WO2009147570A1

WO2009147570A1 - Asynchronous multi-user transmission

Info

Publication number: WO2009147570A1
Application number: PCT/IB2009/052190
Authority: WO
Inventors: Ying Wang; Semih Serbetli
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2008-06-02
Filing date: 2009-05-26
Publication date: 2009-12-10
Also published as: TW201014230A

Abstract

The present invention relates to a transmitter device, a receiver device, a computer program product, and a method of performing multi-user transmission, wherein individual transmission channels from multiple transmission sides to a single receiver side are estimated and multi-user interference cancellation is established based on the channel estimation at the receiver side. Additionally, a characteristic for single user equalization is designed and signaled by the receiver side to all the transmission sides to thereby distribute receive processing among both transmission sides and the receiver side.

Description

Asynchronous Multi-User Transmission

FIELD OF THE INVENTION

The present invention generally relates to a transmitting apparatus, a receiving apparatus, a system and a method of performing multi-user transmission between a transmission end and a plurality of other transmission ends in a transmission systems such as - but not restricted to - a wireless local area network (WLAN).

BACKGROUND OF THE INVENTION

Wireless local area networks (WLANs) as defined e.g. in the IEEE 802.11 specifications are almost omnipresent today. The increase of throughput of the available channel was one major issue, and research has been focused on improving the modulation and coding within the Physical Layer. By employing orthogonal frequency division multiplexing (OFDM) in combination with high-rate signal constellations, up to 54 Mbit/s could be achieved . This huge performance jump - even if achieved only for very limited distances - is caused by to inherent features of OFDM, which have become especially attractive for high bit-rate systems. In OFDM, the given system bandwidth is split into many sub-channels, also referred to as sub-carriers. Instead of transmitting symbols sequentially through one (very broad) channel, multiple symbols are transmitted in parallel. This leads to much longer symbol durations, such that the impact of inter- symbol interference can be reduced significantly, so that no additional measures like costly equalization are necessary. The 802.11 standard makes it mandatory that all stations or user terminals implement a distributed coordination function (DCF) which is a form of carrier sense multiple access with collision avoidance (CSMA/CA). CSMA/CA is a contention-based protocol making certain that all stations first sense the medium before transmitting. The main goal is to avoid having stations transmit at the same time, which results in collisions and corresponding retransmissions. If a station wanting to send a frame senses energy above a specific threshold on the medium (which could mean the transmission of another station), the station wanting access will wait until the medium is idle before transmitting the frame. The collision avoidance aspect of the protocol pertains to the use of acknowledgements that a receiving station sends to the sending station to verify error-free reception. Although somewhat more complex, this process of accessing the medium can be seen as a meeting where everyone is polite and each person only speaks when no one else is talking. In addition, participants who understand what the person is saying nod their head in agreement. Because of its nature, DCF supports the transmission of asynchronous signals. A distinguishing factor of asynchronous signaling is that there are no timing requirements between data carrying frames. For example, the DCF protocol doesn't make any attempt to deliver a series of data frames within any timeframe or at any instant in time. As a result, there is a random amount of delay between each data frame transmission. This form of synchronization is effective for network applications, such as e-mail, Web browsing and VPN access to corporate applications.

A potential for further bit-rate increases is seen in a use of multiple-input multiple-output (MIMO) antenna systems. Hence, a new Medium Access Control (MAC) protocol mechanism has been proposed, which supports multi-user (MU) MIMO transmissions in WLANs according to IEEE 802.11 based standards. The proposed new protocol extends the DCF with single-user (SU) MIMO in such a way that different stations can be destination stations for packets inside a MIMO frame (which is a set of packets transmitted simultaneously on different spatial streams).

The standard includes an optional Request-to-Send (RTS) - Clear-to-Send (CTS) handshake prior to the transmission. In an association procedure prior to data transmissions, stations share among each other the information about their hardware capabilities. Information concerning used antenna elements can be exchanged using an extended form of RTS and CTS control frames. The extended RTS frame - MIMO-RTS (M- RTS) and the extended CTS frame - MIMO-CTS (M-CTS) can be based on the structure of the IEEE 802.11a standard RTS and CTS frames. In order to support multiple antennas, both have a new field, e.g. a bitmap, where each bit stands for one antenna. A bitmap of a length of one byte can thus support up to eight antennas. Of course, the bitmap field can be longer or shorter depending on the number of antennas supported by the mobile stations of a given system. In the M-RTS frame this field may be called Proposed Antenna Bitmap (PAB) and may encode the chosen subset of available antennas proposed for the following transmission. The receiver of the frame confirms which antennas should be active in a Confirmed Antenna Bitmap (CAB) field of the M-CTS frame. The ACK frame is also extended to support per- stream acknowledgements. More specifically, the MIMO-ACK (M-ACK) frame may have a one byte long bitmap field called Acknowledged Packet Bitmap (APB) to confirm the reception of each packet from different streams separately. It contains positive and negative acknowledgements for each spatial stream. It can still be immediately acknowledging, although there are multiple packets being transmitted at a time. The length of the bitmaps (L) can be arbitrary.

Multi-user (MU) MIMO increases the spectral efficiency of wireless networks and will be of high interest for the next generation WLAN systems. The use of cyclic prefix as guard period in communication systems such as OFDM and single carrier block transmission simplifies frequency domain equalization in broadband channels by turning linear convolution with a channel into circular convolution. However, in general for frequency domain approaches to be effective in the presence of interference, e.g., in an uplink multi-user transmission scenario, the interfering users must be synchronized with the desired user, i.e., the packets from multiple users should arrive at the receiver with time differences smaller than the OFDM guard period. In this invention we will refer to the time alignment of multiple uplink users within OFDM guard period as the multi-user synchronization.

However, MAC protocols of future multi-user WLAN systems will probably not support multi-user synchronization due to at least one of hardware limitations (e.g. clock accuracy), increased PHY/MAC overhead of signaling for synchronization purposes, and backward compatibility problems faced when modifying the current MAC protocol to achieve synchronized transmission from multiple users. If the requirement of multi-user synchronization is not satisfied, then a receiver needs long time domain filters to suppress the interfering users and to equalize the desired user, as described for example in M. Breinholt, H. Jung, and M. Zoltowski, "Space-time alignment for asynchronous interference suppression in MIMO OFDM cellular communications", Wirel. Commun. Mob. Comput. 2004. Such a receiver complexity may not be affordable by the WLAN access point while serving multiple users. The WLAN MAC protocols do not and very likely will not support multi-user synchronization.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a multi-user transmission scheme which does not require strict synchronization, while keeping receiver complexity reasonable.

This object is achieved by a receiving apparatus as claimed in claim 1, a transmitting apparatus as claimed in claim 8, and a method as claimed in claim 10.

Accordingly, an access device is allowed to serve multiple asynchronous users by providing a simple and efficient transceiver scheme where the access point is serving multiple users without strict multi-user synchronization requirement, i.e., packets from multiple users can arrive at the access point with time differences e.g. larger than the OFDM guard period. More specifically, it is proposed to separate the receive processing into multiuser interference cancellation and single user equalization. The single user equalization could be further separated into time domain partial equalization (channel shortening) that can be implemented as a transmit filter, and a conventional frequency domain equalization performed at the receiver side. This greatly reduces the processing load at the access point as a receiver since each user as a transmitter contributes to part of the equalization work.

Furthermore, a MAC protocol supporting the uplink MU-MIMO transmission is proposed e.g. for 802.11 based WLAN systems. It adds a MU-MIMO transmission set-up phase during which some control frames are exchanged between the access point and multiple user terminals. This allows the access point to have the channel state information (CSI) of each user available before the actual multi-user transmissions, and also provides a mechanism for the access point to feed back the above filter characteristic single user equalization or some instruction of transmission to the users. An enhanced MAC frame, i.e., the MU-RTS, is defined. This frame is different from the ordinary RTS frame because it has multiple recipient MAC addresses. This enables an improved way of communicating the list of identifications or addresses to the other transmission ends. Although the proposed enhanced MAC frame has specific fields which are only meaningful/understandable to MU devices, the frame can be transmitted in the legacy physical layer and has common fields, that are understandable by all legacy devices. Therefore, legacy devices can decode the bits, interpret common fields and initiate appropriate settings. The interpretation of the enhanced MAC frame may be a pure MAC process, so that no further information is required from the physical layer. Moreover, there is no need to change interpretation rules for corresponding existing or legacy MAC frames. In view of the fact that all other transmission ends can be at least partially interpreted by all other transmission ends, its transmission can be regarded as a broadcast transmission from the physical layer perspective. Consequently, legacy devices and procedures require little modifications.

According to a first aspect, the transmission channels of each user terminal, i.e., each transmission end, may be estimated at the access point (AP), i.e., the receiver end, by processing training sequences received in transmission requests (e.g. M-RTS) from all the transmission ends. Thereby, information provided in an initial set-up phase can be advantageously used without requiring additional processing load. According to a second aspect which may be combined with the first aspect, the receiver processing of extracting the signals of a desired user terminal can be separated into a multi-user interference cancellation stage and a single-user equalization stage.

According to a third aspect which may be combined with at least one of the above first and second aspects, the multi-user interference cancellation is implemented by user-specific interference cancellation filter circuits calculated from the estimated transmission channels.

According to a fourth aspect which may be combined with at least one of the above first to third aspects, the single user equalization may be separated into a time domain partial equalization corresponding to the individual characteristic, and a frequency domain equalization performed at the receiving apparatus (e.g. AP). This distributed equalization approach leads to a reduced processing load at the receiver side.

According to a fifth aspect which may be combined with at least one of the above first to fourth aspects, the individual characteristic may be determined based on estimated transmission channels, and based on a characteristic of the multi-user interference cancellation filter circuits. As the required information is readily available, a straight forward transmit filter design can be implemented.

According to a sixth aspect which may be combined with at least one of the above first to fifth aspects, the individual characteristics may be determined so that a length of non-zero filter taps is smaller than a guard period of the multi-user transmission signal, or so that a power ratio of a first set of filter taps to a second set of filter taps is minimized. Thus, transmit filter can be designed to meet desired properties.

Further advantageous developments are defined in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will now be described based on various embodiments with reference to the accompanying drawings in which:

Fig. 1 shows a schematic block diagram of multi-user MIMO transmission system according to an embodiment; Fig. 2 shows a MU MIMO uplink mechanism according to a fifth embodiment;

Fig. 3 shows a C4T frame structure which can be used in the embodiment;

Fig. 4 shows an M-RTS frame structure which can be used in the embodiment; Fig. 5 shows an MU-CTS frame structure with Tx beamforming vectors which can be used in the embodiment;

Fig. 6 shows an MU-ACK frame structure which can be used in the embodiment; and Fig. 7 shows a schematic signaling and processing diagram according to an embodiment;

DETAILED DESCRIPTIONOF EMBODIMENTS

In the following, preferred embodiments are described on the basis of a MU MIMO system as shown in Fig. 1.

According to Fig. 1. a MU MIMO access point (AP) provides WLAN access for a plurality of N stations comprising respective single-user transmission (TX) chains 10-1 to 10-N, respective user-specific transmit filters 12-1 to 12-N, and respective antennas. The AP comprises N different single-user receiving (RX) chains 20-1 to 20-N for providing different coding and/or modulating schemes, to which input signals are applied via respective user- specific filter circuits 22-1 to 22-N and a plurality of antennas 1 to N₁-.

In general, two types of MIMO techniques can be used in both directions between the AP and each of the stations based on the propagation channel properties, i.e. the structure of the spatial correlation matrix at the receiver's antenna array. In case of high correlation of the received signal different beamforming algorithms can be applied, while in case of low correlation of the received signal - diversity (DIV) and multiplexing (MUX) approaches may give better performance. In MUX schemes, multiple streams are transmitted simultaneously, each using one dedicated antenna. This increases the throughput with a factor equal to the number of streams being transmitted. In DIV schemes, multiple antennas are used in a different way. For the basic DIV scheme the transmitter uses only one antenna. The receiver with multiple antennas receives multiple copies of the transmitted signal so that using an appropriate signal processing algorithm achieves significantly higher signal-to-noise ratios (SNRs). In the schemes combining MUX and DIV, more transmit antennas are active, but the receiver, as in all DIV schemes, may still have more antennas than the number of streams. Multiplexing is present, but the receiver gets more information about the transmitted signal than in the pure MUX case.

By the following embodiments, a simple and efficient physical layer transceiver scheme is implemented e.g. for uplink MU-MIMO WLAN systems, where the AP is serving multiple users without strict multi-user synchronization requirement, i.e., the packets from multiple users can arrive at the access point with time differences larger than the OFDM guard period.

The proposed transceiver scheme provides a simple way to separate the receive processing into a multi-user interference cancellation part, step, or stage and a single user equalization part, step, or stage. The multi-user interference cancellation stage can be implemented by the user-specific filter circuits 22-1 to 22-N at the receiver side. The single user equalization stage can be further separated into a time domain partial equalization (e.g. channel shortening) part, step or stage that can be implemented by the above user-specific transmit filters 12-1 to 12-N at transmitter sides, and a frequency domain OFDM equalization step performed by the user-specific single-user RX chains 20-1 to 20-N at the receiver side. This greatly reduces processing load at the AP - as a receiver - since each user - as a transmitter - contributes to part of the equalization processing.

As an implementation example, a multi-user uplink WLAN scenario is now considered with reference to Fig. 1. There are N stations or user terminals and each user terminal has one transmit antenna. There is one AP with N₁, receive antennas serving the N users. The physical channel from any transmit antenna to any receive antenna can be modeled for example as an FIR filter with L taps.

Without loss of generality, the processing of the upper user terminal 1 in Fig. 1 is now described in more detail. I.e., the signal received from user terminal 1 is therefore treated as the desired signal and the signals received from the other N - I user terminals are treated as interference signals. The receiver at the AP uses its filter circuit 22-1 (which may be a digital filter, such as for example a finite impulse response (FIR) filter) characterized by w⁽¹⁾ with M taps per receive antenna to nullify the interfering users for user terminal 1. The filtering operations of the filter circuit 22-1 can be defined by using a matrix /z⁽¹⁾ and a column vector w⁽¹⁾ , which can be represented as follows:

KO) w(0) h⁽¹⁾ = KL -X) hφ) w⁽¹⁾ = (1) w(M - \) ACV, Xl h(L - l) (L+M-l)xMN_r where h(l) = [Zz₁ (l),...,h_N (I)] , I = 0,...,L - 1 , is a row vector containing the / -th taps of the physical channels from user terminal 1 to the N₁. receive antennas, and w(m) = \w_γ(m),..., w_N (m)]^τ , m = 0,...,M - 1 , is a column vector containing the m -th taps of the filter circuits (which could also be designated as "interference cancellation filters") for the user terminal 1 at the N₁. receive antennas. For the other user terminals , h⁽ⁿ⁾ and w⁽ⁿ⁾ , n = 2,...,N are defined similarly.

During a multi-user transmission set-up phase, the receiver estimates the channel of each user by processing training sequences contained in the M-RTS frames, as described in more detail later. Having the channel state information (CSI) from all users available, the characteristic w⁽¹⁾ can be calculated e.g. by using simple zero-forcing approach by solving equation (2): h ^mw^(l) = 0 , (2)

where /z ⁽¹⁾ =

is an interference channel matrix containing CSFs of user terminals interfering with user terminal 1. The solution of the filter characteristics w⁽¹⁾ can be a null space vector of the above matrix h ⁽¹⁾ . Existence of null space requires that the number of rows is at least one more than the number of columns in h ⁽¹⁾ , which gives M_1111n , the minimum number of taps needed per receive antenna for nullifying the interference of user terminal 1, namely:

(N - I)(I - I) ₊ I

^_mn = (3)

N - N + l

Using the interference channel matrix /z⁽¹⁾ and the calculated filter characteristic w⁽¹⁾ a filter characteristic g⁽¹⁾ of the transmit filter 12-1 can be designed such that the total effective channel of user terminal 1 has some desired properties for ease of equalization at the receiver, e.g., the length of non-zero taps is smaller than the OFDM guard period and/or the power ratio of certain taps to other taps is maximized. Although in the above embodiment, the filter characteristic g⁽¹⁾ relates to a transmit filter, a receive filtering technique for channel shortening, as described for example in P. Melsa, R. Younce, and C. Rohrs, "Impulse response shortening for discrete multitone transceiver", IEEE Trans. Commun., vol. 44, no. 12, Dec. 1996, could be applied similarly to calculate the filter characteristic g⁽¹⁾ . As can be gathered from this document, a channel shortening filter is a bit shorter than the target impulse response. Therefore, in the embodiment, the number of taps needed by the transmit filter 12-1 with characteristic g⁽¹⁾ may be selected roughly correspond to the guard period length. The receiver at the AP thus calculates the transmit filter characteristics g⁽ⁿ⁾ , n = \,...,N , for all user terminals based on the respective estimated channel matrix h⁽ⁿ⁾ and the calculated characteristic w⁽ⁿ⁾ . Finally, the receiver at the AP may utilize an MU-CTS frames proposed in to feed back the determined transmit filter characteristics g^(κ) to each user terminal.

Then, after the exchange of M-RTS and MU-CTS frames, the actual multiuser transmission can start. In the block diagram of Fig. 1, each user terminal may have a conventional single-user transmission chain followed by the additional user- specific transmit filter 12-1 to 12-N informed during the set-up phase by the AP. The AP first uses its user- specific filter circuits 22-1 to 22-N as multi-user interference cancellation filter to extract a concerned user terminal n , and may then apply a conventional single-user receiving chain processing to demodulate a signal received from user terminal n , n = \,...,N . This processing of the respective single user receiving chain 20-1 to 20-N may consist of synchronization, frequency domain channel estimation, OFDM demodulation, frequency domain equalization, etc., since for each user the multi-user interference has already been nullified by a respective one of the filter circuits 22-1 to 22-N with characteristic w⁽ⁿ⁾ , and the total effective channel is made shorter than the guard period by the transmit filter characteristic g^(κ) .

In the following, a MAC protocol enhancement is proposed that supports MU MIMO transmissions with beamforming in the uplink direction of MU transmission systems, such as IEEE 802.11 based WLANS. It is based on a MAC mechanism where a common receiver (e.g. the AP of Fig. 1) initiates a transmissions by broadcasting a call for transmissions (C4T) frame to candidate transmitters (e.g. the user terminals 1 to N of Fig. 1). The addressed candidate transmitters respond by sending M-RTS frames to show their intention to transmit to the common receiver with followed by training sequences for channel estimation at the receiver. As alternative, training sequences may be provided in respective preambles. The receiver estimates the channels from each transmitter and assesses the candidate transmitters according to their channel realizations. The receiver may also find appropriate transmit beamforming or filter vectors for each transmitter and may reply to the M-RTS frames with an MU-CTS frame where it may indicate which transmitters can access the channel by using which transmit beamforming or filter vectors. Then, the MU MIMO transmission can start.

The proposed new mechanism thus provides an uplink channel access mechanism for MU MIMO transmissions supporting transmit beamforming, where a common receiver supports simultaneous multiple packet reception from different transmitters. Thus, the spectral efficiency of the system can be increased. The embedded transmit beamforming mechanism provides a good MU MIMO transmission coordination among the multiple transmitters so that the interference among the spatial streams is minimized. Besides the new mechanism gives the possibility to switch between SU and MU MIMO transmission modes if the channel realizations are not appropriate for a MU MIMO transmission.

More specifically, in the proposed MU MIMO MAC mechanism for the uplink scenario C4T, M-RTS and MU-CTS frames may be used for accessing a channel, and a MU-ACK frame may be used for acknowledging correctly received packets. Optionally, an adaptive MU MIMO transmission may be provided by modifying transmit beamforming vectors for the MU MIMO transmission of only a subset of user terminals whose spatial streams are correctly received to build the next MIMO frame. The decision can be based on information gained from e.g. an error correction code (such as a cyclic redundancy code (CRC)) check of the received packets. In the following, the proposed MU MIMO uplink MAC procedure is described in more detail based on five steps as shown in Fig. 2.

In a first step, a common receiver (e.g. the AP of Fig. 1) broadcasts a call for transmissions (C4T) frame to initiate a MU MIMO transmission in the uplink. In the C4T frame, it indicates the addresses of all uplink MU MIMO capable user terminals, which is a variable number. Alternatively, the AP may decide to poll only a subset of uplink MU MIMO capable user terminals.

The C4T frame may also carry a request for sounding and an indication of the number of spatial dimensions to be sounded. Alternatively, the number of spatial dimensions to be sounded can be standardized to the channel estimation capability of the AP, which can be obtained from the HT capabilities field of the AP, which could be transmitted in beacon frames, association response frames, etc. The duration field may be set to cover the transmission duration up to the start of MU MIMO transmission. As explained above, this duration can be obtained from a summation of the duration of the responses from the user terminals, the duration of the MU-CTS frame and the SIFS/RIFS intervals separating the frames. Because the number of user terminals assigned for the MU MIMO transmission is not known when the C4T frame is constructed, the duration of MU-CTS frame is not known. For the calculation of the duration field of the C4T, the MU-CTS frame is assumed to contain the maximum number of fields and the MCS used is the same as the one used for the C4T frame transmission. It is noted that in the same frame, the format of the CSI report for each user terminal must be conveyed. The mechanism to convey this information can be the same as used in the downlink MU MIMO transmission. The duration must also take into account the CSI reports. The C4T frame can be transmitted regularly. The frequency of recurrence may then depend on the number of uplink MU MIMO capable APs. The frequency may be communicated to other APs in a beacon frame.

Fig. 3 shows an example of a C4T frame structure with multiple transmitter address fields, which can be used in the embodiment.

After each user terminal (transmitter or transmitting end) has received the call, e.g. the C4T frame, they respond in the second step of the procedure by sending an M-RTS packet or frame to indicate their intention to transmit to the indicated receiver. Here, "M-RTS frame" stands for M-DCF RTS frame and contains RTS frame fields and additional fields such as a CSI field. The M-RTS frame could be replaced by an aggregation of an RTS frame and a CSI feedback frame.

The order of M-RTS frames is implicitly determined by the transmitters' order in the list in the C4T frame. The first M-RTS frame is transmitted after an SIFS interval, and the following ones are transmitted after respective RIFS intervals. The physical protocol data unit (PPDU) carrying the M-RTS frame may be a sounding PPDU. The duration may be the summation of two durations, wherein the first duration starts at an SIFS interval after the completion of the M-RTS transmission up to the start of MU MIMO transmission and the second duration is the duration of the data frame transmission if the MCS used for the M- RTS frame would be used to transmit the pending data. From this duration field, the AP can learn about the amount of data to be sent by a user terminal and therefore it can set the duration field in the MU-CTS frame properly.

Fig. 4 shows an M-RTS frame structure which can be used in the embodiment. After the AP (receiver or receiving end) receives the M-RTS frames from the candidate transmitting user terminals and estimates the channel realizations of the user terminals, it assesses in the third step of the procedure the channel realizations of the user terminals for a possible MU MIMO transmission and finds an appropriate transmit beamforming vector for each user terminal or spatial stream. Next, it proceeds to the channel reservation by broadcasting an MU-CTS frame where it indicates which transmitters can access the channel by using which transmit beamforming vectors.

Fig. 5 shows an example of an enhanced MU-CTS frame structure with Tx beamforming vectors which can be used in the embodiment. Alternatively, a more generic MU-CTS frame could be used, which doesn't carry the Tx beamforming vectors. Then, an aggregated steering frame (compressed or non compressed) could be used to carry the Tx beamforming vectors. The MCS to be used in the transmission by the assigned user terminals may also be conveyed in these frames, e.g., in the HTC fields. The duration field may be set to the duration of the longest spatial stream plus an SIFS interval and the time needed to transmit the M-ACK frame.

In the fourth step of the procedure, the user terminals can access the channel by using the Tx beamforming vectors indicated in MU-CTS frame.

Finally, in the fifth step, after the MU MIMO uplink transmission is completed, the AP may transmit an MU-ACK frame where it acknowledges successful receptions of packets transmitted simultaneously by the assigned user terminals.

Fig. 6 shows a corresponding MU-ACK frame structure which can be used in the embodiment. This acknowledgment can be conveyed in the acknowledged packet bitmap (APB) field whose length is equal to the number of Tx addresses in the MU-CTS frame. A successful reception of a packet can be acknowledged e.g. by setting the bit corresponding to the transmitting station to "1".

With the above mechanisms, the receiver can initiate and coordinate the MU MIMO transmission in the uplink by finding the appropriate transmit beamforming vectors and feeding this information to the transmitters, thus, it provides an efficient channel access mechanism and an interference avoidance technique for MU MIMO uplink transmission. In the following, a mechanism is proposed to reduce overhead in MU-DCF.

Most of the overhead in MU-DCF is generated due to multiple M-CTS and M-ACK frame replies with their SIFS intervals and a preamble prior to each frame. Applying a multiple access scheme other then time division multiple access (TDMA) significantly improves the performance of the MU-DCF network. In MIMO systems it is possible to spatially multiplex the frames, but channel knowledge at the transmitter cannot be assumed. In OFDM systems, such as IEEE 802.1 Ia, the use of OFDMA transmissions leads to smallest hardware complexity. However, other schemes such as MC-CDMA or CDMA can have a similar effect. In case of OFMDA, by using e.g. a quarter of sub-carriers, short packets such as the M-CTS and M-ACK frame, are not four times longer - since the major part of the frame is the preamble. Depending on the physical layer, M-CTS and M-ACK frames are several symbols longer. Assuming a packet size of 1024byte, a physical layer mode for data packets of 54Mb/s, a physical layer mode of 36Mb/s (and other relevant parameters as in the IEEE 802.11a standard), the transmission window has a durations of 338//S in the SU mode, 578//S in the MU mode (TDMA), and 362/zs in the MU mode (OFDMA).

Thus, it is proposed to reduce the time needed to transmit the M-CTS and M- ACK frames in MU mode of operation in MU-DCF. This reduces the overhead in the MU almost to that of a SU system, while preserving the above mentioned benefits of MU-MIMO transmissions.

Instead of transmitting the M-CTS and M-ACK frames in a TDMA mode, all the sub-carriers are divided into subsets and each subset is assigned to one user terminal which has to send an M-CTS or M-ACK frame.

The information about the mapping of subcarrier subsets to user terminals can be determined from the order of the receivers in the address list conveyed in MU-RTS frame. Hence, M-CTS and M-ACK frames are transmitted simultaneously, so that the SIFS intervals and preambles which precede each frame in IEEE 802.11 networks and which are the carriers of the overhead in MU-MIMO systems are parallelized. Depending on the physical layer characteristics, the M-CTS and M-ACK frames may be only several symbols longer. Fig. 7 shows a schematic signaling and processing diagram according to an embodiment, wherein a user terminal or station (STA) 10-i communicates with an AP 20.

During an initial set-up stage, the STA 10-i transmits in step SlOl an M(U)- RTS with a training sequence to the AP 20. The AP 20 estimates in step S 102 based on the received training sequences from all the user stations the user channels, and calculates in step S 103 a user-specific characteristic for the interference cancellation filter for STA 10-i (e.g. filter circuits 22-i in Fig. 1) based on derived channel information. Then, in step S 104, a user- specific transmit filter (e.g. transmit filters 12-i in Fig. 1) is designed so that a total effective user channel with desired properties is obtained. The characteristic of the designed transmit filter in then fed back in step S 105 to the STA 10-i and can be used there to set up a user- specific transmit filter. Finally, in step S 107, multi-user transmission can be started based on the obtained filter settings at the STA 10-i and the AP 20.

In summary, a transmitter device, a receiver device, a computer program product, and a method of performing multi-user transmission have been described, wherein individual transmission channels from multiple transmission sides to the single receiver side are estimated and multi-user interference cancellation is established based on the channel estimation. Additionally, a characteristic for single user equalization is designed and signaled to the transmission sides to thereby distribute receive processing among both transmission sides and the receiver side. The channel estimation may be based on a training sequence received with a request for transmission, e.g. MU-RTS frame, MU-CFR frame, or C4T frame, and the characteristic for single user equalization may be signaled in a transmission acknowledgement, e.g. MU-CTS etc. Various advantageous further enhancements and improvements of this underlying general concept have been provided in the above embodiments. It is noted that the present invention is not restricted to the above embodiments and can be used for any multi-user transmission scheme, not only MU MIMO. More specifically, the invention is applicable to all types of MIMO based WLANs, particularly M- DCF systems. The protocol works in both single-user (SU) and MU mode. Performance improvements compared to M-DCF can be expected in highly interconnected systems and in AP downlink, where multiple connections are present. Moreover, the invention is applicable to all multi-user wireless systems. It is expected to facilitate multi-user traffic in wireless networks where the traffic is asynchronous and destined to a single receiver such as an AP in an uplink scenario or a gateway that provides access to an external network.

Variations to the disclosed embodiments can be understood and effected by those skilled in the art, from a study of the drawings, the disclosure and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality of elements or steps. A single processor or other unit may fulfill the functions of Figs. 1 and 5 and several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. A computer program used for controlling processor to perform the claimed method features may be stored/distributed on a suitable medium, such as an optical storage medium or a solid- state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope thereof.

Claims

CLAIMS:

1. A receiving apparatus for receiving multi-user transmission signals from a plurality of transmission ends, said apparatus (20) being adapted: to estimate transmission channels from each of said plurality of transmission ends; to set up multi-user interference cancellation filter circuits (22-1 to 22-N) based on said estimated transmission channels; to determine individual characteristics for single user equalization of said plurality of transmission ends; and to signal a filter information about said individual characteristics to said plurality of transmission ends.

2. An apparatus according to claim 1, wherein said apparatus is adapted to estimate said transmission channels by processing training sequences received in transmission requests from said plurality of transmission ends.

3. An apparatus according to claim 1, wherein said apparatus is adapted to calculate said interference cancellation filter circuits (22-1 to 22-N) to nullify the multi-user interferences for each of said plurality of transmission ends based on said estimated transmission channels.

4. An apparatus according to claim 1, wherein said apparatus is adapted to separate said single user equalization into a time domain partial equalization corresponding to said individual characteristic, and a frequency domain equalization performed at said receiving apparatus (20) .

5. An apparatus according to claim 1, wherein said apparatus is adapted to calculate said individual characteristic based on said estimated transmission channels, and based on a characteristic of said calculated multi-user interference cancellation filter circuits (22-1 to 22-N).

6. An apparatus according to claim 1, wherein said apparatus is adapted to determine said individual characteristics so that a length of non-zero filter taps is smaller than a guard period of said multi-user transmission signal, or so that a power ratio of a first set of filter taps to a second set of filter taps is minimized.

7. An apparatus according to claim 1, wherein said apparatus (20) is adapted to signal said filter information about said determined individual characteristics in a transmission acknowledgement to said plurality of transmission ends.

8. A transmitting apparatus for transmitting a transmission signal to an access point (20), said apparatus(lθ-i) being adapted: to receive a filter information from said access point (20); to set up a transmit filter (12-i) for partial single-user equalization based on said filter information; and to transmit said transmission signal via said transmit filter (12-i).

9. An apparatus according to claim 8, wherein said apparatus (10-1 to 10-N) is adapted to receive said filter information in an acknowledgement sent by said access point (20) in response to transmission requests sent by said apparatus (10-1 to 10-N).

10. A method for performing multi-user transmission between a plurality of transmission ends (10-1 to 10-N) and a single receiver end (20), said method comprising: estimating transmission channels from each of said plurality of transmission ends; setting up multi-user interference cancellation for said plurality of transmission ends at said receiver end based on said estimated transmission channels; determining individual characteristics for single user equalization for said plurality of transmission ends; and performing partial single-user equalization at each of said plurality of transmission ends based on said filter information.

11. A method according to claim 10, further comprising separating said single user equalization into a time domain partial equalization corresponding to said individual characteristic, and a frequency domain equalization performed at said receiver end (20).

12. A method according to claim 10, further comprising calculating said individual characteristic based on said estimated transmission channels, and based on a characteristic of said multi-user interference cancellation.

13. A method according to claim 10, further comprising determining said individual characteristics so that a length of non-zero filter taps is smaller than a guard period of said multi-user transmission.

14. A method according to claim 10, further comprising determining said individual characteristics so that a power ratio of a first set of filter taps to a second set of filter taps is minimized.

15. A computer program product comprising code means for producing the steps (a) to (c) of method claim 10 when run on a computer device.

16. A computer program product comprising code means for producing the step

(d) of method claim 10 when run on a computer device.