U.S. patent application number 11/123054 was filed with the patent office on 2005-09-22 for ofdm communication channel.
Invention is credited to Hadad, Zion.
Application Number | 20050207334 11/123054 |
Document ID | / |
Family ID | 23961192 |
Filed Date | 2005-09-22 |
United States Patent
Application |
20050207334 |
Kind Code |
A1 |
Hadad, Zion |
September 22, 2005 |
OFDM communication channel
Abstract
In an OFDM-based receiver, means for achieving time
synchronization comprising: A. means for extracting pilot signals
contained in the OFDM received signal; B. means for analyzing the
pilot signals in the frequency domain and for issuing a signal
indicative of a synchronization error in the received signal; C.
means for correcting the synchronization error responsive to the
signal indicative of the synchronization error. In an OFDM-based
receiver, automatic frequency correction means in a subscriber unit
comprising: A. an inner frequency correction loop for generating a
LO frequency related to a frequency of a received signal; B. an
outer frequency correction loop for correcting the LO frequency
according to instructions received from a base station. In an
OFDM-based receiver, a channel sounder comprising: A. means for
extracting pilot signals contained in the OFDM received signal; B.
means for analyzing the pilot signals in the frequency domain and
for issuing signals indicative of a distortion in each pilot
signal, wherein each of said pilot distortion signals comprises
both an amplitude and a phase component; C. means for analyzing the
signals indicative of a distortion in each pilot signal and for
computing therefrom corrective signals for correcting distortions
in the received signal.
Inventors: |
Hadad, Zion; (Rishon Lezion,
IL) |
Correspondence
Address: |
Zion Hada
48 Haalmogim street
Rishon Lezion
IL
|
Family ID: |
23961192 |
Appl. No.: |
11/123054 |
Filed: |
May 6, 2005 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
11123054 |
May 6, 2005 |
|
|
|
09493662 |
Jan 28, 2000 |
|
|
|
Current U.S.
Class: |
370/203 ;
370/350 |
Current CPC
Class: |
H04L 25/03057 20130101;
H04L 27/2662 20130101; H04L 27/2657 20130101; H04L 27/2675
20130101; H04L 25/03159 20130101; H04L 2025/03414 20130101; H04L
2025/03522 20130101; H04L 25/022 20130101; H04L 25/0232
20130101 |
Class at
Publication: |
370/203 ;
370/350 |
International
Class: |
H04J 011/00 |
Claims
What is claimed is:
1. In an OFDM-based receiver, means for achieving time
synchronization comprising: A. means for extracting pilot signals
contained in the OFDM received signal; B. means for analyzing the
pilot signals in the frequency domain and for issuing a signal
indicative of a synchronization error in the received signal; and
C. means for correcting the synchronization error responsive to the
signal indicative of the synchronization error:
2. The synchronization means according to claim 1, wherein the
means for extracting pilot signals comprise FFT means and signal
processing means in the frequency domain.
3. The synchronization means according to claim 1, wherein the
means for extracting pilot signals, the means for analyzing the
pilot signals and the means for correcting the synchronization
error operate continuously in real time to keep the OFDM receiver
synchronized.
4. The synchronization means according to claim 1, wherein the
means for analyzing the pilot signals in the frequency domain
include means for measuring the rate of rotation of the pilot
signals.
5. In an OFDM-based receiver, automatic frequency correction means
in a subscriber unit comprising: A. an inner frequency correction
loop for generating a LO frequency related to a frequency of a
received signal; and B. an outer frequency correction loop for
correcting the LO frequency according to instructions received from
a base station.
6. The automatic frequency correction means according to claim 5,
wherein the inner frequency correction loop includes means for
locking to the frequency of the received signal.
7. The automatic frequency correction means according to claim 5,
wherein the outer loop includes DDS means for generating a signal
at a frequency derived from that of the received signal, modified
according to the instructions received from the base station.
8. In an OFDM-based receiver, a channel sounder comprising: A.
means for extracting pilot signals contained in the OFDM received
signal; B. means for analyzing the pilot signals in the frequency
domain and for issuing signals indicative of a distortion in each
pilot signal, wherein each of said pilot distortion signals
comprises both an amplitude and a phase component; and C. means for
analyzing the signals indicative of a distortion in each pilot
signal and for computing therefrom corrective signals for
correcting distortions in the received signal.
9. The channel sounder according to claim 8, wherein the correction
of the received signal is performed in the complex domain, to
include both gain and phase corrections.
10. The channel sounder according to claim 8, further including
means for computing an average distortion of two adjacent pilots
and for using that average to correct the information between these
pilots.
11. The channel sounder according to claim 8, further including
means for computing, for each frequency between two adjacent
pilots, an interpolated value of the distortion, and for using that
interpolated value to correct the information at that
frequency.
12. The channel sounder according to claim 11, wherein the
interpolation is performed in the time domain or the frequency
domain.
13. The channel sounder according to claim 11, wherein the
interpolation is performed using a low pass filter or a FIR or
convolver.
14. In an OFDM-based receiver, a multipath cancellation system
comprising: A. means for extracting pilot signals contained in the
OFDM received signal; B. means for analyzing the pilot signals in
the frequency domain for generating signals indicative of multipath
reflections; and C. equalizer means for reducing multipath, wherein
the parameters of the equalizer are controlled by the signals
indicative of multipath reflections.
15. The multipath cancellation system according to claim 14,
wherein the equalizer means comprise a transversal filter.
16. The multipath cancellation system according to claim 14,
wherein the analyzing means comprise processing in the frequency
domain, followed with an IFFT.
Description
[0001] The present application is a divisional from U.S. patent
application Ser. No. 09/493,662 entitled "OFDM communication
channel" filed on 28 Jan. 2000, and claims priority therefrom.
FIELD OF THE INVENTION
[0002] This invention relates to OFDM communication channels, and
more particularly to improvements in channel performance using
signal processing of pilot signals in the channel.
BACKGROUND OF THE INVENTION
[0003] Advanced communications today may use the Orthogonal
Frequency Division Multiplex (OFDM) modulation for efficient
transmission of digital signals. These signals may include video,
voice and/or data. OFDM is a commonly used implementation of
Multi-Carrier Modulation (MCM).
[0004] The Orthogonal Frequency Division Multiplex (OFDM) is a
modern advanced modulation method, that achieves better use of the
frequency spectrum.
[0005] OFDM has been used in recent years in many applications
where robustness against severe multipath and interference
conditions is required, or a high system capacity, flexibility in
providing variable bit rate services, scalability and a capability
to perform well in Single Frequency Networks (SNF). OFDM forms the
basis for various communication standards, including for example
the Digital Terrestrial Television Broadcasting, wireless LANs and
Wireless Local Loops.
[0006] OFDM requires an advanced signal processing.
[0007] Thus, a block of information is divided among N frequency
channels, so that a portion of the information is transmitted in
each of the abovementioned channels or frequencies. Since each
channel is orthogonal to the others, a better utilization of the
frequency spectrum is achieved.
[0008] In OFDM, since each symbol is N times longer, the percent
overlap between adjacent symbols decreases, hence the Inter-Symbol
Interference ISI is lower. Still better spectrum utilization is
achieved by QAM (Quadrature Amplitude Modulation) on each of the N
carriers.
[0009] An IFFT (Inverse Fourier Transform) is performed on the
modulated carriers, to form the signal in the time domain that
corresponds to the above modulated carriers. The signal is
transmitted as a frame that contains the block of information to be
transmitted.
[0010] A possible problem in the above modulation scheme may be an
error in the time synchronization between signals.
[0011] When there is a time synchronization error, the signals
after FFT in the various subchannels are rotated with respect to
each other.
[0012] This effect creates interference within the subchannel.
[0013] Another problem is a frequency error between the transmitted
signal and the receiver. A frequency error generates a frequency
shift that may change the location of symbols and/or may generate
interference between symbols.
[0014] Because of channel imperfection, a time or phase delay may
be generated between the various parts of the spectrum of the
transmitted signal. This distortion of the frequency spectrum of
the transmitted signal may interfere with the signal reconstruction
in the receiver.
[0015] The problem is further aggravated by multipath.
[0016] Multipath may cause several replicas of a signal to be
received, each possibly having a different time delay, amplitude
and polarity.
[0017] These signals may result in interference between adjacent
transmitted frames.
[0018] Prior art systems apparently are different.
[0019] Thus, Seki et al., U.S. Pat. No. 5,771,224 , discloses an
orthogonal frequency division multiplexing transmission system and
transmitter and receiver therefor. It transmits an OFDM
transmission frame, with null symbols and reference symbols being
placed in the beginning portion of the frame and QPSK symbols are
placed in an information symbol data region in the frame, with
equal spacing in time and frequency.
[0020] The carrier amplitude and phase errors are corrected by a
correction information producing section on the amplitude and phase
variations of the received signal detected by the variation
detector to produce corrected information.
[0021] Baum et al., U.S. Pat. No. 5,802,044 , discloses a
multicarrier reverse link timing synchronization system. A center
station transmits a forward link signal, receives a reverse link
signal, and determines a timing offset for signals received on a
reverse link timing synchronization channel.
[0022] A reverse link symbol timing synchronization can be used in
a system having a plurality of transmitting overlap bandwidth
subscriber units on an OFDM-like spectrally overlapping reverse
channel. The modulation method may comprise M-ary Quadrature Phase
Shift Keying(M-PSK), M-ary Quadrature Amplitude Modulation (QAM) or
other digital modulation method.
[0023] Gudmundson et al., U.S. Pat. No. 5,790,516 , discloses a
method and system for pulse shaping for data transmission in an
orthogonal frequency division multiplexed (OFDM) system.
[0024] Yamauchi et al., U.S. Pat. No. 5,761,190, discloses an OFDM
broadcast wave receiver. An OFDM (Orthogonal Frequency Division
Multiplex) broadcast wave receiver for receiving an OFDM broadcast
wave.
[0025] It automatically discriminates whether the received signal
is of a wide band or a narrow band by determining if a carrier
signal having a predetermined frequency is present among signals of
a plurality of frequencies, acquired by OFDM demodulation of the
reception signal by demodulation means.
[0026] It also controls the demodulating operation of the
demodulation means in accordance with the discrimination result to
thereby acquire a demodulated signal.
[0027] Schmidl et al., U.S. Pat. No. 5,732,113, discloses a a
method for timing and frequency synchronization of OFDM signals. It
relates to a method and apparatus that achieves rapid timing
synchronization, carrier frequency synchronization, and sampling
rate synchronization of a receiver to an orthogonal frequency
division multiplexed (OFDM) signal. The method uses two OFDM
training symbols to obtain full synchronization in less than two
data frames. A first OFDM training symbol has only even-numbered
sub-carriers.
[0028] A second OFDM training symbol has even-numbered sub-carriers
differentially modulated relative to those of the first OFDM
training symbol by a predetermined sequence.
[0029] Synchronization is achieved by computing metrics which
utilize the unique properties of these two OFDM training symbols.
Timing synchronization is determined by computing a timing metric
which recognizes the half-symbol symmetry of the first OFDM
training symbol. Carrier frequency offset estimation is performed
in using the timing metric as well as a carrier frequency offset
metric which peaks at the correct value of carrier frequency
offset. Sampling rate offset estimation is performed by evaluating
the slope of the locus of points of phase rotation due to sampling
rate offset as a function of sub-carrier frequency number.
[0030] Awater et al., U.S. Pat. No. 5,862,182, discloses an OFDM
digital communications system using complementary codes.
[0031] The encoding/transmission of information in an OFDM system
is enhanced by using complementary codes. The complementary codes,
more particularly, are converted into phase vectors and the
resulting phase vectors are then used to modulate respective
carrier signals. The modulated result is then transmitted to a
receiver which decodes the received signals to recover the encoded
information.
[0032] Isaksson, et al. U.S. Pat. No. 5,812,523, discloses a method
and device for synchronization at OFDM-system.
[0033] A method of demultiplexing OFDM signals and a receiver for
such signals.
[0034] The method is concerned with synchronization in an OFDM
receiver. A signal is read into a synchronization unit, in the time
domain, i.e., before Fourier transforming the signal by means of an
FFT processor. In the synchronization unit, a frame clock is
derived for triggering the start of the FFT process and for
controlling the rate at which data is supplied to the FFT
processor. For OFDM reception, it is vital that the FFT process
commences at the right point in time. Once the frame clock has been
recovered, a frequency error can be estimated by the
synchronization unit. The frequency error is used to control an
oscillator which generates a complex rotating vector which is, in
turn, multiplied with the signal to compensate for frequency
errors. The method can be used both with OFDM systems in which
symbols are separated by guard spaces, and with OFDM systems in
which symbols are pulse shaped.
[0035] Kim, U.S. Pat. No. 5,963,592, discloses an adaptive channel
equalizer for use in digital communication system utilizing OFDM
method. An adaptive channel equalizer for use in OFDM receiver is
disclosed. The adaptive channel equalizer comprises a first complex
multiplier for outputting a first in-phase complex multiplication
signal and a first quadrature phase complex multiplication signal;
a reference signal generator for generating a reference signal; an
error calculator for outputting an in-phase error signal and a
quadrature phase error signal; a delay unit for outputting an
in-phase delay signal and a quadrature phase delay signal; a gain
controller for outputting an in-phase gain control signal and a
quadrature phase gain control signal;
[0036] a second complex multiplier for outputting a second in-phase
complex multiplication signal and a second quadrature phase complex
multiplication signal; an adder for outputting updated in-phase and
quadrature phase coefficients; an address generator for generating
a write address signal and a read address signal;
[0037] a storage unit for storing the updated in-phase and
quadrature phase coefficients, and outputting the updated
coefficients; an initial coefficients generator for generating an
initial coefficients; a selecting signal generator for generating a
selecting signal; and a multiplexing unit for selecting one of the
initial coefficients and the updated coefficients according to the
selecting signal.
[0038] Seki et al., U.S. Pat. No. 5,694,389, discloses an OFDM
transmission/reception system and transmitting/receiving apparatus.
The apparatus improves the frequency acquisition range and the
resistance to multipath interference. In a digital signal
transmission system using OFDM, on the transmission side, some or
all of a plurality of equidistant carrier positions are treated as
reference carrier positions. The actual transmitted carriers are
arranged in a predetermined pattern non-equidistant to the
frequency carrier positions to form an OFDM symbol.
[0039] This OFDM symbol is periodically transmitted as frequency
reference symbols. On the reception side, the carrier arrangement
pattern of the frequency reference symbols is detected, a carrier
frequency offset is detected from the detected pattern offset, and
the carrier frequency is compensated based on the frequency
offset.
[0040] Cimini et al., U.S. Pat. No. 5,914,933, discloses a
clustered OFDM communication system. A multicarrier communication
system for wireless transmission of blocks of data having a
plurality of digital data symbols in each block. The communication
system includes a device for distributing the digital data symbols
in each block over a plurality of clusters, each of the clusters
receiving one or more digital data symbols. The digital data
symbols are encoded in each of the cluster; and modulated in each
cluster to produce a signal capable of being transmitted over the
sub-channels associated with each cluster.
[0041] A transmitter thereafter transmits the modulated signal over
the sub-channels. By distributing the modulated signal over a
plurality of clusters, overall peak-to-average power (PAP) ratio is
reduced during transmission and transmitter diversity is
improved.
[0042] Williams et al., U.S. Pat. No. 5,815,488, discloses a
multiple user access method using OFDM. A communication method
enables a plurality of remote locations to transmit data to a
central location. The remote locations simultaneously share a
channel and there is a high degree of immunity to channel
impairments.
[0043] At each remote location, data to be transmitted is coded by
translating each group of one or more bits of the data into a
transform coefficient associated with a frequency in a particular
subset of orthonormal baseband frequencies allocated to each remote
location. The particular subset of orthonormal baseband frequencies
allocated to each location is chosen from a set of orthonormal
baseband frequencies. At each remote location, an electronic
processor performs an inverse orthogonal transform (e.g., an
inverse Fourier Transform) on the transform coefficients to obtain
a block of time domain data. The time domain data is then modulated
on a carrier for transmission to the central location.
[0044] Preferably, the time intervals for data transmission at the
different remote locations are aligned with each other. In one
embodiment of the invention, all of the baseband frequencies are
allocated to a single particular remote location for one time slot.
At the remote location, data is received from a plurality of remote
locations. The data is demodulated to obtain baseband time domain
data. The orthogonal transform is performed on this data to obtain
transform coefficients. Each transform coefficient is associated
with a baseband frequency. The central location keeps track of
which baseband frequencies are allocated to which remote location
for subsequent translation of each transform coefficient.
[0045] Isaksson, U.S. Pat. No. 5,726,973, discloses a method and
arrangement for synchronization in OFDM modulation. A method and an
arrangement for synchronization in OFDM modulation. Frequency
errors of an IF clock and a sampling clock are controlled by
estimating the deviation of the sampling clock and, respectively,
the IF clock for two subcarriers with different frequencies.
[0046] According to the invention, the frequencies are chosen
symmetrically around zero and the absolute phase errors are
detected for both subcarriers.
[0047] Timing errors and phase errors are formed from the absolute
phase errors in order to generate two control signals. The first
control signal is formed from the deviation of the sampling clock
and the timing error for controlling the sampling clock while the
second control signal is formed from the deviation of the IF clock
and the phase error for controlling the IF clock.
[0048] Wright, U.S. Pat. No. 5,838,734, discloses means for
compensation for local oscillator errors in an OFDM receiver. A
receiver for orthogonal frequency division multiplexed signals
includes means for calculating the (discrete) Fourier Transform of
the received signal, and means for calculating the phase error due
to local oscillator errors.
[0049] McGibney, U.S. Pat. No. 5,889,759, discloses an OFDM timing
and frequency recovery system. A synchronizing apparatus for a
differential OFDM receiver that simultaneously adjust the radio
frequency and sample clock frequency using a voltage controlled
crystal oscillator to generate a common reference frequency. Timing
errors are found by constellation rotation. Subcarrier signals are
weighted by using complex multiplication to find the phase
differentials and then the timing errors. The reference oscillator
is adjusted using the timing errors. Slow frequency drift may be
compensated using an integral of the timing error. Frequency offset
is found using the time required for the timing offset to drift
from one value to another.
[0050] Background material on advanced modulation techniques and
related communication topics may be found in the following
articles:
[0051] Scott L. Miller and Robert j. O'Dea, "Peak Power and
Bandwidth Efficient Linear Modulation", IEEE transactions on
communications, Vol. 46, No. 12, pp. 1639-1648, December 1998.
[0052] Kazuki Maeda and Kuniaki Utsumi, "Bit-Error of M-QAM Signal
and its Analysis Model for Composite Distortions in AM/QAM Hybrid
Transmission", IEEE transactions on communications, Vol. 47, No. 8,
pp. 1173-1180, August 1999.
[0053] Kazuki Maeda and Kuniaki Utsumi, "Performance of
Reduced-Bandwidth 16 QAM with Decision-Feedback Equalization", IEEE
transactions on communications, Vol. COM-35, No. 7, pp. 1173-1180,
July 1987.
[0054] Background material on phase noise in advanced communication
systems may be found in the following references:
[0055] Yossi Segal and Zion Hadad, "OFDMA access method for
HIPERACESS", HARNCl.doc, December 1999.
[0056] Naftali Chayat, "Updated Submission Template for
TGa--Revision 2", IEEE 802.11-98/156r2, March 1998.
[0057] Alcatel, Bosch, Ericsson, Lucent, Nokia, Siemens AG and
Siemens ICN, "Proposal for the Adoption of the TDMA Access Scheme
in HIPERACCESS", HA16ERI1a.doc, December 1999.
[0058] Thierry Pollet, Mark Van Bladel and Marc Moeneclaey, "BER
Sensitivity of OFDM Systems to Carrier Frequency Offset and Wiener
Phase Noise", IEEE transactions on communications, Vol. 43, No.
2/3/4, pp. 191-193, February/March/April 1995.
[0059] Luciano Tomba, "On the Effect of Wiener Phase Noise in OFDM
Systems", IEEE transactions on communications, Vol. 46, No. 5, pp.
580-583, May 1998.
[0060] Naftali Chayat, "TGa Comparison Matrix per 98/156r2", IEEE
802.11-98/157r5, May 1998.
[0061] ETSI EP BRAN #16 Athens, Greece November 29- Dec. 3, 1999
HA16RNC1Annex.doc page 3 of 13 22-Nov.-99
SUMMARY OF THE INVENTION
[0062] The present disclosure relates to improvements in OFDM-based
digital communications. The scope and spirit of the invention are
better described with the inclusion of specific applications
thereof.
[0063] A possible problem in the above modulation scheme may be an
error in the time synchronization between several signals appearing
at the receiver, or between transmitter and receiver.
[0064] When there is a time synchronization error, the signals
after FFT in the various subchannels are rotated with respect to
each other.
[0065] This effect creates interference within the subchannel.
[0066] One application of the invention relates to receiver
synchronization using means for Automatic Synchronization Control
(ASC).
[0067] The ASC means use an analysis of pilot signals in the
transmitted signal to implement the ASC loop.
[0068] The analysis is performed continuously, in real time. The
correction of detected errors is also performed continuously in
real time.
[0069] The time synchronization error may be evaluated based on the
rate of rotation of the pilot signals. A correction signal is
generated accordingly, to adjust the timing in the receiver to the
received signal. This is implemented in an ASC loop, to achieve
optimal timing for sampling in the A/D converter.
[0070] Another problem is a frequency error between the transmitted
signal and the receiver.
[0071] A frequency error generates a frequency shift that may
change the location of symbols and/or may generate interference
between symbols. The information may be divided between separate
bins, or may be assigned to other than the desired bins. Some
information may be lost because of the frequency shift. The actual
effect in each case (or at any instant in time) depends on the
measure of frequency deviation.
[0072] Real-time means are used to measure the frequency error and
correct for it in an Automatic Frequency Control (AFC) loop.
[0073] A correction signal is generated accordingly, to correctly
tune the receiver to the received signal.
[0074] Thus, the system will adapt to varying channel
characteristics in real time, to achieve improved
communications.
[0075] This may be useful in DVB-T, for example, where there are a
large number of pilot signals available.
[0076] The frequency resulting from the AFC loop is used as a clock
for the receiver and subsequently for the transmitter. A frequency
error may stem from two possible causes:
[0077] A. an undesired difference between the receiver LO (local
oscillator) and the transmit LO.
[0078] B. a frequency Doppler shift because of the movement of the
mobile subscriber.
[0079] This effect, together with means for its correction using a
dual loop AFC, are detailed elsewhere in the present
disclosure.
[0080] A second application relates to a channel sounder. Using
means for analyzing the received pilot signals, a signal processor
can characterize the communication channel. Using the pilots rather
than the information or noise in the channel may achieve a better
performance system.
[0081] The phase and amplitude of the pilots is measured to
evaluate the channel distortion at different frequencies. The
results are used to apply a correction to the received signal whose
subcarriers are located between the pilot signals.
[0082] In one embodiment, the average distortion of two adjacent
pilots is used to correct the information between these pilots.
When the distortion in each pilot is different, the correction may
be in error.
[0083] A better correction may be achieved using an interpolation
process to correct for phase and amplitude of received signals
between any two adjacent pilots. This corrects the distortion of
the signal frequency spectrum, to improve the receiver
performance.
[0084] Interpolation may be used to arrive at a channel estimate
for each channel frequency, and to correct the signal accordingly.
The correction is made in the complex domain, to include gain and
phase corrections. Interpolation may be implemented either in the
time domain or the frequency domain.
[0085] For example, interpolation may be implemented using a low
pass filter or a FIR or convolver.
[0086] Multipath may interfere with reception of wideband signals.
It may cause several replicas of a signal to be received, each
possibly having a different time delay, amplitude and polarity.
These signals may result in interference between adjacent
transmitted frames.
[0087] A method and system for addressing the multipath problem may
include processing in the frequency domain. Thus, the pilots
spectrum is extracted using FFT for example. Multipath may cause
undesired changes in the amplitude and phase in the pilots, which
are correlated from.one spectral line to the other. These changes
are responsive to the time delay in each multipath signal.
[0088] Using signal processing applied to the spectral picture (the
pilots representation in the frequency domain), each pilot signal
can be reconstructed. The changes in the pilots are indicative of
the multipath effects in the channel. The information thus derived
may be used to correct for multipath. Thus, the interference
because of multipath is reduced.
[0089] Moreover, multipath signals may be added to the main path
signal, to actually increase the signal power to improve the signal
to noise ratio.
[0090] Multipath attenuation or cancellation may be achieved using
the measured characteristics of the channel. Multipath can be
corrected for by using an equalizer or transversal filter. The
parameters for the equalizer are derived from the measured channel
characteristics. For each detected multipath, the filter will
generate a correcting signal of the proper time delay, amplitude
and polarity.
[0091] The equalizer parameters may be computed in the frequency
domain, followed with an IFFT. These parameters may be applied to a
transversal filter.
[0092] The above system and method may be advantageously used in
the physical layer specification proposed as BRAN-HA/PHY, for
example.
[0093] Superior performance may be achieved at lower phase
noise.
[0094] Further objects, advantages and other features of the
present invention will become obvious to those skilled in the art
upon reading the disclosure set forth hereinafter.
BRIEF DESCRIPTION OF THE DRAWINGS
[0095] FIG. 1 illustrates the spectrum of an OFDM signal, with
pilots and data.
[0096] FIG. 2 illustrates the phase of the pilots versus
frequency.
[0097] FIG. 3 details the block diagram of a system for
implementing ASC and AFC.
[0098] FIG. 4A illustrates the phase distortion of pilots in a
communication multipath channel, and FIG. 4B illustrates the
amplitude distortion of the pilots.
[0099] FIG. 5 details a block diagram of a system for correcting
the phase and amplitude distortion of signals in a communication
channel.
[0100] FIG. 6 details a block diagram of a system for correcting
the multipath distortion of signals using a LPF.
[0101] FIG. 7 illustrates the multipath effect on the pilots in the
time domain.
[0102] FIG. 8 details a block diagram of a system for correcting
the multipath distortion of signals using means for pilots
analysis.
[0103] FIG. 9 details a block diagram of a decision feedback
equalizer system.
[0104] FIG. 10 illustrates a dual loop system for implementing
Automatic Frequency Control (AFC).
[0105] FIG. 11 illustrates a conceptual block diagram of the
Downstream Encoding and Modulation subsystem.
[0106] FIG. 12 illustrates a conceptual block diagram of the
Downstream Demodulation and Decoding subsystem.
[0107] FIG. 13 illustrates a conceptual block diagram of the
Upstream Encoding and Modulation subsystem.
[0108] FIG. 14 illustrates a conceptual block diagram of the
Upstream Demodulation and Decoding subsystem.
[0109] FIG. 15 details the Crest Factor versus Roll-Off Factor for
Single Carrier.
[0110] FIG. 16 details the BER/SNR for different Crest Factor
values, as achieved by clipping for a DVB-T 16 QAM OFDM Symbol.
[0111] FIG. 17 details the BER/SNR for different Crest Factor
achieved by clipping for an Upstream 16 QAM OFDM Symbol.
[0112] FIG. 18 illustrates Out-Of-Band Spectrum mask for a 8 MHz
DVB-T transmission
[0113] FIG. 19 illustrates the influence of linear Group-Delay in
Single Carrier system
[0114] FIG. 20 illustrates the BER/SNR of the OFDM and S.C. systems
for different Phase Variance (P.V.) values.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0115] A preferred embodiment of the present invention will now be
described by way of example and with reference to the accompanying
drawings.
[0116] FIG. 1 illustrates the spectrum of an OFDM signal, including
pilots and data in the complex frequency domain, with amplitude
axis (I) 110, amplitude axis (Q) 111 and frequency axis 12. The
spectrum includes the spectrum of data, for example 131, 132, 133
and the pilots 141, 142, 143.
[0117] It is assumed that the transmitted signal includes pilots of
equal amplitude and being in phase. Furthermore, the pilots are
equidistant in the frequency domain. These properties are used in
the present invention, as detailed below. The properties of the
pilots are measured and deviations from the transmitted signal are
indicative of distortions in the communication channel. The
measured distortion are used to compute the correction parameters
for the channel.
[0118] FIG. 2 illustrates a possible distortion in the phase of the
pilots versus frequency.
[0119] The graph indicates an example of phase shift in the
frequency domain, with transmit phase axis 151, receive phase axis
152 and frequency axis 12. The transmit pilots 151 are all in
phase. A time difference may cause a phase shift in the received
pilots 152, as illustrated with the phase of the pilots 141, 142,
143. Such a linear change in the phase of pilots may be caused by a
time error in sampling in the receiver. The slope of the graph is
indicative of the time error.
[0120] This may be used in a receiver to correct for
synchronization errors.
[0121] FIG. 3 details the block diagram of a system for
implementing ASC and AFC. The intermediate frequency (IF) input
channel 211 is transferred to a couple of mixers 21 for quadrature
coherent detection. A delay unit (90 degrees) 22 is used to
generate the quadrature reference from a local oscillator (LO)
23.
[0122] The LO 23 may be implemented, for example, using a voltage
controlled oscillator (VCO). The detected signals (I,Q) are
processed in a pair of low pass filters (LPF) 24 and are converted
to digital words in analog to digital converters (ADC) 25.
[0123] The input wideband signal 211 (time domain), after being
transformed into digital form, is applied to an FFT processor 3.
The FFT processor generates the transformed signal 63 (frequency
domain).
[0124] Signal 63 may include data spectrum and pilots, as
illustrated in FIGS. 1 or 2. A pilots extraction unit 27 extracts
the pilots from the signal 63.
[0125] The system further includes a Phase Lock Loop (PLL) 28,
having a first input 281 from the LO 23 and a second input 282, and
two outputs 283, 284.
[0126] The ASC unit 4 detects the slope of the phase of pilots as
illustrated in FIG. 2 and computes therefrom the synchronization
error.
[0127] A corrected timing signal is applied to a numerically
controlled oscillator (NCO) 26. NCO 26 generates the clock for the
ADC 25. Thus, the timing of the sampling of the analog signals is
adjusted responsive to the measured timing error. This will correct
the timing or synchronization error in the receiver.
[0128] Various embodiments of the invention may be implemented. For
example, sample and hold means (not shown) before the ADC 25 may be
used to correct for synchronization errors.
[0129] Thus, automatic synchronization control is achieved, wherein
ASC unit 4 measures, in real time, the synchronization error and
closes a loop to correct it. The synchronization may have to change
during a communications session. The above loop will change the
synchronization as required, to achieve a system that is adaptive
to changing channel conditions.
[0130] The ASC is performed automatically and without interfering
with the actual communications--no additional synchronization
signals are added and no other changes are required in the
transmitted signals.
[0131] Thus, a possible error in the time synchronization between
signals will be corrected. The reduction or elimination of time
synchronization errors will keep the signals in the various
channels orthogonal to each other, as they should be.
[0132] This may reduce or eliminate a cause of interference between
channels.
[0133] Thus, the system will adapt to varying channel
characteristics in real time, to achieve improved
communications.
[0134] Another problem is a frequency error between the transmitted
signal and the receiver. A frequency error generates a time-varying
phase error, to result in a rotation of the pilot signals of FIG. 2
. Thus, for a frequency error the slope of the pilots will
continuously change at a specific rate.
[0135] The system may detect such a change in the slope of pilots
phase and may compute therefrom the frequency error in the
receiver. This function is implemented in the AFC unit 5. As a
frequency error is detected, AFC unit 5 will issue a correction
signal to VCO 23.
[0136] Thus, automatic frequency control is achieved, wherein AFC
unit 5 measures, in real time, the frequency error and closes a
loop to correct it. The received frequency may change during a
communications session. The above loop will tune the VCO 23 as
required, to achieve a system that is adaptive to changing channel
conditions.
[0137] Real-time means are used for AFC. The frequency error is
evaluated based on the rate of rotation of the pilot signals. A
correction signal is generated accordingly, to correctly tune the
receiver to the received signal.
[0138] The AFC is performed automatically and without interfering
with the actual communications--no additional synchronization
signals are added and no other changes are required in the
transmitted signals.
[0139] The above AFC and ASC systems and methods may be useful in
wideband signals like DVB-T, for example, where there are a large
number of pilot signals available.
[0140] FIG. 4A illustrates the phase distortion of pilots in a
communication channel. Whereas FIG. 2 illustrated a phase
distortion due to a timing delay only, an actual channel may cause
a more complex distortion, where the phase differences between
pilots does not change in a linear fashion.
[0141] Moreover, the amplitude of the pilots may change as well.
That is, for each frequency the channel introduces a distortion
characterized by an amplitude and phase change in the signal. This
channel effect causes a distortion in the transmitted signal and
may reduce the performance of the communication system.
[0142] This channel effect is illustrated in the frequency domain,
with phase axis 15 and frequency axis 12, with pilots 141, 142, 143
each possibly having a different phase.
[0143] The other distortion effect is shown in FIG. 4B, that
illustrates the amplitude distortion of the pilots, with amplitude
axis 11 and frequency axis 12 and pilots 141, 142, 143 of possibly
a different amplitude each.
[0144] Method for Channel Distortion Correction
[0145] The following method may be used to correct for phase and
amplitude distortion in the channel:
[0146] A. measuring the phase of each pilot in a receiver.
Measuring the amplitude of each pilot as well.
[0147] B. computing a correction factor for each pilot, to bring
all the pilots in phase and to an equal amplitude. The correction
factors have a phase shift component and a gain component.
[0148] C. applying the correction factors to the received signals.
Between each two adjacent pilots, the correction factor may be the
average of the factors for these two pilots. Alternately, separate
correction factors may be computed for each frequency using an
interpolation method. This may allow to correct each frequency (or
each output of the FFT) with its individually computed,
corresponding correction factor.
[0149] D. repeating steps A-C all the time, to measure the channel
characteristics in real time and to correct in real time for
changing channel properties.
[0150] End of method.
[0151] Preferably, the above method is implemented after achieving
good frequency lock and synchronization in the receiver. Then,
phase rotation or linear phase change effects are removed and only
remains the distortion caused by the channel to correct.
[0152] FIG. 5 details a block diagram of a system for correcting
the phase and amplitude distortion of signals in a communication
channel.
[0153] This system may be used to implement the method detailed
above for correction of the phase and amplitude distortion in the
channel.
[0154] A receiver 2 may receive and detect a signal, that is
transferred to the FFT processor 3 for computing the spectrum of
the signal. The signal in the frequency domain is transferred to a
pilots extraction and analysis unit 71.
[0155] The unit 71 includes means for:
[0156] A. extracting the pilots from the received signals
[0157] B. analyzing the pilots to detect distortion in phase or
amplitude, as detailed above. These distortions are indicative of
the distortion in the communication channel.
[0158] C. computing the complex correction coefficients for the
various frequencies in the signal, using information derived from
pilots in step (B) above. A possible method may use interpolation.
Averaging of adjacent pilots or other methods may be used as
well.
[0159] D. applying the correction coefficients, as correction
signals 64 (phase and amplitude), to the signal correction unit
72.
[0160] The transformed signal 63 (frequency domain) is transferred
to unit 72, where the correction coefficients are applied to
correct it. This results in the corrected signal 65 (frequency
domain) out of unit 72.
[0161] The above system and method may be used to implement a
channel sounder. Using means for analyzing the received pilot
signals, a signal processor can characterize the communication
channel.
[0162] The phase and amplitude of the pilots is measured to
evaluate the channel distortion at different frequencies. The
results are used to correct the received signal accordingly.
Interpolation may be used to correct for phase and amplitude of
received signals between any two adjacent pilots.
[0163] This system and method corrects the distortion of the signal
frequency spectrum, to improve the receiver performance.
[0164] FIG. 6 details a block diagram of a system for correcting
the multipath distortion of signals using a Low Pass Filter
LPF.
[0165] The system includes a receiver 2 for a received signal. The
input wideband signal (time domain) from receiver 2 is transferred
to an FFT processor 3, that generates a transformed signal in the
frequency domain. This signal is transferred to a pilots extraction
and analysis unit 71, that extracts the pilots from the received
signal. A Low Pass Filter (LPF) 73 is used to measure the
multipath, applying a time-domain processing to the pilots spectrum
that is presented to the LPF as a time-varying signal. Multipath
causes changes in the pilots, that are detected in the LPF.
[0166] The resulting multipath information is applied to a channel
equalizer 74. The channel equalizer 74 also receives the received
signal (in frequency domain) from the FFT processor 3. Unit 74 then
corrects the received signal for multipath. The corrected signal 65
(frequency domain) is the output of the system.
[0167] The above system may be used to correct for multipath, that
may interfere with the reception of wideband signals. It may cause
several replicas of a signal to be received, each possibly having a
different time delay, amplitude and polarity. These signals may
result in interference between adjacent transmitted frames.
[0168] The LPF as detailed is one possible embodiment of means for
time filtering in the frequency domain. The LPF is applied to the
spectral picture (the pilots representation in the frequency
domain), so that each pilot signal can be reconstructed. Multipath
signals are added to the main path signal, to actually increase the
signal power to improve the signal to noise ratio. Furthermore, the
interference because of multipath is reduced.
[0169] FIG. 7 illustrates the multipath effect on the pilots in the
time domain, with amplitude axis 11 and time axis 16. The signal
illustrated is one example of multipath. The pilots are extracted
from the signal and combined in the time domain. A pulse train in
the frequency domain will result in a pulse in the time domain,
this is the pilots pulse 17.
[0170] If there is multipath, it will result in a pulse with a
specific delay, according to the time delay of the multipath
channel in the communication path. Thus, for example, the channel
may have a first multipath pulse 171 and second multipath pulse
172, having a time delay 161 and 162, respectively.
[0171] FIG. 8 details a block diagram of a system for correcting
the multipath distortion of signals using means for pilots
analysis.
[0172] The system may use the above detailed multipath effect, as
detailed with reference to FIG. 7.
[0173] An FFT processor 3 computes the spectrum of the received
signals, that is transferred to unit 71. The pilots extraction and
analysis unit 71 extracts only the pilots in the received signal.
The pilots data undergoes an inverse FFT in IFFT unit 75. The
output 751 of unit 75 may have the shape illustrated in FIG. 7,
that is each multipath path results in a pulse with a
characteristic amplitude, time delay and polarity. Output 751
comprises the channel sounder output of the system.
[0174] The information regarding each multipath is applied to an
equalizer coefficients calculation unit 77.
[0175] Unit 77 computes the coefficients to be used in channel
equalizer unit 76, responsive to the measured channel information
from the channel sounder. The computed coefficients are transferred
to unit 76.
[0176] The unit 76 operates in the time domain to add or subtract
each signal from multipath, to result in a corrected signal 66
(time domain).
[0177] Thus, multipath attenuation or cancellation is achieved
using the measured characteristics of the channel.
[0178] Multipath can be corrected by using an equalizer or
transversal filter. For each detected multipath, the filter will
generate a correcting signal of the proper time delay, amplitude
and polarity.
[0179] As multipath is corrected, two benefits may be achieved: a
signal with no multipath or reduced multipath may result in
improved communications; and, since now the multipath signal is
added in phase, it may actually increase the power of the received
signal, to improve the signal to noise ratio in the system.
[0180] FIG. 9 details a block diagram of a decision feedback
equalizer system (DFE).
[0181] The system implements a multi-stage equalization and error
correction method to be detailed below.
[0182] An input (baseband) 960 is connected to a recording unit
961. This allows the same frame to be played several times into the
processing system. This allows for a simpler, lower cost
implementation. Otherwise, separate units may be used for the
various processing stages, and the unit 961 may not be required in
that case.
[0183] A combiner 962 combines the input signal from unit 961 with
feedback signals from a processor, that may be implemented with FIR
975 and combiner 976.
[0184] A FIR 963 filters the input signals, together with a FIR
combiner/bypass unit 964.
[0185] An equalizer coefficients calculation unit 969 provides the
coefficients for the FIR. Alternately, only the middle tap of the
FIR is output to the FFT 965. To this effect, unit 969 sets all the
FIR coefficients to zero, except the middle tap, that is set to 1
or other nonzero value.
[0186] After the FFT in unit 965, the signal is transferred to
pilot extraction unit 967. This is followed by IFFT 968 and the
equalizer coefficients calculation unit 969, based on the pilots
values in the time domain.
[0187] A switch 971 allows to transfer the equalized received
signal to error detection and correction unit 972 (EDC). The output
973 is the data output of the system, after equalization and error
detection and correction.
[0188] A transmit signal synthesizer 974 is used to generate a
replica of the received signal with the estimated multipath, in
combination with FIR 975 and combiner 976.
[0189] The resulting signal is applied to combiner 962 to remove
multipath to further enhance the received signal.
[0190] Equalization and Error Correction Method
[0191] The system detailed in FIG. 9 may implement a decision
feedback equalizer method comprising the following steps:
[0192] A. record a frame of received data
[0193] B. received data passes through an equalizer (FIR) that is
set to bypass mode, that is all the FIR coefficients are set to
zero, except the middle tap, that is set to 1 or other nonzero
value. This will not filter the signal, however the delay of the
FIR is taken into account.
[0194] C. perform an FFT of the received frame
[0195] D. pilots extraction
[0196] E. IFFT
[0197] F. FIR coefficients calculation and application to the FIR.
Subsequent frames may be used to update the coefficients in a
pipeline fashion. Thus, in future frames the step (B) will use
coefficients computed for the previous frame rather than zero value
coefficients.
[0198] G. the recorded frame is again applied to the system,
however this time the equalizer (FIR) corrects the input data
according to the measured coefficients.
[0199] H. error detection and correction
[0200] I. a replica of the transmitted signal is synthesized, based
on the corrected input signal. The synthesized signal contains the
measured multipath signals, that are generated in a FIR and
combiner.
[0201] J. the recorded frame is again applied to the system,
however this time the replica of the multipath is subtracted from
the input signal.
[0202] K. error detection and correction
[0203] L. output data.
[0204] End of method.
[0205] A possible problem in wireless is a frequency error between
the transmitted signal and the receiver.
[0206] The frequency resulting from the AFC loop is used as a clock
for the receiver and subsequently for the transmitter. A frequency
error may stem from two possible causes:
[0207] A. an undesired difference between the receiver LO (local
oscillator) and the transmit LO.
[0208] B. a frequency Doppler shift because of the movement of the
mobile subscriber.
[0209] This effect, together with means for its correction using a
dual loop AFC, are detailed with reference to FIG. 10.
[0210] FIG. 10 illustrates a dual loop system for implementing
Automatic Frequency Control (AFC).
[0211] The system includes an inner local loop in the subscriber
unit, and an outer loop implemented with components both in the
subscriber unit and the base station.
[0212] The inner loop includes an AFC loop 822 connected to a
subscriber receiver 821 for locking the frequency of receiver 821
to the frequency of the signal received from the base station. For
example, unit 822 may lock the local oscillator to a pilot signal
received from the base station. Accordingly, unit 822 generates a
receiver clock 8221 for the receiver 821. Unit 822 also generates a
transmitter clock 8222 that is transferred to the means for
generating the transmit frequency. In the example as illustrated,
the embodiment of these means is the DDS Tx 826.
[0213] The transmit frequency out of unit 826 is used in the Tx
subscriber 827, that is the transmitter of the subscriber unit, for
transmission to the base station.
[0214] This loop solves the problem of tuning the mobile receiver
to the base station transmissions. The subscriber frequency may be
in error, however, for various reasons. For example, movement of
the subscriber unit may result in a Doppler frequency shift of the
signal received from the base station. The receiver will lock to
the shifted frequency.
[0215] The signal received at the base station will have double
that frequency shift, because of the relative movement between base
and mobile station.
[0216] As various subscriber units will transmit with a frequency
error, the receiver in the base may have difficulty in effectively
separating these receptions.
[0217] To solve these frequency errors, a second (outer) loop is
added, wherein the base stations measures the frequency deviations
of each subscriber and issues instructions to each subscriber to
correct its transmit frequency.
[0218] The outer loop is implemented, in the example as illustrated
in FIG. 10, as follows: The BS Rx 812 (base station receiver)
receives transmissions from mobile subscribers.
[0219] Frequency offset unit 813 measures the frequency error in
the received signal, that is the difference between the actual
received frequency and the precise frequency that was allocated to
that subscriber. The results of the measurement are transferred to
a frequency correction (Up/Down) unit 814. Unit 814 generates
frequency correction messages 815 that are transmitted through the
BS Tx 811 (base station transmitter) to the mobile subscriber.
[0220] In the mobile unit, these messages are received in receiver
821 and are transferred to the information extraction unit 823. The
decoded messages are transferred to the AFC loop closing unit 824,
that controls the instruction from base application unit 825.
[0221] The reconstructed frequency control signals (frequency
correction Up/down instructions) are transferred to the DDS
826.
[0222] The DDS 826 includes means for performing a frequency shift
according to the instructions received from unit 825.
[0223] Thus, the frequency at the output of DDS 826 is derived from
the frequency of the received signal, corrected according to
instructions from the base stations.
[0224] The inner, local frequency control loop sets the frequency
according to that of the received signal.
[0225] The outer frequency control loop corrects the above
frequency setting according to instructions from the base
station.
[0226] The DDS 826 actually forms the transmitter local oscillator.
Its output is transferred to the transmitter 827.
[0227] The above system and method may be advantageously used in
the physical layer specification proposed as BRAN-HA/PHY, for
example. Following is a detailed description of this embodiment of
the invention and its estimated performance.
[0228] It uses an OFDMA access method for the access method for
BRAN-HA /PHY Following is a description of this embodiment of
invention.
[0229] 1. Overview
[0230] Following is a general description of a physical layer
specification proposed as the BRAN- HA/PHY. In order to leverage
existing technology and reduce costs this proposal uses many of the
ETSI Digital Video Broadcasting (DVB) standard for terrestrial
broadcasting in the downstream channel (Base Station to Subscriber
Unit). In addition, this proposal includes physical elements and
implementation aspects that specifically address the challenges to
operating reliably in the 20-60 GHz band.
[0231] 2. Duplexing Technique
[0232] The proposed physical layer is based on Frequency Division
Duplexing (FDD), which provides a separate frequency assignment for
the upstream and down stream channels. We can also use a
modification of the OFDM modulation parameters in order to operate
the system in Time Division Duplexing (TDD) or in Half Frequency
Division Duplexing (H-FDD).
[0233] 3. Multiple Access Method
[0234] The proposed upstream physical layer is based on the use of
a combination of Time Division Multiple Access (TDMA) and
Orthogonal Frequency Division Access (OFDMA). In particular, the
upstream is divided into a number of "time slots" as defined by the
MAC layer. Each time slot (sized to duration of one OFDM symbol) is
then divided in the frequency domain into groups of sub-carriers
referred to as subchannels. The MAC layer controls the assignment
of subchannels and time slots (by bandwidth on demand and Data Rate
on demand). This initial proposal focuses on the efficient
transport of ATM cells and IP packets in the upstream and down
stream channels.
[0235] 4. Downstream Transport Stream and Physical Layer
[0236] The downstream physical layer uses aspects of the
well-proven DVB-T physical layer. This standard uses the OFDM as
its modulation technique. This standard is based on the
transmission of packetized digital video corresponding to MPEG-2.
In particular a transmission convergence layer can be designed to
efficiently transport ATM cells and IP packets (although any frame
structure can be used, the MPEG-2 is widely used today). An OFDM
symbol will be divided (in the frequency domain) into groups. The
first group is a group, which will be dedicated for the broadcast
of MPEG-2 transport and can be used in a SFN as the broadcasting
area.
[0237] The MAC layer for fast feedback or response will use another
group, the last group will be allocated for dedicated channels and
could carry different information in a SFN configuration. We shall
indicate that the broadcasting subcarriers group shall vary as
needed, if there is no need for any broadcasting all of its
subcarriers group shall be used by the dedicated channels. The
encoding and decoding functions for the different group types are
summarized in the next block diagram, the functions for the MPEG-2
data stream and for the dedicated channels are adopted from the
DVB-T standard. (FIG. 1).
[0238] FIG. 11 illustrates a conceptual block diagram of the
Downstream Encoding and Modulation subsystem. The subsystem may be
used for several channels, for example one for broadcasting MPEG-2
850, another for dedicated MPEG-2 851 , and one for MAC messages
852 . The processing in each channel may include a randomization
unit 830, an RS coder (204,188) 831, a convolutional interleaver
832, convolutional encoding and puncturing unit 833, bit
interleaver 834 and a symbol mapper 835.
[0239] The plurality of channels as illustrated (for example one
for broadcasting MPEG-2 850, another for dedicated MPEG-2 851 , and
one for MAC messages 852) are then processed in the IFFT unit 838.
The resulted signal is transmitted over transmission channel
839.
[0240] For the MAC messages 852, the processing preferably includes
an RS coder (26,20) 836 and a small convolutional interleaver
837.
[0241] FIG. 12 illustrates a conceptual block diagram of the
Downstream Demodulation and Decoding subsystem. The signals input
over the reception channel 849 are processed in a FFT unit 848. The
separate resulting data channels are each processed in a symbol
demapper 845, bit deinterleaver 844, convolutional decoding unit
843, convolutional interleaver 842, RS decoder 841 and
randomization unit 840.
[0242] The subsystem is devised to output the data in several
channels as sent, for example one for broadcasting MPEG-2 853,
another for dedicated MPEG-2 854, and one for MAC messages 855.
Some of the channels may include a small convolutional interleaver
847.
[0243] The transport stream is, therefore, very robust and can be
changed as a function of the protection against fading, noise and
distance that should be reached.
[0244] Different modulation schemes QPSK, 16 QAM, 64 QAM and
different puncturing rates 1/2, 2/3, 3/4, 5/6, 7/8 enables an
optimization of the Downstream bit rate and protection. Moreover at
condition of LOS the guard interval needed to mitigate the
multipath affects is very small, therefore a use of a small guard
interval increases the channel capacity. The Guard intervals
supported should then be {fraction (1/256)}, {fraction (1/128)},
{fraction (1/64)} (see calculation section). For a SFN deployment a
larger Guard Interval of {fraction (1/32)}, {fraction (1/16)}, 1/8
can be introduced.
[0245] 5. Upstream Physical Layer
[0246] The upstream physical layer is also based upon OFDM
modulation, the number of subchannels allocated to a specific user
and the timing they will be transmitted in a specified time frame
are controlled by the MAC layer. Since the upstream is TDMA/OFDMA
based the channel can be modeled as a continuos sequence of "time
slots" and each time slot can be modeled as a group of subchannels
that are allocated to different Subscriber Units by Bandwidth On
Demand. By using this technique, QoS requirements and bandwidth
requirements can be managed. The recommended coding and modulation
of upstream packets are summarized in the block diagram shown in
FIG. 13. As shown in the diagram such a coding scheme is used in
order to support a large granularity for the bandwidth on demand
requirements.
[0247] FIG. 13 illustrates a conceptual block diagram of the
Upstream Encoding and Modulation subsystem. The figure illustrates
a reverse channel transmit, for example for MPEG-2 850. The signal
processing includes a de-randomization unit 860, variable RS coder
861, small convolutional interleaver 862, convolutional encoding
and puncturing unit 863, symbol mapper by allocation 865 and IFFT
unit 868.
[0248] The resulting signals are transmitted over the transmission
channel 869.
[0249] FIG. 14 illustrates a conceptual block diagram of the
Upstream Demodulation and Decoding subsystem.
[0250] The figure illustrates an embodiment of signal processing of
signals received over the reception channel 879.
[0251] The signal processing includes a FFT unit 878. From the
outputs of unit 878, a plurality of channels may be formed,
according to the initial carrier allocation at transmission.
[0252] In each channel, the signals are processed in a symbol
de-mapper by sub-channel allocation 875.
[0253] Further means for signal processing include a convolutional
decoding unit 873, small convolutional deinterleaver 872, variable
RS decoder 871 and a de-randomization unit 870.
[0254] The resulting signal is transferred to output the data in
MPEG-2 streaming 854 per user.
[0255] Every subchannel may consist of several carriers (see
calculations part), most are used for data transmission and the
rest are used for pilots transmission.
[0256] 6. Physical Layer Properties
[0257] The next section deals with different aspects of the
physical layer implementation.
[0258] 6.1 Synchronization Technique/Timing Control
[0259] In order to avoid highly accurate frequency source (e.g.,
OCXO) at the Subscriber Unit and satisfy timing requirements for
telephony or other CBR applications (Tl/El), it is highly efficient
to derive the Subscriber Unit's clocks from the Downstream
transmission. This can be achieved by using the Pilots carriers
transmitted by the Base Station, these Pilots can also be used in
order to Synchronize onto the Downstream transmission and achieve
clock extraction. Accurate upstream time slot synchronization shall
be supported through a ranging calibration procedure defined by the
MAC layer using the pilots transmitted by each Subscriber Unit.
[0260] Moreover, the Base Station copes with users transmission not
arriving fully synchronized, and relieving the demand for users
synchronization.
[0261] 6.2 Frequency Control
[0262] The clock extracted from the Downstream (as explained
before) is used as the reference clock of the Subscriber unit, in
particular to produce the RF frequency for the transmission and to
adopt this clock as the Subscriber Unit Base Band clock. Locking on
the Downstream transmission frequency shall allow an accurate
Upstream RF transmission frequency to be produced, that ensures
that all Subscriber Units transmitting shall reach the Base Station
Orthogonal, keeping the OFDM properties.
[0263] 6.3 Power Control
[0264] In order to perform a Upstream power control the Base
Station shall use a calibration and a periodic adjustment
procedures. The adjustment values shall be sent to a Specific
Subscriber Unit via the MAC layer. The Base station shall extract
the adjustment values by monitoring the power on the carriers that
were allocated to the specified user on the specified OFDM symbol.
Controlling the power of the Downstream dedicated channels will
perform another power control mechanism. The specified Subscriber
Unit MAC shall send adjustment values to the base station
correcting the power transmitted on the dedicated channel, and
adjusting it to the demands of a certain SNR. This procedure will
enable an optimized use of the base station Power Amplifier.
[0265] 6.4 Crest Factor
[0266] Much research has been done on the crest factor of OFDM
modulation.
[0267] The maximum crest factor is derived using 10*log(N), where N
is the number of carriers used in the OFDM symbol. Taking into
consideration that in our suggested system we use a 2048 carriers
FFT/IFFT which is very similar to the "2k" mode of the DVB-T we
shall introduce some measurements done on the DVB-T.
[0268] In the DVB-T, 1705 carriers are used for carriers
transmission, a crest factor of 32.3 dB would be expected but in
fact only 9-9.5 dB crest factor (with peaks of 10.5 dB) is actually
measured in any modulation using QPSK, 16 QAM and 64 QAM. These
results are achieved by the randomization of the data sent on the
carriers. In comparison to a single carrier modulation using 64 QAM
and a roll-off factor of 0.25-0.35 we get a crest factor of 8.8-7.8
dB, for 16 QAM we get a 1.4 dB reduction, resulting in 7.4-6.4 dB
see FIG. 15.
[0269] FIG. 15 illustrates the Crest Factor versus Roll-Off Factor
for Single Carrier.
[0270] In order to further reduce and stabilize the crest factor we
can clip the signal in order to achieve a desired crest factor. The
next graph plots BER/SNR for different crest factor limitations for
a DVB-T 16 QAM OFDM symbol, see FIG. 16.
[0271] FIG. 16 illustrates the BER/SNR for different Crest Factor
values, as achieved by clipping for a DVB-T 16 QAM OFDM Symbol
[0272] As we can notice, for a 1-1.5 dB clipping we get no
performance degradation, for a 2-2.5 dB clipping we get only about
0.5 dB degradation. For a 64 QAM modulation a degradation of 0.5 dB
could be achieved when clipping 1.1-1.6 dB, therefore achieving a
steady crest factor of 7.8 dB. By using more sophisticated methods,
more reduction can be achieved.
[0273] For the Upstream where a reduced number of carriers are used
(taking into consideration that all useful carriers are divided
into 16 subchannels), the crest factor achieved is about 7-7.5 dB
for QPSK, 16 QAM and 64 QAM all modulations (with peaks of 9.5
dB).
[0274] Taking the same method as before, for a 16 QAM modulation
clipping the power in such a way that the crest factor is 6.5 dB
will introduce only about 0.2-0.4 dB degradation, see FIG. 17.
[0275] By using more sophisticated methods, more reduction can be
achieved.
[0276] FIG. 17 illustrates BER/SNR for different Crest Factor
achieved by clipping for an Upstream 16 QAM OFDM Symbol
[0277] 6.5 Spectrum Properties
[0278] The spectrum properties of an OFDM modulation are derived
from the FFT/IFFT properties, although there is a natural decay in
the Out-Of-Band frequency domain a much tighter spectrum is
achieved by using additional measures. For an example the
Out-Of-Band spectrum mask for a 8 MHz DVB-T transmission is shown
in the FIG. 18. FIG. 18 illustrates Out-Of-Band Spectrum mask for a
8MHz DVB-T transmission.
[0279] Comparing OFDM to a Single Carrier where using a roll-off
factor of 0.25-0.35, it can be seen that OFDM modulation achieves
much more efficient spectrum properties with no degradation in the
performance, whereas in the Single Carrier there is a degradation
of 0.5-1.5 dB
[0280] 6.6 Power Amplifier Efficiency
[0281] From sections 6.3-6.5, we notice that for high modulation
scheme, the crest factor of an OFDM transmission can be achieved to
be even lower than for single carrier transmission. Furthermore,
considering the spectrum efficiency of the OFDM modulation, we can
derive that the power amplifier usage for an OFDM transmission is
very high, and a power control mechanism allows the better usage of
the Power Amplifier. In particular, these conclusions are enhanced
for an Uplink transmission, while for a Single Carrier transmission
the same power efficiency is achieved.
[0282] For an OFDM transmission, where the user is allocated a
subchannel, the total power transmitted is divided between less
carriers, to achieve an additional power gain of 12 dB (for a case
were the symbol is divided for 16 users).
[0283] 6.7 Timing Sensitivity
[0284] In an OFDM modulation, there is no timing sensitivity within
the sample time and simple phase and channel estimators correct
inaccuracies. Furthermore the Guard Interval of the transmissions
insures immunity in the face of multipath or unsynchronized
reception of OFDM transmission from several sources. In particular
this fact enables the creation of SFN on the DownLink, and of a
very relaxed timing synchronization demands of Subscriber Units in
the Uplink.
[0285] 6.8 Frequency Sensitivity
[0286] OFDM symbol demodulation is sensitive to frequency
inaccuracies. This sensitivity is solved by accurate AFC loops
using DDS. Using the above approach all Subscriber Units lock on
the Base Station frequency as explained in 6.2. In doing so they
ensure that their own transmission is kept orthogonal to other
Subscribers, and the total OFDM symbol shall remain orthogonal.
[0287] 6.9 Equalizations
[0288] While in Single Carrier equalizers are a must, and the
transmission of a training sequence (and the lost of data rate) is
needed, in an OFDM system time sensitivity is relaxed and a channel
estimator is the only thing needed in order to fix the timing
demands and channel imparities.
[0289] 6.10 Group Delay
[0290] The same channel estimators mentioned in 6.7-6.9 can
compensate group Delay caused by filters. The Group Delay
introduced is treated like a channel imparity. Single Carrier
systems are very much influenced by Group Delay as Shown in FIG.
19. In our System, it is expected to be in the 0.15-0.2 (see
calculation and assuming a group delay of 10 nsec).
[0291] In our System, it is expected to be in the 0.15-0.2 Tm/T
(see calculation). FIG. 19 illustrates the influence of linear
Group-Delay in Single Carrier system.
[0292] 6.11 Burst Efficiency
[0293] Upstream bursts of Subscriber User are very efficient
because of a low overhead. Subscriber Unit that has been allocated
to one subchannel has only 14% (16 of 112 carriers) of the carriers
dedicated to pilots (these are used for all receiver demands for
time, power and frequency control, and are also used for channel
estimation). If user is allocated more subchannels there is no need
for further increase of pilots number, so for 2 subchannel
efficiency shall rise and the overhead decreases to 7% (16 of 224
carriers), if all band is given to the user the overhead shall be
less than 1% .
[0294] 6.12 Sectorization, Cross Polarization and Diversity
[0295] Sectorization, Cross Polarization and Diversity can be used
in an OFDMA system as well, and may give many advantages.
[0296] 7. Comparison between OFDMA and Single Carrier TDMA
[0297] The following table is a rough comparison between OFDMA and
a Single Carrier System using TDD, numbers were derived from
experience, simulations and articles.
Table
[0298] 8. Calculations
[0299] The next calculations are for the Downlink/Uplink
transmissions.
[0300] Bandwidth=28 MHz
[0301] OFDM Carriers=2048
[0302] Carriers in use=1792
[0303] Sample Rate=28 MHz * (2048/1792)=32 MHz
[0304] Carriers Distance=Bandwidth/Carriers in use=15625 Hz
[0305] Guard interval 1/128=16 samples=500 nsec
[0306] Frame duration=(2048+16)/32 MHz=64.5 usec
[0307] 8.1 Downlink
[0308] Pilot Carriers per OFDM symbol=80 carriers
[0309] Data carriers in use=1792-80=1712
[0310] Symbol rate=1712 carriers/Frame Duration=26.543 Msps
[0311] Total throuput (QPSK) before ECC=53.085 Mbps
[0312] Total throuput (16 QAM) before ECC=106.17 Mbps
[0313] Total throuput (64 QAM) before ECC=159.26 Mbps
[0314] 8.2 Uplink
[0315] Number of Carriers used for Uplink contention=64
[0316] Number of Subchannels per OFDM frame=16 Subchannels
[0317] Number of carriers per on Sub channel allocation=108
carriers
[0318] Pilot Carriers per Subscriber Unit=16 carriers
[0319] Data carriers assuming n Subchannel for a specified
Subscriber Unit (n ranging from 1 to 16)=108*n-16
[0320] Data carriers assuming 1 Subchannel for a specified
Subscriber Unit=108-16=92 carriers
[0321] Data carriers assuming 16 Subchannel for a specified
Subscriber Unit=1792-64-16=1712 carriers
[0322] Symbol rate assuming best subchannel allocation (all
Subchannels per Subscriber unit)=(1792-64-16) carriers/Frame
Duration=26.543 Msps
[0323] Symbol rate assuming worst subchannel allocation (one per
Subscriber unit)=(1792-64-16*16) carriers/Frame Duration=22.822
Msps
[0324] Symbol rate per subchannel (Worst allocation)=1.4264
Msps
[0325] Total throuput (QPSK) before ECC, worst allocation=45.643
Mbps
[0326] Total throuput (16 QAM) before ECC, worst allocation=91.287
Mbps
[0327] Total throuput (64 QAM) before ECC, worst allocation=136.93
Mbps
[0328] TDMA frame length=16 OFDM symbols
[0329] TDMA frame duration=16 * 64.5 usec=1.032 msec
[0330] 9. Phase Noise Simulations
[0331] The following analysis deals with the influence of phase
noise on OFDM and Single Carrier Systems.
[0332] In order to check the phase noise influence a simulation was
written in MATLAB, using a model suggested in prior art.
[0333] The model simulates the phase noise by using a white
Gaussian process filtered with a single pole low pass filter, the
rational for using this model is a typical behavior of
phased-locked microwave oscillators.
[0334] The spectrum for the phase noise simulation has a Phase
Variance of -26 dB.
[0335] Using this Phase Noise model we tested an OFDM and a Single
Carrier (S.C) system for their BER/SNR performance with different
Phase Variance (P.V) values. The OFDM system used is more precisely
described in prior art. We will just indicate that the system uses
a 28 MHz bandwidth and has 2048 carriers, the system works with a
32 MHz clock. The S.C. system used has the same bandwidth and works
with a 28 MHz clock, no pulse shaping has been applied. Both
systems were tested for a 16 QAM modulation.
[0336] FIG. 20 illustrates the BER/SNR of the OFDM and S.C. systems
for different Phase Variance (P.V.) values.
[0337] 10. Conclusions
[0338] It will be noticed that the difference between the systems
is minor and is in the favor of the OFDM system. For a synthesizer
that has a better Phase Variance than -40 dB, no performance
degradation occurs. For a synthesizer with a Phase Variance of -26
dB a degradation of 0.5-2 dB occurs.
[0339] Such a synthesizer has a phase noise of about -80 dBc at 1
KHz and -90 dBc at 10 KHz.
[0340] These conclusions are different from some results presented
in prior art. However, the results from the simulation are
consistent to those achieved in prior art as summarized in CHAYAT,
May 1998.
[0341] Various modifications of the preferred embodiment are
possible without departing from the scope of the present invention,
and many of these would be obvious to people skilled in the
art.
[0342] Although the invention has been described in connection with
a preferred embodiment, it is to be understood that this
description is not intended to limit the invention thereto. Rather,
the invention is intended to cover all modifications and/or
additions to the abovementioned description, without departing from
the spirit and scope of the invention.
* * * * *