramesh et al. However, it should be noted that the architecture of this autoencoder is adapted so as to produce as output the output vector VOut having the desired dimensions, namely q components, with q = 168. This autoencoder (encoder and decoder parts) has been previously trained so that the reconstructed output is as faithful as possible to the data presented as input. An advantage of using such a neural network to predict the output vector VOut is that it is non-deterministic, in the sense that it is capable of producing several predictions for the same base station, which makes it possible to associate confidence intervals with these predictions.

[0150] It is noted that as a variant, the decoder part of this auto-encoder could be used in combination with another machine learning algorithm than a transformer-type neural network, for example boosted decision trees.

[0151] Finally, in relation to [Fig.8], an example of a hardware structure of a device 100 for managing the resources of an access device such as a base station in a mobile telecommunications network is presented, said device comprising a module for obtaining a geographical coverage area of ​​a radio cell associated with said base station of said network, a module for obtaining demographic and topographical information relating to said coverage area, a module for predicting a time sequence of values ​​of at least one communication data traffic indicator of user terminals attached to said radio cell for a given time period, at least from the information obtained and a module for deciding on an action for managing the resources of the base station as a function of the predicted time sequence for the given time period.

[0152] Advantageously, the device 100 comprises a module for prior learning of a prediction model of the information obtained in said time sequence of values ​​of at least one communication data traffic indicator, said prediction model being used by said prediction module.

[0153] Advantageously, the device 100 comprises, following the execution of the management action, a module for obtaining measurements of the traffic indicator in the radio coverage area of ​​the base station and a module for updating the prediction model using the measurements obtained.

[0154] The term "module" can correspond to a software component as well as to a hardware component or a set of hardware and software components, a software component itself corresponding to one or more computer programs or sub-programs or more generally to any element of a program capable of implementing a function or a set of functions.

[0155] More generally, such a device 100 comprises a random access memory 103 (for example a RAM memory), a processing unit 102 equipped for example with a processor, and driven by a computer program Pgl, representative of the obtaining, prediction and decision modules, stored in a read-only memory 101 (for example a ROM memory or a hard disk). At initialization, the code instructions of the computer program are for example loaded into the random access memory 103 before being executed by the processor of the processing unit 102. The random access memory 103 can also contain for example the demographic and topographical information obtained, the learned prediction model, the traffic indicator measurements collected for the radio coverage area.

[0156] [Fig.8] illustrates only one particular way, among several possible ways, of producing the device 100 so that it performs the steps of the method for managing resources of an access device as detailed above, in relation to [Fig.3], in its different embodiments. Indeed, these steps can be carried out indifferently on a reprogrammable computing machine (a PC computer, 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).

[0157] In the case where the device 100 is produced 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 partially or totally readable by a computer or a processor.

[0158] The different embodiments have been described above in relation to a device 100 integrated into control equipment of a mobile telecommunications network, but it can also be integrated into equipment of this network such as access equipment such as for example a base station.

Claims

Claims

1. Method for managing resources of at least one access equipment (BS) of a telecommunications network (RM), comprising: - obtaining (31) location information (IL) of a geographical area (ZC) served by said access equipment; - obtaining (32) demographic (ID) and topographical (IT) information relating to said geographical coverage area; - predicting (36) a set of values ​​of at least one communication data traffic indicator of terminal equipment connected to said access equipment for a given time period, at least from the demographic and topographical information obtained, said set of values ​​forming a time sequence representative of a variation of said at least one indicator over said time period;and - the decision (37) of at least one action for managing the resources of the access equipment at least as a function of the set of values ​​predicted for the given time period and of technical characteristics relating to the resources of said at least one access equipment, said action for managing the resources of the access equipment comprising a dynamic allocation or deallocation of said resources as a function of said time sequence representative of a variation of said at least one indicator over said time period.;

2. Management method according to claim 1, characterized in that the management action comprises a command to deploy at least one additional resource in the access equipment and belongs to a group comprising at least: - a command to deploy an additional radio antenna in the access equipment, when the access equipment is a base station of a mobile telecommunications network; - a command to replace an existing antenna of the access equipment with a later generation antenna, when the access equipment is a base station of a mobile telecommunications network; - an order to remove an existing antenna from the access equipment, when the access equipment is a base station of a mobile telecommunications network; - an order to deploy the access equipment in the telecommunications network; - an order to move the access equipment in the telecommunications network.

3. Management method according to any one of the preceding claims, characterized in that the management action comprises an adjustment of the resources allocated to the access equipment and belongs to a group comprising at least: - an adjustment of a transmission power level and / or an orientation of at least one antenna of the access equipment, when the access equipment is a base station of a mobile telecommunications network; - an adjustment of a quantity of hardware and / or software resources allocated to the operation of the access equipment; - an adjustment of management functions of the traffic carried by the access equipment.

4. Management method according to any one of the preceding claims, characterized in that the communication data traffic indicator belongs to a group comprising at least: - a downlink and / or uplink communication data rate; - a number of terminal devices connected to the access equipment; - an overall volume of data traffic carried by the access equipment; - a typology of communication data carried by the access equipment.

5. Management method according to any one of the preceding claims, characterized in that the prediction (36) also takes into account said technical characteristics of the resources of said at least one access device.

6. Management method according to any one of the preceding claims, characterized in that it comprises obtaining (34) information (MKPI) of measurements of at least said communication data traffic indicator over the time period given and in that the prediction (36) also takes into account said measurement information.

7. Management method according to any one of the preceding claims, characterized in that it comprises the prior learning (30) of a prediction model (MP) of the information obtained in said set of values ​​of at least one communication data traffic indicator and in that said prediction (36) implements an artificial intelligence module (MPA) configured to use said prediction model (MP).

8. Management method according to any one of the preceding claims, characterized in that it comprises the aggregation (35) of the information obtained for said geographic radio coverage area (ZC) into an input vector (V^) comprising a number (p), called the input number, of first components equal to a number of information obtained, and in that said prediction (36) comprises the transformation (60), from said prediction model (MP), of the input vector (Vin) into an output vector (vout), representative of said set of values ​​of said communication data traffic indicator, comprising a number of components, called the output number (q), equal to a number of time intervals in the given time period.

9. Management method according to claim 8, characterized in that said transformation (60) comprises the discretization (61) of the first components of the input vector and the formation of a vector, called discretized vector (VD), the components of which include the first discretized components, the translation (62) of the discretized vector into a vector (VC) called compressed, representative of said set of values ​​of said communication data traffic indicator and comprising a number of components (n), less than the output number (q), and the decoding (63) of the compressed vector (VD) into the output vector (VOut)-

10. Management method according to any one of claims 7 to 9, characterized in that it comprises, following the execution of the management action, obtaining (38) a new set of measurements of the communication data traffic indicator over a new time period and updating (39) the learning of the data transformation model from said new set.

11. Device (100) for managing resources of at least one access equipment (BS) in a telecommunications network (RM), said device being configured to implement: - obtaining location information (IL) of a geographical area served by said access equipment; - obtaining demographic (ID) and topographical (IT) information relating to said geographical area; - predicting a set of values ​​of at least one communication data traffic indicator of terminal equipment connected to said access equipment for a given time period, at least from the demographic and topographical information obtained, said set of values ​​forming a time sequence representative of a variation of said at least one indicator over said time period;and - the decision of at least one action for managing the resources of the access equipment at least as a function of the set of values ​​predicted for the given time period and of technical characteristics relating to the resources of said at least one access equipment, said action for managing the resources of the access equipment comprising a dynamic allocation or deallocation of said resources as a function of said time sequence representative of a variation of said at least one indicator over said time period.;

12. Control equipment (CTR) of a telecommunications network (RM), configured to control at least one access equipment of said network, said access equipment being configured to connect terminal equipment to said network, characterized in that it comprises a device (100) for managing resources of said access equipment (BS) according to claim 11.

13. Access equipment (BS) of a telecommunications network (RM), configured to connect terminal equipment to said network, characterized in that it comprises a resource management device (100) according to claim 11.

14. System (S) for managing resources in a telecommunications network (RM), characterized in that it comprises at least one access device (BS) of said network, configured to connect terminal equipment for said network and a resource management device for said access equipment according to claim 11.

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

Images