WO2018153439A1

WO2018153439A1 - Full-duplex transceiver and receiving method

Info

Publication number: WO2018153439A1
Application number: PCT/EP2017/053976
Authority: WO
Inventors: Paul FERRAND; Melissa DUARTE GELVEZ
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2017-02-22
Filing date: 2017-02-22
Publication date: 2018-08-30
Also published as: CN110521128B; CN110521128A

Abstract

The present invention provides an in-band Full-Duplex (FD) transceiver (500) and a corresponding method (1300). The FD transceiver (500) comprises a transmitter (501) and a receiver (502). It further comprises a link (503) configured to tap an analog transmit signal of the transmitter (501), and at least two non-equal delay lines (504). Each delay line (504) is configured to receive the tapped signal, and to delay the tapped signal to create an analog reference signal. The FD transceiver (500) also comprises a converter (505) configured to digitalize the at least two analog reference signals and an analog receive signal of the receiver (502), and a digital canceller (506) configured to generate a cancellation signal based on the at least two digitalized analog reference signals. The digital canceller (506) is finally configured to combine the cancellation signal with the digitalized analog receive signal.

Description

FULL-DUPLEX TRANSCEIVER AND RECEIVING METHOD

TECHNICAL FIELD

The present invention relates to an in-band Full-Duplex (FD) transceiver, and to a method of receiving a signal with a FD transceiver. In particular, the present invention is concerned with a case, in which the same physical propagation medium is used for transmitting and receiving data signals.

BACKGROUND

In wireless communication systems, the same physical propagation medium can be used for the transmission and reception of data signals. In this case, the data transmission can create a large self-interference for the data reception. For a long time, wireless communication systems avoided this self-interference problem by separating the transmission and reception of data either in time or in different frequency bands. However, due to recent improvements in digital signal processing, new opportunities for transmitting and receiving data at the same time and in the same frequency band are enabled. The main technical issue is still, however, an efficient removal of self-interference generated by the transmission signal interfering with the reception signal.

For removing the self-interference occurring in a conventional FD transceiver 100 as shown in Fig. 1, a receiver 102 should know an analog signal emitted by the transmitter 101 as precisely as possible. To this end, a reference signal 103 may be tapped from the transmitter 101, before it reaches one or more transmit antennas 104, and may be routed to the receiver 102. The receiver 102 demodulates the tapped signal separately from a desired signal, which is received via one or more receive antennas 105. The receiver 102 thus has both the desired signal plus a self-interference signal, as well as a reference self- interference signal. By subtracting the latter from the former, the receiver 102 can in theory almost perfectly remove the self-interference, and thus recover the desired signal. In practical systems, however, the self-interference reference signal 103 tapped from the transmitter 101 is corrupted by various analog effects - such as additive white noise or multiplicative phase noise - which means that the self-interference signal cannot be known perfectly to the receiver 102. Thus, even after subtracting the self-interference signal, the receiver 102 still has residual self-interference corrupting the desired signal. Since the self- interference signal received wirelessly is of much higher power than the desired signal - notably self-interference is transmitted very close to the receiver 102 - the residual self- interference can easily mask the desired signal, and can thus drastically degrade the performance of the FD transceiver 100.

Fig. 2 shows more details of a conventional FD transceiver 100, which was also considered when making the present invention. The FD transceiver 100 comprises the data transmitter 101 and the data receiver 102 already shown in Fig. 1. The transmitter 101 is specifically composed of a baseband processing unit 209, which is configured to shape an input digital signal into a baseband digital signal, a digital-to-analog (D/A) converter 210, which is configured to transform the digital baseband signal into an analog signal x(t), a local oscillator (LO) and mixer 211, which are configured to up- convert the analog signal to the relevant frequency band, a radio-frequency (RF) processing unit 212, which is configured to process and shape the analog signal, and a power amplifier (PA) 213, which is configured to feed the analog signal to the transmit antenna 104. The FD transceiver 100 also includes an analog link, which is configured to tap the analog signal fed by the PA 213 to the transmit antenna 104, and to provide the tapped signal as the reference signal 103 to the data receiver 102.

The receiver 102 contains the receive antenna 105, an analog canceller 206, which is configured to take the signal tapped from the transmitter 101 (the analog reference signal) and attempt removing a part of the self-interference signal, a first RF processing unit 207, which is configured to process and shape the analog signal fed from the receive antenna 105, and at least two mixers 201 driven by the same LO 202, which are configured to down- convert the received signal and the reference signal to the baseband, in order to obtain signals y(t) and yref(t), respectively. Importantly, the LO 202 may particularly experience phase-noise (receiver phase-noise effects) and frequency offsets. The receiver 102 also includes at least two A/D converters 203, which are configured to convert the received signal and the reference signal from analog into digital, and a digital canceller 204, which is configured to perform self-interference regeneration, in order to obtain a signal for digital self-interference cancellation. The digital canceller 204 is also configured to combine the cancellation signal with the signal coming from the receive antenna 105, and thus to subtract some self-interference from the received signal. Finally, the receiver 102 includes a baseband processing unit 205, which is configured to further process and shape the digital signal after the interference removal operation. The receiver 102 is interfered by the transmitter 101 as already described above with respect to Fig. 1. Performance degrading effects, which arise from this interference, are illustrated in the following model.

For simplicity, the presence of the desired signal is ignored, and only the problematic self- interference signal is considered. Therefore, in Fig. 2 y(t) represents the self-interference. This self-interference consists of multiple delayed copies of the signal x(t) transmitted by the transmitter 101, due to reflections in the environment around the FD transceiver 100. This is modeled through a multipath self-interference channel, represented as

wherein c_i is the attenuation due to path i, and Δ_i is the delay due to path i. The reference channel is represented as a simple delay δ(t - Δ_ref) wherein Δ_ref is in general different from the delays Δ_i. Based on these models, the signal received from the reference channel is equal to x(t - Δ_ref) and the signal received from the self-interference channel is equal to ∑_i c_ix(t - Δ_i).

A key problem of the FD receiver 100 of Fig. 2 is related to the phase-noise generated in the receiving LO 202 (receiver phase noise effects). Thus, including the phase-noise, the signal received from the reference channel is equal to y_ref(t) = e^jΦ(t)x(t - Δ_ref) and the signal received from the self-interference channel is equal to y(f) = e^jΦ(t)∑_i c_ix(t - Δ_i).

From these two equations, it can be seen that although the signals coming from the reference and self-interference channel experience the same phase-noise effects, the phase- noise sample at time t, φ(ΐ) , ends up multiplying different delayed versions of the transmitted signal x(t). Specifically, when observing the reference signal y_ref, the phase- noise at time t multiplies the transmitted signal delayed by Δ_ref. On the other hand, when observing path i of the self-interference signal y, the same phase-noise at time t multiplies the transmitted signal delayed by Δ_i. Thus, after going through the reference channel and the LO 202, the transmitted signal x(t) has the same phase-noise as the signal from the self-interference channel, but the same phase-noise is applied to different delayed versions of the transmitted signal. This mismatch, caused by the same phase-noise affecting the same signal but at different delays, causes a severe problem for the performance of the digital canceller 204 in the FD receiver 100.

Phase-noise processes are highly unpredictable and vary at a very fast rate. In these conditions, the canceller 204 of the conventional FD receiver 100 will not be able to track the equivalent channel between y_ref (t_n) and y(t_n) correctly, and the performance of the self-interference cancellation will degrade significantly. Fig. 4 illustrates this situation.

Namely, in Fig. 4, the larger the delay Δ_I of the self-interference signal, the more difficult it is for the receiver 102 to track the phase-noise process, and in turn cancel the self- interference correctly. For relatively small delays - e.g. hundreds of nanoseconds - the degradation can attain more than 3dB.

One conventional approach of solving the above-described phase-noise issues of the conventional FD receiver 100 is acquiring more knowledge about the phase-noise process. Thus, it may subsequently be possible to better remove the phase-noise from both the reference signal and the received signal.

For example, side information about the phase-noise obtained from the way it affects the transmitted signal may be used, when this transmitted signal uses Orthogonal Frequency Division Multiplexing (OFDM). By means of this side information, some performance lost due to the phase-noise effects (see Fig. 4) can be recovered. However, it is usually not sufficient to cancel out the phase-noise effects completely. Moreover, this approach is quite complex, especially because it requires completely decoding the transmitted signal, in order to recover some partial information about the phase-noise process.

Another approach that may be implemented in conjunction to the above-described approach, consists in using existing pilot signals known in advance, in order to virtually decode the phase-noise on some part of the signal. Using such pilot signals, the receiver knows the phase-noise process at some positions in time, interleaved with data symbols, and is able to reconstruct the phase-noise process by using interpolation between these positions in time.

One extreme generalization of this solution is to even have complete knowledge of the digital signal that was sent by the transmitter 101, for instance, by using a baseband reference signal on top of the existing analog reference signal. Using such baseband reference signal, the whole signal becomes virtually a pilot signal, and it becomes in theory possible to recover the phase-noise process almost perfectly. In practical systems however, it may be rather difficult for the transmitter 101 and the receiver 102 to share information, since they may, for example, be in two different, physically separated boxes.

Thus, the phase-noise issues of the conventional FD receiver 100 are not satisfyingly solved. SUMMARY

In view of the above-mentioned problems and disadvantages, the present invention aims to improve a conventional FD transceiver and a corresponding receiving method. The present invention has the object to provide a FD transceiver and corresponding method, which is configured to perform efficient self-interference cancellation even in the presence of receiver phase-noise effects. Thus, the present invention desires a more effective removal of the self-interference from a received signal. Thereby, the present invention aims at a FD transceiver being well suited for practical systems.

The object of the present invention is achieved by the solution provided in the enclosed independent claims. Advantageous implementations of the present invention are further defined in the dependent claims. In particular the present invention proposes a FD transceiver and a corresponding receiving method for digital self-interference cancellation, wherein multiple analog references are tapped from a transmitter path, and are separated by non-zero delay lines.

A first aspect of the present invention provides a FD transceiver comprising a transmitter and a receiver, a link configured to tap an analog transmit signal of the transmitter, at least two non-equal delay lines, each delay line being configured to receive the tapped signal, and to delay the tapped signal to create an analog reference signal, a converter configured to digitalize the at least two analog reference signals and an analog receive signal of the receiver, and a digital canceller configured to generate a cancellation signal based on the at least two digitalized analog reference signals, and to combine the cancellation signal with the digitalized analog receive signal.

By generating the multiple, differently delayed analog reference signals, phase-noise effects occurring in the receiver and influencing the self-interference signal on the receive signal can be better compensated, and the self-interference can thus be removed from the digital receive signal more effectively. The FD transceiver of the first aspect is furthermore well suitable for practical implementations. In a first implementation form of the FD transceiver according to the first aspect, a plurality of mixers are configured to down-convert the analog reference signals and the analog receive signal to a receiver baseband of the transceiver, wherein the plurality of mixers are driven by a common local oscillator.

In a second implementation form of the FD transceiver according to the first aspect as such or according to the first implementation form of the first aspect, the digital canceller is or comprises at least one adaptive filter. An adaptive filter is a particularly effective and low complexity implementation of the digital canceller.

In a third implementation form of the FD transceiver according to the second implementation form of the first aspect, for generating the cancellation signal, the at least one adaptive filter is configured to perform a regeneration of self-interference, which is caused by the receiver receiving over an air interface the transmit signal from the transmitter.

Thereby, the self-interference signal can be effectively removed from the receive signal, even in the presence of phase-noise effects in the receiver LO.

In a fourth implementation form of the FD transceiver according to the second or third implementation form of the first aspect, the digital canceller is a bank of adaptive filters, preferably implemented as a block filter, each adaptive filter being fed with one of the digitalized analog reference signals.

The digital canceller can thus take into account each delayed version of the tapped reference signal, which allows estimating more precisely the multiple-path self- interference signal, and accordingly leads to a better removal thereof from the receive signal.

In a fifth implementation form of the FD transceiver according to the fourth implementation form of the first aspect, the output signal of the digital canceller is used to drive the adaptation of the adaptive filters. The feedback of the output signal allows to iteratively remove the self-interference signal more efficiently. In a sixth implementation form of the FD transceiver according to the fourth or fifth implementation forms of the first aspect, for generating the cancellation signal, the digital canceller is configured to assign a first weight to at least one tap of each adaptive filter, and to update the first weights by following a Normalized Least Mean Squares, NLMS, procedure.

In a seventh implementation form of the FD transceiver according to the sixth implementation form of the first aspect, for generating the cancellation signal, the digital canceller is configured to assign a second weight to the at least one tap of each adaptive filter, wherein a higher second weight is assigned to an adaptive filter, if the adaptive filter is fed with a digitalized analog reference signal stemming from an analog reference signal with a delay that is closer to a delay of the transmitter signal received over the air interface by the receiver from the transmitter.

By weighting the taps of the adaptive filters in the above-mentioned manner, the phase- noise effects can be taken into account more precisely, and the self-interference signal can be removed most effectively from the digital receive signal.

In an eighth implementation form of the FD transceiver according to the first aspect as such or according to any previous implementation form of the first aspect, the digital canceller is configured to estimate a phase-noise signal based on at least two of the analog or the digitalized analog reference signals, to remove the estimated phase-noise signal from the digitalized analog receive signal to obtain a phase-noise corrected receive signal, and to generate the cancellation signal based on the phase-noise corrected reference signal. Estimating the phase-noise signal using multiple analog references allows for an even better removal of the self-interference signal. Because multiple copies of the phase-noise are provided by virtue of the multiple analog reference signals with different delays, the phase-noise can be estimated better than conventionally. In a ninth implementation form of the FD transceiver according to the eighth implementation form of the first aspect, the transceiver further comprises a link configured to tap at least one baseband signal from a transmitter baseband of the transceiver, in order to obtain at least one digital baseband reference signal, wherein

the digital canceller is configured to generate the cancellation signal based on the at least one digital baseband reference signal and based on the at least two digitalized analog reference signals.

Taking additionally the baseband reference signal further improves the removal of the self- interference signal.

In a tenth implementation form of the FD transceiver according to the eighth or ninth implementation form of the first aspect, the digital canceller is configured to remove the estimated phase-noise signal from at least one digitalized analog reference signal to obtain at least one phase-noise corrected reference signal, and to generate the cancellation signal based on the at least one phase-noise corrected reference signal.

In an eleventh implementation form of the FD transceiver according to the first aspect as such or according to any implementation form of the first aspect, the digital canceller is configured to estimate, before generating the cancellation signal, a phase-noise signal, and to remove the estimated phase-noise signal from each of the digital reference signals and from the digitalized analog receive signal.

Estimating the phase-noise signal allows for an even better removal of the self-interference signal.

In a twelfth implementation form of the FD transceiver according to the eleventh implementation form of the first aspect, the transceiver further comprises a link configured to tap at least one baseband signal from a transmitter baseband of the transceiver, in order to obtain at least one digital baseband reference signal, wherein the digital canceller is configured to estimate the phase-noise signal based on the at least one baseband reference signal and at least one digitalized analog reference signal. Taking additionally the baseband reference signal improves the estimation of the phase- noise signal, and thus effectively leads to a better removal of the self-interference signal.

A second aspect of the present invention provides a method for receiving a signal with a full-duplex, FD, transceiver, the method comprising the steps of tapping an analog transmit signal of a transmitter of the transceiver, providing the tapped signal with at least two non- equal delays to create at least two analog reference signals, digitalizing the at least two analog reference signals and an analog receive signal of a receiver of the transceiver, generating a cancellation signal based on the at least two digitalized analog reference signals, and combining the cancellation signal with digitalized analog receive signal.

In a first implementation form of the method according to the second aspect, the method further comprises down-converting the analog reference signals and the analog receive signal to a receiver baseband of the transceiver, wherein the plurality of mixers are driven by a common local oscillator.

In a second implementation form of the method according to the second aspect as such or according to the first implementation form of the second aspect, the cancellation signal is generated with at least one adaptive filter.

In a third implementation form of the method according to the second implementation form of the second aspect, the cancellation signal is generated by the at least one adaptive filter performing a regeneration of self-interference, which is caused by the receiver receiving over an air interface the transmit signal from the transmitter.

In a fourth implementation form of the method according to the second or third implementation form of the second aspect, the cancellation signal is generated with a bank of adaptive filters, preferably implemented as a block filter, each adaptive filter being fed with one of the digitalized analog reference signals.

In a fifth implementation form of the method according to the fourth implementation form of the second aspect, the method comprises using an output signal of the generation of the cancellation signal for driving the adaptation of the adaptive filters. In a sixth implementation form of the method according to the fourth or fifth implementation forms of the second aspect, for generating the cancellation signal, the cancellation signal is generated by assigning a first weight to at least one tap of each adaptive filter, and to update the first weights by following a Normalized Least Mean Squares, NLMS, procedure.

In a seventh implementation form of the method according to the sixth implementation form of the second aspect, the cancellation signal is generated by assigning a second weight to the at least one tap of each adaptive filter, wherein a higher second weight is assigned to an adaptive filter, if the adaptive filter is fed with a digitalized analog reference signal stemming from an analog reference signal with a delay that is closer to a delay of the transmitter signal received over the air interface by the receiver from the transmitter.

In an eighth implementation form of the method according to the second aspect as such or according to any previous implementation form of the second aspect, the method further comprises estimating a phase-noise signal based on at least two of the analog or the digitalized analog reference signals, removing the estimated phase-noise signal from the digitalized analog receive signal to obtain a phase-noise corrected receive signal, and generating the cancellation signal based on the phase-noise corrected receive signal.

In a ninth implementation form of the method according to the eighth implementation form of the second aspect, the method further comprises tapping at least one baseband signal from a transmitter baseband of the transceiver, in order to obtain at least one digital baseband reference signal, wherein the cancellation signal is generated based on the at least one digital baseband reference signal and based on the at least two digitalized analog reference signals.

In a tenth implementation form of the method according to the eighth or ninth implementation form of the second aspect, the method comprises removing the estimated phase-noise signal from at least one digitalized analog reference signal to obtain at least one phase-noise corrected reference signal, and generating the cancellation signal based on the at least one phase-noise corrected reference signal. In an eleventh implementation form of the method according to the second aspect as such or according to any implementation form of the second aspect, the method further comprises estimating, before generating the cancellation signal, a phase-noise signal, and removing the estimated phase-noise signal from each of the digitalized analog reference signals and from the digitalized analog receive signal.

In a twelfth implementation form of the method according to the eleventh implementation form of the second aspect, the transceiver further comprises tapping at least one baseband signal from a transmitter baseband of the transceiver, in order to obtain at least one digital baseband reference signal, wherein the phase-noise signal is estimated based on the at least one baseband reference signal and at least one digitalized analog reference signal.

A second aspect of the present invention provides a computer program product comprising a computer-readable medium storing instructions which, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations to implement the method of the second aspect as such or according to any implementation form of the second aspect.

With the method of the second aspect and its implementation forms, all effects and advantages of the FD transceiver of the first aspect and its implementation forms, respectively, can be achieved.

It has to be noted that all devices, elements, units and means described in the present application could be implemented in the software or hardware elements or any kind of combination thereof. All steps which are performed by the various entities described in the present application as well as the functionalities described to be performed by the various entities are intended to mean that the respective entity is adapted to or configured to perform the respective steps and functionalities. Even if, in the following description of specific embodiments, a specific functionality or step to be performed by external entities is not reflected in the description of a specific detailed element of that entity which performs that specific step or functionality, it should be clear for a skilled person that these methods and functionalities can be implemented in respective software or hardware elements, or any kind of combination thereof. BRIEF DESCRIPTION OF DRAWINGS

The above described aspects and implementation forms of the present invention will be explained in the following description of general and specific embodiments in relation to the enclosed drawings, in which

Fig. 1 shows a basic wireless communication scenario with a conventional FD transceiver.

Fig. 2 shows a conventional FD transceiver.

Fig. 3 shows a digital canceller of a conventional FD transceiver.

Fig. 4 shows residual self-interference after digital self-interference cancellation in a conventional FD transceiver experiencing phase-noise (PN) effects based on the delay of the self-interference signal. The delay of the reference signal is supposed to be 0.

Fig. 5 shows a FD transceiver according to an embodiment of the present invention.

Fig. 6 shows a FD transceiver according to an embodiment of the present invention.

Fig. 7 shows FD transceivers according to an embodiment of the present invention.

Fig. 8 shows a FD transceiver according to an embodiment of the present invention.

Fig. 9 shows a comparison of the performance of a FD transceiver according to an embodiment of the present invention versus a conventional FD transceiver. Fig. 10 shows a FD transceiver according to an embodiment of the present invention.

Fig. 11 shows a FD transceiver according to an embodiment of the present invention.

Fig. 12 shows a FD transceiver according to an embodiment of the present invention. Fig. 13 shows a method according to an embodiment of the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

The solution of the present invention is based on the insight that, in reality, the phase-noise process does not need to be separated from the receive signal. To illustrate this, reference is made to the following equation of the equivalent channel for the case of the single-tap self-interference channel. As shown in Fig. 3, the signals received from the reference and self-interference channels are affected by the same phase-noise φ(ί). The transmitted signal x(t) is convolved through both the reference channel fr_ref (t) and the multi path self-interference channel h_i (t), and is mixed in the LO 202 with the phase- noise 0(t) to obtain, respectively, the reference signal y_ref(t), and the self-interference signal y(t) . For clarity purposes, the presence of the desired signal is omitted in the equations. Using t_n to denote the digital domain time samples, the inputs to the digital canceller 204 are given by: y(t_n) = e^jΦ(tn) ∑_ic_ix(t_n - Δ_i)

with T_s being the symbol rate, at which the system samples. The goal of the digital canceller 204 is to find the coefficients w_i such that it can reconstruct y(t_n) using its knowledge of y_ref (t_n) . In mathematical terms, it aims at finding the coefficients {wj such that y(t_n) = ∑i w_ty_ref(t_n - iT_s).

The interplay between phase-noise and delay can be better demonstrated using the following example. Under the assumption that there is only one reflection in the self- interference channel, one obtains h_I(t) = δ(t - Δ_I).

The difference between the reference signal delay Δ_I and the self-interference signal delay Δ_ref matters substantially, when phase-noise is present in the system, because in this case the relation between y(t_n) and y_ref(t_n) is

The correct solution for the digital canceller 204 is then to have a single coefficient

Fig. 5 shows a FD transceiver 500 according to a general embodiment of the present invention. The FD transceiver 500 includes a transmitter 501 for transmitting signals, and a receiver 502 for receiving signals. With this FD transceiver 500, the above-mentioned problem of the phase-noise issues is solved in a different way than conventionally.

If Δ_I is equal to Δ_ΚΕΡ, then there is no issue, and the receiver 502 will work perfectly. The problem is thus essentially a problem of delay. As long as the receiver 502 has a reference signal with a delay that matches the delay of the interference channel, the phase-noise will have no effect on the performance. In real systems, however, the self-interference channel is not a single-tap channel, but a multi-tap channel with multiple copies of the transmitted signal being received at different delays. In this case, the delay of the self-interference channel cannot be matched, because there are, in fact, multiple delays for the same signal. Thus, as the core of the present invention, it is ensured that the receiver 502 has access to multiple copies of the analog reference signal at different delays, in order to be able to match the multiple copies of the transmitted signal received on the self-interference channel.

Accordingly, the FD transceiver 500 comprises an analog link 503 configured to tap an analog transmit signal of the transmitter 501, and at least two non-equal delay lines 504, wherein each delay line 504 is configured to receive the tapped signal, and to delay the tapped signal, in order to create an analog reference signal.

Furthermore, the FD transceiver 500 comprises a converter 505 and a digital canceller 506, which are - as shown in Fig. 5 - preferably located in the receiver 502. However, the converter 505 and canceller 506 could also be separate and connected to the receiver 502. The converter 505 is specifically configured to digitalize the at least two analog reference signals and an analog receive signal of the receiver 502, respectively. The digital canceller 506 is configured to generate a cancellation signal based on the at least two digitalized analog reference signals, and further to combine the cancellation signal with the digitalized analog receive signal. By creating the delayed copies of the reference signals in the analog domain, it can be ensured that at each time all the different delayed copies are corrupted with the same phase- noise as the self-interference signal, thus one can use one or several of these copies for digital cancellation. The digital canceller 506 can thus more easily match the channel to provide a better estimate and cancellation of the self-interference signal, and the performance of the FD transceiver 500 is greatly improved.

A block diagram overview of possible components of a FD transceiver 500 according to embodiments of the present invention, which base on the FD receiver 500 shown in Fig. 5, is shown in Fig. 6. The FD receiver 500 comprises the transmitter 501 and the receiver 502. The transmitter 501 may have one or more transmit antennas 602, at least one D/A, and a LO driven mixer (similarly as the FD transceiver 100 shown in Fig. 2). The receiver 502 may have one or more receive antennas 603, a plurality of LO 601 driven mixers 600, a plurality of A/Ds 505, and the digital canceller 506. A link 503 between the transmitter 501 and the receiver 502 taps an analog signal, which is fed into the plurality of delays 504, in order to create the multiple analog reference signals. Optionally, another digital link 604 may provide a baseband reference signal from the transmitter 502 to the receiver 501.

The approach of the present invention with the FD transceiver 500 shown in Figs. 5 and 6 has numerous advantages over conventional FD transceivers. The main advantage, however, is that it is not necessary to decode the phase-noise, or the reference signal at all. In fact, as will be shown with respect to specific embodiments below, the phase-noise does not need to be treated in any specific way. The implementation of the digital canceller 506 naturally matches the delays that are experienced in the self-interference channel, and can thus recover the performance degradation caused by the phase-noise almost completely.

Using the multiple analog references - as shown in Figs. 5 and 6 - can also be used to improve conventional FD transceivers with phase-noise estimation techniques. In fact, since access to multiple copies of the phase-noise is provided by virtue of the added delays 504, phase-noise can be estimated better than in conventional FD transceivers.

Fig. 7 shows different principles of FD transceivers 500 according to embodiments of the present invention. On the left hand side, a FD transceiver 500 with multiple transmit antennas 602 connected to the transmitter 501, and multiple receive antennas 603 connected to the receiver 502 is shown. That is, the transmitted signal may be split to multiple transmit antennas 602, and the received signal may come from multiple reference antennas. On the right hand side, a FD transceiver 500 principle with one antenna 700 used for transmission and reception is shown. This transceiver 500 may be achieved when using a circulator 701. Of course, also multiple antennas 700 could be used for transmission and reception. The different principles of FD transceivers 500 shown in Fig. 7 can be applied to all specific embodiments of FD transceivers 500 of the present invention.

A FD transceiver 500 according to a specific embodiment is shown in Fig. 8. The embodiment bases on the FD transceiver 500 shown in Fig. 5. In this embodiment, N analog references are first converted to the digital domain, wherein each analog reference goes through a separate A/D converter 505. The resulting N digital signals are input to the digital canceller 506, which performs the self-interference regeneration. The digital canceller 506 is or comprises at least one adaptive filter 800, specifically it may be a bank of adaptive filters 800, preferably implemented as a block filter. Each adaptive filter 800 may be fed with at least one of the digitalized analog reference signals. The taps of each adaptive filter 800 are assigned a first weight w₁ , which is updated following an Normalized Least Mean Squares (NLMS) type procedure combined with a further novel weight recalculation w₂ that is a function of (1) the analog delay associated with analog reference j and (2) the total number of analog references N.

The novel custom NLMS implementation of the bank of adaptive filters 800, each with P taps, follows the following procedure. The adaptive filter input is defined as

where y_ref1 (t_n) denotes the sample of signal y_ref1 at time t_n . The filter weights are defined as

Here, w is the diagonal of the product between weights w₁∈ C^1XNP computed via conventional NLMS and weights w₂∈ C^1XNP, which are computed as a function of the analog delays associated with the analog references and the total number of analog references N. The estimate of the self-interference is computed as

The digital canceller 506 output is equal to

which is used to update the weights

The weights w₂ (t_n) can be constant for all t_n hence computed offline. Intuitively, the weights w₂ (t_n) are used to give more weight to the signal coming from analog reference j, when the channel delay is closer to Δ^. In a multi path channel, the weights w₂(t_n) are also applied. In this multipath case if, for example, the channel has paths with delays Δ_i and Δ_j, then signals coming from analog references with delay closest to Δ_i and Δj are given the highest weights. As an example, for three analog references with corresponding delays Δ₁ = 0 , Δ₂ > Δ₁ and Δ₃ > Δ₂ , and three corresponding adaptive filters 800 implemented in the block form explained above, a simple way to compute weights w₂ (t_n) is the following. First, the following preliminary weights are defined. The weights for the first adaptive filter 800 are all equal to 1. For the second adaptive filter 800, the weights that correspond to delays less than (Δ₁ + Δ₂)/2 are set equal to 0, and all others are set to 1. For the third adaptive filter 800, the weights that correspond to delays less than (Δ₂ + Δ₃)/2 are set equal to 0, and all others are set to 1. Finally, it is normalized such that the weights w₂ (t_n) applied to filter taps that correspond to same time instant t_n sum to 1.

In this embodiment, only NLMS is explicitly considered, but the embodiment can also be implemented for any other type of adaptive filters 800, like Recursive Least Squares (RLS). For simplicity of notation, the equations above do not explicitly show the effect of noise, however, the signals y (t_n) and y (t_n) are affected by noise, due to hardware imperfections. Also, other methods for computation of w₂ (t_n) are possible, one trivial case being having all 1. Finally, it is noted that one advantage of this embodiment is that it does not require a baseband reference signal.

Fig. 9 shows the significant performance improvement of the FD transceiver 500 of Fig. 8 compared to a conventional FD transceiver (e.g. FD transceiver 100). On the y-axis, a residual self-interference on the receive signal is plotted in units of dBm. On the x-axis, a delay is plotted in unit of ns. It can be seen that in the embodiment of Fig. 8 (Embodiment 1 in Fig. 9), the residual self-interference is significantly lower than for the conventional FD transceiver 100 (state of the art, with phase-noise) for each delay.

Another specific embodiment of the FD transceiver 500 is shown in Fig. 10. In this embodiment, two analog references with different delay 504 (here 'Analog reference 1 ' is not delayed, while 'Analog reference 2' is delayed) are first converted to the digital domain, wherein each analog reference goes through a separate A/D converter 505. The resulting signals are used to estimate the phase-noise process via standard signal processing

in the digital canceller 506, which may include delay, conjugate, normalization and filtering. Other methods for estimation

from two reference signals are also possible.

This signal

is then used to correct (remove) the phase-noise from the signals that are used for the self-interference regeneration. The self-interference regeneration can be implemented using any kind of adaptive filter 800.

Another specific embodiment of the FD transceiver 500 is shown in Fig. 11. In this embodiment, like in Fig. 10, again two analog references are used to estimate the phase- noise process e via standard signal processing in the digital canceller 506, which may include delay, conjugate, normalization and filtering procedure. The processing can be implemented all in the analog domain (as shown in Fig. 11) or may be split between the analog and digital domain. Other methods for estimation

from two reference signals are also possible. The phase-noise correction signal

is then used to correct (remove) the phase-noise from the signal coming from the receive antenna 603. The phase- noise corrected signal and a baseband reference signal provided via link 604 from the transmitter 501 to the receiver 502 are used for the self-interference regeneration, which can be implemented using any kind of adaptive filter 800.

As an alternative implementation for this embodiment, a case without baseband reference may be considered, similar to the embodiment of Fig. 10. In this case, the embodiment of Fig. 11 is modified such that the phase-noise correction signal

is not only applied to the signal coming from the receive antenna 603, but it is also applied to one of the signals coming from the analog references (as in Fig. 10), and the resulting phase-noise corrected signals are used for self-interference regeneration. Another specific embodiment of the FD transceiver 500 is shown in Fig. 12. This embodiment uses one baseband reference provided via digital link 604 and N analog references provided via another analog link 503. The baseband reference and one of the digitalized analog references are used to obtain an estimate of the phase-noise process

. The signal

is then multiplied with all the N analog references and the signal coming from one or more receive antennas 603, thereby removing all or some part of the phase-noise. The resulting signals are input to adaptive filters 800, which perform the self-interference regeneration. The digital canceller 506 is implemented as a bank of adaptive filters 800, preferably implemented as a block filter. The signal regeneration follows a procedure analogous to the adaptive filter 800 implementation described for the embodiment of Fig. 8.

Fig. 13 illustrates a method 1300 for receiving a signal according to an embodiment of the present invention. The method 1300 is performed by a FD transceiver 500 according to an embodiment of the present invention.

The method 1300 comprises a first step 1301 of tapping an analog transmit signal of a transmitter 501 of the transceiver 500. In a second step 1302, the method 1300 comprises providing the tapped signal with at least two non-equal delays to create at least two analog reference signals. In a third step 1303, the method 1300 comprises digitalizing the at least two analog reference signals and an analog receive signal of a receiver 502 of the transceiver 500. In a fourth step 1304, the method 1300 comprises generating 1304 a cancellation signal based on the at least two digitalized analog reference signals. In a fifth step 1305, the method 1300 comprises combining the cancellation signal with the digitalized analog receive signal

The present invention has been described in conjunction with various embodiments as examples as well as implementations. However, other variations can be understood and effected by those persons skilled in the art and practicing the claimed invention, from the studies of the drawings, this disclosure and the independent claims. In the claims as well as in the description the word "comprising" does not exclude other elements or steps and the indefinite article "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several entities or items recited in the claims. The mere fact that certain measures are recited in the mutual different dependent claims does not indicate that a combination of these measures cannot be used in an advantageous implementation.

Claims

1. Full-Duplex, FD, transceiver (500) comprising

a transmitter (501) and a receiver (502),

a link (503) configured to tap an analog transmit signal of the transmitter (501), at least two non-equal delay lines (504), each delay line (504) being configured to receive the tapped signal, and to delay the tapped signal to create an analog reference signal, a converter (505) configured to digitalize the at least two analog reference signals and an analog receive signal of the receiver (502), and

a digital canceller (506) configured to generate a cancellation signal based on the at least two digitalized analog reference signals, and to combine the cancellation signal with the digitalized analog receive signal.

2. FD transceiver (500) according to claim 1, further comprising

a plurality of mixers (600) configured to down-convert the analog reference signals and the analog receive signal to a receiver baseband of the transceiver (502),

wherein the plurality of mixers (600) are driven by a common local oscillator (601).

3. FD transceiver (500) according to claim 1 or 2, wherein

the digital canceller (506) is or comprises at least one adaptive filter (800).

4. FD transceiver (500) according to claim 3, wherein for generating the cancellation signal, the at least one adaptive filter (800) is configured to perform a regeneration of self- interference, which is caused by the receiver (502) receiving over an air interface the transmit signal from the transmitter (501).

5. FD transceiver (500) according to one of claims 3 to 4, wherein

the digital canceller (506) is a bank of adaptive filters, preferably implemented as a block filter, each adaptive filter (800) being fed with one of the digitalized analog reference signals.

6. FD transceiver (500) according to claim 5, wherein

the output signal of the digital canceller (506) is used to drive the adaptation of the adaptive filters (800).

7. FD transceiver (500) according to claim 5 or 6, wherein for generating the cancellation signal,

the digital canceller (506) is configured to assign a first weight to at least one tap of each adaptive filter (800), and to update the first weights by following a Normalized Least Mean Squares, NLMS, procedure.

8. FD transceiver (500) according to claim 7, wherein for generating the cancellation signal,

the digital canceller (506) is configured to assign a second weight to the at least one tap of each adaptive filter (800),

wherein a higher second weight is assigned to an adaptive filter (800), if the adaptive filter (800) is fed with a digitalized analog reference signal stemming from an analog reference signal with a delay that is closer to a delay of the transmitter signal received over the air interface by the receiver (502) from the transmitter (501).

9. FD transceiver (500) according to one of claims 1 to 8, wherein

the digital canceller (506) is configured to estimate a phase-noise signal based on at least two of the analog or the digitalized analog reference signals, to remove the estimated phase-noise signal from the digitalized analog receive signal to obtain a phase- noise corrected receive signal, and to generate the cancellation signal based on the phase- noise corrected reference signal.

10. FD transceiver (500) according to claim 9, further comprising

a link (604) configured to tap at least one baseband signal from a transmitter baseband of the transceiver (500), in order to obtain at least one digital baseband reference signal, wherein

the digital canceller (506) is configured to generate the cancellation signal based on the at least one digital baseband reference signal and based on the at least two digitalized analog reference signals.

11. FD transceiver (500) according to claim 9 or 10, wherein

the digital canceller (506) is configured to remove the estimated phase-noise signal from at least one digitalized analog reference signal to obtain at least one phase-noise corrected reference signal, and to generate the cancellation signal based on the at least one phase-noise corrected reference signal.

12. FD transceiver (500) according to one of claims 1 to 8, wherein

the digital canceller (506) is configured to estimate, before generating the cancellation signal, a phase-noise signal, and to remove the estimated phase-noise signal from each of the digitalized analog reference signals and from the digital receive signal.

13. FD transceiver (500) according to claim 12, further comprising

the digital canceller (506) is configured to estimate the phase-noise signal based on the at least one baseband reference signal and at least one digitalized analog reference signal.

14. Method (1300) for receiving a signal with a full-duplex, FD, transceiver (500), the method (1300) comprising the steps of

tapping (1301) an analog transmit signal of a transmitter (501) of the transceiver (500),

providing (1302) the tapped signal with at least two non-equal delays to create at least two analog reference signals,

digitalizing (1303) the at least two analog reference signals and an analog receive signal of a receiver (502) of the transceiver (500),

generating (1304) a cancellation signal based on the at least two digitalized analog reference signals, and

combining (1305) the cancellation signal with the digitalized analog receive signal.

15. A computer program product comprising a computer-readable medium storing instructions which, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations to implement the method according to claim 14.