TECHNICAL FIELD Various examples relate to a device configured to transmit a downstream sounding packet and to determine a downstream channel state of a radio channel based on a receive property of the preamble of the downstream sounding packet. The device is further configured to receive an upstream packet and to reevaluate the determined downstream channel state based on a receive property of the preamble of the upstream packet. Further examples relate to a corresponding method.

BACKGROUND Channel sounding helps to tailor transmission properties of data communication on a radio channel. E.g., based on channel sounding, the transmission properties can be tailored depending on the particular transmission environment. Thereby, transmission reliability can be increased. While, generally, channel sounding may be desirable for all kinds of radio channels, one particular field of application is the sounding of a channel comprising a plurality of time-space streams. Such a scenario is often applicable where an antenna array is used to implement multiple input multiple output (MIMO) techniques. MIMO techniques may be combined with multi-user (MU) or single-user (SU) beamforming. Here, channel knowledge is required regarding the setup of the transmitter antennas and the receiver antennas. Antenna weights are determined from the channel sounding.

Typically, for channel sounding, so-called sounding packets (sometimes also referred to as pilot packets or reference packets) are employed which comprise a preamble of well-defined form. Based on one or more receive properties of the preamble, it is then possible to evaluate the channel state. Such techniques are sometimes referred to as acquiring a channel.

To accurately sound the channel, it can be required to implement a high time resolution. This is due to a strong time dependency of the channel state E.g., by communicating sounding packets using a high repetition rate, changes of the channel state can be accurately tracked and, if necessary, compensated for. On the other hand, based on such techniques significant overhead results from the communication of the sounding packets. The data throughput can then be limited. SUMMARY

Therefore, a need exists for advanced techniques of radio channel sounding. In particular, a need exists for techniques which enable accurate channel sounding with limited overhead. This need is met by the features of the independent claims. The dependent claims define embodiments.

According to an example, a device comprises an interface and at least one processor. The interface is configured to transceive on a bi-directional radio channel. The at least one processor is configured to transmit a downstream sounding packet via the interface. The at least one processor is further configured to determine a downstream channel state of the radio channel based on at least one receive property of a preamble of the downstream sounding packet. The at least one processor is further configured to receive an upstream packet via the interface. The at least one processor is further configured to reevaluate the downstream channel state based on at least one receive property of a preamble of the upstream packet.

According to an example a method is executed by a device and comprises transmitting a downstream packet on a bi-directional radio channel. The method further comprises determining a downstream channel state of the radio channel based on at least one receive property of a preamble of the downstream sounding packet. The method further comprises receiving an upstream packet on the radio channel. The method further comprises reevaluating the determined downstream channel state based on at least one receive property of a preamble of the upstream packet. According to an example, a computer program product comprises program code to be executed by at least one processor. Executing the program code causes the at least one processor to perform a method comprising transmitting a downstream packet on a bidirectional radio channel. The method further comprises determining a downstream channel state of the radio channel based on at least one receive property of a preamble of the downstream sounding packet. The method further comprises receiving an upstream packet on the radio channel. The method further comprises reevaluating the determined downstream channel state based on at least one receive property of a preamble

It is to be understood that the features mentioned above and those yet to be explained below may be used not only in the respective combinations indicated, but also in other combinations or in isolation without departing from the scope of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A schematically illustrates a first device and the second device configured to communicate on a bi-directional radio channel according to various embodiments.

FIG. 1 B illustrates a plurality of time-space streams on the radio channel of FIG. 1 A.

FIG. 2 schematically illustrates a transmission protocol stack of a radio access technology (RAT) for communicating on the radio channel.

FIG. 3A illustrates a time evolution of a downstream channel state of the radio channel, wherein the channel state is reevaluated based on a receive property of preambles of upstream packets in a first operational mode.

FIG. 3B illustrates a time evolution of a downstream channel state of the radio channel, wherein the channel state is not reevaluated based on receive properties of preambles of upstream packets in a second operational mode different from the first operational mode of FIG. 3A.

FIG. 4 schematically illustrates a sounding packet implemented by a null data packet comprising a preamble and no payload section. FIG. 5 schematically illustrates a data packet comprising a preamble and a payload section.

FIG. 6 is a signaling diagram of signaling between two devices on a bi-directional radio channel according to various embodiments.

FIG. 7 is a signaling diagram of signaling between two devices on a bi-directional radio channel according to various embodiments. FIG. 8 illustrates antenna weights of an antenna array determined based on a channel state of the radio channel according to various embodiments. FIG. 9 is a flowchart of a method according to various embodiments. FIG. 10 is a flowchart of a method according to various embodiments. FIG. 1 1 is a flowchart of a method according to various embodiments.

DETAILED DESCRIPTION OF EMBODIMENTS

In the following, embodiments of the invention will be described in detail with reference to the accompanying drawings. It is to be understood that the following description of embodiments is not to be taken in a limiting sense. The scope of the invention is not intended to be limited by the embodiments described hereinafter or by the drawings, which are taken to be illustrative only.

The drawings are to be regarded as being schematic representations and elements illustrated in the drawings are not necessarily shown to scale. Rather, the various elements are represented such that their function and general purpose become apparent to a person skilled in the art. Any connection or coupling between functional blocks, devices, components, or other physical or functional units shown in the drawings or described herein may also be implemented by an indirect connection or coupling. A coupling between components may also be established over a wireless connection. Functional blocks may be implemented in hardware, firmware, software, or a combination thereof.

Hereinafter, techniques of channel sounding are described. In detail, the various techniques described herein can be employed to sound a bi-directional radio channel. Hereinafter, one direction of the bi-directional radio channel is denoted as downstream (DS); while the opposing direction is denoted as upstream (US). Generally, the techniques described herein can be applied for any one of the two directions of the bi-directional radio channel; for sake of simplicity, hereinafter, reference is made to the DS direction which, however, may be arbitrarily defined. In some examples, to sound the DS channel, first, at least one receive property of a preamble of a DS sounding packet is evaluated. For this, the DS sounding packet is transmitted by a first device and received by a second device. The second device may provide an US report message back to the first device, the US report message being indicative of the at least one receive property of the preamble of the DS sounding packet. Then, the first device can determine a DS channel state based on the at least one receive property.

The DS sounding packet may be implemented by null data packet (NDP) not comprising a payload section. Hence, the DS sounding packet, in some examples, may only comprise a preamble, but not include any application-layer user data or control data of higher layers. This may reduce the overhead imposed by the DS sounding packet.

The DS channel state may be indicative of elements selected from the group comprising: path loss; multipath fading; phase shift; residual phase error of modulated symbols; interference from external noise sources; bit error rate; packet error rate; signal-to-noise ratio; decoding errors; decoding reliability; etc..

Once the DS channel state has been determined, the first device is configured to receive an US packet communicated on an US channel. The US packet can be used to reevaluate the DS channel state. In detail, in some examples, the DS channel state may be reevaluated based on a receive property of a preamble of the US packet.

Such techniques of reevaluating the DS channel state based on the US packet may be referred to as implicit channel sounding. In one example, said reevaluating comprises updating or refining the DS channel state. In a further example, said reevaluating comprises checking a validity of the DS channel state.

By relying on the receive property of the preamble of the US packet to reevaluate the determined DS channel state, it is possible to reduce the overhead required for sounding of the DS channel. In detail, in may be possible to reduce a frequency of occurrence with which the DS sounding packets are communicated between the first and second devices. Explicit channel sounding can be implemented less frequently. On the other hand, a high temporal resolution of the sounding of the DS channel can be preserved by reevaluating the DS channel state based on the US packet. E.g., the US packet may be a packet communicated anyway on the radio channel, e.g., to deliver application-layer user data or control data of the radio channel; as such, the US packet may be a US data packet such as a US payload data packet or an US control data packet. The control data may, e.g., include one or more of the following: an acknowledgement message of an Automatic Repeat Request (ARQ) protocol, e.g., a block acknowledgement; a Request of Send (RTS) / Clear to Send (CTS) collision avoidance control message; etc.. In some examples, the US packet may not correspond to a dedicated sounding packet; in other examples, the US packet may correspond to an US sounding packet implemented, e.g., implemented by a null data packet (NDP) not comprising a payload section.

The techniques described herein rely on the finding that even for scenarios where reciprocity between the US channel and the DS channel is not given - i.e., where the radio channel is non-reciprocal -, implicit / relative adjustments to the determined DS channel state can still be reliably performed by said reevaluating of the DS channel state based on the US packet. As such, the receive properties of the preamble of the DS sounding packet can provide a reference baseline based on which the reevaluating of the determined DS channel state can be performed taking into consideration the receive property of the preamble of the US packet.

The techniques described herein enable to reduce the overhead on the radio channel. Thereby, it is also possible to reduce energy consumption. As such, techniques described herein may facilitate Internet of Things (loT) applications.

FIG. 1A illustrates a radio channel 120 implemented between a device 101 and a device 1 1 1. In the example of FIG. 1 A, the device 101 implements an IEEE 802.1 1 x Wi-Fi station (STA); also the device 1 1 1 implements a Wi-Fi STA.

While, hereinafter, various techniques will be described with respect to the Wi-Fi RAT, respective techniques may be readily employed for other kinds and types of RATs. Examples include the Third Generation Partnership Project (3GPP) 2G, 3G, 4G, and upcoming 5G RATs, Bluetooth, and Satellite communication.

In the example of FIG. 1A the DS channel 121 is arbitrarily defined from the STA 101 to the STA 1 1 1 ; and the US channel 122 is arbitrarily defined from the STA 1 1 1 to the STA 101 . However, it should be appreciated, that this is an arbitrary definition and that in other examples the DS direction may as well be defined from the STA 1 1 1 to the STA 101 and the US direction may be defined from the STA 101 to the STA 1 1 1. The STA 101 comprises a processor 105 and an interface 106. The interface 106 may comprise a radio transceiver having an analog stage and/or a digital stage. The interface 106 is configured to transmit and/or receive (communicate) data on the radio channel 120. The processor 105 is configured to perform techniques as described herein with respect to channel sounding. In detail, the processor 105 is configured to transmit a DS sounding packet, determine a DS channel state of the radio channel 120 based on a receive property of the preamble of the DS sounding packet, receive an US packet, and to reevaluate the determined DS channel state based on at least one receive property of a preamble of the US packet.

The STA 1 1 1 comprises a processor 1 15 and an interface 1 16. The interface 1 16 may comprise a radio transceiver having an analog stage and/or a digital stage. The interface 1 16 is configured to communicate data on the radio channel 120. The processor 1 15 is configured to perform techniques as described herein with respect to channel sounding. In detail, the processor 1 15 is configured to receive a DS sounding packet, report at least one receive property of the DS sounding packet to the STA 101 , transmit in US packet, etc.

FIG. 1 B illustrates aspects with respect to a plurality of time-space streams 125, 126 implemented on the radio channel 120. In detail, as illustrated in FIG. 1 B, the interface 106 comprises an antenna array 160 comprising two antennas. Differently, in the example of FIG. 1 B, the interface 1 16 does not comprise an antenna array, but comprises a single antenna. In this specific configuration, two time-space streams 125, 126 are defined between the STA 101 and the STA 1 1 1 .

The particular antenna configuration is illustrated in FIG. 1 B is non-limiting. In other examples, it is possible that both devices 101 , 1 1 1 comprise antenna arrays. It is also possible that the each antenna array comprises a larger number of antennas, e.g. more than two antennas, more than ten antennas or even more than 50 antennas. The number of time-space streams 125, 126 generally depends on the number of antennas implemented by the interfaces 106, 1 16.

In general, the DS channel state of the DS channel 121 and the US channel state of the US channel 122 are non-reciprocal. I.e., it is possible that the channel state 121 of the DS channel is at least partly different from the channel state of the US channel 122. I.e., it is possible that the channel state 121 of the DS channel is different from the channel state of the US channel 122 with respect to one or more figures of merit. Because of this lack of reciprocity, in reference implementations, separate and frequent sounding of the DS channel 121 and the US channel 122 based on DS sounding packets and US sounding packets, respectively, is implemented. This increases overhead. Some reference implementations also rely on channel calibration in order to achieve reciprocity. However, it has been found that such calibration of the radio channel 120 is prone to failures and may show significant time-dependencies such that re- calibration is required comparably often. This, in turn, again increases overhead.

Therefore, hereinafter, techniques are described which enable accurate sounding of the DS channel 121 without the need of the radio channel 120 having reciprocity. I.e., in some examples of the techniques described herein, it is possible that the radio channel 120 is non- reciprocal. However, the techniques described herein may also be applied to radio channel 120 being reciprocal.

FIG. 2 illustrates aspects with respect to the transmission protocol stack 130 of the respective RAT employed for communicating on the radio channel 120 - 122. In FIG. 2, the three lowest layers 131 - 133 of the transmission protocol stack 130 are depicted. Layer 1 is the Physical layer (PHY) 131 . Layer 2 is the Data Link Layer (DLL) 132. Layer 3 is the Network layer 133. Higher layers can include, e.g.: the transport layer; and the application layer (both not shown in FIG. 2).

In the example of FIG. 2 - which corresponds to the IEEE 802.1 1 x Wi-Fi protocol - the DLL 132 is subdivided into the Medium Access Control layer (MAC) 132-1 and the Logical Link Control layer (LLC) 132-2. The MAC 132-1 is responsible for implementing the ARQ protocol. Control data may originate from the MAC 132-1 .

In some examples, the packets discussed herein, e.g., the DS sounding packet and/or the US packet, may be defined with respect to the PHY 131. In other examples, the packets, e.g., the DS sounding packet and/or the US packet, discussed herein may be defined with respect to the DLL 132.

FIG. 3A illustrates aspects of reevaluating the DS channel state based on a receive property of a preamble of an US packet 21 1 - 214. In detail, FIG. 3A illustrates DS channel state 499 which is initially determined based on a receive property of a preamble of a DS sounding packet 201 and which is further reevaluated based on at least one receive property of a preamble of US packets 21 1 - 214 over the course of time. In the example of FIG. 3A, the DS sounding packet 201 is initially communicated from the STA 101 to the STA 1 1 1 on the DS channel 121 . The STA 1 1 1 then sends a US report message to the STA 101 , the US report message being indicative of the at least one receive property of a preamble of the DS sounding packet 201 (in FIG. 3A, the report message is not illustrated).

Based on that at least one receive property, the STA 101 may explicitly and accurately sound the DS channel 121 including all time-space streams 125, 126. In particular, the STA 101 may accurately determine the DS channel state 499 at the time of communication of the DS sounding packet 201 .

Then, the time evolution of the DS channel state 499 can be tracked by means of the US packets 21 1 - 214. In detail, the STA 101 receives the US packets 21 1 - 214 communicated from the STA 1 1 1 to the STA 101 on the US channel 122. Based on at least one receive property of the preamble of each one of the US packets 21 1 - 214, the STA 101 can then reevaluate the determined DS channel state.

In the example of FIG. 3A, the radio channel 120 is non-reciprocal. Hence, it is not possible to ab initio conclude back on the DS channel state 499 based on information derived solely from each one of the US packets 21 1 - 214 alone. Therefore, the DS channel state 499 - as determined based on the at least one receive property of the preamble of the DS sounding packet 201 - is used as a reference baseline 251 (indicated in FIG. 3A by the dashed line). Then, based on the at least one receive property of the preamble of the US packet 21 1 - 214, a relative deviation 252 from the reference baseline 251 can be determined. Thereby, it is possible to track the temporal evolution of the DS channel state 499 based on, both, the at least one receive property of the preamble of the DS sounding packet 201 , as well as based on the at least one receive property of the preamble of the US packets 21 1 - 214. Generally, said reevaluating of the determined DS channel state 499 may be based on, both, the at least one receive property of the preamble of the DS sounding packet and the at least one receive property of the preamble of the US packet.

While such a relative deviation 252 of the DS channel state 499 may have a limited accuracy - in particular, the accuracy may be lower if compared to the scenario where the DS channel state 499 is explicitly sounded each time based on a dedicated DS sounding packet - it is possible to reduce the overhead by limiting the number of DS sounding packets. Still, a more or less rough estimate of the temporal evolution of the DS channel state 499 can be obtained based on the US packets 21 1 - 214. As illustrated in FIG. 3A, a temporal resolution of the reevaluating of the determined DS channel state 499 corresponds to a temporal offset 241 between adjacent US packets 21 1 - 214. E.g., the temporal offset 241 may be smaller than 500 milliseconds, preferably smaller than 100 milliseconds, more preferably smaller than 20 milliseconds.

As is apparent from FIG. 3A, the US packet 21 1 is communicated shortly after communication of the DS sounding packet 201 . Then, for a certain duration 242, said reevaluating of the DS channel state 499 continues. E.g., the STA 101 may be configured to receive US packets 21 1 - 214 and reevaluate the DS channel state 499 at least for 5 milliseconds after transmitting the DS sounding packet 201 , preferably for at least 15 milliseconds, more preferably for at least 50 milliseconds.

Different implementations of said reevaluating of the determined DS channel state 499 based on the at least one receive property of the preamble of the US packets 21 1 - 214 are conceivable.

In a first example, it is possible to check a validity of the determined DS channel state 499 based on the at least one receive property of the preamble of the US packets 21 1 - 214. In other words, it could be possible to monitor the magnitude of the deviation 252. If the magnitude of the deviation 252 exceeds a certain threshold, it could be possible to conclude that the initially determined DS channel state 499 is not valid anymore. Then, appropriate action may be taken, such as explicitly sounding the DS channel 121 anew: E.g., it is possible that a further DS sounding packet 202 is selectively transmitted from the STA 101 to the STA 1 1 1 depending on said checking of the validity. Then, a further DS channel state 499 of the DS channel 121 can be determined based on at least one receive property of the preamble of the further DS sounding packet 202. In such a scenario, it is not required to communicate DS data packets on the DS channel 121 according to a value of the DS channel state 499 which is refined based on the at least one receive property of the preamble of the US packets 21 1— 214 if compared to the baseline 251. In other words, it is possible that during the duration 242, communication on the DS channel 120 is performed based on a value of the DS channel state 499 corresponding to the baseline 251 . Here, said reevaluating may be restricted to checking whether this value is still up-to-date.

In a second example, it is, however, also possible to update / refine the determined value of the DS channel state 499 based on the at least one receive property of the preamble of the US packets 21 1 - 214. Here, communication of DS data packets on the DS channel 121 can be continuously adjusted based on the refined values.

The scenario of FIG. 3A corresponds to a first operational mode 391. In the first operational mode 391 , the DS channel state 499 is reevaluated based on the US packets 21 1 - 214. Because such a reevaluation is available, a time sequence 220 of the DS sounding packets 201 , 202 has a comparably low temporal density of DS sounding packets 201 , 202. I.e., the number of DS sounding packets 201 , 202 per time unit is comparably low in the first operational mode 391 . A time-offset 245 between subsequent DS sounding packets 201 , 202 is long.

FIG. 3B corresponds to a second operational mode 392. Here, reevaluating of the DS channel state 499 is not implemented based on US packets. Hence, channel sounding of the DS channel 121 is implemented explicitly and relies solely on DS sounding packets 201 - 204. As is apparent from a comparison of FIGs. 3A and 3B, the time sequence 220 of DS sounding packets 201 - 202 in the first operational mode 391 has a larger number of DS sounding packets 201 , 202 per time than the time sequence 220 in the second operational mode 392. In the second operational mode 392, a shorter time offset 245 between subsequent US sounding packets 245 is obtained. Thus, while overhead for sounding the DS channel 122 is reduced in the first operational mode 391 , the accuracy may be reduced if compared to the channel sounding in the second operational mode 392.

In some examples, it is possible that the STA 101 dynamically switches between the modes of operation 391 , 392. Various decision criteria may be taken into account for said switching, e.g., an absolute value of the DS channel state 499. E.g., the DS channel state 499 is comparably degraded, it is possible that accurate channel sounding is desirable such that the STA 101 operates in the second operational mode 392. Where the channel state 499 is not degraded, the first operational mode 391 may be activated.

In the example of FIGs. 3A and 3B, the DS sounding packets 201 - 204 are implemented by NDPs. A NDP does not comprise a payload section. In other examples, the DS sounding packets 201 - 204 may also comprise a payload section.

FIG. 4 illustrates schematically a NDP that may be used as a DS sounding packet 201 - 204. The NDP 201 - 204 comprises a preamble 421 . The NDP 201 - 204 does not comprise a payload section. The preamble 421 comprises training symbols 422. In detail, the preamble 421 comprises a long training field (LTF) comprising a comparably large number of training symbols. In detail, the number of training symbols 422 comprised in the preamble 421 is not smaller than a number of the time-space streams 125, 126 implemented on the radio channel 120. This allows accurate sounding of the DS channel 121 . E.g., the training symbols may or may not be modulated. The training symbols may not be coded or scrambled. E.g., in some examples the training symbols may be Orthogonal Frequency Duplex Modulated (OFDM) symbols. They may serve for synchronization purposes, e.g., in addition to the channel sounding: here, the beginning of the respective packet may be indicated by the training symbols. The training symbols may be pre-negotiated, i.e., the respective receiver may expect a certain sequence of training symbols to be received.

FIG. 5 schematically illustrates a data packet that may be used as a US packet 21 1 - 214 based on which the determined DS channel state 499 is reevaluated. The data packet - different to the NDP - comprises a payload section 433. The payload section 433 comprises application-layer user data or control data of the radio channel 120. E.g., the control data may correspond to MAC 132-1 control data, e.g., an ARQ or CTS message.

The data packet 21 1 - 214 also comprises a preamble 431 . The preamble 431 comprises training symbols 432. In detail, the preamble 431 comprises a training field (TF) comprising a comparably small number of training symbols if compared to the LTF. In detail, the number of training symbols 432 comprised in the preamble 431 is smaller than a number of time-space streams 125, 126 implemented on the radio channel 120. This allows to limit overhead introduced by the preamble 431. While, in the example of FIG. 5, a scenario is illustrated where the number of training symbols 432 of the TF of the preamble 431 of the data packet 21 1 - 214 is smaller than the number of time-space streams, in other examples, it is also possible that a larger number of training symbols 432 is included in the preamble 431 . E.g., it would be possible that the number of training symbols 432 included in the preamble 431 is not smaller than the number of time- space streams 125, 126 of the radio channel 120. In such an example, accurate reevaluating of the determined DS channel state 499 based on the at least one receive property of the preamble 431 of the US packet 21 1 - 214 becomes possible.

Generally, the number of training symbols 432 included in the preamble 431 of the US packet 21 1 - 214 may be smaller, equal to, or larger than the number of training symbols 422 included in the preamble 421 of the DS sounding packet 21 1 - 204. It is generally not required that the US packet 21 1 - 214 used for reevaluating the determined DS channel state 499 is a data packet. In further examples, it is possible that a NDP is employed as the US packet for reevaluating the DS channel state 499. Here, a particularly large number of training symbols 432 may be included in the preamble 431 which facilitates accurate reevaulating.

FIG. 6 is a signaling diagram illustrating aspects of reevaluating the DS channel state 499 based on at least one receive property of the preamble 431 of an US packet 21 1 , 212.

First, at 1001 , a NDP implementing a DS sounding packet 201 is communicated from the STA 101 to the STA 1 1 1 . The STA 1 1 1 then determines the feedback matrix 1002 and reports the feedback matrix as a US report message 209 back to the STA 101 , see 1002 and 1003. The feedback matrix corresponds to the at least one receive property of the preamble 421 of the NDP 201 .

Based on the feedback matrix indicated by the report message 209, the DS channel state 499 is determined by the STA 101 . In detail, at 1004, the STA 101 determines the steering matrix. The steering matrix defines antenna weights of different antennas of an antenna array of the interface 106. The steering matrix is used in order to implement beamforming at the STA 101. Beamforming employs MIMO capabilities of the STA 101 to bundle energy into the direction at which the STA 1 1 1 is positioned. This increases a transmission reliability.

Meanwhile, US communication of data is ongoing: in detail, at 1005, and US data packet 21 1 is communicated from the STA 1 1 1 to the STA 101 . Based on the at least one receive property of the preamble 431 of the US data packet 21 1 , the steering matrix is refined, 1006. At 1007, a further US data packet 21 1 is communicated from the STA 1 1 1 to the STA 101 . Based on the at least one receive property of the preamble 431 of the US data packet 212, the steering matrix is further refined, 1008.

Then, a further DS sounding packet 202, again implemented as a NDP, is transmitted by the STA 101 and received by the STA 1 1 1 . 1010 - 1012 correspond to 1002 - 1004, respectively.

E.g., the further DS sounding packet 202 may be triggered by timeout of a timer initialized at 1001 . E.g., the further DS sounding packet 202 may be triggered by a timing pattern of communicating DS sounding packets. In another example, refining of the steering matrix at 1008 may be based on the deviation 252 which is larger than a threshold (cf. FIG. 3A).

In the example of FIG. 6, the US data packets 21 1 , 212 are communicated in response to the need of delivering US data to the STA 101. Transmission occurrences of the US data packets 21 1 , 212 may be controlled by the MAC 132-1 according to reference techniques. If no US data is available, zero padding may be employed.

In other examples, it is also possible that the US packets 21 1 - 214 used for reevaluating the DS channel state 499 are scheduled. Here, the persistently scheduled timing of repetitive transmission occurrences or a dedicated scheduled transmission occurrence defined for each individual US packet 21 1 - 214 can be employed.

FIG. 7 is a signaling diagram illustrating aspects of reevaluating the DS channel state 499 based on at least one receive property of the preamble 431 of an US packet 21 1 , 212. FIG. 7 generally corresponds to FIG. 6. However, instead of using US data packets, NDPs 21 1 and 212 are used at 1 105, 1007 for reevaluating the determined DS channel state 499. The timing 245 of the transmission occurrences of the US NDPs 21 1 , 212 is persistently scheduled, e.g., as part of transmission setup 201 between the STAs 101 , 1 1 1 at 1 100. The respective timing 241 is thus predefined.

In other examples, the STA 101 may transmit respective dedicated request messages to the STA 1 1 1 , the request message is prompting transmission of a US NDP. This corresponds to a dedicated scheduled transmission occurrence.

1 101 - 1 104 corresponds to 1001 - 1004. 1 106 corresponds to 1006. 1 108 corresponds to 1008. 1 109 - 1 1 12 corresponds to 1009 - 1012, respectively.

FIG. 8 illustrates aspects with respect to the interfaces 106, 1 16. In detail, in the example of FIG. 8, the interfaces 106, 1 16 comprise an antenna array 160. Each antenna 160 is associated with a certain antenna weight 161 defining amplitude and phase of the respective signal handled by that antenna. The antenna weights correspond to the steering matrix. The antenna weights 161 may be determined based on the DS channel state. The antenna array 160 may facilitate MIMO techniques. In detail, a plurality of time-space streams may be implemented based on the antenna array 160. Thereby, SU or MU beamforming can be implemented. FIG. 9 is a flowchart of a method according to various examples. At 2001 , a DS sounding packet is transmitted, e.g., from the STA 101 to the STA 1 1 1 or vice versa. The DS sounding packet may be transmitted according to a predefined timing. The DS packet may be a NDP.

Then, at 2002, the DS channel state is determined based on the at least one receive property of a preamble of the DS sounding packet communicated at 2001. Here, a US report message indicative of the at least one receive property may be received.

Next, 2003, an US packet is received, e.g., a NDP or a data packet. Based on at least one receive property of the preamble of the US packet received at 2003, the DS channel state is reevaluated at 2004.

FIG. 10 is a flowchart of a method according to various examples. In detail, the flowchart of FIG. 10 illustrates a possible implementation of reevaluating the DS channel state determined at 2004.

Here, at 201 1 , it is checked whether the deviation 252 from a reference baseline 251 associated with the at least one receive property of the preamble of the DS sounding packet is smaller than a threshold (cf. FIG. 3A). If this is the case, 2003 is re-executed, i.e., a further US packet is received and evaluated. If this is not the case, transmission of a further DS sounding packet is triggered by re-executing 2001.

In the example of FIG. 10, it is not required that the value of the DS channel state 499 as determined at 2002 is adjusted at 201 1 . Rather, said reevaluating of the DS channel state 499 may be restricted to whether the channel state changes significantly such that new channel sounding based on a further DS sounding packet is required.

FIG. 1 1 is a flowchart of a method according to various examples. In detail, the flowchart of FIG. 1 1 illustrates a possible implementation of reevaluating the DS channel state determined at 2002. Here, at 2021 , the value of the DS channel state 499 is refined / updated. I.e., the value of the DS channel state 499 can be changed. Hence, subsequent communication on the DS channel 121 uses new parameters, e.g., as defined by the steering matrix. At 2022, it is then checked whether the communication of a further US packet is expected, e.g., according to a persistently scheduled timing 241 . If this is not the case, a further DS sounding packet is transmitted at 2001 in order to determine a new DS channel state 499. Otherwise, the further US packet is received at 2003 and 2021 , 2022 are executed anew. E.g., a dedicated request message may be communicated to trigger the further US packet in some examples. Summarizing, above techniques of implicit channel sounding have been illustrated. The implicit channel sounding is based on using, at the transmitter, original channel knowledge from direct sounding of the channel together with transmission coming back on the radio channel from the receiver. These techniques enable to implement reliable communication on a bi-directional channel without a need of error-prone calibration techniques. It is not required that the bi-directional channel is reciprocal.

These techniques enable to reduce the overhead due to channel sounding. In detail, sounding packets may be transmitted less frequently. Nonetheless, the channel state may be updated at a high temporal resolution based on implicit techniques. Mutual interference and deteriorated user reception may be avoided.

Although the invention has been shown and described with respect to certain preferred embodiments, equivalents and modifications will occur to others skilled in the art upon the reading and understanding of the specification. The present invention includes all such equivalents and modifications and is limited only by the scope of the appended claims.