Method and apparatus for radio communication

Abstract

Methods and apparatus for radio communication are provided. The method comprises: monitoring at least one downlink channel (DLCH); determining at least one downlink channel quality (Q # 1, Q # 2) associated with the monitored at least one downlink channel (DLCH); determining at least one future downlink channel quality indicator (fQ # 1, fQ # 2) characterizing an estimated future quality associated with the downlink channel (DLCH) based at least on the at least one monitored downlink channel quality (Q # 1, Q # 2); determining at least one estimated quality indicator (e # 1, e # 2) characterizing an estimated quality of the at least one associated future downlink channel indicator (fQ # 1, fQ # 2) based at least on the at least one monitored downlink channel quality (Q # 1, Q # 2); transmitting the at least one estimated quality indicator (e # 1, e # 2); and transmitting the at least one future downlink channel quality indicator (fQ # 1, fQ # 2).

Description

Technical Field

[0001] The present invention relates to a method and an arrangement for radio communications. Background Art

[0002] Modern communication systems adapt the used modulation and coding schemes to the channel conditions in order to optimize the spectral efficiency of the system.

[0003] Since the 3GPP framework uses a multi-carrier waveform, namely OFDM, the modulation and coding scheme used can also be optimized for resource blocks. As a result, accurate CSI makes efficient use of the channel. In NR, downlink measurements are periodically reported by the UE to the BS. The measurement report typically includes a channel quality indicator (CQI), a precoding matrix indicator (PMI), and a rank indicator (RI).

[0004] In addition, non-terrestrial networks (NTNs) using satellite-based gNBs are envisioned. Summary of the invention

[0005] One aspect of the specification is directed to the following subject matter: a method, in particular according to an estimation support mode, comprising: monitoring at least one downlink channel; determining at least one downlink channel quality associated with the at least one monitored downlink channel; determining at least one future downlink channel quality indicator characterizing an estimated future quality associated with the downlink channel based at least on the at least one monitored downlink channel quality; determining at least one estimated quality indicator characterizing an estimated quality of at least one associated future downlink channel indicator based at least on the at least one monitored downlink channel quality; transmitting the at least one estimated quality indicator; and transmitting the at least one future downlink channel quality indicator.

[0006] Advantageously, for example, a radio terminal is enabled to support a radio access node with an expected future quality associated with a downlink channel that the radio terminal is expected to experience at a future point in time. On the one hand, privacy and security benefit because only an abstract quality associated with the downlink channel is transmitted. On the other hand, spectral efficiency is increased because the radio access node is enabled to adapt its downlink transmission parameters so that the radio terminal can quickly receive downlink data with a reduced probability of duplication.

[0007] The estimation quality indicator advantageously provides the radio access network with an indication of the prediction accuracy. Thus, the radio access network can determine appropriate transmission parameters such as MCS, transmission power, subchannels, etc. The UE can help the BS to adapt, for example, the modulation and coding scheme based on the reported information so as not to violate the channel capacity. An erroneous prediction report may even degrade the system performance.

[0008] An advantageous example is characterized by comprising receiving a support configuration indicating at least one support mode, in particular an estimation support mode.

[0009] Advantageously, the supporting configuration enables both the radio terminal and the radio access node to react flexibly to environmental conditions affecting the radio domain.

[0010] A beneficial example is characterized in that the method further comprises: determining at least one future propagation delay associated with the at least one downlink channel; determining at least one estimation range based on the determined at least one future propagation delay, the estimation range characterizing the time period during which the at least one future downlink channel quality indicator and its associated estimated quality indicator or the at least one supporting configuration is valid.

[0011] For example, benefits may be achieved by supporting a configuration indicating an auxiliary support mode, wherein, according to the auxiliary support mode, the method comprises: monitoring the at least one downlink channel; determining at least one downlink channel quality associated with the at least one monitored downlink channel; and transmitting the at least one downlink channel quality.

[0012] Advantageously, transmitting the downlink channel quality provides a legacy mode for UEs that are not able to determine a downlink quality prediction.

[0013] Another example is characterized in that the at least one estimated quality indicator is valid for an indeterminate period of time, at least when it is transmitted, in particular for a period of time during which the RRC state is active.

[0014] Advantageously, the validity of the estimated quality is transmitted once at the beginning and is valid in the absence of a contrary signal from the radio terminal.

[0015] Another example is characterized in that the method further includes: determining at least one additional future downlink channel quality indicator characterizing the estimated future quality associated with the downlink channel based at least on the at least one monitored downlink channel quality; determining at least one additional estimated quality indicator characterizing the estimated quality of the at least one associated future downlink channel indicator based at least on the at least one monitored downlink channel quality; comparing the at least one additional estimated quality indicator with at least one previously determined estimated quality indicator; and transmitting the at least one additional estimated quality indicator based on the comparison.

[0016] Advantageously, the further estimation quality indicator enables signaling of a change in the estimation quality. Advantageously, the further estimation quality indicator replaces a previously determined estimation quality indicator. For example, the further estimation quality indicator is transmitted based on a sudden change of the further estimation quality indicator.

[0017] The benefits are achieved by the method further comprising: determining a first potential code based on the at least one downlink channel quality via a machine-trained downlink embedding function; determining a second potential code based on self-sensor based data via a machine-trained auxiliary data embedding function; and determining the at least one future downlink channel indicator and the at least one estimated quality indicator based on at least the first potential code and based on the second potential code via a machine-trained prediction function.

[0018] When it comes to the question of downlink CSI, the UE naturally has more information than the gNB since the measurements are made in the UE. Furthermore, the UE has access to additional self-sensor information (such as cameras, route information, etc.) which may have stricter (e.g. privacy) constraints and cannot be shared with the BS. Therefore, the UE will be better qualified to provide a prediction of the channel quality. However, the UE's prediction will be used to help the BS choose its modulation scheme. Therefore, an accuracy measure attached to the prediction is needed to not trigger an adverse decision by the gNB, which in turn results in worse performance than without predictive CQI. In other words, the provided machine learning architecture is able to: fuse self-sensor information to improve downlink quality, such as CQI in radio communication networks.

[0019] Since NTN can be considered as an extension of the traditional TN - terrestrial network - in certain areas (underground, tunnels, bridges, etc.), the coverage of NTN may suddenly degrade. These occasional events depend on the geographical location of the UE and its relative position to the satellite. Therefore, probabilistic channel models as used in the literature will not be sufficient for designing appropriate predictors. Therefore, data-driven methods do not require closed-form channel models, but instead can be trained to realistic propagation scenarios.

[0020] Another example is characterized in that the supporting configuration includes at least one estimation range, which characterizes the time period during which the at least one future downlink channel quality indicator and its associated estimated quality indicator or the at least one supporting configuration is valid, and wherein the at least one future downlink channel quality indicator and its associated estimated quality indicator are determined based on the at least one estimation range.

[0021] The benefits are achieved by the method further comprising: receiving an estimated capability request; after receiving the estimated capability request, determining an estimated capability response that in particular at least indicates at least one supported mode, in particular an estimated support mode; transmitting the estimated capability response; and wherein receiving a support configuration as a reply to the transmitted estimated capability response.

[0022] According to the assistance support mode, the method further comprises the following to achieve the benefit: determining assistance information characterizing at least one present or future condition of the downlink channel receiving device or a physical entity to which the receiving device is associated or installed; and transmitting the assistance information.

[0023] Advantageously, the assistance information provides an alternative to the transmission of future downlink quality indicators and their associated estimated quality indicators.The radio access node is enabled to internally determine future downlink channel quality indicators, in particular if the radio terminal is not enabled or configured to determine previous indicators.

[0024] For example, a benefit may be achieved by the at least one further estimated quality indicator indicating a difference relative to at least one estimated quality indicator of a previous transmission.

[0025] Advantageously, by transmitting the difference, a few bits are sufficient to adapt the estimation quality on the gNB / radio access node side and increase the radio efficiency for submitting control information.

[0026] Furthermore, an example is advantageous in that the at least one estimated quality indicator and the associated future downlink channel quality indicator are transmitted in pairs.

[0027] Advantageously, the radio access node always has an indication of the estimated quality.Thus, sudden changes in the estimated quality are received together with the estimated downlink channel quality and enable an immediate adaptation of the downlink channel.

[0028] A beneficial example is characterized in that the method further includes: observing at least one parameter associated with the sidelink channel; determining a third potential code based on at least one parameter associated with the sidelink channel via a machine-trained sidelink embedding function; and wherein the at least one future downlink channel indicator and the at least one estimated quality indicator are further determined based on the at least one third potential code.

[0029] Utilizing the sidelink information has the following benefits: the sidelink acts as a "further view" of the radio terminal's local environment.

[0030] One aspect of the present specification is directed to the following subject matter: an apparatus, in particular a radio terminal or user equipment or a module thereof, in particular according to an estimation support mode, comprising: a monitoring component configured to monitor at least one downlink channel; a determination component configured to determine at least one downlink channel quality associated with the at least one monitored downlink channel; a determination device configured to determine at least one future downlink channel quality indicator characterizing an estimated future quality associated with the downlink channel based at least on at least one monitored downlink channel quality; a determination component configured to determine at least one estimated quality indicator characterizing an estimated quality of at least one associated future downlink channel indicator based at least on the at least one monitored downlink channel quality; a transmission component configured to transmit the at least one estimated quality indicator; and a transmission component configured to transmit the at least one future downlink channel quality indicator.

[0031] One aspect of the specification is directed to the following subject matter: a method, in particular according to an estimation support mode, comprising: receiving at least one future downlink channel quality indicator, which characterizes an estimated future quality associated with a downlink channel; receiving at least one estimated quality indicator, which characterizes the estimated quality of the at least one associated future downlink channel indicator; determining at least one transmission parameter based at least on the received at least one future downlink channel quality indicator and based on the at least one estimated quality indicator; and transmitting downlink data via the downlink channel based on the determined at least one transmission parameter.

[0032] Furthermore, an example is advantageously provided that it comprises transmitting a support configuration indicating at least one supported mode, in particular an estimated supported mode.

[0033] Advantageously, the supporting configuration enables radio terminals and radio access nodes to react flexibly to environmental conditions affecting the radio domain.

[0034] In addition, an example is advantageous in that a configuration indicating an auxiliary support mode is supported, wherein, according to the auxiliary support mode, the method includes: receiving at least one downlink channel quality; receiving auxiliary information characterizing at least one current or future condition of a downlink channel receiving device; determining additional transmission parameters based on the at least one downlink channel quality and based on the at least one auxiliary information; and transmitting downlink data via the downlink channel based on the additional transmission parameters.

[0035] Advantageously, the assistance information provides an alternative to the transmission of future downlink quality indicators and their associated estimated quality indicators. In particular, if the radio terminal is not enabled or configured to determine a previous indicator, the radio access node is enabled to internally determine the future downlink channel quality indicator and then determine the further transmission parameters.

[0036] Furthermore, an example is advantageous in that the at least one estimated quality indicator is valid for an indeterminate period of time, at least upon its reception, in particular for a period of time during which the RRC state exists.

[0037] Advantageously, the validity of the estimated quality is transmitted once at the beginning and is valid in the absence of a contrary signal from the radio terminal.

[0038] The benefit is achieved by the method further comprising: determining, based at least on the at least one monitored downlink channel quality, at least one additional future downlink channel quality indicator characterizing an estimated future quality associated with a downlink channel; determining, based at least on the at least one monitored downlink channel quality, at least one additional estimated quality indicator characterizing an estimated quality of the at least one associated future downlink channel indicator; comparing the at least one additional estimated quality indicator to at least one previously determined estimated quality indicator; and transmitting the at least one additional estimated quality indicator based on the comparison.

[0039] A beneficial example is characterized by receiving at least one further estimated quality indicator indicating a difference relative to at least one previously received estimated quality indicator; and determining a further transmission parameter based at least on the at least one previously received estimated quality indicator and on the at least one received further estimated quality indicator; and transmitting another downlink data via the downlink channel according to the determined further transmission parameter.

[0040] Advantageously, by transmitting the difference, a few bits are sufficient to adapt the estimation quality on the gNB / radio access node side and increase the radio efficiency for submitting control information.

[0041] Advantageously, the further estimation quality indicator enables signaling of a change in the estimation quality. Advantageously, the further estimation quality indicator replaces a previously determined estimation quality indicator. For example, the further estimation quality indicator is transmitted based on a sudden change of the further estimation quality indicator.

[0042] Benefits are achieved by the supporting configuration comprising at least one estimation range, the estimation range characterizing the at least one future downlink channel quality indicator and its associated estimation quality indicator or at least one time period over which the supporting configuration is valid, and wherein the at least one future downlink channel quality indicator and its associated estimation quality indicator are determined based on the at least one estimation range.

[0043] A beneficial example is characterized in that the method further comprises: determining an estimation capability request based on the current need to support downlink channel parameter adaptation; transmitting the estimation capability request; after transmitting the estimation capability request, receiving an estimation capability response which in particular at least indicates at least one supported mode, in particular an estimation support mode; wherein the support configuration is determined and transmitted as a reply to the transmitted estimation capability response.

[0044] One aspect of the present specification is directed to the following subject matter: an apparatus, in particular a radio access node of a radio access network or a module thereof, in particular according to an estimation support mode, comprising: a receiving component configured to receive at least one future downlink channel quality indicator characterizing an estimated future quality associated with a downlink channel; a receiving component configured to receive at least one estimated quality indicator characterizing an estimated quality of the at least one associated future downlink channel indicator; a determining component configured to determine at least one transmission parameter based at least on the received at least one future downlink channel quality indicator and on the at least one estimated quality indicator; and a transmitting component configured to transmit downlink data via a downlink channel according to the determined at least one transmission parameter.

[0045] One aspect of the specification is directed to the following subject matter: a method comprising: transmitting first information characterizing at least one quality associated with a downlink channel received by a radio terminal; transmitting second information characterizing a current or future environmental condition of a physical entity associated with the radio terminal.

[0046] Advantageously, to perform the prediction, the processing is shifted towards the gNB. The provided machine learning architecture is able to perform predictions of, for example, CQI on the radio access network side.

[0047] The benefits are achieved by the first information transmitted being a first potential code, wherein the method further comprises: determining at least one measured downlink channel quality from observing a downlink channel; and determining the first potential code based on the at least one measured downlink channel quality via a machine-trained downlink embedding function.

[0048] Advantageously, the first latent code can include more relevant information as input for further training of the function with less data usage. Thus, not only the prediction quality at the gNB side is increased, but also the spectral efficiency is increased.

[0049] Benefits are achieved by the transmitted second information being a second latent code, wherein the method further comprises: determining self-sensor based data based on at least one sensor measurement associated with the physical entity; and determining the second latent code via a machine trained auxiliary information embedding function.

[0050] Advantageously, the second latent code can include more relevant information as input for further training of the function with less data usage. Thus, not only the prediction quality at the gNB side is increased, but also the spectral efficiency is increased.

[0051] Another example is characterized in that the method further comprises: transmitting third information characterizing at least one parameter associated with the sidelink channel.

[0052] Furthermore, an example is advantageous in that the transmitted first information is at least one measured downlink channel quality, wherein the method further comprises determining the at least one measured downlink channel quality from observing the downlink channel.

[0053] Advantageously, the first information may be QCI, 5QI or another parameter.

[0054] Another example is characterized in that the transmitted second information is ego sensor based data, wherein the method further comprises determining the ego sensor based data based on at least one sensor measurement of a sensor associated with the physical entity.

[0055] Advantageously, the processing work is shifted towards the gNB. Therefore, energy efficiency is increased on the side of the radio terminals which may be mobile.

[0056] A beneficial example is characterized in that the transmitted third information is a third latent code, wherein the method further includes: observing at least one parameter associated with the sidelink channel; and determining the third latent code based on the at least one parameter associated with the sidelink channel via a machine-trained sidelink embedding function.

[0057] Advantageously, the third latent code can include more relevant information as input for further training of the function with less data usage at the gNB. Thus, not only the prediction quality at the gNB side is increased, but also the spectral efficiency is increased.

[0058] The method further comprises achieving a benefit by observing at least one parameter associated with a sidelink channel, wherein the transmitted third information is at least one sidelink parameter associated with the sidelink channel.

[0059] Advantageously, the processing work is shifted towards the gNB. Therefore, energy efficiency is increased on the side of the radio terminals which may be mobile.

[0060] One aspect of the present specification is directed to the following subject matter: an apparatus, in particular a radio terminal or user equipment or a module thereof, comprising: a transmission component for transmitting first information characterizing at least one quality associated with a downlink channel received by the radio terminal; a transmission component for transmitting second information characterizing current or future environmental conditions of a physical entity associated with the radio terminal.

[0061] One aspect of the present specification is directed to the following subject matter: a method comprising: receiving first information characterizing at least one quality associated with a downlink channel received by a device; receiving second information characterizing current or future environmental conditions of a physical entity associated with the device; determining at least one downlink transmission parameter based on the at least one first information and based on the at least one second information via at least one machine training function; and transmitting downlink data via a downlink channel based on the at least one determined downlink transmission parameter.

[0062] Advantageously, the downlink channel transmission is adapted according to environmental conditions on the UE side.

[0063] A beneficial example is characterized in that the received first information is a first potential code and in that the second information is a second potential code, wherein the method further comprises: determining at least one downlink transmission parameter based on the received first potential code and based on the received second potential code via a machine trained predictor function.

[0064] Advantageously, the second latent code can include more relevant information as input for further training of the function with less data usage. Thus, not only the prediction quality at the gNB side is increased, but also the spectral efficiency is increased.

[0065] The benefit is achieved by the received first information being at least one downlink channel quality, and wherein the second information is a second potential code, wherein the method further comprises: determining, via a machine-trained downlink embedding function, at least one first potential code based on the received downlink quality; determining, via a machine-trained predictor function, the at least one downlink transmission parameter based on the determined first potential code and based on the received second potential code.

[0066] Advantageously, the first information may be QCI, 5QI or another parameter.

[0067] Another example is characterized in that the received first information is at least one first potential code, and wherein the second information is based on self-sensor data, wherein the method further includes: determining at least one second potential code based on the received self-sensor based data via a machine-trained auxiliary data embedding function; determining the at least one downlink transmission parameter based on the determined first potential code and based on the received second potential code via a machine-trained predictor function.

[0068] Advantageously, the processing work is shifted towards the gNB. Therefore, energy efficiency is increased on the side of the radio terminals which may be mobile.

[0069] An advantageous example is characterized in that the method further comprises receiving third information characterizing at least one parameter associated with the sidelink channel.

[0070] For example, benefits may be achieved by the method further comprising: determining, via the at least one machine training function, at least based on the first information and at least based on the second information, at least one future downlink channel quality indicator characterizing an estimated future quality associated with the downlink channel, and at least one estimated quality indicator characterizing an estimated quality of at least one associated future downlink channel indicator; and determining the at least one transmission parameter based at least on the at least one future downlink channel quality indicator and on the at least one estimated quality indicator.

[0071] Another example is characterized in that the received third information is a third potential code, wherein the method further comprises: determining the at least one downlink transmission parameter based on at least one third potential code via the at least one machine training function.

[0072] Advantageously, the third latent code can include more relevant information as input for further training of the function with less data usage at the gNB. Thus, not only the prediction quality at the gNB side is increased, but also the spectral efficiency is increased.

[0073] The benefit is achieved by the received third information being at least one sidelink parameter associated with a sidelink channel, wherein the method further comprises determining at least one third potential code via a sidelink embedded machine training function, and determining the at least one downlink transmission parameter based at least on the third potential code via the at least one machine training function.

[0074] Advantageously, the processing work is shifted towards the gNB. Hence, energy efficiency is increased on the radio terminal side. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A schematic sequence diagram is depicted.

[0076] Figure 2 Depicted Figure 1 Configuration section CP.

[0077] Figure 3 Signaling for downlink transmission is depicted.

[0078] Figure 4 A schematic block diagram of a machine trained function or estimator 400 is depicted.

[0079] Figure 5 A schematic sequence diagram is depicted.

[0080] Figure 6 Depicted is the reporting of a sudden difference in reliability compared to the previously valid reliability signaled via e#1.

[0081] Figure 7 Depicted from Figure 1 ASM auxiliary support mode.

[0082] Figure 8 A non-terrestrial cellular communication network CN#1 is depicted.

[0083] Fig. 9 A terrestrial cellular radio communications network CN#2 is depicted.

[0084] Fig.10 The schematic sequence diagram depicts Figure 1 Another example of auxiliary support mode ASM.

[0085] Fig.11 Depicted Figure 4 An example of a machine trained function 400.

[0086] Fig.12 Depicted is an example of the distribution of different functional entities between UE#1 and gNB#1.

[0087] Fig.13 Another example of the distribution of different functional entities is depicted.

[0088] Fig.14 Another example of a function distribution is depicted.

[0089] Fig.15 Another device UE#2 is depicted transmitting SL#1 to gNB#1.

[0090] Fig.16 Another example of a function distribution is depicted.

[0091] Fig.17 It is depicted that another device UE#2 transmits the latent code LC#C instead of UE#1.

[0092] Fig.18 Function 400 depicting machine training is arranged at device gNB#1.

[0093] Fig.19 Several devices UE#1, UE#2, UE#3 are depicted transmitting their respective measured downlink channel qualities Q#1, Q#2, Q#2 together with auxiliary information in the form of respective ego sensor based data esD#1, esD#2, esD#3. DETAILED DESCRIPTION

[0094] In the following, reference is made to the figures.

[0095] Figure 1 A schematic sequence diagram is depicted. The device gNB#1, in particular a radio access node or gNB, transmits 202 a support configuration sc towards the device UE#1, in particular a radio terminal or user equipment UE. The UE#1 receives 102 a support configuration sc indicating a support mode, in particular an estimation support mode ESM.

[0096] An example relates to a method, in particular according to an estimation support mode ESM, the method comprising: monitoring 106 a downlink channel DLCH; determining 108 downlink channel qualities Q#1, Q#2 associated with the monitored downlink channel DLCH; determining 110 future downlink channel quality indicators fQ#1, fQ#2 characterizing estimated future qualities associated with the downlink channel DLCH based at least on the monitored downlink channel qualities Q#1, Q#2; determining 112 estimated quality indicators e#1, e#2 characterizing estimated qualities of associated future downlink channel indicators fQ#1, fQ#2 based at least on the monitored downlink channel qualities Q#1, Q#2; transmitting 114; 164 the estimated quality indicators e#1, e#2; and transmitting 116; 166 the future downlink channel quality indicators fQ#1, fQ#2.

[0097] The estimation support mode ESM is intended for supporting a radio access node in determining transmission parameters of transmissions on a downlink channel DLCH. In the estimation support mode ESM, the determination and transmission of the estimation quality indicators e#1..n is performed at least once after receiving a support configuration sc.

[0098] It is depicted that the method further comprises: determining 150 a future propagation delay associated with the downlink channel; determining 152 an estimation range EH based on the determined future propagation delay, the estimation range EH characterizing a time period for which a future downlink channel quality indicator fQ#1 and its associated estimated quality indicator e#1 or a support configuration sc is valid.

[0099] It is assumed that there is sufficient knowledge of the relative distance on the UE side. The UE can calculate the future propagation delay and estimate the range EH by itself. In this case, the corresponding field in the support configuration sc remains empty or is ignored.

[0100] According to one example, the future propagation delay is determined based on at least one of the current location of UE#1, the heading of UE#1, the speed of UE#1, and the future location of gNB#1.

[0101] The diagram illustrates a support configuration sc indicating an auxiliary support mode ASM, wherein according to the auxiliary support mode ASM the method comprises: monitoring 106 a downlink channel DLCH; determining 110 a downlink channel quality Q#3 associated with the monitored downlink channel DLCH; and transmitting 132 the downlink channel quality Q#3.

[0102] It is shown that according to the assistance support mode ASM the method further comprises determining 130 assistance information AS characterizing a present or future condition of the downlink channel receiving device RT or a physical entity to which the receiving device RT is associated or installed and transmitting 134 the assistance information AS.

[0103] According to an example, the estimated quality indicator e#1, e#2 is valid for an indeterminate period of time, in particular for a period of time during which the RRC state is active, at least when it is transmitted. In this case, the estimated quality indicator e#1, e#2 is valid during a period of time associated with a session according to which the radio terminal is connected to the radio access node. Upon receiving a new fQ#2, gNB#1 replaces or updates its predicted valid value in the sense of fQ#2.

[0104] The method is illustrated as further comprising: determining 160 an additional future downlink channel quality indicator fQ#2 representing the estimated future quality associated with the downlink channel DLCH based at least on the monitored downlink channel quality Q#2; determining 160 an additional estimated quality indicator e#2 representing the estimated quality of the associated future downlink channel indicators fQ#1, fQ#2 based at least on the monitored downlink channel quality Q#2; comparing 162 the additional estimated quality indicator e#2 with the previously determined estimated quality indicator e#1; and transmitting 164 the additional estimated quality indicator e#2 based on the comparison.

[0105] According to an example, the further estimated quality indicator e#2 indicates a difference with respect to the previously transmitted estimated quality indicator e#1. In other words, an update of the estimated quality is performed on the gNB#1 side in order to appropriately determine the next downlink transmission parameter. For example, at the beginning of a connection establishment or during a connection, e#1 is transmitted as a starting value. gNB#1 provides a first lookup table to determine the MSE and / or CE based on the received e#1 as a function of the downlink transmission parameter p#2.

[0106] According to another example, e#2 represents a pointer to a second lookup table. gNB#1 provides the second lookup table to determine a positive or negative difference relative to the associated e#1.

[0107] For example, if the comparison indicates that the difference or change in the estimated quality indicator e#2 is significant, the estimated indicator e#2 is transmitted. For example, a threshold operation determines whether to transmit e#2. In another example, a predefined or trainable function takes both e#1 and e#2 as input and outputs a transmission decision.

[0108] Steps 106-112 are referenced with reference numeral 160. After the determination 160, at least one further future downlink channel quality indicator fQ#2 is transmitted 166 towards the device gNB#1. As indicated by the dashed line, the transmission 164 of the further estimated quality indicator e#2 is optional and may be omitted.

[0109] In other words, the estimated future quality associated with the downlink channel is the quality that the UE is expected to experience at a future point in time or during a future time span.For example, the future downlink channel quality indicator represents the determination accuracy of the associated future downlink channel quality indicator.

[0110] An apparatus UE#1, in particular a radio terminal or user equipment or a module thereof, in particular according to an estimation support mode ESM, is shown, the apparatus comprising: a monitoring component configured to monitor 106 a downlink channel DLCH; a determining component configured to determine 108 a downlink channel quality Q#1, Q#2 associated with the monitored downlink channel DLCH; a determining component configured to determine 110 a future downlink channel quality Q#1, Q#2 characterizing an estimated future quality associated with the downlink channel DLCH based at least on the monitored downlink channel qualities Q#1, Q#2; Channel quality indicators fQ#1, fQ#2; a determination component configured to determine 112 estimated quality indicators e#1, e#2 based at least on the monitored downlink channel qualities Q#1, Q#2, the estimated quality indicators e#1, e#2 characterizing the estimated quality of the associated future downlink channel indicators fQ#1, fQ#2; a transmission component configured to transmit 114; 164 the estimated quality indicators e#1, e#2; and a transmission component configured to transmit 116; 166 the future downlink channel quality indicators fQ#1, fQ#2.

[0111] Also shown is a method for operating gNB#1, in particular according to the estimation support mode ESM, the method comprising: receiving 216 future downlink channel quality indicators fQ#1, fQ#2, which characterize the estimated future quality associated with the downlink channel DLCH; receiving 214 estimated quality indicators e#1, e#2, which characterize the estimated quality of the associated future downlink channel indicators fQ#1, fQ#2; determining 218 a transmission parameter p#2 based at least on the received future downlink channel quality indicators fQ#1, fQ#2 and on the estimated quality indicators e#1, e#2; and transmitting 220 downlink data d#2 via the downlink channel DLCH according to the determined transmission parameter p#2.

[0112] The illustrated example comprises transmitting 202 a supported configuration sc indicating a supported mode, in particular an estimation support mode ESM.

[0113] According to an example, the estimated quality indicators e#1, e#2 are valid for an indeterminate period of time at least upon their reception, in particular for a period of time during which the RRC state exists.

[0114] Depicted is receiving 264 a further estimated quality indicator e#2 indicating a difference relative to a previously received estimated quality indicator e#1; and determining 265 a further transmission parameter based at least on the previously received estimated quality indicator e#1 and on the received further estimated quality indicator e#2; and transmitting another downlink data via a downlink channel according to the determined further transmission parameter.

[0115] For example, if the comparison indicates that the difference or change in the estimated quality indicator e#2 is significant, the estimated indicator e#2 is transmitted. For example, a threshold operation determines whether to transmit. In another example, a predefined or trainable function takes both e#1 and e#2 as input and outputs a transmission decision.

[0116] The previously received first downlink data d#1 is transmitted according to the first transmission parameter p#1 which is different from the transmission parameter p#2.

[0117] According to one example, the method further includes: determining 160 an additional future downlink channel quality indicator fQ#2 representing the estimated future quality associated with the downlink channel DLCH based at least on the monitored downlink channel quality Q#2; determining 160 an additional estimated quality indicator e#2 representing the estimated quality of the associated future downlink channel indicators fQ#1, fQ#2 based at least on the monitored downlink channel quality Q#2; comparing 162 the additional estimated quality indicator e#2 with the previously determined estimated quality indicator e#1; and transmitting 164 the additional estimated quality indicator e#2 based on the comparison.

[0118] According to an example, the support configuration sc includes an estimation range EH, which characterizes a future downlink channel quality indicator fQ#1 and its associated estimated quality indicator e#1 or a time period during which the support configuration sc is valid, and wherein the future downlink channel quality indicator fQ#1 and its associated estimated quality indicator e#1 are determined based on the estimation range EH.

[0119] Also depicted is a support configuration sc indicating an auxiliary support mode ASM, wherein, according to the auxiliary support mode ASM, the method comprises: receiving 232 a downlink channel quality Q#3; receiving 234 auxiliary information AS characterizing a present or future condition of a downlink channel receiving device RT; determining 236 an additional transmission parameter p#3 based on the downlink channel quality Q#3 and based on the auxiliary information AS; and transmitting 237 downlink data d#3 via the downlink channel DLCH according to the additional transmission parameter p#3.

[0120] A device, in particular a radio access node gNB#1 of a radio access network or a module thereof, in particular according to an estimation support mode ESM, is shown, the device comprising: a receiving component configured to receive 216 future downlink channel quality indicators fQ#1, fQ#2, which characterize the estimated future quality associated with a downlink channel DLCH; a receiving component configured to receive 214 estimated quality indicators e#1, e#2, which characterize the estimated quality of the associated future downlink channel indicators fQ#1, fQ#2; a determining component configured to determine 218 a transmission parameter p#2 based at least on the received future downlink channel quality indicators fQ#1, fQ#2 and on the estimated quality indicators e#1, e#2; and a transmitting component configured to transmit 220 downlink data d#2 via the downlink channel DLCH according to the determined transmission parameter p#2.

[0121] Figure 2 Depicted Figure 1 It is shown that the method further comprises: determining 238 an estimation capability request ecRq based on the current need to support downlink channel parameter adaptation; transmitting 240 the estimation capability request ecRq; after transmitting the estimation capability request ecRq, receiving 242 an estimation capability response eqResp, the estimation capability response eqResp in particular indicating at least a supported mode, in particular an estimation support mode ESM; wherein a supported configuration sc is determined and transmitted as a reply to the transmitted estimation capability response ecResp.

[0122] A method for operating UE#1 is depicted comprising: receiving 140 an estimation capability request ecRq; after receiving the estimation capability request ecRq, determining 142 an estimation capability response eqResp, the estimation capability response eqResp in particular indicating at least a supported mode, in particular an estimation support mode ESM; transmitting 144 the estimation capability response ecResp; and wherein receiving a support configuration sc as a reply to the transmitted estimation capability response ecResp.

[0123] The illustrated example is characterized in that the support configuration sc includes an estimation range EH, which characterizes the time period during which the future downlink channel quality indicator fQ#1 and its associated estimated quality indicator e#1 or the support configuration sc is valid or should be valid, and wherein the future downlink channel quality indicator fQ#1 and its associated estimated quality indicator e#1 are determined based on the estimation range EH.

[0124] For example, in the case where gNB#1 is embodied as a satellite, gNB#1 uses the knowledge of its own trajectory and constellation to define the prediction range and signals it in the support configuration sc.

[0125] Alternatively or additionally, the device UE#1 determines an estimated range EH.

[0126] Figure 3 Signaling for downlink transmission is depicted. Device gNB#1 performs downlink transmission 302, 304. Device UE#1 measures the downlink channel quality, e.g. in terms of RSRP / RSRQ / RSSI, and derives a future downlink channel quality indicator in the form of estimates of CQI, PMI and RI. The future downlink channel quality indicator is sent back to the BS.

[0127] The transmission starts at t0. A measurement of Q#1 is taken at t1. At time t2, an estimated downlink channel quality indicator f#Q1 is received at gNB#1. Note that the time between transmission and reception of data on the downlink is denoted D1, and the transmission delay on the uplink is denoted D2. Note that in a non-terrestrial network (NTN), gNB#1 (i.e., satellite) is moving very fast. Similarly, the UE is moving quite fast in a vehicular application.

[0128] Therefore, D1!=D2. Furthermore, for the next downlink transmission, a different delay D3 is assumed, so D1!=D2!=D3. In the prior art, the DL transmission is changed based on D1. However, since D2+D3 can no longer be ignored, it is expected that D1 and D2 will be significantly different.

[0129] In contrast to sending the actual measurement at t1, the expected channel conditions in the future are transmitted towards gNB#1 in the form of fQ#1.

[0130] Figure 4 A schematic block diagram of a machine trained function or estimator 400 is depicted. The trainable estimator 400 receives input data, for example including Q#1. Based on the input data Q#1, the machine trained estimator 400 derives fQ#1, for example a predicted CQI. In addition, the device also outputs e#1 associated with Q#1, where e#1 may also be referred to as a prediction accuracy identifier PAI. According to another example, fQ#1 and e#1 are determined separately.

[0131] Figure 5 A schematic sequence diagram is depicted. At connection establishment, device gNB#1 transmits ecRq to UE#1 / UE. If UE#1 provides prediction capability / estimation capability, it sends a response in the form of rqResp. If UE#1 does not provide estimation capability, it returns failure to gNB#1.

[0132] Assuming that an ACK is sent to gNB#1 via rqResp, gNB#1 informs UE#1 of the actual prediction task via support configuration sc. Support configuration sc contains at least one of: task type, such as classification or regression; estimation range; request to signal the estimation accuracy class; dedicated parameter for which prediction / estimation is desired, where the parameter is one of the following: channel quality indicator CQI, 5QI, 6QI, precoding matrix indicator PMI, rank indicator RI. A classification task is, for example, the determination of a discrete value, such as QCI. A regression task is, for example, the prediction of a continuous value, such as RSSI.

[0133] gNB#1 checks the supported configuration sc and evaluates its capabilities based on the requirements. gNB#1 then transmits e#1, which may be referred to as Session-PAI S-PAI. S-PAI may be, for example, an X-bit indicator that gives some granularity regarding:

[0134] 1. The expected mean squared error (MSE), which is the variance of the regression task; and

[0135] 2. Cross entropy loss CE, which is the probability of incorrect category assignment.

[0136] According to the following table, gNB#1 is able to derive MSE or CE based on the received or currently valid estimated quality e#n. MSE, CE or any other comparable value represents a measure of how reliable the associated downlink quality indicator is.

[0137]

[0138]

[0139] Table 1

[0140] The example shown involves the estimated quality indicator e#1 and the associated future downlink channel quality indicator fQ#1 being transmitted in pairs.

[0141] This example can also be called predicted reliability report per time. gNB#1 sends request ecRq at connection establishment. However, e#1 is not attached to the session. Upon receiving ACK, gNB#1 issues support configuration sc. gNB#1 does not report e#1 for the session, but the predicted measurement fQ#1 includes e#1 as an extra field in the measurement report.

[0142] Figure 6Depicts the reporting of a sudden difference in reliability compared to the previously valid reliability signaled via e#1. Because the radio conditions and environmental conditions change unexpectedly, and the requirements indicated in e#1 can no longer be met, UE#1 signals e#2, which represents the difference from e#1. e#2 can be a small integer, represented by a few bits. For example, it can represent:

[0143] [-2] reduces the accuracy from e#1 by 2,

[0144] [-1] reduces the accuracy from e#1 by 1,

[0145] [+1] Increase the accuracy from e#1 by 1,

[0146] [+2] Increase the accuracy from e#1 by 2.

[0147] Figure 7 Depicted from Figure 1 Assisted Support Mode ASM. Under certain conditions, the radio terminal UE#1 / UE has no prediction capability. In this case, UE#1 responds to ecRq with a failure F. In this case, gNB#1 may reply with a support configuration sc or a separate request to assist the prediction process on the gNB#1 side by transmitting assistance information AS.

[0148] According to an example, conditional prediction on the gNB#1 side enhances the UE prediction capability. Let us consider a case where the UE / RT is able to provide fQ#1, but the associated e#1 may not allow the intended operation on the gNB#1 side. In this case, gNB#1 may additionally request assistance information, although the pCQI request is responded to. In this case, Figure 1 The blocks of ESM and ASM are executed sequentially, not alternately.

[0149] If ecRq replies with a failure, and the request for auxiliary information also results in a failure, a legacy device may be detected.

[0150] Figure 8 A non-terrestrial cellular communication network CN#1 is depicted. A radio access node gNB#1 is attached to a satellite. As indicated by the arrows, a radio terminal UE#1 is located on the ground and moves relatively slowly compared to gNB#1. Due to the relative movement to each other, both the uplink radio channel ULCH and the downlink radio channel DLCH suffer from possible quality degradation.

[0151] In future communication systems, the combination of terrestrial and non-terrestrial networks improves coverage and thereby improves the availability of corresponding communication networks, especially in rural areas. In contrast to traditional terrestrial networks, deploying parts of the network in space leads to a variety of new challenges. The satellites serving terrestrial UEs are not located in geostationary orbits, but in lower orbits. Compared to geostationary satellites, this allows for a reduction in round-trip time delays due to the relatively short distance. However, the delay is still quite large compared to the propagation delay observed in typical terrestrial deployments.

[0152] Since CN#1 can be viewed as an extension of the terrestrial network in certain areas, underground, in tunnels, bridges, etc., the coverage of CN#1 may suddenly degrade. These occasional events depend on the geographical location of the UE and its relative position to the satellite. Therefore, probabilistic channel models as used in the literature will not be sufficient to design an appropriate predictor. Therefore, data-driven methods do not require closed-form channel models, but instead can be trained to realistic propagation scenarios.

[0153] Fig. 9 A terrestrial cellular radio communication network CN#2 is depicted. In this case, gNB#1 has a fixed position on the ground and UE#1 moves as part of vehicle V#1. Furthermore, in this case, ULCH and DLCH may suffer from potential degradation due to the relative movement. Furthermore, UE#1 may lose its connection with gNB#1 if V#1 enters a tunnel or moves near additional obstacles that affect the downlink quality, such as trees or buildings.

[0154] like Figure 8 and 9 The relative mobility and possible future degradation of the downlinks as shown in motivates the approach described herein.

[0155] Fig.10 The schematic sequence diagram depicts Figure 1 Another example of auxiliary support mode ASM.

[0156] According to step 1002, UE#1 determines first information i#1, such as measuring the quality Q#1 of the received or observed downlink channel DLCH. In other words, the first information i#1 may be based on or passively based on radio measurements, such as RSRP, RSSI, RSRQ or Doppler measurements.

[0157] According to step 1004, UE#1 determines second information i#2, such as self-sensor information. The self-sensor information includes or is based on at least one of image data from lidar / radar / camera, vehicle heading, vehicle speed, route information, waypoints, and map information.

[0158] The figure relates to a method for operating UE#1, comprising: transmitting 1110 first information i#1, which characterizes at least one quality associated with a downlink channel DLCH received by the radio terminal UE#1; transmitting 1112 second information i#2, which characterizes the current or future environmental situation of a physical entity E#1 associated with the radio terminal UE#1.

[0159] In other words, the second information i#2 is determined using at least one sensor arranged at or near the physical entity E#1. The physical entity E#1 may be a car or industrial equipment to which the UE#1 is attached or at least functionally coupled. The second information may also be referred to as assistance information, since the second information assists the gNB in ​​determining the transmission parameters for the downlink transmission.

[0160] The second information i#2 characterizing the environmental situation may be for currently determined sensor information or information based at least in part on sensor information from a sensor attached to a physical entity or a radio terminal. In addition, the second information may be associated with the current situation and / or the future situation. In the case of route information, the second information includes the future situation. In addition, in the case where the sensor information (such as a camera picture) depicts a tunnel entrance, the second information includes the future situation, i.e., possible communication degradation after entering the tunnel entrance.

[0161] One example relates to a method for operating a gNB#1, comprising: receiving 2110 first information i#1 characterizing a quality associated with a downlink channel DLCH received by an apparatus UE#1; receiving 2112 second information i#2 characterizing a current or future environment situation of a physical entity E#1 associated with the apparatus UE#1. Determining 2120 a downlink transmission parameter p#A based on the first information i#1 and based on the second information i#2 via a machine-trained function 400; and transmitting 2130 downlink data d#A via the downlink channel DLCH according to the determined downlink transmission parameter p#A. As a result, UE#1 receives 1130 downlink data d#A with an increased downlink quality.

[0162] The diagram illustrates an apparatus gNB#1, in particular a radio access node of a radio access network or a module thereof, comprising: a receiving component for receiving 2110 first information i#1, which characterizes at least one quality associated with a downlink channel DLCH received by the apparatus UE#1; a receiving component for receiving 2112 second information i#2, which characterizes a current or future environmental situation of a physical entity E#1 associated with the apparatus UE#1; a determining component for determining 2120 a downlink transmission parameter p#A based on the first information i#1 and based on the second information i#2 via a machine-trained function 400; and a transmitting component for transmitting 2130 downlink data d#A via the downlink channel DLCH according to the determined downlink transmission parameter p#A.

[0163] A device UE#1, in particular a radio terminal or user equipment or a module thereof, is depicted, comprising: a transmission component for transmitting 1110 first information i#1, which characterizes at least one quality associated with a downlink channel DLCH received by the radio terminal or device UE#1; a transmission component for transmitting 1112 second information i#2, which characterizes a current or future environmental situation of a physical entity E#1 associated with the radio terminal or device UE#1.

[0164] Fig.11 Depicted Figure 4 An example of a machine trained function 400, which may be implemented at UE#1, gNB#1, or in a distributed manner between one or more UE#1..n and gNB#1, as illustrated below.

[0165] The method illustrated includes: determining a first potential code LC#A based on a downlink channel quality Q#1 via a machine-trained downlink embedding function 410; determining a second potential code LC#B based on self-sensor-based data esD#1 via a machine-trained auxiliary data embedding function 420; and determining a future downlink channel indicator fQ#1 and at least one estimated quality indicator e#1 based on at least the first potential code LC#A and on the second potential code LC#B via a machine-trained prediction function 490.

[0166] The method is depicted as further comprising: observing 1006 a parameter SL#1 associated with a sidelink channel SLCH; determining a third latent code LC#C based on the parameter SL#1 associated with the sidelink channel SLCH via a machine-trained sidelink embedding function 430; and wherein future downlink channel indicators fQ#1, fQ#2 and estimated quality indicators e#1, e#2 are further determined based on the third latent code LC#C.

[0167] In addition to steps 1002 and 1004, step 1006 is also shown. According to step 1006, a parameter SL#1 associated with the sidelink channel is determined. SL#1 may also be referred to as sidelink assistance information and may originate from, for example, a roadside unit (RSU) and / or a nearby vehicle and may include sensor sharing. Examples of SL#1 include: relative location information, CQI at a UE connected via a sidelink.

[0168] In an exemplary scenario, the self-sensor provides additional information in the form of esD#1 by observing the environment. For example, a camera or lidar detects a tree or a tunnel entrance, which may cause a very sudden and considerable change in the channel characteristics. Therefore, esD#1 is considered to represent or describe the current or future environmental conditions of the physical entity to which the device UE#1 is connected or coupled.

[0169] In front of the prediction function 490, an embedding layer em is arranged. The embedding layer em maps the corresponding high-dimensional features into the corresponding low-dimensional latent space. Raw inputs such as Q#1, esD#1, and SL#1 have high inter-feature correlation. For example, RSSI, RSRP, and RSRQ are closely related. Similarly, objects observed by a lidar may also be detected by a camera. The embedding layer em is used as a pre-filtering step to extract relevant information. The raw inputs Q#1, esD#1, and SL#1 are separated from the predictor function 490. Subsequently, the prediction function 490 is more robust to changes in the raw input. For example, if different types of lidar sensors or the like are used. The prediction function 490 is implemented, for example, as a recurrent neural network, which is very suitable for capturing time input sequences. Specific implementations of the prediction function 490 include long short-term memory LSTM or continuous-discrete recurrent Kalman network CDRKN.

[0170] The machine trained downlink embedding function 410 determines the first latent code LC#a based on Q#1. Function 410 is embodied as a deep neural network.

[0171] The machine-trained auxiliary information embedding function 420 determines the second latent code LC#B based on esD#1. Function 420 is embodied as a deep neural network.

[0172] The machine-trained sidelink embedding function 430 determines the third latent code LC#C based on SL#1. Function 430 is embodied as a deep neural network.

[0173] There are several ways to train the network structure of function 400. For example, for offline supervised training, the training device has access to a huge dataset and can therefore capture different sensor inputs for different setups as well as radio measurements between satellites and UEs in case of non-terrestrial applications to provide a well-trained model.

[0174] In case of online self-supervised training, UE#1 and / or gNB#1 starts with a warm-up phase where no predictions are made. It saves its own observations from the past to build a local dataset. This dataset will grow over time and match the available sensors on the UE#1 side. UE#1 then checks its own predictions in the form of fQ#1 and e#1 with the associated samples Q#2 actually observed in the future. Thus, it can perform self-supervised training.

[0175] As another example, training federated learning is performed in a distributed manner. Instead of training on a large global dataset, training is performed at different entities based on local datasets. Training can then be performed in parallel. This shortens the time-consuming process of large amounts of data collection / data storage and data sharing prior to training.

[0176] The example shown relates to the method further comprising: determining, via a machine-trained function 490, future downlink channel quality indicators fQ#1, fQ#2 characterizing an estimated future quality associated with a downlink channel DLCH based at least on the first information and at least on the second information, and estimated quality indicators e#1, e#2 characterizing an estimated quality of the associated future downlink channel indicators fQ#1, fQ#2; and determining 2120 a transmission parameter p#A based at least on the future downlink channel quality indicators fQ#1, fQ2 and based on the estimated quality indicators e#1, e#2. In other words, according to one example, the processing of LC#C and / or SL#1 is optional.

[0177] Determine 2120, for example, a mapping function that maps the received fQ#1 and e#1 pair to the corresponding p#A. Therefore, a lookup table can be applied to determine 2120. In another example, another machine-trained mapping function is used to determine 2120. In addition, the transmission parameter p#A includes at least one of the following: modulation and coding scheme, transmission power, etc.

[0178] Fig.12 Depicted is an example of the distribution of different functional entities between UE#1 and gNB#1. Data transmission from UE#1 to gNB#1 is via the uplink channel ULCH.

[0179] It is depicted that the transmitted first information i#1 is the first latent code LC#A, wherein the method further includes: determining 1002 the measured downlink channel quality Q#1 from observing the downlink channel DLCH; and determining the first latent code LC#A based on the measured downlink channel quality Q#1 via a machine-trained downlink embedding function 410.

[0180] The transmitted second information i#1 is illustrated as a second latent code LC#B, wherein the method further comprises: determining 1004 self-sensor based data esD#1 based on sensor measurements associated with the physical entity; and determining the second latent code via a machine trained auxiliary information embedding function 420.

[0181] In one example it is illustrated that the method further comprises: transmitting third information characterizing a parameter SL#1 associated with a sidelink channel SLCH.

[0182] It is shown that the third information transmitted is the third latent code LC#C, wherein the method further includes: observing 1006 parameters associated with the sidelink channel SLCH; and determining the third latent code LC#B based on the parameter SL#1 associated with the sidelink channel SLCH via a machine-trained sidelink embedding function 430.

[0183] The figure relates to a case where the received first information i#1 is a first latent code LC#A, and where the second information i#2 is a second latent code LC#B, wherein the method further comprises: determining a downlink transmission parameter p#A based on the received first latent code LC#A and based on the received second latent code LC#B via a machine-trained predictor function 490.

[0184] Figures 12 to 17 Examples of different implementations of the machine trained function 400 are shown. However, since some implementations at least partially overlap, corresponding wording may be applied to corresponding parts of another implementation.

[0185] Fig.13 Another example of the distribution of different functional entities is depicted. Fig.12 In contrast, Fig.13 A further device UE#2 is included, which determines a third latent code LC#C and transmits LC#C towards gNB#1.

[0186] The diagram illustrates that the received first information i#1 is the downlink channel quality Q#1, and wherein the second information i#2 is the second latent code LC#B, wherein the method further includes: determining the first latent code LC#A based on the received downlink quality Q#1 via a machine-trained downlink embedding function 410; determining 490 the downlink transmission parameter p#A based on the determined first latent code LC#A and based on the received second latent code LC#B via a machine-trained predictor function.

[0187] It is illustrated that the received third information is a third LC#C potential code, wherein the method further comprises: determining a downlink transmission parameter based on at least one third potential code LC#C via a machine-trained function.

[0188] Fig.14 Another example of a function distribution is depicted. Fig.12 and 13 In contrast, Fig.14 Depicted is the transmission of only latent code LC#B, where latent codes LC#A and LC#C are determined at gNB#1.

[0189] Fig.14 Another example of a function distribution is depicted. Fig.12 and 13 In contrast, Fig.14 Depicted is the transmission of only latent code LC#B, where latent codes LC#A and LC#C are determined at gNB#1.

[0190] It is shown that the method for operating UE#1 further comprises observing 1006 a parameter SL#1 associated with a sidelink channel SLCH, wherein the transmitted third information i#3 is the sidelink parameter SL#1 associated with the sidelink channel SLCH.

[0191] The figure relates to the transmitted first information Q#1 being the measured downlink channel quality Q#1, wherein the method for operating UE#1 further comprises: determining the measured downlink channel quality Q#1 from observing the downlink channel DLCH.

[0192] It is shown that the received first information is the first latent code LC#A, and wherein the second information is self-sensor based data esD#1, wherein the method for operating UE#1 further includes: determining the second latent code LC#B based on the received self-sensor based data esD#1 via a machine-trained auxiliary data embedding function 420; determining the downlink transmission parameter p#A based on the determined first latent code LC#A and based on the received second latent code LC#B via a machine-trained predictor function 490.

[0193] It is shown that the method for operating gNB#1 further includes: receiving third information characterizing a parameter SL#1 associated with a sidelink channel SLCH.

[0194] It is shown that the received third information is a sidelink parameter SL#1 associated with the sidelink channel SLCH, wherein the method for operating gNB#1 further includes: determining a third latent code LC#C via a function 430 of sidelink embedded machine training, and determining a downlink transmission parameter p#A based at least on the third latent code LC#C via a function 490 of machine training.

[0195] Compared with other features Fig.14 In contrast, Fig.15Another device UE#2 is depicted transmitting SL#1 to gNB#1.

[0196] Fig.16 Another example of function distribution is depicted. Latent code LC#A is transmitted to gnB#1, while LC#B is determined at gNB#1 side.

[0197] The illustrated example relates to the transmitted second information esD#1 being ego sensor based data esD#1, wherein the method for operating UE#1 further comprises determining 1004 ego sensor based data esD#1 based on at least one sensor measurement of a sensor associated with the physical entity.

[0198] and Fig.16 In contrast, Fig.17 It is depicted that another device UE#2 transmits the latent code LC#C instead of UE#1.

[0199] and Figure 3 In contrast, Fig.18 Function 400 depicting machine training is arranged at the device gNB#1. For example, if UE#1 is not able or configured to perform prediction of Q#1, Q#2, etc., UE#1 signals gNB#1 that it cannot provide prediction. Assuming gNB#1 has more processing power, gNB#1 can calculate fQ#1..n itself based on the provided at least one Q#1.

[0200] In the context of non-terrestrial networks (NTN), gNB#1 is a satellite and computational resources may also be limited. Therefore, the prediction of fQ#1..n is at least partially envisaged at the radio access network, edge server or the like. That is, in general, the determination of fQ#1 may be distributed between terminal devices such as UE#1 and devices of the radio access network, depending on the available resources.

[0201] Apart from Fig.18 In addition, Fig.19 Several devices UE#1, UE#2, UE#3 are depicted transmitting their respective measured downlink channel qualities Q#1, Q#2, Q#2 together with auxiliary information in the form of respective ego sensor based data esD#1, esD#2, esD#3. The previous data serves as input to a function 400 at the device gNB#1. The transmission 304 is based on transmission parameters, which are determined based on fQ#1.

[0202] In contrast to UE#1, gNB#1 is not expected to have direct access to UE#1's own sensors. However, UE#1 may report additional assistance information in the form of esD#1 or latent code LC#C if possible. Thus, the own sensor information may be pre-processed and / or compressed before transmission. This allows us to also share possibly sensitive private sensor information. Furthermore, gNB#1 may utilize reports sent from several nearby devices UE#1-3.

Claims

1. A method, in particular according to an Estimation Support Model (ESM), comprising: monitoring (106) at least one downlink channel (DLCH); determining (108) at least one downlink channel quality (Q#1, Q#2) associated with at least one monitored downlink channel (DLCH); determining (110) at least one future downlink channel quality indicator (fQ#1, fQ#2) characterizing an estimated future quality associated with a downlink channel (DLCH) based at least on at least one monitored downlink channel quality (Q#1, Q#2); determining (112) at least one estimated quality indicator (e#1, e#2) characterizing an estimated quality of at least one associated future downlink channel indicator (fQ#1, fQ#2) based at least on at least one monitored downlink channel quality (Q#1, Q#2); transmitting (114; 164) the at least one estimated quality indicator (e#1, e#2); and The at least one future downlink channel quality indicator (fQ#1, fQ#2) is transmitted (116; 166).

2. The method according to claim 1, comprising: A support configuration (sc) indicating at least one support mode, in particular an estimation support mode (ESM), is received (102).

3. The method according to claim 1 or 2, wherein the method further comprises: receiving (140) an estimate capability request (ecRq); Upon receiving an estimate capability request (ecRq), determining (142) an estimate capability response (eqResp) which in particular indicates at least one supported mode, in particular an estimation support mode (ESM); transmitting (144) an estimated capabilities response (ecResp); and Therein a supported configuration (sc) is received as a reply to the transmitted estimated capabilities response (ecResp).

4. The method according to any one of claims 1 to 3, wherein the support configuration (sc) indicates an Assisted Support Mode (ASM), wherein: According to the Assisted Support Mode (ASM), the method comprises: monitoring (106) the at least one downlink channel (DLCH); determining (110) at least one downlink channel quality (Q#3) associated with at least one monitored downlink channel (DLCH); and The at least one downlink channel quality (Q#3) is transmitted (132).

5. The method according to claim 4, wherein: According to Assisted Support Mode (ASM), the method further comprises: determining (130) auxiliary information (AS) characterizing at least one present or future condition of a downlink channel receiving device (RT) or a physical entity to which the receiving device (RT) is associated or installed; and Transmit (134) auxiliary information (AS).

6. The method according to any one of claims 1 to 5, wherein the at least one estimated quality indicator (e#1, e#2) is valid for an indeterminate period of time, at least when it is transmitted, in particular for a period of time in which the RRC state is active.

7. The method according to any one of claims 1 to 6, wherein the method further comprises: determining, via a machine trained downlink embedding function (410), a first latent code (LC#A) based on the at least one downlink channel quality (Q#1); Determining a second latent code (LC#B) based on the ego sensor based data (esD#1) via a machine trained auxiliary data embedding function (420); At least one future downlink channel indicator (fQ#1, fQ#2) and at least one estimated quality indicator (e#1, e#2) are determined based on at least a first potential code (LC#A) and based on a second potential code (LC#B) via a machine-trained prediction function (490).

8. An apparatus, in particular a radio terminal or user equipment (UE#1) or a module thereof, in particular according to an estimation support mode (ESM), comprising: a monitoring component configured to monitor (106) at least one downlink channel (DLCH); a determining means configured to determine (108) at least one downlink channel quality (Q#1, Q#2) associated with at least one downlink channel (DLCH) being monitored; Determining means configured to determine (110) at least one future downlink channel quality indicator (fQ#1, fQ#2) characterizing an estimated future quality associated with a downlink channel (DLCH) based at least on at least one monitored downlink channel quality (Q#1, Q#2); Determining means configured to determine (112) at least one estimated quality indicator (e#1, e#2) characterizing an estimated quality of at least one associated future downlink channel indicator (fQ#1, fQ#2) based at least on at least one monitored downlink channel quality (Q#1, Q#2); Transmitting means configured to transmit (114; 164) the at least one estimated quality indicator (e#1, e#2); as well as Transmitting means are configured to transmit (116; 166) the at least one future downlink channel quality indicator (fQ#1, fQ#2).

9. A method, in particular according to an Estimation Support Model (ESM), comprising: receiving (216) at least one future downlink channel quality indicator (fQ#1, fQ#2) characterizing an estimated future quality associated with a downlink channel (DLCH); receiving (214) at least one estimated quality indicator (e#1, e#2) characterizing an estimated quality of at least one associated future downlink channel indicator (fQ#1, fQ#2); determining (218) at least one transmission parameter (p#2) based at least on the received at least one future downlink channel quality indicator (fQ#1, fQ#2) and based on at least one estimated quality indicator (e#1, e#2); as well as Downlink data (d#2) is transmitted (220) via a downlink channel (DLCH) according to the determined at least one transmission parameter (p#2).

10. The method according to claim 9, comprising: A support configuration (sc) indicating at least one support mode, in particular an estimation support mode (ESM), is transmitted (202).

11. The method according to claim 9 or 10, wherein the method further comprises: determining (238) an estimated capability request (ecRq) based on a current requirement to support downlink channel parameter adaptation; transmitting (240) an estimate capability request (ecRq); After transmitting said estimation capability request (ecRq), receiving (242) an estimation capability response (eqResp) indicating in particular at least one supported mode, in particular an estimation support mode (ESM); Wherein the supported configuration (sc) is determined and transmitted as a reply to the transmitted estimated capabilities response (ecResp).

12. The method according to any one of claims 9 to 11, wherein the support configuration (sc) indicates an Assisted Support Mode (ASM), wherein: According to the Assisted Support Mode (ASM), the method comprises: receiving (232) at least one downlink channel quality (Q#3); receiving (234) auxiliary information (AS) characterizing at least one current or future condition of a downlink channel receiving device (RT); determining (236) a further transmission parameter (p#3) based on the at least one downlink channel quality (Q#3) and based on the at least one auxiliary information (AS); and Downlink data (d#3) is transmitted (237) via a downlink channel (DLCH) according to the further transmission parameters (p#3).

13. The method according to one of claims 9 to 12, comprising: receiving (264) at least one further estimated quality indicator (e#2) indicating a difference relative to at least one previously received estimated quality indicator (e#1); as well as determining (265) a further transmission parameter based at least on at least one previously received estimated quality indicator (e#1) and based on at least one received further estimated quality indicator (e#2); as well as According to the determined further transmission parameter, another downlink data is transmitted via the downlink channel.

14. An apparatus, in particular a radio access node (gNB#1) of a radio access network or a module thereof, in particular according to an estimation support mode (ESM), the apparatus comprising: Receiving means configured to receive (216) at least one future downlink channel quality indicator (fQ#1, fQ#2) characterizing an estimated future quality associated with a downlink channel (DLCH); Receiving means configured to receive (214) at least one estimated quality indicator (e#1, e#2) characterizing an estimated quality of at least one associated future downlink channel indicator (fQ#1, fQ#2); determining means configured to determine (218) at least one transmission parameter (p#2) based at least on the received at least one future downlink channel quality indicator (fQ#1, fQ#2) and on the at least one estimated quality indicator (e#1, e#2); as well as The transmission means is configured to transmit (220) downlink data (d#2) via a downlink channel (DLCH) according to the determined at least one transmission parameter (p#2).

15. Use of one of the methods or devices (RT, gNB#1) according to one of the preceding claims.

Images