The present invention relates generally to multi-antenna wireless communication systems and, more particularly, to calibrating system frequency responses for joint antenna processing in such a system.
The phenomenal growth in the demand for wireless communications has put persistent pressure on wireless network operators to improve the capacity of their communication networks. To improve the spectral efficiency of these networks, scarce radio resources have to be reused aggressively in neighboring cells. As a result, inter-cell interference has become a main source of signal disturbance, limiting not only the service quality to users at the cell edges, but also the overall system throughput.
A multi-antenna wireless communication system, such as one employing Coordinated Multipoint (CoMP) transmission and/or reception, can mitigate inter-cell interference. In such a system, multiple antennas are deployed across a plurality of geographically contiguous cells—referred to as sub-cells—and connected to a central controller, e.g., via a fast backhaul. This architecture enables the central controller to jointly process downlink signals transmitted from and/or uplink signals received at the multiple antennas, in order to mitigate inter-cell interference.
To jointly process downlink signals transmitted from the multiple antennas, in particular, the central controller must have available to it information characterizing the associated downlink channel responses. If the system employs frequency-division duplexing (FDD), the served mobile terminals must measure these downlink channel responses and explicitly feed them back to the central controller. If the system instead employs time-division duplexing (TDD), the downlink channel responses can advantageously be estimated from uplink signals received at the multiple antennas, based on the assumption that for TDD the downlink channel can be inferred from the uplink channel.
This assumption of reciprocity between the downlink and uplink channel, however, may sometimes be inaccurate in a multi-antenna system. For instance, each of the multiple antennas is connected to a corresponding transceiver that comprises a transmit-receive radio frequency (RF) chain. These transmit-receive RF chains may have different frequency responses, due for example to differences in the transfer characteristics of the components (e.g., analog filters, power amplifiers, etc.) making up those RF chains. If the RF chains of the multiple antennas do not have identical frequency responses, the assumption of reciprocity no longer proves accurate, which in turn prevents advantageous estimation of the downlink channel responses from uplink signals. Accordingly, the RF chains must be initially, and perhaps periodically, calibrated with one another to account for differences in their frequency responses.
Known approaches to calibrating the RF chains of antennas in a multi-antenna system rely on mobile terminals to assist with the calibration process. These approaches must identify at least one mobile terminal that is sufficiently stationary (i.e., has a relatively low Doppler spread) and that has a sufficient quality of service (i.e., a strong channel quality indicator, CQI) to adequately measure and feed back the downlink channel responses. If such a suitable mobile terminal can be identified, calibration factors can be computed based on the provided feedback and then applied to the RF chains. Of course, a suitable mobile terminal may not always be available, meaning that these approaches are inherently undependable. Further, by requiring feedback of the downlink channel responses, the approaches necessarily diminish the very advantage sought by calibrating the RF chains in the first place. Still further, these and other approaches limit the scalability of the calibration process to a relatively small number of antennas.
Teachings herein include a large scale, self-calibrating multi-antenna wireless communication system for calibrating the transmit-receive RF chains of antennas in the system. The system is self-calibrating in the sense that it advantageously does not rely on mobile terminals to assist with the calibration process, and does not require feedback of downlink channel responses. Rather, the system transmits pilot signals amongst the antennas themselves, and then calculates RF-chain calibration factors for the antennas based on those received pilot signals.
In one embodiment, for example, each of a plurality of antennas in the system belongs to one of three or more time-reuse groups. The time-reuse groupings are geographically dispersed, to geographically separate all antennas belonging to the same time-reuse group. Due to this separation, pilot signals concurrently transmitted from antennas belonging to the same time-reuse group do not significantly interfere with one another.
With the system configured in this way, a calibration controller, during a first time interval, transmits a pilot signal from at least those antennas in a first time-reuse group, so that the pilot signals are received at antennas in one or more other time-reuse groups. The calibration controller further, during a second time interval, transmits a pilot signal from at least those antennas in a second time-reuse group, so that the pilot signals are received at antennas in one or more other time-reuse groups. The calibration controller also, during a third time interval, transmits a pilot signal from at least those antennas in a third time-reuse group, so that the pilot signals are received at antennas in one or more other time-reuse groups.
Based on these pilot signals received amongst the antennas, the calibration controller is configured to calculate an RF-chain calibration factor for each of one or more of the antennas. The calibration factor calculated for a given antenna accounts for a frequency response difference between the RF chain of that antenna and the RF chain of another one of the antennas, e.g., an antenna used as a reference for the calibration.
Having calculated the calibration factors in this way, the calibration controller in some embodiments adjusts the RF chain of the one or more antennas based on the corresponding calibration factors. Alternatively, the calibration controller may apply the corresponding calibration factors to a downlink data transmission being sent from those antennas, or an uplink data transmission being received at those antennas, and thereby adjust the frequency response of the data transmission rather than the transmit-receive chains themselves. In any case, the RF chains are effectively calibrated with one another to account for differences in their frequency responses.
With the RF chains calibrated, the system can accurately assume reciprocity between the downlink and uplink channels, and thereby advantageously estimate downlink channel responses from uplink signals for joint processing of multi-antenna signals.
Of course, the present invention is not limited to the above features and advantages. Indeed, those skilled in the art will recognize additional features and advantages upon reading the following detailed description, and upon viewing the accompanying drawings.
With signals being transmitted or received in geographically contiguous cells 30 during the same time interval, interference between the cells 30 (i.e., inter-cell interference) can occur. Accordingly, each cell communicates, e.g., via fiber optic cable, with the central control node 10. The central control node 10 coordinates and/or jointly processes the signals being transmitted or received so as to minimize inter-cell interference. To jointly process downlink signals transmitted from the multiple antennas 20, in particular, the central control node 10 estimates the downlink channel responses from uplink signals received at the multiple antennas 20, based on the assumption of reciprocity between the downlink and uplink channel.
Of course, each of the multiple antennas 20 is connected to a corresponding transceiver that comprises a transmit-receive radio frequency (RF) chain.
H
TX1(f)H*RX1(f)=HTX2(f)H*RX2(f)=HTX3(f)H*RX3(f)= . . . (1)
for all RF chains. See, e.g., Jiann-Ching Guey and L. Daneil Larsson, “Modeling and Evaluation of MIMO Systems Exploiting Channel Reciprocity in TDD mode,” Proceeding, IEEE VTC 2004 Fall. The frequency responses H(f) of the RF chains however may not satisfy this relation if, for example, mismatches in the hardware components making up the RF chains cause the RF chains to have different transfer characteristics.
The operating parameters of these and other hardware components of a transmit-receive RF chain may differ within any given antenna 20, as well as between different antennas 20. Such mismatches may cause the RF chains of the antennas 20 to have different transfer characteristics, so that the frequency responses H(f) of the RF chains fail to satisfy the relation in equation (1).
Accordingly, the system 100 includes a calibration controller configured to calibrate the RF chains of the antennas 20 (e.g., by adjusting one or more operating parameters of the components in the RF chains) so that the RF chains' frequency responses H(f) do in fact satisfy the relation in Equation (1). Moreover, the system 100 is self-calibrating in the sense that it advantageously does not rely on mobile terminals 40 to assist with the calibration process, and does not require feedback of downlink channel responses. Rather, the system 100 transmits pilot signals amongst the antennas 20 themselves, and then calculates RF-chain calibration factors for the antennas 20 based on those pilot signals.
The calibration controller, which may be implemented at the central control node 10 or at some other node in the system 100 (e.g., at one of the antennas 20), coordinates the transmission and reception of pilot signals amongst the antennas 20 during different time intervals. These time intervals may be different time intervals of the original TDD cycle, such that instead of transmitting downlink signals to or receiving uplink signals from mobile terminals 40 during a given time interval of the original TDD cycle, the antennas 20 transmit and receive pilot signals amongst themselves for calibration.
In one embodiment shown in
Based on these pilot signals received amongst the antennas 20, the calibration controller is configured to calculate an RF-chain calibration factor ρ(f) for each of one or more of the antennas 20. The calibration factor ρ(f) calculated for a given antenna 20 accounts for a frequency response H(f) difference between the RF chain of that antenna 20 and the RF chain of another one of the antennas 20.
Consider, for instance,
R
12(f)=HTX1(f)H12(f)HRX2(f), (2)
assuming negligible noise. Likewise, with respect to antenna 20-3, the pilot signal propagates through a channel having a frequency response H13(f), and then propagates through the RX chain of antenna 20-3, having a frequency response HRX3(f), back to the central control node 10. The pilot signal received from the RX chain of antenna 20-3 when antenna 20-1 transmits a pilot signal during the first receive interval 510 is therefore given by:
R
13(f)=HTX1(f)H13(f)HRX3(f) (3)
Similarly, the pilot signals received from the RX chains of antennas 20-1 and 20-3 when antenna 20-2 transmits a pilot signal during the second receive interval 530 are given by:
R
21(f)=HTX2(f)H21(f)HRX1(f)
R
23(f)=HTX2(f)H23(f)HRX3(f) (4)
And finally, the pilot signals received from the RX chains of antennas 20-1 and 20-2 when antenna 20-3 transmits a pilot signal during the third receive interval 540 are given by:
R
31(f)=HTX3(f)H31(f)HRX1(f)
R
32(f)=HTX3(f)H32(f)HRX2(f) (5)
Based on these received pilot signals, the calibration controller in one embodiment calculates RF-chain calibration factors ρ12(f), ρ13(f) for antennas 20-2 and 20-3, respectively, using antenna 20-1 as a reference for the calibration. That is, the calibration controller calculates an RF-chain calibration factor ρ12(f) for antenna 20-2 that accounts for a frequency response difference between the RF chain of antenna 20-2 and the RF chain of antenna 20-1. Likewise, the calibration controller calculates an RF-chain calibration factor ρ13(f) for antenna 20-3 that accounts for a frequency response difference between the RF chain of antenna 20-3 and the RF chain of antenna 20-1. Specifically, the calibration controller calculates ρ12(f) and ρ13(f) as:
Of course, much like the downlink and uplink channels between an antenna 20 and a mobile terminal 40, the transmit and receive channels between any two antennas 20 are reciprocal. Accordingly, H12(f)H21(f) and H13(f)=H31(f), meaning that ρ12(f) and ρ13(f) simplify to:
Then, assuming that each RF chain has unit gain, introduces only time delay and phase rotation, and thus satisfies HTXi(f)H*TXi(f)=HRXj(f)H*RXj(f)=1 for all TX chains i and all RX chains j:
Having calculated the calibration factors in this way, the calibration controller in some embodiments adjusts the RF chain of antenna 20-2 and 20-3 based on the corresponding calibration factors, ρ12(f) and ρ13(f), respectively. For example, the calibration controller may adjust the TX chains of antennas 20-2 and 20-3 according to:
{tilde over (H)}
TX2(f)=ρ12(f)HTX2(f),
{tilde over (H)}
TX3(f)=ρ13(f)HTX3(f), (9)
such that:
H
TX1(f)H*RX1(f)={tilde over (H)}TX2(f)H*RX2(f)={tilde over (H)}TX3(f)H*RX3(f)= . . . (10)
Alternatively, the calibration controller may adjust the RX chains of antennas 20-2 and 20-3 according to:
{tilde over (H)}*
RX2(f)=ρ12(f)H*RX2(f),
{tilde over (H)}*
RX3(f)=ρ13(f)H*RX3(f), (11)
such that:
H
TX1(f)H*RX1(f)=HTX2(f){tilde over (H)}*RX2(f)=HTX3(f){tilde over (H)}*RX3(f)= . . . (12)
Those skilled in the art will nonetheless appreciate that the above represents a non-limiting example of the calibration process, and that the calibration controller may therefore calculate RF-chain calibration factors ρ(f) for each of one or more of the antennas 20-1, 20-2, and 20-3, using any one of those antennas 20 as a reference for the calibration.
Further, those skilled in the art will appreciate that the calibration controller is not limited to adjusting the TX or RX chains of the antennas 20 based on the corresponding calibration factors ρ(f), as described above. Rather, the calibration controller may instead be configured in some embodiments to apply the corresponding calibration factors ρ(f) to a downlink data transmission being sent from those antennas 20, and thereby adjust the frequency response of the outgoing data transmission rather than the TX chains themselves. Likewise, the calibration controller may be configured in some embodiments to apply the corresponding calibration factors ρ(f) to an uplink data transmission being received at those antennas 20, and thereby adjust the frequency response of the received data transmission rather than the RX chains themselves.
Still further, those skilled in the art will appreciate that transmitting and receiving pilot signals amongst the antennas 20 during different RX intervals of the original TDD cycle in this way may, at least temporarily, reduce the uplink capacity of the system 100. That is, the calibration controller may temporarily suspend scheduling of uplink data transmissions from the mobile terminals 40 to the antennas 20 during these RX intervals, so that the antennas 20 can transmit and receive the pilot signals amongst themselves during that time. This, of course, may be acceptable during periods of relatively low utilization of the system's uplink capacity, but may be unacceptable during periods of relatively high utilization of the uplink capacity—especially if the system's downlink capacity is experiencing relatively low utilization.
According to some embodiments, therefore, instead of coordinating the transmission and reception of pilot signals amongst the antennas 20 during different RX intervals of the original TDD cycle, the calibration controller may coordinate such transmission and reception during different TX intervals of the original TDD cycle.
Similar to previous embodiments, transmitting and receiving pilot signals amongst the antennas 20 during different TX intervals of the original TDD cycle in this way may, at least temporarily, reduce the downlink capacity of the system 100. That is, the calibration controller may temporarily suspend scheduling of downlink data transmissions from the antennas 20 to the mobile terminals 40 during these TX intervals, so that the antennas 20 can transmit and receive the pilot signals amongst themselves during that time. This, again, may be acceptable during periods of relatively low utilization of the system's downlink capacity, but may be unacceptable during periods of relatively high utilization of the downlink capacity—especially if the system's uplink capacity is experiencing relatively low utilization.
Accordingly, in some embodiments, the calibration controller is configured to dynamically switch between transmitting and receiving pilot signals during different RX intervals, and transmitting and receiving pilot signals during different TX intervals. This dynamic switching may be based, at least in part, on the current utilization level of the system's downlink and/or uplink capacity.
Regardless, the calibration controller may, in some embodiments, perform the calibration described above with respect to the three exemplary antennas 20-1, 20-2, and 20-3 for a relatively large number of antennas 20. That is, due to the above described coordination of transmission and reception of pilot signals amongst the antennas 20, the calibration process scales even to a relatively large number of antennas 20. These antennas 20 may be for instance all, or substantially all, of the antennas 20 jointly processing a given signal, or all, or substantially all, of the antennas 20 under the control of a given central control node 10 in the system 100.
Regardless of the number of antennas 20 per se, though, the calibration controller may in some embodiments coordinate the transmission and reception of pilot signals from the antennas 20 so as to further reduce the possibility of interference between those pilot signals. As shown in
The same can be said for the second and third time-reuse groups T2 and T3, as illustrated in
In some embodiments, the pilot signals for the different subsets of antennas 20 in a given time-reuse group (and thereby the pilot signals for different frequency-reuse groups) are sent using different contiguous blocks of frequencies. In other embodiments, these pilot signals are sent using different sets of frequencies, where each set includes a number of non-contiguous frequencies spread out across a given frequency band. In these latter embodiments, the calibration controller is advantageously able to calculate the calibration factors ρ(f) based on estimates determined over the entire frequency band, rather than just some sub-block of the frequency band.
Consider, for example,
The same can be said of the pilot signal transmissions from different subsets of time-reuse groups T2 and T3. As also seen in
Of course, much in the same way as described above, the calibration controller is not limited to coordinating the transmission and reception of pilot signals during different RX intervals, but rather may coordinate such transmissions and reception during different TX intervals.
The calibration controller has been described above as being configured to, during a given time interval, transmit pilot signals from only those antennas belonging to one time-reuse group, and to receive those pilot signals at the other, e.g., two, time-reuse groups. In other embodiments, however, the calibration controller is configured to, during a given time interval, transmit pilot signals from those antennas belonging to any of two or more time-reuse groups, and to receive those pilot signals at the other, e.g., one, time-reuse group.
In
Those skilled in the art will readily appreciate that while the calibration controller described above with respect to
Those skilled in the art will also appreciate that the calibration process depicted in the various figures described above may be continually, or periodically, carried out as part of a larger process. The time intervals depicted above therefore may be repeated continually, or periodically, as needed to carry out network-wide calibration.
With the above points of variation and implementation of the self-calibrating multi-antenna system 100 in mind, those skilled in the art will appreciate that a calibration controller of the system 100 generally performs the method illustrated in
Having transmitted these pilot signals, the calibration controller then calculates RF-chain calibration factors ρ(f) for each of one or more antennas 20 in the system 100 based on the received pilot signals (Block 1140). The calibration controller uses these calibration factors ρ(f) in joint processing of the signals, e.g., the downlink data signals, transmitted from the multiple antennas 20 in the system 100 (Block 1150).
Those skilled in the art will appreciate that the calibration controller may include various “circuits,” whether analog, digital, or some combination of both. These circuits may include one or more processors configured with software and/or firmware (e.g., stored in memory) that, when executed by the one or more processors, perform as described above. One or more of these processors, as well as the other digital hardware, may be included in a single application-specific integrated circuit (ASIC), or several processors and various digital hardware may be distributed among several separate components, whether individually packaged or assembled into a system-on-a-chip (SoC).
If the calibration controller 1200 is implemented in a node of the system 100 other than the central control node 10 (e.g., in one of the antennas 20), the controller 1200 may have a control node interface 1230 that communicatively couples the controller 1200 to the central control node 10. Regardless, the controller 1200 includes a TX/RX interface 1240 that communicatively couples the controller 1200 to the transmit-receive RF chains of the antennas 20.
Indeed, all of the foregoing description and the accompanying drawings represent non-limiting examples of the methods and system taught herein. As such, the present invention is not limited by the foregoing description and accompanying drawings. Instead, the present invention is limited only by the following claims and their legal equivalents.