Method for predicting the packet error probability of a radio link using deep neural networks
Patent Information
- Application Number
- EP2023792990
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-10-27
- Filing Date
- 2023-10-20
- Publication Date
- 2025-09-03
AI Technical Summary
Existing methods for predicting packet error probability in wireless communication networks are inefficient, requiring expensive simulation time and limited precision, especially when trained on small databases, and do not provide significant improvements in inference performance compared to traditional methods.
A method using a deep neural network architecture to determine calibration parameters from an SNR vector, processing them with a differentiable mathematical function to obtain a predicted equivalent signal-to-noise ratio, which is then used to calculate the packet error probability, allowing for improved inference performance even with a small database by learning from pairs of SNR vectors and associated equivalent signal-to-noise ratios.
This approach enables accurate prediction of packet error probability with zero or very low error for any SNR vector, enhancing learning and inference performance through the generalization capacity of neural networks, and can account for channel attenuation invariance, leading to better resource allocation and network simulation.
Smart Images
Figure 1.1
Abstract
Description
DESCRIPTION Title of the invention: Method for predicting the packet error probability of a radio link using deep neural networks Technical field
[0001] The present invention relates to a wireless telecommunications network and more particularly to a method for predicting the quality of data transmission on this telecommunications network in terms of packet error probability or equivalent SNR. Two applications are particularly targeted by the invention, namely resource allocation for allocating the best resources to the transmitter (power, modulation and coding scheme (MCS), etc.) and link abstraction for simulating the telecommunications network. Prior art
[0002] Figure 1 discloses a known method for predicting a PEP (for Packet Error Probability) noted ^̂, of a radio link between a transmitter and a receiver. This method comprises: - a step of receiving a vector of signal-to-noise ratios SNR (for "Signal to Noise Ratio" in English) ^ measured on the link. For example, in the case of OFDM (for "Orthogonal Frequency-Division Multiplexing") or frequency hopping systems, the SNRs are measured or calculated (depending on the intended application) on each of the dimensions, per carrier or on each level respectively, then the PEP of the link can be deduced. In the prior art of Figure 1, the method further comprises a step of processing said SNR vector ^ by an EESM (for "Exponential Effective SINR Mapping") function, said two-parameter EESM function being written as follows: −^ ^ ^^ ^ ^ ^ ^^ with ^ ^ and ^ ^calibration parameters, ^ the ith value ^ ur of ^, ^ ranging from 1 to ^, with ^ a non-zero natural integer representing the length of the SNR vector. This EESM function is notably disclosed in the article I. Latif, F. Kaltenberger and R. Knopp, "Link abstraction for multi-user MIMO in LTE using interference-aware receiver," 2012 IEEE Wireless Communications and Networking Conference (WCNC), 2012, pp.842-846, doi: 10.1109 / WCNC.2012.6214489, known as document Latif et al.
[0003] The processing by the EESM function makes it possible to estimate a scalar of signal-to-noise ratio called (or called) "equivalent" ^^ (translation of "effective SNR" in English). The predicted PEP ^̂ is then determined from a "Gaussian mapping" function which associates a PEP value with an SNR value in a Gaussian channel for a given modulation and coding scheme (MCS). This mapping function is constructed from a set of ^ ^points ^ = ^(^ ^ (^), ^ ^ (^)) for ^ = 1, … , ^ ^ ' obtained by simulation of the radio link and known as LUT (for " Table » in English), where ^ ^ (^) is the signal-to-noise ratio in the Gaussian channel and ^ ^ (^) the error probability measured for this SNR value for a given MCS. When the values of ^^ do not belong to all the points in the set ^, it is necessary to carry out an interpolation to calculate the PEP value. There are then two ways to carry out the mapping: 1) a segment-based interpolation approach, 2) a global interpolation approach on all the points. In case 1) we can for example use a linear interpolation, that is, if ^ ^ (^) ≤ ^^ < ^ ^ (^ + 1) then we calculate ^̂ = +^^ + , with + =-.(^ / ^)0-.(^)1.(^ / ^)01.(^)and , = ^ ^ ( ^ ) − +^ ^ ( ^) . In case 2) we construct an interpolation function 2 ^3456 (7), for example by a Lagrange polynomial interpolation passing through the points of the set ^. In this case, the estimated PEP is given by ^̂ = 2 ^3456 (^^). The resulting function 2 ^3456 (7) is differentiable, an assumption which will be necessary to describe a particular mode of learning an embodiment of the invention.
[0004] The prediction process of Figure 1 requires a model calibration phase to determine (^ ^ , ^ ^), calibration carried out from a database of ^8 pairs ^(^(^), ^(^)), ^ = 1, … , ^8' obtained by simulations or measurements, ^(^) being the packet error probability associated with the parameter vector ^(^) of the ^th element of the database. Obtaining this database is very expensive to obtain due to a significant simulation time. Finally, the precision of the predictions obtained is limited.
[0005] Figure 2 illustrates an alternative solution known in the state of the art using a deep neural network trained from a database (not shown). This deep neural network is adapted to receive as input the SNR vector ^ and to provide as output the predicted PEP ^̂. The use of a network of neurons in a method for predicting the performance of a radio link is notably disclosed in the publication E. Chu, J. Yoon, and BC Jung, “A Novel Link-to-System Mapping Technique Based on Machine Learning for 5G / IoT Wireless Networks”, Sensors 2019, 19, 1196. https: / / doi.org / 10.3390 / s19051196.
[0006] The prediction method in Figure 2 has an imprecise inference power if the deep neural network is trained on a frugal database, i.e. of small size (cardinal of the order of a few thousand) and does not provide any gain in terms of inference performance compared to the method without neural network as explained in paragraph [3].
[0007] There is therefore a need to propose a method for predicting a parameter representative of the quality of the radio link between the transmitter and the receiver which is simple and practical to implement and whose inference performance is better than the state-of-the-art solutions mentioned previously in the case of a frugal database.
[0008] Statement of the invention
[0009] The present invention aims to address this need. More particularly, the present invention aims to cover a method for predicting a predicted packet error probability of a radio link between a transmitter and a receiver in a wireless communication network for a plurality of modulation and coding schemes, said prediction method comprising: - a step of receiving an SNR vector; - a step of determining a vector of calibration parameters from said SNR vector, said step of determining the vector of calibration parameters being carried out from a deep neural network architecture; - a step of processing said calibration parameters by a differentiable mathematical function to obtain the predicted equivalent signal-to-noise ratio; - a Gaussian mapping step for determining the predicted packet error probability from the predicted equivalent signal-to-noise ratio.
[0010] In a particular embodiment, the Gaussian mapping is carried out by segment interpolation and said prediction method comprises: - a step of learning the weights of the deep neural networks from pairs comprising an SNR vector and an associated equivalent signal-to-noise ratio, said pairs being determined for the plurality of modulation and coding schemes, said pairs being previously stored in a learning database, the equivalent signal-to-noise ratios of the pairs being determined by an inversion of the Gaussian mapping, said Gaussian mapping being obtained by a simulation on a Gaussian channel for a given modulation and coding scheme.
[0011] In a particular embodiment, the Gaussian mapping is carried out by global interpolation and said prediction method comprises: - a step of learning the weights of the deep neural networks from pairs comprising an SNR vector and an associated packet error probability, said pairs being determined for the plurality of modulation and coding schemes, said pairs being previously stored in a learning database.
[0012] The method thus makes it possible to predict an equivalent SNR and then to deduce the PEP of a link from a vector of measured / estimated data, here contained in the SNR vector ^ which constitutes a characteristic of the link. In addition, it may be possible to take into account an invariance of the channel attenuation vectors of the radio link. The deep neural network no longer predicts the PEP directly as in the prior art but from calibration parameters. The conventional method of the prior art without a neural network predicts the PEP from the SNR vector through a function (for example EESM) parameterized by one or two coefficients, coefficients which are constant regardless of the SNR vector presented at the input of the device. Thus, since the two approaches of the prior art have the same inputs and outputs, the invention cannot in any way be deduced by merging the two approaches of the prior art.In the invention, the calibration parameters are generated by the neural network for each SNR vector presented at the input of the device. According to the inventors, what makes the implementation of this solution possible is a property that has never been disclosed in the prior art. Indeed, it can be shown that one can always find a set of calibration parameters that allows predicting the PEP without error for any SNR vector ^, so that the neural networks produce as many different calibration parameters as input SNRs, each with zero or very low prediction error. Thus, the proposed architecture that relies on this. property that has never been revealed in the prior art constitutes an undeniable source of progress. The proposed architecture makes it possible to converge towards very good learning, then leading to good inference performances thanks to the generalization capacity of neural networks when their weights are correctly trained. These calibration parameters are then intended to be applied to the mathematical function differentiable with respect to said parameters. This learning is carried out by a conventional neural network learning algorithm by giving itself a criterion to optimize between the output of the prediction system and the corresponding value in the database, for example the mean squared error. We will subsequently distinguish two learning modes depending on whether the Gaussian mapping function is carried out by segment interpolation (point 1 of paragraph [3]) or by global interpolation (point 2 of paragraph [3]).
[0013] In a particular embodiment, the SNR vector is determined from a step of processing a vector of channel attenuations of the radio link.
[0014] In a particular embodiment, the step of processing the channel attenuation vector comprises a processing step for taking into account a number of antennas at transmission, the transmitter having several antennas, and a number of antennas at reception, the receiver having several antennas, in order to obtain a signal-to-noise ratio vector.
[0015] In a particular embodiment, the step of processing the channel attenuation vector comprises a step of invariance processing by permutation of the signal-to-noise ratio vector to obtain the SNR vector.
[0016] In a particular embodiment, the permutation invariance processing is a sorting operation.
[0017] In a particular embodiment, the permutation invariance processing is a discretization operation.
[0018] In a particular embodiment, the permutation invariance processing is performed by other deep neural networks.
[0019] In a particular embodiment, the SNR vector is concatenated with an average signal-to-noise ratio.
[0020] In a particular embodiment, the SNR vector is concatenated with a 1 of 9 encoding vector in the case of 9 modulation and coding schemes (MCS).
[0021] In a particular embodiment, the differentiable mathematical function (:(;, ^)) corresponds to the following equation: ^ ^ ( ^ ) = −^ ^ ^ : ;, ^ ^ a first calibration parameter and ^ ^ a second the ith value of the SNR vector ^, ^ ranging from 1 to ^.
[0022] In a particular embodiment, the first calibration parameter ^ ^ is equal to 1 and said differentiable mathematical function (:(;, ^)) corresponds to the following equation: −^ ^ ^ ^ ^ 0<^
[0023] In a particular embodiment, the second calibration parameter ^ ^ is equal to 1 and said differentiable mathematical function (:(;, ^)) corresponds to the following equation: ^ ^^ ^
[0024] The present invention will be better understood upon reading the detailed description of embodiments taken as non-limiting examples and illustrated by the appended drawings in which:
[0025] [Fig 1] Figure 1 illustrates a method according to the prior art for predicting a PEP of a radio link between a transmitter and a receiver;
[0026] [Fig 2] Figure 2 illustrates another method according to the prior art for predicting a PEP of a radio link between a transmitter and a receiver using a deep neural network architecture;
[0027] [Fig 3] Figure 3 illustrates a method for predicting a PEP according to a first embodiment of the invention;
[0028] [Fig 4] Figure 4 illustrates more specifically a learning step of the prediction process of Figure 3 when the Gaussian mapping is carried out by segment interpolation;
[0029] [Fig 5] Figure 5 illustrates more specifically a learning step of the prediction process of Figure 3 when the Gaussian mapping is carried out by global interpolation;
[0030] [Fig 6] Figure 6 illustrates a method for predicting a packet error probability according to a second embodiment of the invention;
[0031] [Fig 7] Figure 7 illustrates a method for predicting a packet error probability according to a third embodiment of the invention;
[0032] [Fig 8] Figure 8 details a processing step of the prediction method of Figure 7;
[0033] [Fig 9] Figure 9 illustrates an operation performed by deep neural networks for permutation invariance processing in the prediction process of Figure 8;
[0034] In the following description, vectors are referred to in bold relative to scalars. Similarly, in the figures, vectors are referred to in bold and the associated lines have a thickness greater than the thickness of the lines associated with the scalars.
[0035] Figure 3 illustrates a method for predicting a packet error probability ^̂ according to a first embodiment of the invention.
[0036] In this first embodiment, the prediction method comprises: - a step of receiving an SNR vector ^; - a step of determining a vector; of calibration parameters from the SNR vector ^ using a deep neural network; - a step of processing the calibration parameters; by a differentiable mathematical function :(;, ^) to obtain the predicted equivalent signal-to-noise ratio ^^. - a step of determining the packet error probability ^̂ from the predicted equivalent signal-to-noise ratio ^^ using a Gaussian mapping function.
[0037] The SNR vector ^ is made up of discriminating parameters of the radio link. In the case of link prediction for systems experiencing multiple SNRs during packet transmission (carriers in OFDM, steps for frequency hopping, etc.), the input vector is necessarily made up of the SNRs. These SNR can be expressed in linear or dB, or even in general after undergoing any predistortion. The SNR values ^ used in the function :(;, ^) are not necessarily predistorted or are not necessarily predistorted in the same way as the SNRs at the input of the deep neural network. For example in the case of EESM, the SNR values used in the ^ ^ formula: ( ;, ^ ) = −^ ^ log ^ ^ ^ ^ ∑ ^^^ ^ ^^ are expressed linearly while the SNRs at the input of the deep neural network can be expressed in dB.
[0038] Figure 5 illustrates a particular embodiment in which information referenced = is concatenated. This information is, for example, an average signal-to-noise ratio and / or information of the modulation and coding scheme considered. A possible implementation to inform the network of the MCS is the use of a "one-hot" vector >, i.e. an encoding of 1 among 9, in the case of 9 MCS. This makes it possible to take into account several MCS with the same neural network architecture, i.e. with the same weights. The training database is thus constituted by the pairs (^, ^) of the different MCS considered. For example, for 9 = 3 MCS, we have > = @100B or > = @010B or > = @001B.
[0039] It would be possible to associate other information if necessary depending on the problem to be solved.
[0040] Alternatively, it would be possible to replace SNRs with Channel State Information (CSI) or Channel Quality Indicator (CQI) data.
[0041] The step of determining the vector of calibration parameters is carried out using a deep neural network architecture (DNN). By "deep neural network" we mean a neural network composed of at least three layers, including two hidden layers and one output layer. The number of neurons in each layer can be different.
[0042] The weights C of the different deep neural networks are determined by a learning algorithm A from a plurality of vectors ^ and the associated packet error probabilities previously stored in a training database BD and a Gaussian mapping. The database is consisting of a set of ^8 pairs ^(^(^), ^(^)), ^ = 1, … , ^8' obtained by simulations or measurements, ^(^) being the packet error probability associated with the parameter vector ^(^) of the ^th element of the database. Depending on the implementation of the Gaussian mapping, either by segment interpolation or by global interpolation, two different learning schemes are applied.
[0043] In the case of Gaussian mapping by segment interpolation, the learning scheme is illustrated in Figure 4. Since the Gaussian mapping function with segment interpolation is not differentiable, we cannot apply neural network learning algorithms by taking the PEP value at the output of the Gaussian mapping as the criterion to be optimized. To overcome this problem, we perform the learning by considering the equivalent SNR as the criterion to be optimized. It is necessary to create a new training database BD consisting of ^8 pairs ^(^(^), ^(^(^)), ^ = 1, … , ^8' where the ^(^ ( ^ ) ) are deduced from the ^(^) by inverting the Gaussian mapping, for example by linear interpolation.
[0044] The learning algorithm A is adapted to receive a plurality of predicted equivalent signal-to-noise ratios ^^ from the output of the prediction method.
[0045] The learning algorithm A will then modify the weights of the different deep neural networks so as to satisfy the equation below: − ^
[0047] The weights of deep neural networks are saved.
[0048] In the case of a Gaussian mapping by global interpolation, the learning scheme is illustrated in Figure 5. Since the Gaussian mapping is carried out by a differentiable function, we apply the neural network learning algorithms, taking as the criterion to be optimized the value of the PEP at the output of the Gaussian mapping.
[0049] The learning algorithm A is adapted to receive a plurality of predicted packet error probabilities ^̂ from the output of the prediction method.
[0050] The learning algorithm A will then modify the weights of the different deep neural networks so as to satisfy the equation below: − ^
[0052] The weights of deep neural networks are saved.
[0053] The vector ; of calibration parameters is processed by a differentiable mathematical function :(;, ^) to obtain the predicted equivalent signal-to-noise ratio ^^.
[0054] This differentiable mathematical function can, for example, be the EESM function which has two calibration parameters; = @^ ^ , ^ ^ B such that: ^ ^ ^ ^
[0056] It should be noted that any other differentiable SNR compression function can be used, such as, for example, a function approaching mutual information called the MIESM function known from the document Latif et al.
[0057] In the case of OFDM (regardless of the channel) or step-flat fading in the case of frequency hopping, the SNR vector ^ is determined from a processing step T of a vector of channel attenuations L of the radio link.
[0058] Figures 7 and 8 disclose such a processing step T. Thus, this step T comprises a processing step (PA) for taking into account a number of antennas at transmission ^4≥ 1 and a number of antennas at reception ^6≥ 1 in order to obtain a signal-to-noise ratio vector ^ N .
[0059] Thus the channel attenuation vector L can be written:
[0060] L = @O^^P(Q ^ ), O^^P(Q ^ ), … , O^^P(Q ^R )B
[0061] with ^ S the number of dimensions over which the equivalent SNR is evaluated, for example the number of subcarriers in OFDM or the number of steps in frequency hopping, Q ^a matrix (^6, ^4) of the channel attenuations for the dimension ^ considered (subcarrier or step), and T ^ (U, V) the element (U, V) of the matrix Q ^ .
[0062] In the case of a SISO mode (for "Single Input Single Output"), ^4= ^6= 1, we have the following relationships:
[0063] L = @T ^ (1, 1), T ^ (1, 1), … , T ^R (1, 1)B YZ[ ^th element of vector L.
[0065] In the case of a SIMO mode (for "Single Input Multiple Output"), where ^4> 1 and ^6= 1, we have for example for ^4= 2, the following relations:
[0066] L = @T ^ (1, 1), T ^ (2, 1), T ^ (1, 1), T ^ (2, 1), … , T ^R (1, 1), T ^R (2, 1)B
[0067] In option 1, in which there is no information about the receiver: ^ N YZ[ ^th element of vector L.
[0069] In option 2, where the receiver is known, for example an MRC (for “Maximum Ratio Combining”) type receiver: ^ YZ[
[0071] In the case of a MIMO mode (for "Multiple Input Multiple Output"), where ^4> 1 and ^6> 1, we have:
[0072] L = @T ^ (1,1), T ^ (1,2), … , T ^ (1, ^4), T ^ (2,1), … , T ^ (2, ^4), … , T ^ (^6, 1), … , T ^ (^6, ^4),
[0073] In option 1, in which there is no information about the receiver: ^ YZ[ ^th element of vector L.
[0075] In option 2, in which there is an optimal receiver:
[0076] ^ ^ N = a ^ ^ bc ∕ e f ^ with a ^ the largest value of Q ^ .
[0077] In the particular case of frequency-selective fading channels acting on each step in frequency hops, the components of the vector L are no longer scalars but vectors of the impulse responses of said channels. In this case the receiver carries out an equalization on each of the steps and the elements ^ ^ N of the vector ^' are then equal to the post-equalization SNRs of the component ℎ ^ calculated or measured according to the application.
[0078] In Figure 8, the signal-to-noise ratio ^ N is processed during a step TI of invariance processing by permutation of said signal-to-noise ratio vector ^', for obtain the SNR vector ^. This TI processing step allows us to take into account the permutation invariance of the input SNR vector. Indeed, the quality of the link does not depend on the order of the SNR values in the vector, i.e. ^ ( ^ ) = ^He( ^ ) I regardless of the permutation σ(). In addition, this TI processing step makes it possible to compensate for a relatively small size of the BD training database (a few thousand representatives).
[0079] In a particular embodiment, this processing step TI and a sorting step in which we have:
[0080] ^' = @ ^ ^ N , ^ ^ N , … , ^ ^ N B whose components are ordered according to the vectorization operations of the Q matrices ^ or according to the processing carried out with the knowledge of the receiver,
[0081] and ^ = @^ i N ^ , ^ i N ^ , … , ^ i N j B which verifies the order relation ^ i N ^ ≥ ^ i N ^ ≥ … ≥ ^ i N j with U ^ ≠ U l for ^ ≠ m and the set ^U^ , ^ = 1, … , ^' ∈ ^1, … , ^'.
[0082] Such a step of sorting the input data, here in ascending order, is simple to implement. It allows to give good results in inference on a frugal BD training database.
[0083] In a particular embodiment, the processing step TI is a discretization in which ^ is a vector of dimension (^ + 1) > ^.
[0084] We have the equation:
[0085] n = ∑ ^ ^ ^^ o ^
[0086] where o ^ for ^ ∈ ^1, … , ^' is determined according to the following equations:
[0087] We define ^ real values ^ O i , U = 1, … , ^ ' checking O ^ ≤ O ^ ≤ ⋯ ≤ O Y
[0088] for U = 2, … , ^, o ^ ( U ) = 1 if (O i − 1) ≤ ^ ′ ^ ˂ Oi , otherwise o ^ ( U ) = 0
[0089] o ^ (1) = 1 if ^ ^ N ˂ O ^ , otherwise o ^ (1) = 0
[0090] o ^ (^ + 1) = 1 if ^ ^ N > O ^ , otherwise o ^ (^ + 1) = 0
[0091] In a first option, ^ = min(n, 1), hence we have ^ elements at 1 and (^ − ^) elements at 0.
[0092] In a second option, ^ = n, then we have r ≤ ^ non-zero elements and (^ − r) elements at 0.
[0093] In an embodiment illustrated in Figure 9, the TI permutation invariance processing is an operation performed by deep neural networks called DeepSet. This processing consists of processing the vector ^' = @ ^ ^ N , ^ ^ N , … , ^ ^ N B by a deep neural network architecture DNN' which set of vectors ^n ^ , n ^ , … , n ^ ' whose components are added term by term to form the SNR vector ^. The scalar inputs ^ i N produce the vector outputs n i for U = 1, … , ^, through the same deep neural network architecture DNN' of the same weights.
[0094] The invention thus allows: - the introduction of a mathematical model associated with the DNN architecture for the prediction of radio link performance in terms of packet error probability and the associated architecture; - the introduction of the mathematical model allows to be much more precise because the DNN learns as many vectors; as input vectors ^ of the training database whereas in conventional methods a single vector is determined; for the entire database. It is thus possible to show that for certain mathematical models, there are vectors; of dimension two and dimension one allowing to predict exactly (i.e. without error) the equivalent SNR for any input vector ^; - a prediction of the performances of the radio links precise in inference; - to obtain good performances even on a frugal database.
[0095] The invention relates more particularly to operators and manufacturers of wireless communication systems.
[0096] The method for predicting the predicted packet error probability ^̂ can advantageously be implemented in a base transmission station of a wireless communication network.
Claims
CLAIMS 1. Method for predicting a predicted packet error probability (^̂) of a radio link (L R) between a transmitter and a receiver in a wireless communication network for a plurality of modulation and coding schemes (MCS), said prediction method comprising: - a step of receiving an SNR vector (^); - a step of determining a vector (;) of calibration parameters from said received SNR vector (^), said step of determining the vector (;) of calibration parameters being carried out from a deep neural network (DNN) architecture; - a step of processing said calibration parameters determined for the received SNR vector (^), by a differentiable mathematical function (:(;, ^)) to obtain the predicted equivalent signal-to-noise ratio (^^); - a Gaussian mapping step for determining the predicted packet error probability (^̂) from the predicted equivalent signal-to-noise ratio (^^). 2.
3. Prediction method according to claim 1 wherein the Gaussian mapping is carried out by segment interpolation and in that said prediction method comprises: - a step of learning the weights of the deep neural networks (DNN) from pairs comprising an SNR vector (^) and an associated equivalent signal-to-noise ratio (^), said pairs (^, ^) being determined for the plurality of modulation and coding schemes (MCS), said pairs (^, ^) being previously stored in a training database (BD), the equivalent signal-to-noise ratios (^) of the pairs (^, ^) being determined by an inversion of the Gaussian mapping, said Gaussian mapping being obtained by a simulation on a Gaussian channel for a given modulation and coding scheme.Prediction method according to claim 1 in which the Gaussian mapping is carried out by global interpolation and in that said prediction method comprises: - a step of learning the weights of the deep neural networks (DNN) from pairs comprising an SNR vector (^) and an associated packet error probability (^), said pairs (^, ^) being determined for the plurality of. modulation and coding schemes (MCS), said pairs (^, ^) being previously stored in a learning database (BD).
4. Prediction method according to any one of claims 1 to 3, in which the SNR vector (^) is determined from a processing step (T) of a channel attenuation vector (L) of the radio link (L R).
5. Prediction method according to claim 4, wherein the step (T) of processing the channel attenuation vector (L) comprises a processing step (PA) for taking into account a number of antennas at transmission (^4), the transmitter having several antennas, and a number of antennas at reception (^6), the receiver having several antennas, in order to obtain a signal-to-noise ratio vector (^').
6. Prediction method according to claim 5, wherein the step (T) of processing the channel attenuation vector (L) comprises a step (TI) of invariance processing by permutation of the signal-to-noise ratio vector (^') to obtain the SNR vector (^).
7. Prediction method according to claim 6, wherein the invariance processing by permutation (TI) is a sorting operation.
8. A prediction method according to claim 6, wherein the permutation invariance (TI) processing is a discretization operation. 9.A prediction method according to claim 6, wherein the permutation invariance (TI) processing is performed by other deep neural networks (DNN').
10. A prediction method according to any one of claims 1 to 9, wherein the SNR vector (^) is concatenated with an average signal-to-noise ratio ( ).
11. A prediction method according to any one of claims 1 to 10, wherein the SNR vector (S) is concatenated with an encoding vector (>) 1 out of 9 in the case of 9 modulation and coding schemes (MCS).
12. A prediction method according to any one of claims 1 to 11, wherein the differentiable mathematical function (:(;, ^)) corresponds to the following equation:. ^ ^ with ^ ^ a first calibration parameter and ^ ^ a second calibration parameter, ^ ^the ^th value of the SNR vector (^), ^ ranging from 1 to ^ and N designating the number of SNR vectors(^).
13. Prediction method according to claim 12, in which the first calibration parameter c1 is equal to 1 and said differentiable mathematical function (:(;, ^)) corresponds to the following equation: :(;, ^) = −^ ^ log ^ ^ 0< ^ ∑ ^^^ ^ ^ 14. Prediction method according to claim 12, in which the second calibration parameter ^ ^ is equal to 1 and said differentiable mathematical function (:(;, ^)) corresponds to the following equation: ^ ^ ^ ^