FR3038089A1 - METHOD AND DEVICE FOR PRODUCING COMPLEX DATA ACQUISITION PROGRAM BY CONSTELLATION OF ELEMENTARY DATA ACQUISITION DEVICES - Google Patents

Abstract

The invention relates to a device and a method for generating a complex data acquisition program by a constellation of acquisition devices, said method comprising: a step of receiving (10) a plurality of client requests; a step of cutting (20) each client request into a plurality of elementary requests; a determination step (30), for each elementary request, of at least one acquisition time window of this elementary request by at least one device for acquiring the constellation; an assignment step (40) to each acquisition device of said constellation, by a cooperative multi-agent system (50, 51, 52), a list of elementary requests each associated with an acquisition slot.

Description

BACKGROUND OF THE INVENTION 1. Technical Field of the Invention The invention relates to a method and a device for producing a data acquisition program. a complex data acquisition program by a constellation of data acquisition devices. The invention particularly relates to a method and a device for planning a data acquisition mission, in particular the planning of an earth observation mission by a constellation of satellites. The invention also relates to a computer program product for implementing the method according to the invention and a corresponding storage means. 2. Background Art There are many technical fields in which a constellation of elementary data acquisition devices is implemented to carry out the acquisition of complex data in a predetermined period of time and according to predetermined constraints. Throughout the text, complex data is a data item formed by a plurality of elementary data and can not be acquired directly by a single elementary data acquisition device or by a single elementary data acquisition. In other words, an elementary datum is data that can be acquired in full by a single data acquisition device in a single acquisition whereas a complex datum is a datum formed of a plurality of elementary data and thus requiring a plurality of acquisitions by one or more data acquisition devices. For example, in the case of an application of earth observations by a satellite constellation, a request to observe a portion of the earth 30 can form a complex piece of data which can not be acquired by a single 3038089 2 acquisition of images by a single satellite, the dimensions of the geographical area to be observed being, for example, larger than the dimensions of the swath of the different satellites of the constellation. In the context of such an application, complex data can also correspond to a request for observations of a plurality of disjunct zones of the earth, and can only be realized by the combination of a plurality of acquisitions of given by one or more separate satellites. Throughout the remainder of the text, the invention is essentially described in connection with a constellation of satellites for observation of the earth. That being so, the invention extends to all types of elementary data acquisition devices, such as for example flying robots, better known under the name of drones, equipped with image acquisition devices, with a view to to carry out specific complex data acquisition missions, or according to another example, stratospheric balloons equipped with meteorological data acquisition instruments for example. In general, the invention applies to all constellation types of elementary data acquisition devices to maximize the number of complex data that can be acquired within a predetermined period of time and under a set of predetermined constraints. In the case of an earth observation constellation, it may be acquisition of ultraviolet signals, visible, infrared. A constellation according to the invention can also be dedicated to the acquisition of radiofrequency signals. In the case of a possibly heterogeneous constellation of satellites, the scheduling of the observation requests received by the constellation manager is a difficult task which aims to optimize in particular the number and / or the quality of the observations made by the constellation. within a predetermined period of time. The difficulty of planning (in other words, the development of the observation program) comes from the fact that many parameters and constraints that are often contradictory must be taken into account, such as the number of available satellites and their characteristics and constraints. technical (maneuverability, memory capacity, on-board power, etc.), 3038089 3 the volume of requests received (which are the requests received by the customers of the satellite constellation manager), the specific characteristics of each customer request (type of shooting, priority of the client, quality requested, time range of validity, etc.) and external constraints 5 (cloud cover, availability of a base station to recover the acquired data, etc.). In addition, the determination of the constellation planning must be done in a limited time. Planning must also be able to be dynamic to integrate priority urgent requests. There is therefore a need for a system and method for scheduling earth observation requests, and generally for planning complex data acquisitions, by an observation satellite constellation. , and by extension a constellation of data acquisition devices. The planning of complex data (also referred to as queries throughout the rest of the text) 15 is therefore to select a list of requests from the set of requests received in order to allocate them to different data acquisition devices. In particular, there is a need to be able to program daily on a dozen satellites of the order of 1000 earth observation requests chosen from more than 10 000 requests received.

The determination of a mission plan by a satellite constellation is to select requests from a set of requests and to assign each selected request to a satellite of the constellation. One of the methods commonly used today to determine a mission plan for a satellite constellation is the so-called "chronological Wolverine" method, better known by the English name of "Greedy Algorithm". This method consists in chronologically scheduling the requests according to their accessibility date and implementing a local heuristic in case of conflict between several requests. The quality of the results provided by this method therefore strongly depends on the local heuristics used to handle programming conflicts. It is therefore necessary to adapt this heuristic to each type of problem encountered. 3038089 4 In other words, the method can not be used easily to plan various missions. The heuristic is very closely related to the type of missions encountered. In addition, this method has the disadvantage of not being dynamic. If a request, urgent or not, arrives while the planning is already in progress, it is necessary to resume the entire programming to be able to program this new request. In practice, any request that arrives while the programming of the satellite constellation is already launched, will be saved in a memory and taken into account only during the next programming of the constellation. Moreover, this method does not make it possible to manage the balance of the load of each of the satellites of the constellation. The inventors have therefore sought to propose a new method of planning terrestrial observation requests by a satellite constellation, and by extension, a new method of planning complex data acquisition requests by a plurality of acquisition devices. basic data. 3. OBJECTIVES OF THE INVENTION The present invention aims at overcoming at least some of the disadvantages of the methods commonly used at present to plan the missions of a constellation of data acquisition devices, such as a data acquisition device. constellation of earth observation satellites. In particular, the invention also aims to provide, in at least one embodiment of the invention, a method and a device for developing a complex data acquisition program by a constellation of data acquisition device. basic data.

In other words, the invention aims to provide a method and a device for scheduling the missions of a constellation of acquisition devices. The invention also aims to provide, in at least one embodiment of the invention, a method and a device that allow a dynamic programming of the mission, that is to say a programming that allows the taking into account immediate receipt of a new application for acquisition, including during the mission's programming and / or which authorizes the immediate recognition 3038089 5 of any evolution of the constellation (loss of an acquisition device or a feature of an acquisition device). The invention also aims to provide, in at least one embodiment, a mission scheduling method and device that makes it possible to manage the load balance of each of the constellation acquisition devices. The invention also aims to provide, in at least one embodiment of the invention, a method and a device for breadmaking missions, which make it possible to maximize the number of acquisitions made by the constellation of satellites, and consequently of to minimize the average cost of data acquisition by the constellation. The invention also aims to provide, in at least one embodiment of the invention, a mission planning method and device adapted for earth observation missions by a satellite constellation. The invention also aims to provide a computer program product 15 for carrying out the method according to the invention and a corresponding storage means. 4. DISCLOSURE OF THE INVENTION To this end, the invention relates to a method for producing a complex data acquisition program by a plurality of elementary data acquisition devices 20, called constellation of acquisition devices. said method comprising: a step of receiving a plurality of complex data acquisition requests, known as client requests, each client request comprising acquisition constraints of said complex data, a step of splitting each client request into one plurality of requests, called elementary requests, for acquiring an elementary data item that can be acquired in a single acquisition by a single acquisition device of said constellation, a determination step, for each elementary data item, of at least one acquisition time window - preferably 3038089 6 each time window of acquisition - of this data item p ar at least one device for acquiring said constellation, a step of assigning to each acquisition device of said constellation, by a cooperative multi-agent system, a list of elementary data to be acquired, each associated with a slot; acquisition data located within a time window of acquisition of this data item, said elementary data thus assigned forming a plurality of complex data of said customer requests, and defining said acquisition program. A method according to the invention thus makes it possible to calculate a program for acquiring a plurality of complex data by programming the various elementary data elements forming the complex data on the various data acquisition devices of the constellation. This programming of the different elementary data implements a cooperative multi-agent system which has the capacity to take into account a large number of distinct entities and to emerge a global solution by solving a plurality of problems at the scale local. Thus, according to the invention, the overall constrained programming problem of data acquisition requests on the constellation of acquisition devices is broken down into a number of particular problems, which are solved locally. A cooperative multi-agent system notably allows an exchange of information between the different agents and a decision-making guided by predetermined criteria.

In addition, a cooperative multi-agent system is a self-adaptive artificial intelligence system. Also, a method according to the invention implementing such a mufti-agents system is dynamic and allows consideration of any new programming request, including during programming. Advantageously and according to the invention, the step of assigning each acquisition device of said constellation a list of elementary data items to be acquired each associated with an acquisition slot comprises the following steps: a step of calculating, for each elementary request, a programming criticality for this basic request, a step of selecting, for each elementary request, a list of acquisition devices compatible with at least one acquisition time slot associated with this basic request, called authorized list, a programming request step of said elementary requests on each acquisition device of said authorized list associated with each elementary request, a calculation step, for each acquisition device, a cost programming of each elementary request received by this acquisition device, a selection step, in for each elementary request, of the acquisition device having the lowest cost, a selection step, in the event of a programming conflict, for each acquisition device, of the elementary request having the strongest programming criticality, a step of confirming the programming, for each selected elementary request, of the selected acquisition device. These steps are implemented by the multi-agent system and make it possible to guide the programming of elementary requests on each of the acquisition devices of the constellation according to two specific criteria.

The first specific criterion guiding the programming of elementary requests is the criticality of each elementary request, that is to say an information representative of the distance that separates the current state of the elementary request from the state in which its objective local is reached, that is to say its programming.

The invention aims to promote the programming of elementary requests with a strong criticality. This results in the selection step, in the case of a programming conflict, for each acquisition device, of the basic request having the strongest programming criticality. Thus, the acquisition devices will spontaneously choose to program the most critical requests in the event of a programming conflict between several elementary requests. According to a variant of the invention, the criticality calculation step assigns each request identical criticality. Such a variant makes it possible to perform iso-critical programming. Advantageously and according to the invention, this programming criticality 10 of an elementary request is a function of information representative of the urgency of the planning of this basic request, and / or of the number of acquisition time windows available for this request. basic request, and / or a percentage of realization of the client request on which this basic request depends.

Preferably, the programming criticality is a function of all three parameters mentioned above. According to this variant, the programming criticality of an elementary request depends on three parameters. The first parameter is to translate the urgency of planning the basic query. This urgency is typically provided by a client with the client request. The value of this parameter may for example be dependent on the price paid by the customer for the operator of the constellation to take charge of his request. The second parameter is an indication of the number of remaining acquisition time windows to program the basic query. Indeed, the invention provides that the elementary request having the smallest number of available acquisition time windows has priority over the elementary request having a greater number of available acquisition time windows. The third parameter aims to translate the interest of programming the basic request considering the programming of the other basic requests forming the client request.

The second specific criterion guiding the programming of elementary requests is the cost of programming on a device for acquiring an elementary request. This cost is an indicator of the difficulty of programming an elementary request by an acquisition device. The invention therefore makes it possible to promote the programming of elementary requests on the acquisition devices which are the least affected by this programming. This results in the step of selecting, for each basic request, the acquisition device having the lowest cost. Advantageously and according to the invention, the cost of programming an elementary request by an acquisition device is a function of the load of this acquisition device and / or the number of programming requests received for this device. acquisition and / or criticality of this elementary query. Preferably, the programming cost of an elementary request depends on all the three parameters mentioned above. Alternatively, the programming cost may depend on other parameters. A method according to the invention therefore makes it possible to provide an acquisition planning program which is guided by the criticality of the basic requests and by the cost of each programming. If the number of acquisition devices changes, these criteria are automatically updated. Also, a method according to the invention can be used with different types of constellation, without requiring the development of a specific heuristic.

According to a variant of the invention, the step of selecting, in the event of a programming conflict, for each acquisition device, of said elementary request having the strongest programming criticality comprises, in the case of the presence of several requests. with equivalent criticality, a substep of attempting to adapt the programming of said requests programmed in the vicinity of the conflicting requests to be able to program all the conflicting requests. This variant makes it possible to manage the cases where several basic queries with equivalent criticality (the notion of equivalence being determined by a predetermined metric) seek to be programmed on the same acquisition device.

In this case and according to this variant, the acquisition device will attempt to move each preprogrammed basic request in the vicinity of the conflicting requests 3038089 10 in order to be able to program all the conflicting elementary requests. For this, the acquisition device will determine whether the movement of the elementary requests in the vicinity of the conflicting requests allows this programming while respecting the different programming constraints of each of the elementary requests. Advantageously and according to the invention, said acquisition constraints associated with client requests comprise constraints, spatial, temporal, and / or technical. These spatial constraints typically include a geographical delineation of the data to be acquired. For example, in the case of a constellation of earth observation satellites, the spatial constraint delimits the area from which images are to be acquired. Time constraints typically include a time window of data acquisition, for example, a 24-hour window to provide the result to the client. The technical constraints include, for example, indications of the type of instrument to be used to acquire the data, or tolerances on the quality of the data that can be acquired. For example, according to an advantageous variant of the invention, said technical constraints comprise a cloud coverage rate tolerated during the acquisition of the elementary data associated with said complex request. In particular, in the case of a constellation of earth observation satellites, a client can provide with his observation request a tolerance on the cloud cover, for example by accepting a cloud cover of 20% maximum. If the cloud cover exceeds this value, the request can not be programmed. According to an advantageous variant of the invention, said devices for acquiring said constellation are image acquisition devices mounted on earth observation satellites, said complex data to be acquired are zones of the terrestrial globe, and said elementary data are images that can be acquired directly by a single satellite of said constellation.

According to this variant, the invention allows the planning of an earth observation mission, including optical observations, by satellites. The invention also relates to a device for producing a complex data acquisition program by a plurality of elementary data acquisition devices, called constellation of acquisition devices, said device comprising: a reception module a plurality of complex data acquisition requests, called client requests, each client request including acquisition constraints of said complex data, a module for cutting each client request into a plurality of requests, called basic requests, of acquisition of an elementary data item that can be acquired in a single acquisition by a single acquisition device of said constellation, a module for determining, for each elementary data item, at least one temporal window for acquiring this item of data by at least one constellation acquisition device, a multi-agent cooperative system configured d, in order to be able to assign to each acquisition device of said constellation a list of elementary data items to be acquired each associated with an acquisition slot located within a temporal window of acquisition of this item of data, said data thus allocated elementary elements forming a plurality of complex data of said client requests, and defining said acquisition program. Advantageously, a device according to the invention implements a method according to the invention and a method according to the invention is advantageously implemented by a device according to the invention.

Throughout the text, a module is defined as a software element, a subset of a software program, which can be compiled separately, either for independent use, or to be assembled with other modules of a program, or a hardware element, or a combination of a hardware element and a software subprogram. Such a hardware element may include an application-specific integrated circuit (ASIC) or a programmable logic circuit (FPGA) for the English name Field. -Programmable Gate Array) or a specialized microprocessor circuit (better known by the acronym DSP for the English name Digital Signal Processor) or any equivalent hardware. In general, therefore, a module is an element (software and / or hardware) that provides a function. A cooperative multi-agent system is a set of agents that cooperates with each other to bring out a solution. An agent is a module as defined above, capable of deciding and interacting with other agents depending on specific parameters characterizing a situation. Advantageously and according to the invention, said cooperative multi-agent system comprises a plurality of agents, called acquisition agents, configured to pursue an objective of acquiring basic data, a plurality of agents, called elementary request agents, configured to pursue a purpose of its programming and a plurality of agents, called client request agents, configured to pursue an objective of programming all the basic requests it contains, each agent comprising communication means and The invention also relates to a computer program product downloadable from a communication network and / or recorded on a computer readable medium and / or executable by a processor, characterized by comprising program code instructions for carrying out the method of detecting terminating a complex data acquisition program by a plurality of elementary data acquisition devices according to the invention, when the program is run on a computer.

The invention also relates to a computer readable storage medium, totally or partially removable, storing a computer program comprising a set of instructions executable by a computer to implement the method of determining a program of complex data acquisitions by a plurality of elementary data acquisition devices, according to the invention. The invention also relates to a method and a device for determining a complex data acquisition program by a constellation of acquisition devices, a computer program product and a corresponding storage means, characterized in combination by all or part of the characteristics mentioned above or below. 5. List of Figures Other objects, features and advantages of the invention will appear on reading the following description given solely by way of non-limiting example and which refers to the appended figures in which: FIG. 1 is a diagrammatic view of FIG. a method of developing an image acquisition program by a satellite constellation, FIGS. 2a and 2b are views of the principle of splitting an image into a plurality of sub-images, called meshes, implemented during the image acquisition process according to one embodiment of the invention, FIG. 3 is a diagrammatic view of the agents of the cooperative multiagent system implemented by the method of production of data. acquisition of images according to an embodiment of the invention, and integrated in a device for producing an image acquisition program according to one embodiment of the invention, FIGS. 4a, 4b, 4c , and 4d are i Illustrative of an example of interactions between mesh agents and satellite agents implemented by a method of developing an image acquisition program according to an embodiment of the invention, the figures 5a, 5b, 5c, and 5d are illustrations of an example of interactions between mesh agents and satellite agents implemented by a method of developing an image acquisition program according to an embodiment of the invention. 6. DETAILED DESCRIPTION OF AN EMBODIMENT OF THE INVENTION The detailed description which follows is concerned with a constellation of optical satellites whose purpose is to establish a program of observations. In other words, according to the embodiment described here in detail, the constellation of data acquisition devices is formed by a constellation of earth observation satellites. The complex data is formed by a set of images and the elementary data is formed by elementary images, hereinafter referred to as the meshes. A complex data according to the embodiment of the figures is therefore a set of elementary images, in other words, a set of meshes.

As shown in FIG. 1, the method for determining an earth observation program comprises a step of receiving a plurality of image acquisition requests, each request comprising acquisition constraints. of these images. These constraints include, for example, the priority of the request, the type of shot desired, the validity of the request, etc. The method further comprises a step of splitting each request into a plurality of meshes, each mesh being acquired in a single acquisition by a single shot. In other words, a client earth observation request is divided into a plurality of elementary cells, each mesh defining a geographical area of the earth composing the client request and that a satellite of the constellation can acquire in one only shot. The step of cutting a client request into a plurality of elementary cells will not be described here in detail. There are many algorithms in the literature that allow this clipping.

Depending on the type of request received, it is possible that some client requests do not need to be split into a plurality of distinct meshes, since the field of view of the geographical area of the request is already in place. compatible with satellite view fields. In this case, the mesh and the client request are combined. Figures 2a and 2b schematically illustrate the principle of a division of a customer request. FIG. 2a represents a geographical zone 7 defined by the client request whose acquisition is requested. This geographical zone 7 is larger than the field of view of each of the satellites of the constellation. Also, the cutting step 20 cuts the geographic area 7 of Figure 2a into three areas 7a, 7b, 7c shown in Figure 2b. Each zone 7a, 7b, 7c of FIG. 2b can be acquired by a single satellite during a single shot. This operation is repeated for each client request received in step 10 and which one seeks to program. The method further comprises a step 30 of determining, for each cell, all of its time windows of visibility, that is to say of all the periods during which the mesh can be acquired by at least a satellite. This step is not described here in detail and can be implemented by the joint use of an algorithm for calculating satellite trajectories around the earth, and an algorithm for calculating solar ephemerides, for example in the case of the acquisition of images in the visible domain. A method according to the invention also comprises a step 40 of allocating to each satellite of the constellation, by a cooperative multi-agent system, a list of meshes to acquire each associated with an acquisition slot located inside. a time window of visibility of this mesh 25, the mesh thus assigned to form later by combining the mesh at least some of the received client requests that have been programmed well. Preferably, the allocation step 40 comprises a calculation step 41, for each cell, a programming criticality for this mesh. It also includes a selection step 42, for each mesh, of a list of satellites, said list authorized for this mesh, having at least one window 3038089 16 visibility associated with this mesh. It also comprises a request step 43 for programming said meshes on each satellite of said authorized list associated with each cell. It also comprises a calculation step 44, for each satellite, of a programming cost of each mesh on this satellite. It also includes, for each mesh, a selection step 45 of the satellite having the lowest cost. It also comprises a selection step 46, in the event of a programming conflict on a satellite, and for each satellite, the mesh presenting the strongest criticality. Finally, it includes a confirmation step 47 of the programming of each selected mesh on each selected satellite. Although these different steps are presented here in a sequential and orderly manner, in practice these steps are implemented in a nested, potentially iterative manner by the use of a cooperative multi-agent system.

In particular, the invention uses a cooperative multi-agent system to assign to each observation satellite of the constellation a list of meshes to be acquired, each associated with an acquisition slot located inside one of the temporal windows of visibility of this mesh. In the following, we will detail the implementation and operation of this cooperative multi-agent system. A multi-agent system is a distributed artificial intelligence system described, for example, in MP Gleizes's document, entitled "Self-adaptive Complex Systems", published in 2012 in volume 7541, pages 114128 of the minutes of the conference "European Workshop on Multi-Agent Systems (EUMAS), Maastricht, the Netherlands. According to a preferred embodiment of the invention, the cooperative multi-agent system is based on the generic agent model known by the term "AMAS4Opt" and adapted to provide design patterns for solving optimization problems. This model is described in particular in the doctoral thesis of E. Kaddoum, entitled "Optimization under constraints of distributed complex problems using cooperative self-organization".

FIG. 3 schematically illustrates the interactions between the different agents used in a cooperative multi-agent system according to the invention. A multi-agent system according to the invention is formed by a set of modules, here called agents. It will be recalled that a module is a software element, a subset of a software program, which can be compiled separately, either for independent use, or to be assembled with other modules of a program, or an element. hardware, or a combination of a hardware element and a software subprogram. The agents are subject to a common environment and interact locally with one another according to predetermined rules. The rules and interactions between the agents make it possible to carry out and implement the various steps described in connection with the allocation step 40. In particular, according to the embodiment of the figures, the multi-agent system comprises a plurality of agents. said satellite agents 50 (forming in this embodiment the data acquisition agents), a plurality of agents, so-called mesh agents 51 (forming in this embodiment the basic query agents) and a plurality of agents, so-called query agents 52. The principle of a cooperative multi-agent system is that each agent interacts with a plurality of distinct entities to arrive at bringing out a global solution by a plurality of solutions to local problems. According to the embodiment of the invention, nine entities have been identified: the three families of cooperative agents: the query agents 52, the satellite agents 50 and the mesh agents 51, 25 three active entities: a cloud cover module 66 , a solar ephemeris calculation module 65 and download stations 62 acquired images (the difference between these entities and the cooperative agents comes from the fact that they have no purpose to satisfy), three passive entities ( system resources): a satellite data storage subsystem, said memory module 63, a satellite energy management subsystem 3038089, said energy module 64, and a computing computation subsystem trajectory and orbital attitude, said trajectory calculation module 61. The embodiment of the various modules is not described here in detail and does not pose particular problems to the skilled person. In the following, we will describe the interactions between these different entities. The use of the AMAS4Opt model makes it possible to define the behavior and the interactions between the agents. Satellite agents have a "service" role, that is, they have the means to solve problems. Mesh agents and queries have a "constrained" role, that is, they raise problems that need to be solved. A satellite agent has the following characteristics: Objective: to acquire shots, - Interactions with the following entities: o the mesh agents 51, o a trajectory management module 61, o a module for calculating the cloud cover 66, the energy module 64, the memory module 63, the calculation module of the solar ephemeris 65, the download stations 62, Negotiations with: the mesh agents. A mesh agent has the following characteristics: Objective: to be programmed, Interactions with the following entities: satellite agents 50, query agents 52, cloud coverage calculation module 66, calculation module of solar ephemeris 65, Negotiations with: 3038089 19 o satellite agents 50, o the query agents 52. A request agent has the following characteristics: - Objective: programming all its meshes, 5 - Interactions with the following entities: o the mesh agents 51, - Negotiations with: o mesh agents 51. The multi-agent system will allow the programming of observation requests 10 by the cooperation of its agents thus defined. This cooperation is ensured by the exchange of messages between the agents and is guided by two indicators: the criticality of the elementary requests (the meshes in the case of an application of observation of the ground), and the cost of programming of the requests by the acquisition devices (the satellites in the case of an earth observation application). The mesh agents send their programming requests to all satellite agents having visibility access to them, and subsequently they prefer the satellite agent at the lowest cost. The satellite agents process all mesh agents that have made a programming request, giving priority to those with the highest criticality. Query agents influence how satisfied their meshes are with the progress of planning the query as a whole. The satellite agent is configured to send mainly two messages: a message to provide an estimate of the cost of planning a mesh, and a message to confirm or not the programming of the mesh. The mesh agent is configured to send mainly two messages: a programming request and a program confirmation request.

The request agent is configured to send mainly a message: a transmission of new information to the meshes that compose it.

This various information is exchanged during the steps 41, 42, 43, 44, 45, 46, 47 described in connection with FIG. 1. FIGS. 4a, 4b, 4c, 4d illustrate a cycle of negotiation and exchanges. information between two mesh agents AM1, AM2 and two satellite agents AS1 and AS2 capable of programming the meshes AM1 and AM2. FIG. 4a is the beginning of the cycle corresponding to the programming request step 43 described in connection with FIG. 1. The AM1 mesh only has a visibility window, hereinafter referred to as access, and referenced al. 1. The AM2 mesh has two ports, referenced a2.1 and a2.2. Each mesh 10 AM1 and AM2 transmits programming requests to satellite agents AS1, AS2 for each of its accesses. FIG. 4b illustrates the next step of the cycle corresponding to the programming cost calculation step 44. The satellite agents AS1 and AS2 calculate the programming costs C1, C2, C3 of each of the links AM1, AM2. These costs are transmitted to the mesh agents AM1 and AM2. Figure 4c illustrates the next step of the cycle corresponding to step 45 of selecting the least expensive satellite. Agents AM1 and AM2 each select the least expensive satellite agents and request confirmation. Finally, FIG. 4d illustrates the validation of the programming by the satellite agents AS1 and AS2. In the case where the satellites encounter a programming conflict, the cycle further comprises a step of selecting the mesh having the highest criticality. According to the invention, these different cycles are repeated until the system is considered as little evolutive. To do this, it is possible to introduce an observer module that regularly comes to determine the stable or evolving character of the programming. According to another variant, the cycles are repeated for a predetermined period of time by experiment. FIGS. 5A, 5b, 5c, 5d illustrate a more dynamic embodiment in which a new AM2 mesh agent is introduced into the system, following, for example, the receipt of a new urgent request from a client. On the 3038089 21 figures, the mesh agent AM1 is already programmed on the satellite AS1. FIG. 5a illustrates the first cycle during which the new agent AM2 makes a programming request to the satellite agent AS1, whereas this agent already has, in this time slot, the programming of the mesh agent AM1.

FIG. 5b illustrates the following cycle in which the satellite agent AS1 compares the cryicity of the AM1 mesh agent with the criticality cr2 of the AM2 mesh agent. It appears that the mesh agent AM2 is more critical than the mesh agent AM1. The satellite agent AS1 then calculates the programming cost of the AM2 mesh agent at the access previously assigned to the mesh agent AM1 and transmits it to the mesh agent AS2. FIG. 5c illustrates the following cycle during which the mesh agent AM2 requests a programming confirmation from the satellite agent AS1. FIG. 5d illustrates the following cycle during which the satellite agent AS1 programs AM2 and cancels the programming of the mesh agent AM1 at access a1.

The mesh agent AM1 is therefore reintegrated in the list of meshes to be programmed. A method according to this embodiment is therefore dynamic and can take into account evolutionarily any new requests submitted to the system. In general, a method according to the invention is dynamic, not only by the possibility of taking into account any new request submitted, but more generally by taking into account any modification of any entity composing the multi system. cooperative agents according to the invention. In particular, if a satellite originally present in the constellation becomes inoperative and can no longer, as such, ensure the acquisition of initially programmed images, the requests previously programmed on this satellite are again submitted to the system and the overall program will evolve for each satellite remaining according to the various criteria of cost and criticality described above of the different requests to reprogram and those already programmed on the various satellites.

Likewise, any evolution of any entity (energy module, memory module, etc.) can be taken into account by the multi-agent system, thus making a process and a device according to the invention dynamic. A method according to the invention can be implemented in the form of a sequence of instructions of a computer program. The method can also be implemented in hardware form or in a mixed hardware and software form. In the case where the invention is implemented partially or totally in the software form, the corresponding instruction sequence can be stored in a removable storage means, such as a floppy disk, a CD-ROM, a DVD-ROM, a USB key, etc., or non-removable storage means, these various means being read partially or totally by a computer or a microprocessor. The invention is not limited to the embodiments described. In particular, the invention has been essentially described in connection with a constellation of earth observation satellites. That said, the method of determining a data acquisition program can be applied to all types of acquisition devices and to all types of data. For example, the invention can be applied to a constellation of drones, or a constellation of balloons probes.

Claims (12)

REVENDICATIONS1. A method of developing a complex data acquisition program by a plurality of elementary data acquisition devices, said constellation of acquisition devices, said method comprising: - a step of receiving (10) a plurality complex data acquisition requests (7), called client requests, each client request comprising acquisition constraints of said complex data (7), - a step of cutting (20) each client request into a plurality of requests, said elementary requests, for acquiring an elementary data element (7a, 7b, 7c) that can be acquired in a single acquisition by a single acquisition device of said constellation, - a determination step (30), for each elementary data item at least one temporal window for acquiring this elementary data item by at least one device for acquiring the constellation, - an allocation step (40) for each positive acquisition of said constellation, by a cooperative multi-agent system (50, 51, 52), a list of elementary data to acquire each associated with an acquisition slot located within a time window acquisition of this elementary data, said elementary data thus assigned forming a plurality of complex data of said client requests, and defining said acquisition program. 2. Method according to claim 1, characterized in that said assigning step comprises: a calculation step (41), for each elementary request, a programming criticality for this elementary request, a selection step (42) , for each basic request, a list of acquisition devices compatible with at least one acquisition time window associated with this elementary request, said authorized list, - a programming request step (43) of said requests 5 elementary elements on each acquisition device of said authorized list associated with each elementary request, - a calculation step (44), for each acquisition device, of a programming cost of each elementary request received by this acquisition device A selection step (45), for each elementary request, of the acquisition device presenting the lowest cost, - a selection step n (46), in the event of a programming conflict, for each acquisition device, of the elementary request presenting the strongest programming criticality, 15 - a confirmation step (47) of the programming, for each selected elementary request of the selected acquisition device. 3. Method according to claim 2, characterized in that said step (46) for selecting, in case of programming conflict, for each acquisition device, said basic request having the strongest programming criticality comprises, in case of presence of several queries with equivalent criticality, a substep of attempting to adapt the programming of said queries programmed in the vicinity of the conflicting queries to be able to program all the queries in conflict. 25 4. Method according to one of claims 2 or 3, characterized in that said cost (c1, c2, c3) of programming a basic request by an acquisition device is a function of the load of this device. acquisition and / or the number of programming requests received for this acquisition device, and / or the criticality of this basic request. 30 5. Method according to one of claims 2 to 4, characterized in that said programming criticality of an elementary request is a function of an information representative of the urgency of the planning of this basic request, and / or the number of acquisition time windows available for this basic request, and / or a percentage of completion of the complex request on which this elementary request depends. 5 6. Method according to one of claims 1 to 5, characterized in that said acquisition constraints associated with acquisition requests include constraints, spatial, temporal, and / or technical. 7. Method according to claim 6, characterized in that said technical constraints comprise a cloud coverage rate tolerated during the acquisition of the elementary data associated with said complex request. 8. Method according to one of claims 1 to 7, characterized in that said acquisition devices of said constellation are optical acquisition devices mounted on earth observation satellites and in that said data complex to Acquire are regions of the earth to be acquired and in that said elementary data are images that can be acquired directly by a single satellite of said constellation. 9. A device for developing a complex data acquisition program by a plurality of elementary data acquisition devices, called constellation of acquisition devices, said device comprising: a reception module of a plurality complex data acquisition requests, known as client requests, each client request including constraints for acquiring said complex data, - a module for splitting each client request into a plurality of requests, called basic requests, for acquiring data. an elementary datum that can be acquired in a single acquisition by a single acquisition device of said constellation, a determination module, for each elementary data element, of at least one temporal window for acquisition of this elementary data by minus a constellation acquisition device, a cooperative multi-agent system (50, 51, 52) configured in order to be able to assign to each acquisition device of said constellation, a list of elementary data items to be acquired, each associated with an acquisition slot located within a temporal window of acquisition of this piece of data, said data thus allocated elementary elements forming a plurality of complex data of said client requests, and defining said acquisition program. 10. Device according to claim 9, characterized in that said cooperative multi-agent system (50, 51, 52) comprises a plurality of agents, called acquisition agents (50), configured to pursue an acquisition objective. elementary data item, a plurality of agents, said elementary request agents (51), configured to pursue an objective of its programming and a plurality of agents, said client request agents (52), configured to pursue an objective of 15 programming all the basic requests it contains, each agent comprising means of communication and being able to interact with at least some of the other agents of the system. 11. Computer program product downloadable from a communication network and / or recorded on a computer readable medium and / or executable by a processor, characterized in that it comprises program code instructions for the implementation the method of developing a complex data acquisition program by a plurality of elementary data acquisition devices, according to one of claims 1 to 8, when the program is executed on a computer. 25 12. A computer readable storage medium, totally or partially removable, storing a computer program comprising a set of instructions executable by a computer to implement the process of developing a program of complex data acquisition by a plurality of elementary data acquisition devices according to one of claims 1 to 8.