Method for allocating time-frequency resources in a satellite communication system with beamforming, apparatus and computer program therefor

Description

Technical field :

[0001] The invention is in the field of satellite telecommunications.

[0002] More precisely, the invention relates to satellite telecommunications systems implementing a MF-TDMA (Multi-Frequency and Time-Division multiple access) or OFDM (Orthogonal Frequency Division Multiplexing) waveform, based on single-user elementary resources (also called block resources), a digital payload with beamforming capabilities to dynamically center each of the telecommunications radiofrequency beams on the user terminal (UE for "User Equipment") selectively associated with the beam, for example in satellite systems implementing telecommunications according to the 4G or 5G New Radio Non-Terrestrial Network (5G NR NTN) standard.

[0003] More specifically, the invention relates to a method for allocating telecommunications time-frequency resources in a satellite wireless telecommunications system, said satellite wireless telecommunications system comprising a satellite, user terminals UE, the satellite being adapted to simultaneously implement several telecommunications beams each associated with a distinct user and to dynamically center each beam on the associated UE; each telecommunications beam being established according to time-frequency resources selectively allocated to the UE associated with the beam. Prior art :

[0004] Time-frequency resources in 5G NR NTN telecommunications are organized, in a known manner, in the form of blocks of elementary resources selectively allocated by a scheduler module to exchanges with a UE. In the time domain, the resource is divided into OFDM symbols / slot (1 slot = temporal succession of 14 symbols) / subframe / frame (see FIG 6 ). When allocating resources to a UE, a so-called "scheduler" module decides on an allocation in terms of PRB ('Physical Resource Block') for the frequency domain, and symbols and / or slots in the time domain.

[0005] If beam management is not coordinated, it may happen that the PRBs associated with two beams formed in directions that are too close give rise to interference levels resulting in a prohibitive degradation of reception throughput performance. US 2021 / 273703 describes a satellite system with several grids, and each UE located in a grid uses a portion of the resources allocated to the grid. Summary of the invention :

[0006] To this end, according to a first aspect, the present invention describes a method for allocating telecommunications time-frequency resources in a satellite wireless telecommunications system, said satellite telecommunications system comprising a satellite, user terminals UE, the satellite being adapted to simultaneously implement several telecommunications beams each associated, at a given instant, with a distinct user terminal UE and to dynamically refocus each beam on the UE associated with said beam; each telecommunications beam being established as a function of time-frequency resources selectively allocated to the associated UE; said method being characterized in that it comprises the following steps implemented by an electronic resource allocation block: obtaining the position of each UE;given a set of grids each representing at least a portion of the Earth's surface and each comprising a plurality of non-contiguous zones distributed in the grid, the distance between any two zones of a grid being greater than a determined non-zero threshold associated with the grid, the allocation of time-frequency resources is carried out, relative to an instant T, by application of at least the following rule, at least one same time-frequency resource being previously associated with the grid: said time-frequency resource associated with the grid can be allocated to each of 2 UEs only if the positions obtained from said 2 UEs are located in distinct zones of said plurality of non-contiguous zones distributed in the grid. ;

[0007] In embodiments, such a method will further comprise at least one of the following features: if during a first step of allocating time-frequency resources using said set of grids each associated with a first determined threshold, it has not been possible to allocate a time-frequency resource to at least one UE in accordance with the rule, an additional allocation step is carried out using at least one set of additional grid(s) representing at least said portion of the Earth's surface and each comprising a plurality of non-contiguous zones distributed in the grid, the distance between any two zones of a grid being greater than a second determined non-zero threshold associated with the grid, less than the first threshold;the allocation of time-frequency resources relative to a time T is further carried out by applying at least the following rule: given a first grid, respectively a second grid distinct from the first grid, of the set of grids associated with a first time-frequency resource, respectively a second time-frequency resource distinct from the first time-frequency resource: when a UE is determined to be located both in a first zone of said plurality of non-contiguous zones distributed in the first grid and a second zone of said plurality of non-contiguous zones distributed in the second grid, it is determined whether the UE is closer to the center of the first zone or to the center of the second zone; and if the UE is determined to be closer to the center of the first zone respectively of the second zone, the first time-frequency resource, respectively the second time-frequency resource, is allocated to it;the position of a UE is obtained by implementing the following steps: calculating values ​​of the beamforming weights maximizing the power of predefined signals received from a UE; estimating the Direction of Arrival, known as DOA, of a signal coming from a UE as a function of said beamforming weights determined for the UE by applying a regression algorithm linking beamforming weights and DOAs; determining the position of the UE as a function of at least the intersection of the Earth's surface and the estimated DOA.

[0008] According to another aspect, the invention describes a computer program intended to be stored in the memory of an electronic resource allocation block in a satellite wireless telecommunications system comprising a satellite, user terminals UE, the satellite simultaneously implementing several telecommunications beams each associated with a distinct user and for dynamically refocusing each beam on the associated UE; each telecommunication beam being established as a function of time-frequency resources selectively allocated to the associated UE; said time-frequency resource allocation 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.

[0009] According to another aspect, the invention describes a device for allocating telecommunications time-frequency resources for a satellite wireless telecommunications system, said satellite telecommunications system comprising a satellite, user terminals UE, the satellite being adapted to simultaneously implement several telecommunications beams each associated, at a given instant, with a distinct user terminal UE and to dynamically refocus each beam on the UE associated with said beam; each telecommunications beam being established as a function of time-frequency resources selectively allocated to the associated UE;said resource allocation device (50) being characterized in that it is adapted to obtain the position of each UE and, given a set (80) of grids each representing at least a portion of the Earth's surface and each comprising a plurality of non-contiguous zones distributed in the grid, the distance between any two zones of a grid being greater than a determined non-zero threshold associated with the grid, to carry out the allocation of time-frequency resources relative to an instant T, by application of at least the following rule, at least one same time-frequency resource being previously associated with the grid: said time-frequency resource associated with the grid can be allocated to each of 2 UEs only if the positions obtained from said 2 UEs are located in distinct zones of said plurality of non-contiguous zones distributed in the grid. ;

[0010] In embodiments, such a device will further comprise at least one of the following features: the device is adapted to, if during a first step of allocating time-frequency resources using said set of grids each associated with a first determined threshold, it has not been able to allocate a time-frequency resource to at least one UE in accordance with the rule, carry out an additional allocation step using at least one set of additional grid(s) representing at least said portion of the Earth's surface and each comprising a plurality of non-contiguous zones distributed in the grid, the distance between any two zones of said plurality of zones of a grid being greater than a second determined non-zero threshold associated with the grid, less than the first threshold;the time-frequency resource allocation device is adapted to carry out the allocation of time-frequency resources relative to a time T, by further applying at least the following rule: given a first grid, respectively a second grid distinct from the first grid, of the set of grids associated with a first time-frequency resource, respectively a second time-frequency resource distinct from the first time-frequency resource: when a UE is determined to be located both in a first zone of said plurality of non-contiguous zones distributed in the first grid and a second zone of said plurality of non-contiguous zones distributed in the second grid, it is determined whether the UE is closer to the center of the first zone or to the center of the second zone;and if the UE is determined to be closer to the center of the first zone respectively of the second zone, the first time-frequency resource, respectively the second time-frequency resource, is allocated to it; the time-frequency resource allocation device is adapted to obtain the position of a UE by calculating values ​​of the beamforming weights maximizing the power of predefined signals received from a UE, then by estimating the Direction of Arrival, called DOA, of a signal coming from a UE according to said beamforming weights determined for the UE by applying a regression algorithm linking beamforming weights and DOAs and by determining the position of the UE according to at least the intersection of the surface of the Earth and the estimated DOA. ; Brief description of the figures :

[0011] The invention will be better understood and other characteristics, details and advantages will appear more clearly on reading the following description, given without limitation, and thanks to the appended figures, given by way of example. [ Fig. 1 ] There Figure 1 is an illustration of a satellite telecommunications system in one embodiment of the invention; [ Fig. 2 ] There Figure 2 schematizes part of the processing chain in reception of the satellite of the satellite telecommunications system of the Figure 1 in one embodiment of the invention; [ Fig. 3 ] There Figure 3 schematizes part of the satellite transmission processing chain of the satellite telecommunications system of the Figure 1 in one embodiment of the invention; [ Fig. 4 ] There Figure 4 illustrates steps of a time-frequency resource allocation method implemented in the satellite telecommunications system of the Figure 1 in one embodiment of the invention; [ Fig. 5 ] There Figure 5 represents a set of grids in one embodiment of the invention; [ Fig. 6 ] There Figure 6 illustrates the division of the time-frequency resource into slots of 14 OFDM symbols and PRB of 12 subcarriers respectively according to the frame numerology for the 5G NR standard; [ Fig. 7 ] There Figure 7 illustrates a decomposition of treatments in an embodiment of the invention with a hybrid DBFN; [ Fig. 8 ] There figure 8 represents a mesh in one embodiment of the invention; [ Fig. 9 ] There Figure 9 illustrates the calculation of the EU direction vector; [ Fig. 10 ] There Figure 10 illustrates the calculation of the satellite direction vector; [ Fig. 11 ] There Figure 11 illustrates the projection of the direction vector into the satellite antenna frame.

[0012] Identical references may be used in different figures when they designate the same or comparable elements. Detailed description :

[0013] There Figure 1 is an illustration of a satellite telecommunications system 10 in one embodiment of the invention.

[0014] The satellite telecommunications system 10 comprises a plurality of user terminals (UE) 20_1, 20_2, ... 20_p, p strictly greater than 1 (4 UEs have been represented in Figure 1 ) and a satellite platform 11.

[0015] Each UE comprises a satellite transmission and reception processing block and a satellite transmission and reception antenna and is adapted to implement satellite communications with the satellite 11 based on an MF-TDMA or OFDM waveform for example. They can be of the fixed terminal type (example: 20_2) or mobile (example: 20_1). Each UE 20_1, ..., 20_p and the satellite 11 which overhangs it are in direct visibility of each other.

[0016] The satellite platform 11 comprises a satellite antenna 12, for example MIMO, and a transmission and reception processing device. It is suitable for implementing satellite communications with the UEs 20_1, ..., 20_p.

[0017] In embodiments, the satellite telecommunications system 10 comprises a ground processing device 14 connected to the satellite platform 11 via a “Feeder” link implemented via the antenna 12 (or via another antenna) of the platform 11 and a satellite antenna of the ground processing device 14. The ground processing device 14 is adapted in these embodiments, to carry out, for the satellite platform 11, processing operations for example requiring a lot of computation or storage volume (for example the calculation of the complex coefficients of the beamforming laws (weights) and / or the modulation / demodulation of the symbols, etc.).

[0018] The telecommunications implemented between the satellite and each UE terminal are for example compliant with the 5G NR NTN standard (and for example Et NB-loT / eMTC (loT NTN)): data are exchanged between the satellite 11 and each UE, respectively 20_1, ..., 20_p, then in communication by the implementation of a specific beam, respectively B_1, ..., B_p, generated between the satellite and this UE and oriented specifically towards this UE. There are therefore at least as many distinct telecommunications beams as there are UEs currently communicating with the satellite at the time considered (because there are also beams used to carry the signaling channels). The frequency width of a beam is variable, one or more unit time-frequency resources being able to be assigned at a given time to a UE according to its needs.The satellite 11 is adapted to implement beamforming techniques to generate these beams and is adapted to, in real time, automatically refocus the beam on the targeted UE based on the updated position of this UE and the updated position of the satellite in the event of movement of the latter.The useful data are transmitted on these beams in a coded manner, in the form of OFDM symbols, using time-frequency resources which are specifically allocated to each UE, in the form of PRB, (“Physical Resource Block”) in the frequency domain and in the time domain, in the form of symbols and / or slots (the present invention applies to these specifically allocated resources; that said, it will be noted that there also exists in the network, in one embodiment, data sent to all the UEs, for example on the physical broadcast channels, and the time-frequency resources allocated to these exchanges are not single-user).

[0019] In a known manner and as represented in Figure 6 , a time slot comprises 14 successive OFDM symbols. A PRB corresponds to a block of 12 successive subcarriers. The spacing referenced Δf on the Figure 6 is equal to the value of SCS (“Subcarrier Spacing”).

[0020] When data is to be exchanged between satellite 11 and a UE 20_i, time-frequency resources are specifically allocated to these exchanges, in a centralized and supervised manner, to control the level of interference between the beams.

[0021] A control block 50, in the satellite platform 11, contributes to the determination of this allocation, in the manner indicated later. During each allocation period, the allocated time-frequency resources are single-user.

[0022] The control block 50 comprises a base 70 comprising the definition data of at least one set of grids, a PRB allocation block 51 and a UE location estimation block 52.

[0023] The calculation of the complex coefficients of the beamforming laws at the PRBs is carried out PRB by PRB, and this, for each OFDM symbol. The coefficient by which a PRB is multiplied is calculated in particular as a function of the location of the UE to which it has been allocated, the known position of the satellite and the central frequency associated with the PRB. Depending on the case, the coefficients are calculated locally in the satellite platform 11 or in the ground processing device 14.

[0024] There Figure 2 schematizes a part of the processing chain 30 in reception of the satellite 11 of the satellite telecommunications system of the Figure 1 in one embodiment of the invention.

[0025] There Figure 3 schematizes a part of the transmission processing chain 60 of the satellite of the satellite telecommunications system of the Figure 1 in one embodiment of the invention.

[0026] In a known manner, the indication of the PRBs to be used for the exchanges between each UE and the satellite platform 11, during a next transmission (or reception) on their part, is delivered dynamically by the control block 50, for example to the gNB, and is then dynamically notified to the UEs (via control signals), and also to the processing chains 30 in reception and transmission of the satellite platform, for example at least every T milliseconds, with T included in the range <0 ms; 10 ms > for example).

[0027] In reference to the Figure 2 , the reception processing chain is adapted to receive, via radiating elements of the antenna 12 of the satellite platform 11, signals transmitted by a plurality of UEs.

[0028] The reception processing chain 30 comprises N parallel processing channels and is further connected to the control block 50 of the satellite system 30.

[0029] Depending on the embodiments, the control block 50 is embedded in the satellite platform 11, or is located on the ground, in the ground processing device 14, or its elements are distributed between the ground and the satellite platform. For example, the control block is in the gNB / eNB.

[0030] The connection(s) between the receiving processing chain and the control block 50 are, depending on the case, wired (for example when block 50 is on board) and / or wireless (in particular if at least part of block 50 is on the ground).

[0031] The reception processing chain notably comprises N parallel processing channels VRi, i = 1 to N, each supplied by the signal captured by a respective radiating element (RE) 40_i of the satellite antenna 12.

[0032] Each VRi processing path includes: a Radio Frequency RF block 31 which is adapted to carry out a frequency change to lower the signal in frequency, an analog-to-digital converter (ADC) block 32 which digitizes the signal, a CP (Cyclic Prefix) suppression block 36; a series-to-parallel (S / P) transposition block 37; an FFT block 33 which is adapted to apply a fast Fourier transform to the signals received at the input, and thus allows to then work in the frequency domain and in particular which allows to manipulate the complex coefficients transmitted on each subcarrier; a parallel-to-series (P / S) transposition block 38; an equalization block (EQLZR) 39; a Demapping block 34 which is adapted to filter the PRBs according to the current PRB allocation list which has been provided to it by the control block 50, to transmit at the output of the Demapping block only the PRBs actually allocated to a UE;the Demapping block uses to do this the indication of the PRBs to be used for the exchanges between each UE and the satellite platform 11 determined dynamically by the control block 50 (in one embodiment, the control block 50 provides the determined allocations to the gNB, which then effectively allocates the resources according to these determined allocations). a DBFN block 35 ("Digital Beamforming Network") is adapted to apply complex beamforming coefficients to the PRBs (in fact, the symbols transmitted on the PRBs allocated in phase and amplitude are weighted) and also uses, to do this, the indication of the PRBs to be used (function of each UE) for the exchanges from each UE to the satellite platform 11 provided dynamically by the control block 50. This processing makes it possible to selectively obtain all the symbols transmitted per transmitting UE (and therefore per receiving beam)). ;

[0033] These sets of symbols at the output of the DBFN 35 are then modulated / demodulated (via the gNB function) according to the digital modulation carried out, locally at the satellite platform 11 or by the ground processing block 14, and according to the PRB allocation information.

[0034] Now referring to the Figure 3 , the transmission processing chain 60 is adapted to shape and deliver to the radiating elements RE of the antenna 12 of the satellite platform 11 useful data intended for UEs among the UEs 20_1, ..., 20_p.

[0035] The transmission processing chain 60 comprises the DBFN block 35 and N parallel processing channels VE1, ..., VEN.

[0036] The transmission processing chain 60 is further connected to the control block 50 of the satellite system 30.

[0037] The DBFN block 35 is adapted to apply complex beamforming coefficients to the symbols from the upstream coding chain which will then be transmitted on the PRBs allocated to this transmission and also uses, to do this, the indication of the PRBs to be used, depending on each UE, for the exchanges from the satellite platform 11 to the UEs and provided dynamically by the control block 50.

[0038] The output data of block DBFN 35 are provided as input to each processing channel VEi, i=1 to N.

[0039] Each VEi treatment pathway includes: a series to parallel (S / P) transposition block 61; an iFFT block 62 which is adapted to apply the inverse of a fast Fourier transform to the signal received at input from the DBFN 35, and thus allows to pass from the frequency domain to the time domain; a parallel to series (P / S) transposition block 65; a cyclic prefix (CP) insertion block 66; a digital-to-analog converter (DAC) block 63 which transforms the signal into analog; a 64 RF Radio Frequency block which is suitable for performing a frequency change (upconversion) of the signal.

[0040] The signal at the output of the VEi channel, i=1 to N, is then delivered to a radiating element RE 60_i of the satellite antenna 12, for transmission.

[0041] The transmission by the radiating elements RE 60_1 to 60_N thus produces a plurality of radiocommunication beams, each beam, using the PRBs allocated to a UE, being dedicated to this UE alone and being selectively centered on this UE.

[0042] Thus the allocation of time-frequency resources, in particular PRBs, in combination with intelligent beamforming, is a sensitive task that directly affects the quality of the exchanged signals, in particular the Signal to Interference Ratio (C / I). It is therefore important to ensure that two nearby beams, for example beams B_1 and B_2, do not use identical PRBs on the same time resources (symbols / slots), which would give rise to interference levels between the beams that would significantly degrade the throughput performance in reception and demodulation.

[0043] The invention thus proposes to take advantage of the knowledge of the location of the UEs to distribute them on a regular grid when allocating the PRBs, the allocation of PRBs on the same frequency can only be done for UEs that are sufficiently far apart.

[0044] In one embodiment of the invention, a set 80 of grids is defined. The base 70 of the control block 50 stores the definition data of the set 80 of grids.

[0045] The set of grids 80 comprises in the case considered a plurality of grids, used as sieves, to sort the UEs to which the same PRB can be allocated.

[0046] Each grid represents the surface of the Earth visible from the satellite platform 11. Each grid point therefore corresponds to a position on the surface of the Earth. Although this arrangement is not mandatory, the point at the top left of each grid corresponds, for example, to the same point on the Earth at the corresponding end of the visible surface of the Earth and the same reference point is used for the different grids, as well as the same scale.

[0047] Each grid comprises discrete non-contiguous areas, which will hereinafter be referred to as "holes". In one embodiment, the holes are arranged in a regular two-dimensional mesh based on square cells (the holes corresponding to the vertices of the cells).

[0048] The pattern of arrangement of the holes on the grid is a function of simple geometric criteria (minimum distance between the UEs using the same PRB = distance between the holes of a grid) and / or is a function of the beam patterns: the shape of the beams is not necessarily always regular, especially in the case of significant misalignment: the grid is then adapted to modify / widen the mesh.

[0049] In this case, the holes are circles, here all of the same size. The size of the hole (here the diameter of the circle) represents the hysteresis that is granted on the position of the EU potentially associated with this hole. Each hole corresponds to a portion of the Earth's surface.

[0050] In embodiments, a hole has a shape other than a circle, e.g., a square, a triangle.

[0051] In one embodiment, the mesh of the holes is defined so that a beam directed at the center of the hole illuminates the entire hole satisfactorily (greater than a fixed minimum threshold) and that two beams illuminating two neighboring holes do not overlap (by setting a maximum beam power threshold at the boundary of the hole it illuminates).

[0052] For example, the diameter of the circle corresponds to a distance from a few km to a few tens of km, for example between 0 km and 100 km, and the minimum distance between two neighboring holes corresponds to a distance for example between a few tens and a few hundreds of km, for example in the range [10 km; 900 km].

[0053] The sizing of the shape of the holes and the associated mesh is done in such a way as to guarantee a certain level of isolation in terms of C / I between the beams likely to point at the holes of the mesh. This sizing will take into account the beam formation diagram which can be generated with the satellite antenna 12 as well as the power allocation carried out by the satellite between the beams served. With reference to the Figure 5 , a portion of each grid G1, ..., G16 (the portions corresponding to the same portion of the Earth's surface) is represented: the holes are represented in white, the rest of the grid portion in gray.

[0054] In one embodiment, in the different grids G1 to G16, the mesh of the points is the same, but offset, i.e. arranged differently each time on the grid.

[0055] In reference to the Figure 4, in one embodiment, the control block 50, which comprises the PRB allocation block 51, the UE location estimation block 52 and the base 70 of the set of grids is adapted to implement the steps of a PRB allocation method 100 for the next time interval considered (i.e. for one or more slots) according to the invention.

[0056] Thus in a step 101, among all the UEs to be served in the coverage, the UEs to be served simultaneously on the next slot are identified and the UE location estimation block 52 obtains the location of the UEs of the system 10 which are in visibility of the satellite platform 11.

[0057] In a step 102, the PRB allocation block 51 determines the allocation of PRBs per UE visible from the satellite platform 11 as a function of the location of the UEs, the allocation of PRBs per UE defining the PRBs to be used for the next time interval considered (excluding those not allocated) for the transmission and reception of data between this UE and the satellite platform.

[0058] Each grid Gi, i = 1 to 16, of the set of grids 80 is associated with a PRB (or a set of PRBs). The same PRB (or set of PRBs) cannot be associated with two distinct grids of the set of grids 80. Distinct PRBs (or sets of PRBs) can be associated independently with the same grid (i.e. without being part of a set of PRBs allocated in block to a UE); these grids are then considered as two distinct grids in the remainder of step 102.

[0059] The allocation of PRB resources is carried out in accordance with the following rule: the same PRB associated with a grid can only be allocated to each of several UEs if the UEs are in distinct holes in the grid (i.e. the PRB cannot be allocated to two or more UEs which would be in the same hole; and the PRB cannot be allocated to a UE not appearing in a hole).

[0060] There are multiple ways to determine how to pair PRBs to grids in step 102, following this rule:

[0061] Based on specified geometric criteria, the PRB allocation block 51 associates UEs with holes in the grids, ensuring that a UE is ultimately associated with only one grid and only one hole in that grid. There may be multiple UEs in the same hole of the same grid. It selects the UEs that are allowed to transmit / receive simultaneously in the next time slot. To do this, it may take into account one or more criteria, including: traffic demand and / or minimizing the number of grids associated with the selected UEs (minimizing the number of different grids maximizes spectrum reuse and therefore maximizes capacity).

[0062] Then the PRB allocation block 51 allocates one or more PRBs first for each grid associated with the selected UEs then for each UE selected and associated with this grid while respecting the following rules and the constraints on the traffic demand: PRBs allocated to different grids must be different. UEs that are in the same hole of the same grid must have different PRBs.

[0063] For example, the process implemented in step 102 is the following to select the UEs authorized to transmit / receive simultaneously in the next time interval; a criterion of minimizing the number of different grids associated with the groups of selected UEs will for example also be taken into account (minimizing the number of different grids makes it possible to maximize the reuse of the spectrum and therefore to maximize the capacity): for each UE considered successively: - in a sub-step 102_1, the PRB allocation block 51 determines, as a function of the location obtained in step 101, among the grid holes remaining to be selected, the grid holes in which the UE is located (i.e. the grid hole corresponding to the portion of the Earth's surface in which the UE is located), and it selects, among these, the grid hole whose center is closest to the UE; in a sub-step 102_2, the hole thus selected in step 102_1 is deleted from the set of grid holes remaining to be selected; in a sub-step 102_3: the sub-steps 102_1 and 102_2 are iterated as long as PRBs remain to be allocated to the UE in question; then the steps sub-102_1 to 102_3 are then carried out for the next UE in question.

[0064] The association of a PRB or a set of exclusive PRBs to the grid is done for each grid according to the traffic demand associated with the UEs (those with the most demand) on each of the grids. Then, the PRB associated with the grid is allocated to each UE associated with the grid; or when a set of several PRBs has been associated with the grid, the allocation of the PRBs among the set of exclusive PRBs for each UE associated with the same grid is done according to the volume of the traffic demand of this UE: for example, considering the PRBs named PRB1, PRB2 and PRB3 associated with a grid, itself associated with three terminals UE1, UE2, UE3 corresponding to respective demand volumes V1, V2, V3, with V1 <V2<V3, PRB1 sera par exemple affecté à chacun de UE1, UE2, UE3, PRB2 sera affecté à UE2 et UE 3 et PRB3 sera affecté à UE3.

[0065] In one embodiment, the allocation of PRBs between grids is performed based on the location of the UE, its associated grid and an overview of the grid association and concurrently served UEs.

[0066] Finally, in a step 103, the allocation thus defined for the next exchanges is delivered by the control block 50 to the different elements of the network 10 needing to exploit this allocation: the gNB, the transmission and reception processing chains of the satellite 30, 60, the visible UEs.

[0067] Then steps 101 to 103 are repeated to determine the allocation of the next time interval (e.g. every T milliseconds).

[0068] Each grid associated with a PRB (or a set of PRBs) is thus associated with a group of UEs to which the PRB (or at least one PRB from the set of PRBs) has been allocated. In one embodiment, for each UE, the grid which is chosen as a priority is the one which includes the hole to which the UE is closest (for example the one which includes the hole whose center is located at the smallest distance from the UE).

[0069] In one embodiment, in step 102, if after the implementation of step 102 (1st pass), there remain PRBs to be allocated and UEs that have not received the necessary allocation, which means that the constraints imposed by the grid pattern can no longer be respected, step 102 is implemented again (2nd pass) for the PRBs still to be allocated relative to the UEs still to be served by considering another set of grid(s). Each grid of this other set of grid(s) has holes closer to each other than the grids of the set of grids 80 (worse C / I). Other passes can be implemented with other sets of grids that are successively more “tight” until the resources are exhausted or a C / I set as too degraded is reached.

[0070] In one (degraded) embodiment, there are points on the Earth's surface covered by the satellite platform that lie outside all holes in all grids. In such a case, the grid hole closest to one of these points will be assigned to said point.

[0071] In one embodiment, the location of each UE in step 101 is obtained by the location estimation block 52 based on one or more locations of the UE among: a location of the UE provided by the UE itself (for example by the GPS of the UE or example a location derived from the GNSS receiver of the UE); a location of the UE directly or indirectly provided by the gNB function in the ground processing device 14, (for example via information of the “Channel State Information” (CSI) type); a location of the UE determined according to an estimate of the direction of arrival (DOA), at the satellite antenna 12, of the signal coming from each UE.

[0072] The sounding signals (SRS) transmitted by the UEs and received by the satellite platform 11 are known: in accordance with the 5G NR standard, these signals are known in advance and personalized for each user. Regular measurements of these signals are carried out, in accordance with the 5G NR standard, at the level of the satellite reception chain. The application of an LMS ("Least Mean Square") type algorithm makes it possible to update the value of the beamforming weights associated with each radiating element of the antenna so as to maximize the signal-to-noise ratio of the received SRS signals. The newly obtained beamforming laws thus make it possible to estimate the direction of arrival of these same signals coming from the UE... This update is carried out according to the embodiments in the satellite platform 10 or in the ground processing device 14 (and in the latter case, the necessary data are exchanged on the Feeder 13).

[0073] In one embodiment of the invention, in step 101, the location estimation block 52 (which, as seen previously, is located in the satellite platform 11 or in the ground processing device 14 and then exchanges data with the satellite platform 11 via the Feeder 13) applies an LMS ("Least Mean Square") type algorithm to determine the value of the beamforming weights associated with each radiating element of the antenna maximizing the power of the sounding signals (received SRS signals). These newly obtained beamforming laws then make it possible to estimate the direction of arrival of these same signals coming from the UE: the estimation block 52 then applies a non-linear regression to find, from the weights recently determined, the direction of arrival (DOA) in terms of elevation and azimuth.

[0074] These aspects are now clarified.

[0075] Using satellite 11 and the center of the steered cell position, the beam direction is calculated and applied to the beam by the DBFN (Digital Beam Froming Network).

[0076] The center of the cellular direction vector is considered as the origin; the reference 80 indicates the center of the Earth.

[0077] The direction relative to the normal of the DRA panel of the antenna makes it possible to determine the phase laws to be applied to form the beam which will illuminate the cell.

[0078] In reference to the figures 1 , 2 And 3 we have: d UE → = P x UE − P x SAT , P y UE − P y SAT , P z UE − P z SAT

[0079] d SAT is the direction of the antenna normal to a center-earth reference frame, such as of EU.

[0080] As with beamforming, the important thing is direction d relative to the normal of the antenna, it is necessary to put the vector of the EU in the satellite antenna reference frame.

[0081] So, d = d UE - d SAT .

[0082] First, two angles from the antenna normal must be determined: α which is the angle between the antenna normal (z axis) and the projection of d on the plane (xy) called xy project either the azimuth; β which is the angle between the antenna normal (z axis) and the projection of d on the plane (yz) called YZ project or the elevation.

[0083] Using these two angles, it is then possible to calculate the direction angles of a rectangular antenna using the following algorithm. Let S be the distance between the center of consecutive radiating elements in the two axes of the antenna (x and y) Let N x RE the number of radiating elements in the x axis. Let N y RE the number of radiating elements in the y axis. Let λ be the wavelength of the signal to be formed.

[0084] The weights are calculated as follows: for k x = 0 to N x RE − 1 for k y = 0 to N y RE − 1 d x = k x . S . sin α d y = k y . S . sin β d = d x 2 + d y 2 W x , y = e − 2 πd . i / λ

[0085] This allows weights to be calculated from known angles.

[0086] This being said, let us now consider the reverse path, that is to say how to find α And β from the W x,y: this is where the regression mentioned above comes in. Let W x , y tg the set of weights for each radiating element for which we seek to find α tg< And β tg< . Regression allows us to find α tg< And β tg< which allows to minimize the squared error, that is to say to minimize ∑ x , y W x , y tg − W x , y 2 .

[0087] Regression can be performed using various algorithms well known to those skilled in the art, for example non-linear regression or the gradient algorithm. In all cases, the objective of these algorithms is to find α tg< And β tg< which minimizes the squared error. This allows us to determine the vector d , then knowing d SAT which is given by the satellite platform which controls the attitude of the satellite, of the EU is then determined based on d SAT And d .

[0088] The location estimation block 52 then determines the position of a UE as the intersection of the Earth's surface and the DOA thus estimated.

[0089] It should be noted that the evaluation of the DOA and / or location of a UE as described above based on the updated beamforming weights can be implemented independently of any use of DOA or location in the allocation of PRBs.

[0090] The invention has been described above in a particular embodiment, with particular reference to the 5G NR NTN standard. It is of course applicable to other technologies allocating, for example, OFDM time-frequency resources, OFDM-based FO, but also DVB RCS2, MF-TDMA, etc.

[0091] In the embodiment below, the allocation has been described with reference to a satellite. In another embodiment, the allocation according to the invention is implemented at the level of a satellite constellation comprising several satellites, in order to avoid interference also between telecommunications using different satellites of the constellation.

[0092] The described method can be implemented by executing software instructions on a processor. Alternatively, it can be implemented by dedicated hardware, typically a digital integrated circuit, either specific (ASIC) or based on programmable logic (e.g. FPGA / Field Programmable Gate Array).

[0093] Depending on the embodiments, the control block is located on the ground (for example in the device 14) or in the satellite platform 11.

[0094] For example, the prior art architecture of a transparent satellite is modified to include this RU function in place of the DTP (“Digital Transparent Processor”) in the satellite platform 11. The payload then becomes regenerative and the Feeder link 13 then transports the samples in the frequency domain in accordance with the functional split between the RU on board and the DU (“Distributed Unit”) on the ground. A good compromise is for example a split of type 7.3 on the downlink DL and a split of type 7.2 on the uplink UL, which makes it possible to limit the complexity on board while limiting the overheads on the Feeder 13.

[0095] Some types of satellites do not have digital beamformers: large satellites with MFPB (“Multi-Feed Per Beam”) output beamformers for example, or satellites with ABFN beamformers (Ka constellations for example). An approach proposed in the millimeter bands makes it possible to maintain a certain agility while limiting consumption by implementing ABFNs, shown schematically in Figure 7 . Overall, ABFNs allow a subset of radiating elements to be assembled into sub-panels delivering one or two beams. These sub-beams on each sub-panel are then assembled together by a DBFN allowing the number of beams to be multiplied. This hybrid approach makes it possible to implement the invention even on payloads that do not implement a pure DBFN.

[0096] In the case of a VHTS satellite solution based on MFPB, the invention makes it possible, for example, to shift the mesh of the beams to enable them to be refocused on the users to be served and to “plug” the areas between the beams which present roll-offs between 3 and 6 dB, depending on the layout: cf figure 8 , which compared to the grids of the FIG 5 corresponds to hexagon-shaped holes. On the figure 8 , we see the representation of two grids (one in solid line and the other in dotted line).

Claims

1. An allocation method for allocating time-frequency telecommunication resources in a wireless satellite telecommunication system (10), said satellite telecommunication system comprising a satellite (11), user terminals UE (20_1, 20_2), the satellite being suitable for simultaneously implementing multiple telecommunication beams (B_1, B_2) that are each associated, at a given time, with a separate user terminal UE (20_1, 20_2) and for dynamically re-centring each beam on the UE associated with said beam; each telecommunication beam (B_1, B_2) being established according to time-frequency resources selectively allocated to the associated UE (20_1, 20_2); said method comprising the following steps implemented by an electronic resource allocation unit (50): - obtaining the position of each UE (20_1, 20_2); and being characterised in that: - given a set (80) of grids (G1, G2) each representing at least a portion of the Earth's surface and each containing a plurality of non-contiguous areas distributed in the grid, the distance between any two areas of a grid being above a determined non-zero threshold associated with the grid, the allocation of the time-frequency resources relative to a time T is performed by applying at least the following rule, at least one and the same time-frequency resource being associated with the grid beforehand: said time-frequency resource associated with the grid may be allocated to each of 2 UEs (20_1, 20_2) only if the obtained positions of said 2 UEs are in separate areas from said plurality of non-contiguous areas distributed in the grid.

2. The time-frequency resource allocation method according to claim 1, wherein, if during a first step of allocating time-frequency resources using said set (80) of grids (G1, G2) that are each associated with a first determined threshold, it has not been possible to attribute time-frequency resources to at least one UE (20_1, 20_2) in accordance with the rule, an additional allocation step is performed using at least one set of additional grid(s) representing at least said portion of the Earth's surface and each containing a plurality of non-contiguous areas distributed in the grid, the distance between any two areas of a grid being above a second determined non-zero threshold associated with the grid, which is below the first threshold.

3. The time-frequency resource allocation method according to claims 1 or 2, wherein the allocation of the time-frequency resources is performed, relative to a time T, further by applying at least the following rule: given a first grid (G1), respectively a second grid (G2) separate from the first grid, from the set (80) of grids associated with a first time-frequency resource, respectively a second time-frequency resource separate from the first time-frequency resource: when a UE (20_1, 20_2) is determined as being located in both a first area from said plurality of non-contiguous areas distributed in the first grid and a second area from said plurality of non-contiguous areas distributed in the second grid, it is determined whether the UE is closer to the centre of the first area or to the centre of the second area; and if the UE is determined to be closer to the centre of the first area, respectively of the second area, the first time-frequency resource, respectively the second time-frequency resource, is allocated thereto.

4. The time-frequency resource allocation method according to one of the preceding claims, wherein the position of a UE (20_1, 20_2) is obtained by implementing the following steps: - calculating values of the beamforming weights maximising the power of predefined signals received from a UE; - estimating the direction of arrival, referred to as DOA, of a signal from a UE according to said beamforming weights determined for the UE by applying a regression algorithm linking beamforming weights and DOAs; - determining the position of the UE according to at least the intersection of the Earth's surface and the estimated DOA.

5. A computer program intended to be stored in the memory (70) of an electronic unit (50) for allocating resources in a wireless satellite telecommunication system (10) comprising a satellite (11), user terminals UE (20_1, 20_2), the satellite simultaneously implementing multiple telecommunication beams (B_ 1, B_2) that are each associated with a separate user (20_1, 20_2) and dynamically re-centring each beam on the associated UE; each telecommunication beam (B_1, B_2) being established according to time-frequency resources selectively allocated to the associated UE (20_1, 20_2); said time-frequency resource allocation unit (50) further comprising a microcomputer; said computer program comprising instructions that, when executed on the microcomputer, implement the steps of a method according to one of the preceding claims.

6. A device (50) for allocating time-frequency telecommunication resources for a wireless satellite telecommunication system (10), said satellite telecommunication system comprising a satellite (11), user terminals UE (20_1, 20_2), the satellite being suitable for simultaneously implementing multiple telecommunication beams (B_1, B_2) that are each associated, at a given time, with a separate user terminal UE (20_1, 20_2) and for dynamically re-centring each beam on the UE associated with said beam; each telecommunication beam (B_1, B_2) being established according to time-frequency resources selectively allocated to the associated UE (20_1, 20_2); said resource allocation device (50) being suitable for obtaining the position of each UE (20_1, 20_2) and being characterised in that: given a set (80) of grids (G1, G2) each representing at least a portion of the Earth's surface and each containing a plurality of non-contiguous areas distributed in the grid, the distance between any two areas of a grid being above a determined non-zero threshold associated with the grid, for performing the allocation of the time-frequency resources relative to a time T, by applying at least the following rule, at least one and the same time-frequency resource being associated with the grid beforehand: said time-frequency resource associated with the grid may be allocated to each of 2 UEs (20_1, 20_2) only if the obtained positions of said 2 UEs are in separate areas from said plurality of non-contiguous areas distributed in the grid.

7. The time-frequency resource allocation device (50) according to claim 6, suitable for, if during a first step of allocating time-frequency resources using said set (80) of grids (G1, G2) that are each associated with a first determined threshold, it has not been possible to attribute time-frequency resources to at least one UE (20_1, 20_2) in accordance with the rule, performing an additional allocation step using at least one set of additional grid(s) representing at least said portion of the Earth's surface and each containing a plurality of non-contiguous areas distributed in the grid, the distance between any two areas from said plurality of areas of a grid being above a second determined non-zero threshold associated with the grid, which is below the first threshold.

8. The time-frequency resource allocation device (50) according to claim 6 or 7, suitable for performing the allocation of the time-frequency resources relative to a time T, by further applying at least the following rule: given a first grid (G1), respectively a second grid (G2) separate from the first grid, from the set (80) of grids associated with a first time-frequency resource, respectively a second time-frequency resource separate from the first time-frequency resource: when a UE (20_1, 20_2) is determined as being located in both a first area from said plurality of non-contiguous areas distributed in the first grid and a second area from said plurality of non-contiguous areas distributed in the second grid, it is determined whether the UE is closer to the centre of the first area or to the centre of the second area; and if the UE is determined to be closer to the centre of the first area, respectively of the second area, the first time-frequency resource, respectively the second time-frequency resource, is allocated thereto.

9. The time-frequency resource allocation device (50) according to one of claims 6 to 8, suitable for obtaining the position of a UE (20_1, 20_2) by calculating values of the beamforming weights maximising the power of predefined signals received from a UE, and then by estimating the direction of arrival, referred to as DOA, of a signal from a UE according to said beamforming weights determined for the UE by applying a regression algorithm linking beamforming weights and DOAs and by determining the position of the UE according to at least the intersection of the Earth's surface and the estimated DOA.