FR3139688A1 - Method and device for adaptive configuration of wireless telecommunications - Google Patents

Abstract

The invention relates to a method for adaptive configuration of telecommunications in a mission, by an aircraft (30), on wireless links (52), with a platform (80), said method comprising these steps: - prediction of the state future of the links by a connectivity predictor (22) depending on the current state of the links and mission conditions; - depending on at least said prediction of the future state of the wireless links, prediction, by a predictor of the exchange templates (21), of an operating mode of at least one function of the set of processing function(s) among several operating modes of said function associated with distinct exchange templates; - following said prediction, when said predicted operating mode of the function is different from an operating mode of the function currently implemented, triggering a switch of at least said function to said predicted operating mode. Figure for abstract: Fig. 1

Description

Method and device for adaptive configuration of wireless telecommunications

The invention is located in the field of wireless telecommunications.

The conditions in which tactical missions, for example aerial, take place are varied, and can change quickly (for example the jamming of a radio link between the aircraft and another platform, voluntary or due to a mask for example by the aileron or drift of the plane, to a cloud for certain frequency bands etc.), leading to a loss of connectivity for the planes, on parts of their mission. Indeed, their wireless telecommunications systems, particularly radio, cannot adapt, once the change is detected, their parameters as quickly as the speed of change in conditions. This delay in switching to another transmission mode (for example a satellite link) leads to a loss of communication in the adaptation interval, which can nevertheless be crucial.

There is therefore a need to ensure that wireless communications are maintained from aircraft, and more generally flying platforms, during their aerial missions.

To this end, according to a first aspect, the present invention describes a method for adaptive configuration of wireless telecommunications implemented, in a mission, by a mobile machine, on one or more wireless links, with at least one platform. a set of platforms, said method comprising the set of following steps, iterated during successive moments of the mission, an electronic predictor of the exchange templates and an electronic connectivity predictor being on-board, in the mobile machine, in an electronic processing block comprising a set of processing function(s):
- prediction of the future state of the wireless links by the connectivity predictor over at least one predefined time horizon, depending on at least the current state of the wireless links and one or more mission condition parameters among weather data, the current position of the mobile device, the current attitude of the mobile device, the trajectory followed, the speed of the mobile device, the relief, the heading, friendly positions, enemy threats, electromagnetic environment;
- depending on at least the profile of said mission and/or depending on an estimate of future exchange needs over said time horizon indicating at least one volume of data to be exchanged by all the processing functions via wireless links and depending on said prediction of the future state of the wireless links, prediction, by the exchange template predictor, of an operating mode of at least one function of the set of function(s) ) processing among several operating modes of said function associated with distinct exchange templates;
- following said prediction, when said predicted mode of operation of the function is different from an operating mode of the function currently implemented, triggering a switch of at least said function to said predicted mode of operation.

• une sélection a préalablement eu lieu, en fonction du profil de ladite mission, d’un prédicteur électronique de connectivité parmi plusieurs prédicteurs électroniques de connectivité et d’un prédicteur électronique des gabarits d’échanges parmi plusieurs prédicteurs électroniques des gabarits d’échanges, lesdits prédicteurs embarqués étant ceux sélectionnés ;
• le bloc de traitement de l’engin mobile , en fonction d’au moins la prédiction de l’état futur des liaisons sans fil, effectue au moins une action parmi :

l'adaptation de la trajectoire de l’engin mobile ou d’une plateforme pour maintenir une liaison sans fil existante ou créer une nouvelle liaison sans fil ;
la re-direction de flux au sein desdites liaisons sans fil ;
l’ adaptation des caractéristiques desdites liaisons ;

• l’engin mobile est un aéronef et la mission est une mission aérienne qui comporte plusieurs phases parmi au moins une phase de décollage, une phase de transit, une phase de théâtre d’opération et une phase d’atterrissage et la prédiction de l’état futur des liaisons sans fil par le prédicteur de connectivité et la prédiction des gabarits d’échanges futurs par le prédicteur des gabarits d’échanges sont chacune fonction de la phase courante de la mission ;
• au moins l’un des prédicteurs parmi le prédicteur de connectivité et le prédicteur des gabarits d’échanges comporte un réseau de neurones ; et ladite sélection de prédicteur comprend la sélection d’un jeu de poids et biais du réseau de neurones parmi un ensemble de jeux de poids et biais disponibles, en fonction du profil de mission ;
• au moins l’un des prédicteurs parmi le prédicteur de connectivité et le prédicteur des gabarits d’échanges comporte un réseau de neurones ; et

lors de la mission, des données indiquant l’état courant de connectivité des liaisons, des données indiquant les échanges réalisés par les fonctions sur les liaisons sans fil et des données indiquant les conditions courantes de mission sont collectées et mémorisées ;
et lesdites collectées et le profil de mission sont utilisées pour poursuivre l’apprentissage dudit réseau de neurones.In embodiments, such a method will also include at least one of the following characteristics:

• a selection has previously taken place, depending on the profile of said mission, of an electronic connectivity predictor among several electronic connectivity predictors and of an electronic predictor of exchange templates among several electronic predictors of exchange templates, said embedded predictors being those selected;
• the processing block of the mobile machine, based on at least the prediction of the future state of the wireless links, performs at least one action among:

adapting the trajectory of the mobile machine or a platform to maintain an existing wireless link or create a new wireless link;
re-direction of flows within said wireless links;
adapting the characteristics of said connections;

• the mobile machine is an aircraft and the mission is an aerial mission which comprises several phases among at least one take-off phase, a transit phase, a theater of operation phase and a landing phase and the prediction of the future state of wireless links by the connectivity predictor and the prediction of future exchange templates by the exchange template predictor are each a function of the current phase of the mission;
• at least one of the predictors among the connectivity predictor and the exchange template predictor comprises a neural network; and said predictor selection comprises selecting a set of weights and biases of the neural network from a set of available sets of weights and biases, depending on the mission profile;
• at least one of the predictors among the connectivity predictor and the exchange template predictor comprises a neural network; And

during the mission, data indicating the current state of connectivity of the links, data indicating the exchanges carried out by the functions on the wireless links and data indicating the current mission conditions are collected and stored;
and said collected data and the mission profile are used to continue learning said neural network.

According to another aspect, the present invention proposes a computer program intended to be stored in the memory of an electronic processing unit on board a mobile vehicle to adaptively configure telecommunications implemented during a mission , by the mobile machine, on one or more wireless links; the processing block further comprising a microcomputer, said computer program comprising instructions which, when executed on the microcomputer, implement the steps of a method according to the first aspect of the invention.

According to another aspect, the invention describes an electronic processing block adapted to be embarked in a mission, in a mobile machine adapted to implement, on one or more wireless links, telecommunications with at least one platform of a set of platforms, said electronic processing block comprising a set of processing function(s), an electronic predictor of exchange templates and an electronic connectivity predictor; in which :
the connectivity predictor is adapted to, during successive moments of the mission, predict the future state of the wireless links at least a predefined time horizon, as a function of at least the current state of the wireless links and one or more mission conditions parameters from weather data, the current position of the mobile device, the current attitude of the mobile device, the trajectory followed, the speed of the mobile device, the relief, the heading, friendly positions, enemy threats, the electromagnetic environment;
- the predictor of the exchange templates is adapted for, during successive moments of the mission, according to at least the profile of said mission and/or according to an estimate of future exchange needs over said time horizon indicating at least one volume of data to be exchanged by all the processing functions via wireless links and depending on said prediction of the future state of the wireless links, predicting an operating mode of at least one processing function the set of processing function(s) among several operating modes of said function associated with distinct exchange templates;
- the processing block being adapted to, following said prediction, when said predicted operating mode of the function is different from an operating mode of the function currently implemented, trigger a switch of at least said function to said mode predicted operation.

• Il est adapté pour, en fonction d’au moins la prédiction de l’état futur des liaisons sans fil, effectuer au moins une action parmi :

l'adaptation de la trajectoire de l’engin mobile (30) ou d’une plateforme (70, 80) pour maintenir une liaison sans fil existante ou créer une nouvelle liaison sans fil ;
la re-direction de flux au sein desdites liaisons sans fil ;
l’ adaptation des caractéristiques desdites liaisons ;

• l’engin mobile (30) est un aéronef et la mission est une mission aérienne qui comporte plusieurs phases parmi au moins une phase de décollage, une phase de transit, une phase de théâtre d’opération et une phase d’atterrissage et la prédiction de l’état futur des liaisons sans fil par le prédicteur de connectivité (22) et la prédiction des gabarits d’échanges futurs par le prédicteur des gabarits d’échanges (21) sont chacune fonction de la phase courante de la mission.

In embodiments, such a processing block will also include at least one of the following characteristics:

• It is adapted to, depending on at least the prediction of the future state of the wireless links, perform at least one action among:

adapting the trajectory of the mobile machine (30) or a platform (70, 80) to maintain an existing wireless link or create a new wireless link;
re-direction of flows within said wireless links;
adapting the characteristics of said connections;

• the mobile machine (30) is an aircraft and the mission is an aerial mission which comprises several phases among at least a take-off phase, a transit phase, a theater phase of operation and a landing phase and the prediction of the future state of the wireless links by the connectivity predictor (22) and the prediction of future exchange templates by the exchange template predictor (21) are each a function of the current phase of the mission.

The invention will be better understood and other characteristics, details and advantages will appear better on reading the description which follows, given on a non-limiting basis, and thanks to the appended figures, given by way of example.

There is an illustration of an air mission processing system 1 implementing an embodiment of the invention;

There represents the steps of an aerial mission processing method 1 implementing an embodiment of the invention.

Identical references may be used in different figures when they designate identical or comparable elements.

An air mission processing system 1 is represented in incorporating an embodiment of the invention.

The air mission processing system 1 comprises, in the case shown, a configuration platform 10, a fleet of mobile machines equipped with wireless transmission means (for example fighter planes, reconnaissance planes, refueling planes), in particular in the example considered, aircraft including an airplane 30, an airplane 80 and a rolling machine 70.

The air missions are implemented by the air mission processing system 1, in an embodiment subsequently described with reference to the with reference to an aerial mission involving several aircraft, in particular the aircraft 30 alone represented in .

The configuration platform 10 includes a server 11 and a database 12.

The server 11 includes an electronic configuration module 13.

The database 12 includes a set of sets of exchange template predictor configuration parameters and a set of sets of connectivity state predictor configuration parameters.

The plane 30 includes a processing block 20.

The processing block 20 of the aircraft 30 comprises an electronic block predicting exchange templates 21, an electronic block predicting connectivity state 22, a control block 23, a block of sensors 24, a block for calculating trajectory 25 and a wireless telecommunications block 26.

• un ensemble de sous-bloc(s) de télécommunication sans fil adapté chacun pour, en phase de réalisation de mission, établir une liaison de télécommunication sans fil (par exemple de type VHF, UHF, satellite) avec une plateforme distante, par exemple comme détaillé plus tard, une liaison 51 avec l’engin roulant 70, une liaison 52 avec l’avion 80.

The wireless telecommunications block 26 comprises a plurality of wireless telecommunications sub-blocks, for example:

• a set of wireless telecommunications sub-block(s) each adapted to, in the mission execution phase, establish a wireless telecommunications link (for example of the VHF, UHF, satellite type) with a remote platform, for example as detailed later, a connection 51 with the rolling machine 70, a connection 52 with the aircraft 80.

The wireless telecommunications links implemented by the wireless telecommunications block 26 are constrained: the telecommunications capacity is limited (compared to gigabit optical fiber networks) and not very stable: the telecommunications links in the context of aerial missions are frequently deployed locally and occasionally in a health or military intervention area (a geological or other disaster zone, a conflict zone, etc.); these transmission links are frequently broken without notice (for example by intentional or unintentional jamming), some are very limited in terms of bandwidth or experience occurrences, much more frequent than in traditional consumer networks, of bandwidth variations or latency.

Each or some at least of one or more blocks among the control blocks 23, sensors 24, trajectory calculation 25 implement for example one or more functions, in particular software applications, which execute during the realization of an air mission, and which, during this execution, transmit and/or receive data (measurement, analysis, command, alert, observations, etc.) via one or more established wireless telecommunications links by the wireless telecommunications block 26.

For example, a software application of a video sensor in block 24 captures images and transmits them remotely, etc.

Some of these blocks, and/or some of the functions of the blocks are further adapted to operate selectively according to an operating mode selected from several of their operating modes, for example nominal mode, degraded mode 1, etc., mode degraded n (n≥1) associated respectively with distinct characteristics, volumes of data transmitted and/or lifespan of the information (which will condition the maximum latency that the network must respect between the source and the recipient) and/ or the tolerated jitter, the error rate supported, the sovereignty of this data, the level of protection associated with the transmission of the data, etc. (in fact, anything that will affect the conditions of transport can be taken into account). Each operating mode is thus associated with a corresponding transport template, which translates the characteristics of the operating mode into transport characteristics by the network.

Distinct modes of operation of a function of a block are distinguished for example by the format of the transmission data considered (in particular compression mode, coding) delivered by the function, by the transmission protocol used by the function, by the precision of the processing carried out by the function (example: resolution of the transmitted images), by a level of filtering affecting the completeness of the transmitted data etc., which result in distinct characteristics regarding the transmission: certain less relevant data may therefore example not be transmitted if resources were to be lacking, corresponding for example to a narrowing of the field of observation), or an adaptable selectivity of the information transmitted (certain information could be deleted for reasons of priority or level of security For example).

The configuration platform 10 is adapted to determine the configuration parameters of the connectivity state predictor 22 and exchange predictor 21 blocks, during the preparation of the mission, in the manner described below with reference to the .

The connectivity state predictor block 22, hereinafter called connectivity predictor 22, is adapted, once configured as described below, to predict, when carrying out the mission, the changes in connectivity states of the connections established by the wireless telecommunications block 26, such as breakdown of a wireless telecommunications link implemented in the mission being carried out, reduction of the bandwidth below a predefined threshold, or increase in the error rate above a predefined threshold, i.e. before these changes of state occur.

In the considered embodiment, the connectivity state predictor 22 comprises a neural network which performs the prediction based on current input data provided to the connectivity state predictor 22.

The exchange template predictor block 21, once configured as described below, is adapted, when carrying out the mission, to depending on input data including in particular a connectivity state predicted by the state predictor of connectivity 22, deduce, from this predicted connectivity state related to the exchange needs of blocks 24 and 25 of the processing block 20 (indicated for example by the mission profile and/or the mission phase), modes of operation of the (functions of) sensor blocks 24 and/or trajectory calculation 25 (associated with respective exchange templates), to be applied in anticipation of a future change, then deduce one or more corresponding commands intended to this or these blocks of sensors 24, for calculating the trajectory 25 to adapt, where appropriate, their operating modes so as to adapt their emitted and/or received flows, for example for at least a predetermined time: the exchange template predictor block 21 thus predicts the exchange needs of block 24 and block 25 to anticipate the operating mode to be applied and thus adapt the flows emitted to the transport conditions observed or predicted.

For each function one or more operating modes have been defined, as indicated above. An operating mode brings together a set of characteristics making it possible to quantify and qualify the data flows that the application generates. An operating mode is therefore specific to an application.

At a given time, for each application, one and only one operating mode is active.

An exchange template (also called a transport template) is the counterpart of a mode of operation, from the point of view of the telecommunications network and reflects the way for the network to take this mode of operation into account. It expresses the point of view of the container, when the mode of operation expresses the point of view of the content. For example, generated data will have a lifetime (content point of view) and for the data to still be relevant upon receipt, its transport through the network must be carried out with maximum latency (container point of view). Each operating mode of the application is associated with one and only one network exchange template.

The characteristics of an operating mode are used to define the characteristics of the corresponding template, via a translation. Then, based on the observed transport conditions, the network determines, for each data flow, the template to apply among those that have been predefined. Knowledge of the template then makes it possible to identify the operating mode with which it is associated. The network thus knows how to restore to the function the mode of operation that it must implement for optimal operation given the transport conditions of the moment.

The exchange template predictor makes it possible to determine the templates that it would be relevant to respect given the network events that are anticipated, and by transitivity the operating modes to be implemented by the functions.

In the embodiment considered, the exchange template predictor 21 comprises a neural network which determines the prediction based on the current input data supplied to the exchange template predictor 21.

There represents a method 100 of mission planning and adaptive configuration of telecommunications implemented in the air mission in one embodiment of the invention.

The method 100 comprises 3 phases: a mission communications preparation phase 200 which takes place before the mission, a mission execution phase 300 which takes place during the mission and a mission restitution phase 400 which takes place after the mission.

Phase of mission preparation

During the mission communications preparation phase 200, which therefore takes place before the mission, the configuration module 13 in the server 10 receives the mission profile as input.

• le plan de vol/la trajectoire planifiée de chacun des avions/véhicules impliqués dans la mission, dont l’avion 30 ;
• le type de mission : observation, interception, surveillance, combat, besoin de discrétion, … ;
• les différentes phases de mission, leur durée, leurs localisations … ;
• les besoins d’échanges entre les différents avions/véhicules, et ce dans le cas considéré, en fonction des phases respectives de la mission ; ces échanges (télécommunications sans fil) seront mis en œuvre via des fonctions, y compris des applications logicielles, s’exécutant dans le module de traitement 20 de l’avion 30 lors de phase(s) respective(s) de la mission ;
• identification des interlocuteurs (avions/véhicules) impliqués dans lesdits échanges ;
• la date de début et la date de fin de la mission ;
• le mode d’activation des communications, i.e. à la demande d’un opérateur ou par défaut au démarrage de la mission ;
• la priorité de chacun des besoins d’échanges ;
• ….

The mission profile indicates in particular the following elements, or at least some of them:

• the flight plan/planned trajectory of each of the aircraft/vehicles involved in the mission, including aircraft 30;
• the type of mission: observation, interception, surveillance, combat, need for discretion, etc.;
• the different mission phases, their duration, their locations, etc.;
• the needs for exchanges between the different planes/vehicles, in the case considered, depending on the respective phases of the mission; these exchanges (wireless telecommunications) will be implemented via functions, including software applications, executing in the processing module 20 of the aircraft 30 during respective phase(s) of the mission;
• identification of the interlocutors (planes/vehicles) involved in said exchanges;
• the start date and end date of the mission;
• the mode of activation of communications, ie at the request of an operator or by default at the start of the mission;
• the priority of each exchange need;
• ….

Typically, the mission phases successively include a take-off phase, a transit phase (to reach a target zone, the so-called theater zone), a theater phase (at the level of the theater zone, which can itself include one or more phases each associated with an action such as observation of the theater zone and/or interception and/or combat etc.), a transit phase and a landing phase.

With each phase of this mission, depending on the mission profile, specific functions are thus associated, then executing in the blocks of the processing block 20 of the aircraft 30 (trajectory calculation, sensors) and then generating data exchanges. (in reception and/or transmission) with interlocutors, including for example the plane 80 and/or the vehicle 70.

Exchange needs and connectivity are of course dependent on the mission profile.

The configuration module 13 of the server 10, running in the server 11 selects, in step 200, based on the mission profile received, a set of configuration parameters of the exchange template predictor from the set of sets of configuration parameters of template predictors stored in the database 12 and selects a set of configuration parameters of the connectivity predictor from the set of sets of configuration parameters of connectivity predictors stored in the database 12 ( for example, the sets of parameters are each associated with a typical mission profile and the selection rules include the identification (possibly by mission phase) of the typical profile that is “closest” to the mission profile received as input, according to a proximity criterion for example resulting from the comparison between each (or at least some) of the elements of the mission profile received as input and the corresponding element of the standard profiles.

In the case considered, the configuration module 13 delivers the sets of configuration parameters of the selected predictors. The configuration parameters of each predictor include, in the embodiment considered, respective values of the weights and biases for the neural network of the predictor.

In the embodiment considered, the set of configuration parameters includes a specific subset of configuration parameters per mission phase included in the mission for the exchange template predictor and/or the connectivity predictor. The use of the network is thus segmented according to the mission phase, with for each phase taking into account measurable conditions likely to influence the upcoming transmissions of each platform: terrain models, weather, enemy threats, environment electromagnetic with disturbances, etc.

The configuration module 13 of the server 10 further determines the configuration parameters of the wireless transmission links used during the mission: allocation of frequencies for the radio links involved. It also determines the configuration parameters of the network supported by the transmission links (routing rules, QoS policies, security associations, filtering rules, etc.) as well as the configuration parameters of the communication services to be implemented. works to ensure exchange needs (names of services, operating modes, rules for identifying exchanges, etc.). For example, it also determines one or more elements among: the TRANSEC and COMSEC secret keys, the addresses to use (MAC and network), the parameters for implementing the waveforms, the frequency jump laws if applicable, …

In one embodiment, these wireless link configuration parameters are themselves delivered by a link configuration predictor comprising a neural network receiving the mission profile as input. This predictor in a learning phase exploiting the learning base comprising the connectivity data, the mission conditions and the exchanges carried out will have learned to identify the configuration parameters of the transmission modes relevant to the mission.

At the end of this preparation phase 200 for the upcoming mission, each of the predictors of the exchange templates 21 and connectivity 22 is configured with the respective set of parameters selected.

It is therefore a question of anticipating the type of exchanges necessary (priority, volume, typology of exchanges, etc.) by mission phase as well as the state of connectivity, in the sense of end-to-end transport conditions). .

Initial learning

An initial learning phase, prior to step 200, made it possible to obtain the sets of sets of exchange template predictor configuration parameters and the sets of sets of connectivity predictor configuration parameters.

It is described below.

Typical air mission profiles have been defined. For each typical air mission profile, multiple missions of this type have been carried out, in real deployment or by simulations; and connectivity data, exchange data and mission conditions were collected forming a learning database, referred to below as the learning database.

Initial training of the neural network to obtain the connectivity predictor 22.

Training of the neural network intended for the connectivity predictor 22 is carried out for each type of mission profile.

• des valeurs relatives à la connectivité constatée ou calculée considérée à chaque instant θi (d’une pluralité d’instants successivement considérés) et issues de mesures (et éventuellement antérieures à θi, à l’intérieur d’une fenêtre temporelle Fθi de durée Δθ se terminant à θi ; par exemple Δθ compris entre quelques dizaines de ms et quelques centaines de ms ou quelques secondes) ; ces mesures relatives à la connectivité indiquent typiquement pour chaque liaison de télécommunication sans fil impliquée dans la mission la valeur d’une ou de plusieurs de ses caractéristiques de transmission : bande passante, taux d’erreur, latence, gigue ; et au-delà des mesures effectuées pour chaque liaison de transmission, les conditions de transport bout-en-bout sont également fournies en entrée du réseau de neurones, i.e. le nœud local vers tous les nœuds du réseau, incluant l’identification des nœuds non atteignables ; il s’agit également de fournir les niveaux d’engorgement des files d’attente du réseau et le nombre de paquets perdus par débordement de files d’attente ;
• la phase de mission courante à l’instant θi ; et
• des indications de conditions de mission à l’instant θi (et éventuellement antérieures à θi, à l’intérieur de la fenêtre temporelle Fθi) ; par exemple ces indications de mission indiquent au moins certaines des données parmi des données météo, position et attitude courantes de l’avion, trajectoire suivie (virage, montée, descente), vitesse de l’avion, accélération, relief (modèle de terrain), cap, positions amies et menaces ennemies, environnement électromagnétique dont brouillage …. On notera que ces indications peuvent comprendre des indications relatives à l’instant courant ou échantillonnées pour la fenêtre Fθi.

The input data of the neural network intended for the connectivity predictor 22, extracted from the learning base, for a given mission, include:

• values relating to the connectivity observed or calculated considered at each instant θi (of a plurality of instants successively considered) and resulting from measurements (and possibly prior to θi, within a time window Fθi of duration Δθ se ending at θi; for example Δθ between a few tens of ms and a few hundred ms or a few seconds); these measurements relating to connectivity typically indicate for each wireless telecommunications link involved in the mission the value of one or more of its transmission characteristics: bandwidth, error rate, latency, jitter; and beyond the measurements carried out for each transmission link, the end-to-end transport conditions are also provided as input to the neural network, ie the local node to all nodes in the network, including the identification of non-local nodes. attainable; it also involves providing network queue congestion levels and the number of packets lost by queue overflow;
• the current mission phase at time θi; And
• indications of mission conditions at time θi (and possibly before θi, within the time window Fθi); for example these mission indications indicate at least some of the data among weather data, current position and attitude of the aircraft, trajectory followed (turn, climb, descent), speed of the aircraft, acceleration, relief (terrain model) , heading, friendly positions and enemy threats, electromagnetic environment including jamming…. Note that these indications may include indications relating to the current instant or sampled for the window Fθi.

Learning makes it possible to successively adjust the value of the weights and biases in the neural network intended for the predictor 22, for a given number of layers of neurons and a given number of neurons in each layer, until the error between the result output by the neural network and the state of connectivity, as indicated in the learning base, which actually followed any instant ti in a time window of non-zero duration ΔT starting at ti or less at a predefined threshold. For example, the prediction concerns several tens of seconds (for example over k seconds, with k between 10 and 90), or even over a few minutes (for example k’ minutes, with k’ between 1 and 10).

Typically, the predicted connectivity states indicate for example for each wireless telecommunications link involved in the mission the value (or selectively indicate a subrange of values in which this value is located, within a range of values comprising several sub-ranges) of one or more of its transmission characteristics: bandwidth, error rate, latency, jitter.

This results in a set of parameters (weight and learned bias), associated with each mission type profile, to configure the neural network of the connectivity state predictor 22.

Initial training of the neural network to obtain the predictor of exchange templates 21

Similarly in one embodiment, training of the neural network intended for the exchange template predictor 21 was carried out for each type of mission profile.

• l’état de connectivité prédit par le prédicteur d’état de connectivité 22 à partir des données d’entrée relatives à l’instant initial ti ; et
• la phase de mission courante à l’instant ti (qui définit la nature et la volumétrie des échanges, i.e. le besoin d’échanges pour le type de mission considérée, pendant la phase considérée) ;
• optionnellement : types et caractéristiques des échanges constatés à l’instant ti (ces éléments pourraient en effet conditionner les futurs besoins d’échanges en cas de variation de l’état de connectivité).

The input data of the neural network intended for the exchange template predictor 21, extracted from the learning base, for a given mission, comprise, relative to each instant ti of a plurality of instants ti:

• the connectivity state predicted by the connectivity state predictor 22 from the input data relating to the initial time ti; And
• the current mission phase at time ti (which defines the nature and volume of exchanges, ie the need for exchanges for the type of mission considered, during the phase considered);
• optionally: types and characteristics of exchanges observed at time ti (these elements could in fact condition future exchange needs in the event of a variation in the state of connectivity).

Learning makes it possible to successively adjust the value of the weights and biases in the neural network intended for the predictor 21, for a given number of layers of neurons and a given number of neurons in each layer, until the error between the result output by the neural network (i.e. the predicted templates of the functions) and the observed exchange templates of the functions as indicated in the learning base, which actually followed at any instant t'i in a time window of duration ΔT' (for example ΔT'<ΔT) starting at t'i (corresponding to the predetermined time window for predicting the predictor 21) or less than a predefined threshold.

• volumétrie des échanges indiquant le volume de données échangées ;
• débit utilisé ;
• durée des échanges ;
• destinataire(s) (notamment si l’échange est point-à-point ou point-à-multipoint) ;
• priorité relative des flux typologie des échanges (typiquement s’agit-il de phonie claire, de phonie chiffrée, de messages tactiques, d’images, de vidéo, de transferts de fichiers, d’échanges de base de données ;
• …

The exchange templates which followed the instant ti during the duration Fti, are deduced from one or some of the following information present, directly or indirectly, in the learning base because associated with the type and phase of mission considered (or derived from them), for the period of duration Fti after the time ti:

• volume of exchanges indicating the volume of data exchanged;
• flow rate used;
• duration of exchanges;
• recipient(s) (in particular if the exchange is point-to-point or point-to-multipoint);
• relative priority of flows typology of exchanges (typically it involves clear speech, encrypted speech, tactical messages, images, video, file transfers, database exchanges;
• …

The exchange template predictor 21 makes it possible to identify the most relevant template adjustment rules (and therefore operating modes) which will make it possible to minimize the impact of a variation in transport conditions.

This results in a set of parameters (weight and learned bias), associated with each mission type profile, to configure the neural network of the exchange predictor 21.

Phase of realisation of assignment

The mission begins, the aircraft successively executing the mission phases, wireless telecommunications being implemented, for example in the present case between the aircraft 30 and the control tower during the take-off phase by means of wireless links implemented by the telecommunications block 26 and subsequently in the intervention zone, between the aircraft 30 and the rolling machine 70. Throughout the execution of the mission, the set 300 of steps, comprising the steps 301 to 307, is implemented.

• les conditions de mission courantes ; comme vu précédemment, elles comportent au moins certaines des données parmi : données météo, position et attitude courantes de l’avion, trajectoire suivie (virage, montée, descente), vitesse de l’avion, accélération, relief (modèle de terrain), cap, positions amies et menaces ennemies, environnement électromagnétique dont brouillage) et
• des valeurs relatives à la connectivité considérée à un instant Ti (et éventuellement antérieures à Ti, à l’intérieur d’une fenêtre temporelle Fti de durée Δt se terminant à Ti ; ces mesures relatives à la connectivité indiquent typiquement pour chaque liaison de télécommunication sans fil impliquée dans la mission la valeur d’une ou de plusieurs de ses caractéristiques de transmission : bande passante, taux d’erreur, latence, gigue.

At each instant Ti of a set of successive instants Ti, in an evaluation step 301, the control block 23 evaluates:

• current mission conditions; as seen previously, they include at least some of the following data: weather data, current position and attitude of the aircraft, trajectory followed (turn, climb, descent), speed of the aircraft, acceleration, relief (terrain model), heading, friendly positions and enemy threats, electromagnetic environment including jamming) and
• values relating to the connectivity considered at a time Ti (and possibly prior to Ti, within a time window Fti of duration Δt ending at Ti; these measurements relating to connectivity typically indicate for each telecommunications link without wire involved in the mission the value of one or more of its transmission characteristics: bandwidth, error rate, latency, jitter.

This evaluation is carried out as a function of obtaining, by the control block 23, measurements carried out in the aircraft 30 and, where appropriate, carried out in the platforms in connection with the aircraft 30 and then transmitted to the latter (signaling internal to the network complements the local measurements, the local evaluation is however not conditioned on the reception of network signaling and takes place in all cases with a variable level of precision).

The data indicating the current mission conditions and the data relating to the connectivity considered at a time Ti evaluated are transmitted to the connectivity state predictor 22.

In a connectivity anticipation step 302, based on this transmitted data and furthermore on the indication of the current mission phase that it receives as input, the connectivity predictor 22 determines the future connectivity state, in a time window of duration ΔT starting at Ti, of the wireless links involved in the mission: for example, the connectivity state of link 51, in the theater phase. As indicated previously in relation to learning, the predicted connectivity states indicate for example, for each wireless telecommunications link involved in the mission, the value (or selectively indicates a subrange of values in which this value is located, at within a range of values comprising several sub-ranges) of one or more of its transmission characteristics: bandwidth, error rate, latency, jitter.

In a step 303 which is for example parallel to step 301 all the data indicating, at time Ti, the current mission conditions and the connectivity data considered in step 301 as well as all the data characterizing the exchanges at time Ti (volume of exchanges (volume of data exchanged, flow rate used, duration of exchanges, etc.), relative priority of flows, typology of exchanges, etc.), are stored by the control block 23 in a memory of the processing module 20.

In a step 304, the exchange template predictor 21 receives as input the current mission phase indication and the future connectivity state predicted by the predictor 22 in step 302 (and optionally the types and characteristics of the exchanges observed at time T). Then depending on this current mission phase and furthermore depending on the predicted connectivity state, the exchange template predictor 21 predicts the exchange adjustment rules (in the duration window ΔT' starting at Ti) , in particular the operating modes of the sensor blocks 24 and/or trajectory calculation blocks 25. The exchange template predictor 21 or the control block 23 determines, where appropriate, notifications intended for a block or blocks among the blocks sensors 24 and/or trajectory calculation 25, when their respective predicted operating mode differs from their current operating mode, to switch them to their predicted mode, and thus adapt the flows emitted by the aircraft.

In a step 305, each of the blocks receiving such a notification switches from its current operating mode to the predicted operating mode.

If the predicted connectivity state is a drop in bandwidth below a first predefined threshold and above a second predefined threshold, the exchange predictor block 21 or the control block 23 determines for example that it would be suitable for a video sensor to switch to its degraded operating mode 1, a notification then being transmitted to this effect to the video sensor; if the predicted connectivity state is a drop in bandwidth below the second predefined threshold and above a third predefined threshold, the exchange predictor block 21 or the control block 23 determines that it would be suitable for a sensor video to switch to its degraded mode 2, a notification then being transmitted to this effect to the video sensor (or that it would be appropriate to switch another application from its nominal operating mode to its degraded mode 1) etc. Conversely, if the predicted connectivity state is an increase in bandwidth, the opposite changes are notified. Hysteresis mechanisms are also implemented to maintain a certain stability.

The predictor block 21 provides the operating modes to be applied, in particular to the sensors to minimize the impact of variations in the available resources. Each sensor will thus anticipate a future reduction in resources: for example will prepare a buffer with compressed files (image/video/other sensors) with the quality adapted to the transmission mode to be applied from now on or will reduce its transmission frequency...

In the event of an anticipated breakdown in communication (i.e. no operating mode applicable), the sensor can buffer the data it produces while communication is reestablished. It can then release the stored data according to the operating mode which will be active when communication returns. Switching to a degraded operating mode may be accompanied by a change in encoding to reduce the throughput or to accommodate higher latency, filtering of the data exchanged to reduce the throughput, a filtering of a certain type of information to be compatible with an exchange on a civil network (as opposed to a heritage network). In a step 306 of trajectory adaptation (for example parallel to step 304 or 305), the control block 23, depending on the connectivity state predicted in step 302, determines a trajectory adaptation of the aircraft 30. For example, an adaptation to maintain satisfactory transport conditions is to extract itself from 'a masking zone linked to the terrain for example; another adaptation is to bypass an enemy to avoid detection due to the transmissions implemented in the mission. This then involves providing the avionics system with elements enabling it to modify the trajectory of the aircraft.

In a step 307 of adaptation of the telecommunications links (for example parallel to step 304 or 305 or 306), the control block 23, depending on the connectivity state predicted in step 302 adapts the communication links. wireless telecommunications involved in the mission: for example, to avoid a loss (or drop) of connectivity predicted in the connectivity state and relating to the link 51, flows having to pass normally on the radio link 51 are redirected, depending on their relative priority, towards a satellite link implemented by the transmission block 26 of the aircraft 30; or the frequency used (or any other transmission resource) to implement the radio link 51 is modified; or else an adaptation of the network topology by the control block is carried out by creating a new wireless link (for example, the plane 86 is then sent close to the plane 30 to serve as a wireless telecommunications relay, for example the implementation of a wireless link 52 with the aircraft 30, to compensate for a predicted break on the link 51).

The connectivity conditions of each mission phase also imply respective maximum distances to ensure transmission, the mission platforms will be positioned in such a way as to respect these distances, and therefore to ensure a certain topology of the aircraft network during the phase. respective to the mission. These topologies vary from one phase of the mission to another, the planes will have to be controlled to follow their main mission but also to ensure this connectivity.

Steps 304, 305, 306, 307 are implemented in one embodiment also depending on at least one level of QoS desired in the operation in the mission therefore impacting the telecommunications implemented wirelessly, and more generally functions performing in aircraft (including Aircraft 30) and involved in these mission telecommunications.

In one embodiment, the orchestration of the different steps is carried out by the control block 23. For example, the control block includes a processor and a memory storing software instructions which, when executed on the processor, put implement the orchestration of these steps.

• mettant en œuvre une liaison de remplacement, et/ou
• en basculant sur un mode de fonctionnement dégradé des fonctions ; et/ou
• en adaptant les liaisons etc.

Thus the aim of the invention is the following: before the occurrence of an event which could break a wireless telecommunications link implemented in the mission, or affect such a link in such a way that a result rendered by a function executing in the aircraft cannot be delivered by the latter, the event is predicted and an action is triggered according to this connectivity event prediction and according to the prediction of the need for exchanges to allow the continuity of service, in:

• implementing a replacement link, and/or
• by switching to a degraded operating mode of the functions; and or
• by adapting connections etc.

The invention, by combining connectivity state prediction and exchange prediction, anticipates variations in connectivity with a view to maintaining the overall QoS of the system per phase of the mission and jointly anticipates resource allocations and adaptations of modes of communication. operation, before transmission losses occur.

Post-mission phase

In mission restitution, the learning of the predictors is continued, with the connectivity, exchange and mission condition data collected during the mission, so as to continuously improve the relevance of the anticipation.

The invention makes it possible to anticipate variations in connectivity and the need for exchanges (configuration of network parameters and source and channel transmission modes, anticipation of changes in topology and link states, evaluation of exchange needs, trajectory prediction) during these mission phases and make the necessary adaptations to maintain connectivity, based on these predictions, during the mission. It makes it possible to ensure exchanges as best as possible, to reduce communication losses (by implementing alternative solutions in advance to maintain communication in the event of loss of connection), to manage transmission resources as best as possible, to reduce switching times between two configurations and being more resilient to the different hazards of the mission.

The invention has been described above with reference to aerial missions using in particular an airplane, it is of course applicable to aerial missions relating to any type of flying machine (airplane, helicopter, drone, etc.) and to a coordinated set of devices, some mobile or not, comprising means of wireless transmission. More generally, the invention is also applicable to collaborative tactical missions in environments other than the air environment, for example the naval environment or the land environment.

The invention has been described above in an embodiment using Machine Learning techniques to define the predictors 21, 22, by learning from mission data or data obtained in simulation

In another embodiment, the predictions are determined by deterministic functions (rules) from known models implemented by the predictors 21, 22.

In step 200 described above, a set of configuration parameters for each predictor was selected; in another embodiment, each connectivity predictor, respectively of exchanges, is selected from several connectivity predictors, respectively of exchanges, according to the mission profile.

Claims (10)

Method for adaptive configuration of wireless telecommunications implemented, in a mission, by a mobile machine (30), on one or more wireless links (51, 52), with at least one platform (70, 80) of a set of platforms, said method comprising the set of following steps, iterated during successive moments of the mission, an electronic predictor of exchange templates (21) and an electronic connectivity predictor (22) being on board, in the mobile machine, in an electronic processing block (20) comprising a set of processing function(s):
- prediction of the future state of the wireless links by the connectivity predictor (22) over at least one predefined time horizon, depending on at least the current state of the wireless links (51, 52) and a or several mission conditions parameters among weather data, the current position of the mobile device, the current attitude of the mobile device, the trajectory followed, the speed of the mobile device, the relief, the heading, the friendly positions, enemy threats, the electromagnetic environment;
- depending on at least the profile of said mission and/or depending on an estimate of future exchange needs over said time horizon indicating at least one volume of data to be exchanged by all the processing functions via wireless links and depending on said prediction of the future state of the wireless links, prediction, by the exchange template predictor (21), of an operating mode of at least one function of the set of processing function(s) among several operating modes of said function associated with distinct exchange templates;
- following said prediction, when said predicted operating mode of the function is different from an operating mode of the function currently implemented, triggering a switch of at least said function to said predicted operating mode. Method for adaptive configuration of wireless telecommunications according to claim 1, according to which a selection has previously taken place, depending on the profile of said mission, of an electronic connectivity predictor (22) among several electronic connectivity predictors and of a electronic predictor of exchange templates (21) among several electronic predictors of exchange templates, said on-board predictors being those selected. Method for adaptive configuration of wireless telecommunications according to claim 1 or 2, according to which the processing block (20) of the mobile machine (30), as a function of at least the prediction of the future state of the wireless links , performs at least one action among:
adapting the trajectory of the mobile machine (30) or a platform (70, 80) to maintain an existing wireless link or create a new wireless link;
re-direction of flows within said wireless links;
adapting the characteristics of said connections. Method for adaptive configuration of wireless telecommunications according to any one of the preceding claims, according to which the mobile machine (30) is an aircraft and the mission is an aerial mission which comprises several phases among at least one take-off phase, a phase transit, a theater phase and a landing phase and the prediction of the future state of the wireless links by the connectivity predictor (22) and the prediction of future exchange templates by the template predictor exchanges (21) are each a function of the current phase of the mission. Method for adaptive configuration of wireless telecommunications according to any one of the preceding claims, according to which at least one of the predictors among the connectivity predictor (22) and the exchange template predictor (21) comprises a neural network ; and said predictor selection comprises selecting a set of weights and biases of the neural network from a set of available sets of weights and biases, depending on the mission profile. Method for adaptive configuration of wireless telecommunications according to any one of the preceding claims, according to which, at least one of the predictors among the connectivity predictor (22) and the exchange template predictor (21) comprises a network of neurons; And
- during the mission, data indicating the current state of connectivity of the links, data indicating the exchanges carried out by the functions on the wireless links and data indicating the current mission conditions are collected and stored;
- and said data collected and the mission profile are used to continue learning said neural network. Computer program intended to be stored in the memory of an electronic processing unit on board a mobile vehicle to adaptively configure telecommunications implemented during a mission, by the mobile vehicle, on one or multiple wireless links; the processing block further comprising a microcomputer, said computer program comprising instructions which, when executed on the microcomputer, implement the steps of a method according to one of the preceding claims. Electronic processing block (20) adapted to be embarked on a mission, in a mobile machine (30) adapted to implement, on one or more wireless links (51, 52), telecommunications with at least one platform (70 , 80) of a set of platforms, said electronic processing block (20) comprising a set of processing function(s), an electronic predictor of exchange templates (21) and an electronic connectivity predictor (22); in which :
the connectivity predictor (22) is adapted to, during successive moments of the mission, predict the future state of the wireless links at least one predefined time horizon, as a function of at least the current state of the wireless links (51, 52) and one or more mission condition parameters from weather data, the current position of the mobile machine, the current attitude of the mobile machine, the trajectory followed, the speed of the machine mobile, terrain, heading, friendly positions, enemy threats, electromagnetic environment;
- the exchange template predictor (21) is adapted for, during successive moments of the mission, as a function of at least the profile of said mission and/or as a function of an estimate of future exchange needs on said time horizon indicating at least one volume of data to be exchanged by all the processing functions via wireless links and depending on said prediction of the future state of the wireless links, predicting an operating mode of at least a function of the set of processing function(s) among several operating modes of said function associated with distinct exchange templates;
- the processing block (20) being adapted to, following said prediction, when said predicted operating mode of the function is different from an operating mode of the function currently implemented, trigger a switch of at least said function towards said predicted mode of operation. Electronic processing block according to claim 8, adapted for, depending on at least the prediction of the future state of the wireless links, to perform at least one action among:
adapting the trajectory of the mobile machine (30) or a platform (70, 80) to maintain an existing wireless link or create a new wireless link;
re-direction of flows within said wireless links;
adapting the characteristics of said connections. Electronic processing block according to claim 8 or 9, in which the mobile machine (30) is an aircraft and the mission is an aerial mission which comprises several phases among at least one take-off phase, a transit phase, a phase of theater of operation and a landing phase and the prediction of the future state of the wireless links by the connectivity predictor (22) and the prediction of future exchange templates by the exchange template predictor (21) are each a function of the current phase of the mission.