Note: Descriptions are shown in the official language in which they were submitted.
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
SYSTEMS AND METHODS FOR TRANSMITTING AND RECEIVING ADDITIONAL
DATA OVER LEGACY SATELLITE DIGITAL AUDIO RADIO SIGNALS
CROSS-REFERENCE TO OTHER APPLICATIONS:
This application claims the benefit of and hereby incorporates by reference
U.S.
Provisional Patent Application No. 61/124,626, entitled "SYSTEMS AND METHODS
FOR
VIDEO TRANSMISSION OVER DIGITAL SATELLITE RADIO LEGACY SIGNALS
('BACKSEAT TV TECHNOLOGY')", and filed on April 18, 2008.
This application also claims priority to and the benefit of, and hereby
incorporates by
reference, co-pending U.S. Nonprovisional Patent Application No. 12/184,659,
entitled
"OVERLAY MODULATION TECHNIQUE FOR COFDM SIGNALS BASED ON
AMPLITUDE OFFSETS", and filed on August 1, 2008.
This application also claims priority to and the benefit of, and hereby
incorporates by
reference, co-pending U.S. Nonprovisional Patent Application No. 12/416,027,
entitled "
OVERLAY MODULATION OF COFDM USING PHASE AND AMPLITUDE OFFSET
CARRIERS", and filed on March 31, 2009.
This application also claims priority to and the benefit of, and hereby
incorporates by
reference, co-pending U.S. Nonprovisional Patent Application No. 12/183,980,
entitled
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
"METHOD AND APPARATUS TO JOINTLY SYNCHRONIZE A LEGACY SDARS
SIGNAL WITH OVERLAY MODULATION", and filed on July 31, 2008.
This application also claims priority to and the benefit of, and hereby
incorporates by
reference, co-pending U.S. Nonprovisional Patent Application No. 12/079,782,
entitled
"HIERARCHICAL OFFSET COMPENSATION TO IMPROVE SYNCHRONIZATION AND
PERFORMANCE", and filed on March 28, 2008.
TECHNICAL FIELD:
This application relates to satellite broadcast communications, and more
particularly to
systems and methods for transmitting and receiving additional data over pre-
existing
("legacy") satellite digital audio radio signals.
BACKGROUND OF THE DISCLOSURE:
Satellite radio services, such as, for example, the Satellite Digital Audio
Radio Service
("SDARS") provided by Sirius Satellite Radio, Inc. ("Sirius"), successfully
broadcast audio
programs to millions of users. There is now a demand for these services to
additionally
provide video programming. However, given the existing bandwidth limitations
and the
large number of legacy receivers currently in the hands of subscribers, it is
therefore
desirable to allow new programming and new services (e.g., a new video
service) to be
2
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
provided over digital satellite radio legacy signals while not interfering
with the existing
legacy audio service.
SUMMARY OF THE DISCLOSURE:
Systems and methods for transmitting and receiving additional data, such as
video data,
over legacy satellite digital audio radio signals are provided.
Hierarchical modulation may be used in a satellite broadcast communications
system to
transmit additional data, such as video data, over pre-existing signals by
angularly
offsetting and/or changing the amplitude of data symbols. Systems and methods
are
provided for receiving the transmitted signals and for removing the offsets
resulting from
the hierarchical modulation schemes to improve synchronization and performance
in both
legacy and hierarchical decoders and receivers of the satellite broadcast
communications
system. The overlay modulation system may exist in parallel with the legacy
system so
as to provide a robust overlay data rate while avoiding legacy reception
degradation.
For example, according to some embodiments, there is provided a method of
transmitting
information over a satellite digital audio radio service ("SDARS") system. The
method
includes first modulating a legacy signal using a first modulation scheme to
encode the
legacy signal into a data stream of symbols, and second modulating the first-
modulated
symbols using a second layer of modulation to encode at least one additional
signal. The
method also includes receiving the twice modulated symbols at a receiver, and
first
3
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
demodulating the symbols to extract the legacy signal. The method may also
include
second demodulating the second layer of modulation to extract the at least one
additional
signal. The at least one additional signal may be a video signal. The
receiving may
include using antenna diversity. The first modulation scheme may be a Time
Division
Multiplexing ("TDM") scheme, such as a Quadrature Phase Shift Keying ("QPSK")
scheme, or
a Frequency Division Multiplexing ("FDM") scheme, such as a Coded Orthogonal
Frequency
Division Multiplexing ("COFDM") scheme. The second modulating may include
phase
modulating or amplitude modulating the first-modulated symbols. The receiver
may include
a legacy receiver.
According to some embodiments, there is provided a method of transmitting
information
over a SDARS system. The method includes error correction encoding an
additional
signal, interleaving the error correction encoded additional signal with an
overlay
identification marker ("OIM"), first modulating a legacy signal using a first
modulation
scheme to encode the legacy signal into a data stream of symbols, and second
modulating the first-modulated symbols using a second layer of modulation to
encode the
interleaved additional signal. The OIM may convey a characteristic of the
second layer of
modulation, such as an amount of an overlay offset modulation.
According to some embodiments, a method of transmitting information over a
SDARS is
provided that includes first modulating a legacy signal using a first
modulation scheme to
encode the legacy signal into a first data stream of symbols, second
modulating the first-
modulated data stream of symbols using a second modulation scheme to encode an
4
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
additional signal into a second data stream of symbols, and time delaying the
twice
modulated second data stream of symbols by a first period of time. The method
also
includes third modulating the legacy signal using a third modulation scheme to
encode
the legacy signal into a third data stream of symbols, fourth modulating the
third-
modulated data stream of symbols using a fourth layer of modulation to encode
the
additional signal into a fourth data stream of symbols, fifth modulating the
legacy signal
using a fifth modulation scheme to encode the legacy signal into a fifth data
stream of
symbols, sixth modulating the fifth-modulated data stream of symbols using a
sixth layer
of modulation to encode the additional signal into a sixth data stream of
symbols, and
time delaying the twice modulated sixth data stream of symbols by the first
period of time.
BRIEF DESCRIPTION OF THE DRAWINGS:
The above and other aspects of the invention, its nature, and various features
will
become more apparent upon consideration of the following detailed description,
taken in
conjunction with the accompanying drawings, in which like reference characters
refer to
like parts throughout, and in which:
FIG. 1 illustrates a Satellite Digital Audio Radio Service ("SDARS") system
architecture, in
accordance with some embodiments of the invention;
FIG. 1A illustrates a SDARS system architecture, similar to the system
architecture of
FIG. 1, but in greater detail, in accordance with some embodiments of the
invention;
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 2 illustrates an overlay encoding and modulation process, in accordance
with some
embodiments of the invention;
FIG. 3 illustrates an encoding circuit for a Bose Chaudhuri Hocquenghem
("BCH")
encoder, in accordance with some embodiments of the invention;
FIG. 4 illustrates a physical frame shuffler, in accordance with some
embodiments of the
invention;
FIG. 5 illustrates an overlay data scrambler, in accordance with some
embodiments of the
invention;
FIGS. 6A and 6B illustrate Overlay Identification Marker ("OIM") bit Maximum
Length Shift
Register ("MLSR") generation, in accordance with some embodiments of the
invention;
FIG. 6C illustrates an outline of an OIM format, in accordance with some
embodiments of
the invention;;
FIG. 6D illustrates the data fields of a portion of the OIM format of FIG. 6C,
in accordance
with some embodiments of the invention;
6
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 7 illustrates a channel interleaver structure, in accordance with some
embodiments
of the invention;
FIG. 8 illustrates the format of various data streams, in accordance with some
embodiments of the invention;
FIG. 8A illustrates an exemplary uplink data format, in accordance with some
embodiments of the invention;
FIG. 9 illustrates overlay framing, in accordance with some embodiments of the
invention;
FIG. 10 illustrates an exemplary angular offset over a Quadrature Phase Shift
Keying
("QPSK") hierarchical modulation scheme, in accordance with some embodiments
of the
invention;
FIG. 11 illustrates overlay offset compensation, in accordance with some
embodiments of
the invention;
FIG. 12 illustrates overlay framing, in accordance with some embodiments of
the
invention;
7
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 13 illustrates an exemplary constellation resulting from single carrier
QPSK
modulation, such as found in legacy bit streams, in accordance with some
embodiments
of the invention;;
FIG. 14 illustrates an exemplary constellation resulting from an exemplary
modulation
scheme of the symbols shown in FIG. 13, in accordance with some embodiments of
the
invention;
FIG. 15 illustrates an exemplary fan blade type distortion resulting from a
multipath
distortion manifested on a differentially modulated QPSK signal, in accordance
with some
embodiments of the invention;
FIG. 16 illustrates an exemplary set of data rings, and a unit circle shown as
a dotted line,
resulting from an exemplary modulation scheme, in accordance with some
embodiments
of the invention;
FIG. 17 illustrates Coded Orthogonal Frequency Division Multiplexing ("COFDM")
framing, in
accordance with some embodiments of the invention;
FIG. 18 illustrates Station and Local Content ID ("SLID") insertion, in
accordance with
some embodiments of the invention;
FIG. 19 illustrates SLID framing, in accordance with some embodiments of the
invention;
8
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 20 illustrates an overlay receiver, in accordance with some embodiments
of the
invention;
FIG. 21 illustrates a physical layer functional diagram for an overlay
receiver, in
accordance with some embodiments of the invention;
FIG. 22 depicts an exemplary COMM received QPSK constellation in the absence
of channel
equalization, in accordance with some embodiments of the invention;
FIG. 23 depicts an exemplary COMM overlay channel equalization architecture,
in
accordance with some embodiments of the invention; and
FIG. 24 depicts an exemplary implementation of COMM overlay channel
equalization, in
accordance with some embodiments of the invention.
DETAILED DESCRIPTION OF THE DISCLOSURE:
In certain broadcast communications systems, such as, for example, satellite
radio
systems, hierarchical modulation ("HM") can be used to overlay additional data
on top of
a legacy transmission so as to obtain additional bandwidth. Such a scheme can
be used,
for example, to offer additional channels or services. For example, in the
Sirius XM Radio
Inc. ("Sirius") Satellite Digital Audio Radio Service ("SDARS"), video
channels can be sent
9
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
over existing audio channels via such an overlay modulation scheme. In such
exemplary
embodiments, a video signal, or any other additional type of information, can,
for
example, be sent in an overlay modulation layer or "layer 2" bit stream, on
top of an
existing audio service, known as the "legacy" signal or "layer 1" bit stream.
Satellite broadcast communication systems, such as, for example, Sirius'
SDARS, can employ
multiple redundant transmitted signals, and can, for example, use various
forms of modulation
to transmit information over a carrier signal, such as, for example, Time
Division Multiplexing
("TDM") and Frequency Division Multiplexing ("FDM").
Quadrature Phase Shift Keying ("QPSK") is an exemplary TDM technique that can
allow for
the transmission of digital information across an analog channel. In QPSK,
data bits can be
grouped into pairs, with each pair represented by a particular waveform,
commonly referred to
as a symbol. There are four possible combinations of data bits in a pair, and
a unique symbol
is required for each possible combination of data bits in a pair. For example,
QPSK can create
four different symbols, one for each pair, by changing the I gain and Q gain
for cosine and sine
modulators. The symbol can then be sent across an analog channel after
modulating a signal
carrier. A receiver can demodulate the signal and look at the recovered symbol
to determine
which combination of data bits in a pair was sent.
Coded Orthogonal Frequency Division Multiplexing ("COFDM") is an exemplary FDM
technique. COFDM can distribute a single digital signal across several (e.g.,
one thousand or
more) signal carriers simultaneously, where coded data can be modulated and
inserted into
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
orthogonal carriers in the frequency domain. Because signals may be sent at
right angles to
each other, the signals effectively do not interfere with one another. In
general, the term "multi-
path effects" refers to the scattering of a signal due to obstructions such as
canyons, buildings,
and the like, that can cause a signal to take two or more paths to reach its
final destination.
COMM is highly resistant to multi-path effects (also known as "ghosting")
because it uses
multiple carriers to transmit the same signal.
Additional information on legacy SDARS transmission and reception of signals
using multiple
modulation schemes, such as TDM and COMM, may be found in Riazi et al. U.S.
Patent
No. 6,580,705, Riazi et al. U.S. Patent No. 6,618,367, and Riazi et al. U.S.
Patent
No. 6,798,791, each of which is hereby incorporated by reference herein in its
entirety.
Hierarchical modulation can utilize a further modulation of a transmitted
legacy bit or
symbol, for example, as to amplitude, phase, or a combination of the two, to
encode
additional information on top of the legacy information. For example,
additional data can
be transmitted over legacy differential COMM signals by changing the amplitude
of
legacy data symbols. Thus, the possible states a symbol can have may be
interpreted
differently in a system employing hierarchical modulation schemes than in a
system using only
conventional modulation techniques (e.g., Sirius' legacy TDM and COMM signals
without any
hierarchical modulation). Thus, using hierarchical modulation two separate
data streams can,
for example, be transmitted over a single transmission channel. In systems
employing
hierarchical modulation schemes, one data stream can be used, for example, as
a secondary
data stream while the other can be used, for example, as a primary data
stream.
11
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Systems and methods for using hierarchical modulation to transmit and receive
additional
data over legacy satellite digital radio signals, while maintaining backward
compatibility
for legacy systems using only conventional modulation techniques, are provided
and next
described with reference to FIGS. 1-24.
FIG. 1 shows an illustrative implementation of a satellite digital audio radio
service ("SDARS")
system 10. System 10 may include a broadcast studio 12 that may generate one
or more
channels of program information (e.g., audio data for one or more radio shows,
video data for
one or more video programs, etc.) and control information (e.g., information
as to who may
access certain audio and video programs, etc.). System 10 may then broadcast
this information
over one or more transmission paths. For example, system 10 may broadcast the
studio
information as three signals, over three different transmission paths 42, 44,
and 46, to one or
more receivers 50.
As shown in FIG. 1, for example, two of the three signals may be broadcast
over paths 42
and 44 via respective satellites 32 and 34 to receiver 50. These two signals
may first be
transmitted from studio 12 to a satellite uplink site 20. Satellite uplink
site 20 may then transmit
the two signals to respective satellites 32 and 34 over uplinks 22a and 24a,
for example, via
transmitters 22 and 24.
The third signal may be broadcast over path 46 via one or more terrestrial
repeaters 36 to
receiver 50. This third signal may first be transmitted from studio 12 to a
very small aperture
12
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
terminal ("VSAT") satellite uplink site 60. VSAT satellite uplink site 60 may
then transmit the
signal to a VSAT satellite 76 over a VSAT uplink 66a, for example, via
transmitter 66. VSAT
satellite 76 may then transmit the signal to each terrestrial repeater 36 over
a VSAT
downlink 76a. It is to be appreciated that, in other embodiments, satellite 76
may be any other
suitable type of satellite instead of a VSAT satellite and that uplink site 60
may be any other
suitable type of uplink site instead of a VSAT satellite uplink site, such as
a dedicated land line
or via the internet.
Satellites 32 and 34 may broadcast the signals received from uplinks 22a and
24a, respectively,
over transmission paths 42 and 44 to receiver 50, and each terrestrial
repeater 36 may
broadcast the signal received from downlink 76a over transmission path 46 to
receiver 50.
These three signals may be transmitted using two or more different modulation
schemes. For
example, the first and second signals may be transmitted via satellites 32 and
34 to receiver 50
over paths 42 and 44 using a TDM mode, and the third signal may be transmitted
via terrestrial
repeater 36 to receiver 50 over path 46 using a COFDM mode.
The combination of three transmission paths using both TDM and COFDM modes may
provide
for time, frequency, and space diversity within system 10. For example, the
transmission of the
signal over path 44 may be delayed with respect to the transmission of the
signal over path 42
to provide time diversity. Moreover, terrestrial repeater 36 and satellites 32
and 34 may be
physically spaced apart from one another to provide space diversity between
the signals
transmitted over paths 42, 44, and 46, while the difference between the TDM
and COFDM
modes may provide for frequency diversity.
13
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
The signals transmitted to receiver 50 over each one of transmission paths 42,
44, and 46 may
occupy various portions of the frequency band available to the SDARS. For
example, the signal
transmitted to receiver 50 over each one of transmission paths 42, 44, and 46
may occupy
approximately one-third (e.g., 4.167 MHz) of a 12.5 MHz band available to the
SDARS.
However, the signals transmitted over each uplink 22a and 24a, as well as the
signals
transmitted over VSAT uplink 66a and VSAT downlink 76a, may each occupy a
frequency band
other than the frequency band available to the SDARS (e.g., frequencies within
the Ku band).
FIG. 1A shows an exemplary embodiment of a system 110 in greater detail.
System 110 may
be similar to system 10 of FIG. 1, and is discussed below in conjunction with
FIGS. 2-22.
FIG. 2 shows an illustrative block diagram of an entire overlay processing
portion 200 of
system 10 (see, e.g., FIG. 1) for generating hierarchically modulated signals
to be
transmitted to receiver 50 of system 10. For example, a legacy data bit stream
201
(i.e., layer 1 or L1) may be hierarchically modulated in process 200 by an
overlay bit
stream 202 (i.e., layer 2 or L2) and transmitted through system 10. A
discussion of
process 200 is divided into three parts: (1) Forward Error Correction ("FEC")
encoding; (2)
transport to uplink modulators; and (3) physical layer modulation. Overlay FEC
encoding
is common to both satellite signal paths (e.g., transmission paths 42 and 44
of FIG. 1)
and terrestrial signal paths (e.g., transmission path 46 of FIG. 1), which
shall be
performed at the location of the FEC encoding portions (e.g., OFECs 106 of
FIG. 1A and
portion 210 of FIG. 2 for overlay portions and SPACE 103 of FIG. 1A and unit
203 of
14
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 2). The composite overlay bit stream 202 (i.e., layer 2 or L2) may be
provided at any
suitable rate, such as at a rate of 2.771085 Mbits/second, for example. After
the FEC
encoding, composite overlay bit stream 202 may be transported to the satellite
uplink
facility and terrestrial repeater network of system 10. Then, the overlay
framing and
modulation of overlay bit stream 202 with respect to the legacy data bit
stream 201 may
be handled, for example, by each modulator unit of the system.
Additional overlay transport capacity may be required if non-real-time, and
even additional
overlay data may be supported by the system (e.g., layer 3 data). The
additional
capacity, which may only be found on the satellite signals and hence not
required from
the VSAT terminal, for example, may be about 869.864Kbits/second.
FEC Encoding:
Overlay bit stream 202 may be FEC encoded by an FEC encoding portion 210 of
processing portion 200. As shown in FIG. 2, FEC encoding portion 210 may
include a
Bose Chaudhuri Hocquenghem ("BCH") coding portion 212, a Low Density Parity
Check
("LDPC") encoding portion 214, a shuffling portion 216, a data scrambling
portion 218, an
interleaving portion 220, and a synchronizing portion 222.
To improve system performance, an outer BCH code may be incorporated into the
process of FEC encoding portion 210 by BCH coding portion 212. This code may
be
used to lower the residual error floor of the LDPC decoder at the receive
side, and to
minimize errors due to severe code puncturing (e.g., via fading or multipath).
It is to be
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
understood that in other embodiments, any other suitable error-correcting code
may be
incorporated into coding portion 212 of FEC encoding portion 210. The outer
BCH code
of BCH coding portion 212 may have any suitable correction factor, such as,
for example,
a correction factor of 12 bit errors. The BCH block code shall accept a number
of source
information bits, such as 12256 source information bits from overlay bit
stream 202. The
output of BCH encoding portion 212 may be the original source bits followed by
a
particular number of parity bits, such as, for example, 168 parity bits, for
forming a BCH
block code of 12424 bits. Therefore, the coding rate of this outer code may be
12256/12424 (i.e., 0.986), for example. The encoded block size of 12424 bits
can be
achieved using the encoding procedure outlined below. Data may be applied to
and
output from BCH encoding portion 212 at any suitable rates. In some
embodiments, data
may be applied to the BCH input of BCH coding portion 212 at a rate of
917.3075295 Kbits/second, and the output of BCH coding portion 212 may be
929.8815884 Kbits/second, for example.
FIG. 3 shows an illustrative encoding circuit 300 that may be used for BCH
coding
portion 212 of FIG. 2. Encoding circuit 300 may be for an (n, k) BCH code with
the
following parameters: "m" may be any suitable galois field size, such as 14;
"n" = 2m - 1
and may be any suitable length of codeword in bits, such as 12424 bits; "t"
may be the
maximum number of error bits that can be corrected by encoding circuit 300,
such as 12
error bits; "k" >_ n - m * t and may be any suitable number of information
bits in a
codeword, such as 12256 bits; "dm;,," >_ 2 * t + 1 and may be any suitable
minimum
16
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
distance, such as 25; and "m * t" may be any suitable number of parity bits,
such as
168 parity bits.
The BCH encoding operation of BCH coding portion 212 may begin as follows for
each
code block. At the start of each code block generation, all registers in
encoding
circuit 300 of FIG. 3 may be reset to zero. Information bits 1 to k may be
applied to
encoding circuit 300 with both switches S1 and S2 in their respective position
2.
Modulo 2 arithmetic may be used at each adder stage A of encoding circuit 300.
After the
last input data bit of input i(x) is applied, both switches S1 and S2 may be
placed in their
respective position 1. Then, circuit 300 may be clocked an additional m * t
times to
generate the m * t parity bits, for completing the BCH code block length of
data plus parity
bits. For example, in accordance with the above given exemplary values,
circuit 300 may
be clocked an additional168 times to generate the 168 parity bits, for
completing the BCH
code block length of 12424 data plus parity bits.
The coefficients g; that may represent the tap weights of encoding circuit 300
of FIG. 3
may be those listed in the following table for a circuit with a galois field
size of 14, for
example.
g i (z) 1 +z+Z +Z +Z
g2(z) 1 +z +Z +Z +Z-1 4
g3(z) 1 +z+Z +z +Z +Z -14 -14
g4(z) 1 +Z +Z +z +z +Z +z -11 g5(z) 1 +Z +Z +z +Z +z +z -13 -14
17
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
g6(z) 1 +z-'+z-'+z-'+z +2 -14
g7(z) 1 +z-'+z-'+z-'+z- +2 +2 +2 -14 -14
g8(z) 1 +Z-,+Z-,+Z-9+z +2 +z
g9(z) 1 +Z+2 +2 +2 +2 -14
gio(z) 1 +2 +2 +Z- +Z- +2 -14 -14
11(z) 1+z +2 +2 +z
g 12(z) 1 +z+2 +2 +2 +2 +2 +2 -10 -13 -14
The generator polynomial g(z) of the t error correcting BCH code for circuit
300 may be
obtained by multiplying the first t polynomials in the above table. This may
result in the
following Standard Generator Polynomial:
g(z)= 1 +Z2+Z5+Z7+Z8+z10+z16+z19+z20+z24+z28+z30+z31 +z32+
Z33+Z34+ Z36+Z38+Z39+Z40+Z41 +Z42+Z45+Z46+Z47+Z48+Z49+Z50+
Z51 +Z55+Z57+z60+z62+Z64+Z67+z69+z70+Z76+Z79+Z80+Z81 +Z85+z
87+288+289+293 + 296+298+299+2102+2103+2105+2109+2110+Z113+2
116+2117+2119+Z 120+Z 123 +2125+2126+2131+Z 132 +2135+2137+2139+2141
+ 2142+ 2143+ 2144+ 2145+ 2147 + 2148+ 2150+ 2151 + 2153+ 2157+ 2158+ 2166+ z
168
In compact form, this encoding polynomial may be expressed as the following:
G =[10100101101000001001100010001011111
01011111001111111000101001010100101
100000100111000101110001001011001101
18
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
000110010011011001011000011001010101
111101101101000110000000101].
Output c(x) of BCH encoding circuit 300 of BCH coding portion 212 may be
provided as
the input to LDPC encoding portion 214 of FEC encoding portion 210 of FIG. 2.
It is to be
understood that in other embodiments, any other suitable error-correcting code
may be
incorporated into encoding portion 214 of FEC encoding portion 210. The inner
code
used by LDPC encoding portion 214 for the overlay FEC may include any suitable
code,
such as, for example, an Extended Irregular Repeat Accumulate ("eIRA") LDPC
code.
The inner code rate may be any suitable rate, such as 1/3 (or, more
particularly,
0.335838), with any suitable block size (i.e., data + parity), such as a block
size of
36994 bits. The LDPC encoder of LDPC encoding portion 214 may first output
parity bits,
such as 24570 parity bits, followed by encoded data bits, such as 12424
encoded data
bits. The LDPC encoder of LDPC encoding portion 214 may accept data at any
suitable
data rate, such as at an input rate of 929.8815884 Kbits/second. The final
output rate of
the LDPC encoder of LDPC encoding portion 214 may be at any suitable rate,
such as,
for example, at a rate of 2.768837692 Mbits/second.
Each LDPC code block of LDPC encoding portion 214 (e.g., of length 36994) may
be
distributed over any suitable number of frames, such as, for example, 49
physical frames.
The first 48 physical frames may contain a first number of bits, such as 755
bits, and the
last physical frame may contain a second number of bits, such as 754 bits. To
ease
hardware state machine design, the last physical frame of each LDPC code block
of
19
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
LDPC encoding portion 214 may be amended by a number of zero bits (e.g., 1
zero bit)
with zero pad insertion line 215 of FIG. 2, which may force the last physical
frame to
contain the same amount of bits as the preceding physical frames of the code
block. This
process may be repeated for each of the code blocks within a transmission
frame. The
final output rate of the LDPC encoder of LPDC encoding portion 214 with zero
padding
may be any suitable rate, such as, for example, 2.768912537 Mbits/second. In
some
embodiments, each LDPC code block may include 49 physical frames, and each
transmission frame may include 26 code blocks, for example.
Additional randomness to the channel interleaving process may be provided via
a suitable
shuffling portion 216. System performance may be improved by feeding
interleaving
portion 220 with a random selection of physical frames from each LDPC code
block
produced by encoding portion 214, as opposed to providing a straight feed of
physical
frames from each LDPC code block to interleaving portion 220, which may lead
to
consecutive data block errors under long fade intervals.
For example, shuffling portion 216 may be provided in the data path between
encoding
portion 214 and interleaving portion 220. In some embodiments, shuffling
portion 216
may provide an S-random physical frame shuffler. An S-Random shuffler may
shuffle
each of the physical frames per LDPC code block, for example, by using a
different
random pattern for each of the code blocks ("CB") within a transmission frame.
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Shuffling portion 216 of layer 2 FEC encoding 210 may start a shuffling
process at the
beginning of each transmission frame. Each CB may be permuted as dictated, for
example, by any suitable S-random table. An S-random table may consist of any
suitable
number of elements, grouped into any suitable number of columns of non-
repeating
random numbers within any suitable range. For example, in some embodiments, a
suitable S-random table may include 1274 elements, grouped into 26 columns
(e.g., the
number of LDPC CBs in a transmission frame), and each column may include non-
repeating random numbers ranging from 1 to 49 (e.g., the number of physical
frames per
LDPC CB). Each column may represent the physical frame permutation pattern
applied
to its respective CB within a transmission frame. For example, the first 49
elements of a
first column of the table may be used to shuffle the first CB with respect to
a transmission
frame boundary, and the second CB may be shuffled by the second column and so
on.
This process may be repeated for all of the code blocks within a transmission
frame.
Each transmission frame may shuffle every CB using the same S-random table.
FIG. 4 shows an illustrative diagram of an S-random physical frame shuffler
400 that may
provide shuffling portion 216 of FEC encoding process 210 with a suitable
shuffling
operation. Physical frames from each CB may be written via input line 401 into
input
buffer 402 of shuffler 400 in normal order from the LDPC encoder of encoding
portion 214. The entire buffer may be any suitable length. For example, in
accordance
with the above given exemplary values, the entire buffer may be 755 * 49 bits
(i.e., 36995 bits) in length, which may include the zero pad bits. Shuffler
400 may
permute each CB based on physical frame boundaries using permutation table
403, and
21
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
may retain the original bit order within each physical frame. FIG. 4 outlines
the shuffling
process for a first CB. The first physical frame that may be read out of
output buffer 404
of shuffler 400 on line 405 is shown to be 26, followed by 16, and ending with
physical
frame 37. No bit level reordering may be performed within each permuted
physical frame.
To ensure a random and even distribution of logic 0 and 1 overlay data bits,
original
overlay data stream 202 or the output of shuffler 216 may be applied to data
scrambling
portion 218 of FEC encoding process 210. Scrambling layer 2 data may avoid
long
strings of logic 1 or 0 bits that may interfere with legacy radio carrier
recovery loops.
Scrambling portion 218 may be provided with an overlay data scrambler 500 of
FIG. 5.
Data may be scrambled by data scrambler 500 via a Maximum Length Shift
Register
("MLSR") 502 that may include any suitable number of registers 504. For
example, as
shown in FIG. 5, MLSR 502 may include 23 shift registers 504 and a feedback
tap into an
adder 507. At the beginning of each transmission frame, each one of registers
504 may
be reset to an all 1's pattern. Any suitable operation 506, such as an XOR
operation, for
example, may be performed between the output of MLSR 502 and original data
stream 202. All addition operations are over galois field length 2 (i.e.,
GF(2)), and
MLSR 502 may be clocked only when overlay modulation is performed. Overlay
data
scrambler may provide a scrambled output 508.
An Overlay Identification Marker ("OIM") may be inserted in each transmission
frame. For
example, an OIM may be inserted as the last physical frame of each
transmission frame,
22
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
such as after the last physical frame of the last LDPC code block. As shown in
FIG. 2, an
OIM insertion line 219 may provide the insertion of OIMs into the data stream
through
FEC encoding portion 210 of processing portion 200. An OIM may indicate
whether or
not a legacy signal has been overlay modulated. Additionally, an OIM may
convey the
amount of overlay offset modulation being used as well as the number of
transmission
frame delays that may be required for proper Maximal Ratio Combining ("MRC")
of the
data at a receiver component. An OIM may include any suitable number of
sections
having any suitable number of bits. For example, in some embodiments and as
shown in
FIG. 6C, an OIM 620 may include 5 sections, such as a short 64 bit MLSR
section
repeated twice as first section 621 and last section 625 within OIM 620, a
long 523 MLSR
middle section 623, and a 52 bit data field section repeated twice within the
OIM as
second section 622 and fourth section 624 separating the middle section from
each of the
first and last sections, respectively. The total of all these OIM sections may
equal 755
bits (i.e., 2*64+523+2*52).
As mentioned, an OIM may be inserted as the last physical frame of each
transmission
frame. In some embodiments, the OIM pattern may not be FEC protected or
scrambled
and may be mapped as encoded data. The first and last sections of an OIM may
include
the first bits provided by an MLSR. For example, as shown in FIG. 6A, an MLSR
600
may include any suitable number of registers and may generate the first and
last sections
of an OIM. For example, in accordance with the above given exemplary values,
the first
and last 64 bits of a 755 bit OIM may be the first 64 bits generated by MLSR
600 of
FIG. 6A. MLSR 600 may include any suitable number of shift registers. For
example,
23
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
MLSR 600 may include 7 shift registers 602 and a feedback tap into an adder
607, and
each one of registers 602 may be reset to an all 1's pattern after each
transmission
frame. MLSR 600 may provide an output 604..
The middle section of an OIM may include the first bits provided by an MLSR.
For
example, as shown in FIG. 6B, an MLSR 610 may include any suitable number of
registers and may generate the middle sections of an OIM. For example, in
accordance
with the above given exemplary values, the middle 523 bits of a 755 bit OIM
may be the
first 523 bits generated by MLSR 610 of FIG. 6B. MLSR 610 may include any
suitable
number of registers. For example, MLSR 610 may include 10 shift registers 612
and a
feedback tap into an adder 617, and each one of registers 612 may be reset to
an all 1's
pattern after each transmission frame. MLSR 610 may provide an output 614.
After generation of a complete OIM bit pattern, each MLSR may be reset back to
its initial
state. The center 523 bit section provided by MLSR 610 may be inverted for the
next
transmission frame. This bit section may be the only section within the OIM
that inverts
from one transmission frame to the next. This may add an additional level of
unambiguity
between time markers. This process may be repeated indefinitely, the non-
inverted bit
pattern may be identified as "OIM+" and the inverted bit pattern may be
identified as
"OIM-". If this inverting 523 MLSR bit pattern is not detected by a receiver
equipped to
handle hierarchically modulated signals (i.e., an "overlay receiver)", the
receiver may not
include the respective signal in an MRC process. Detection of the OIM in a
COMM
receiver path may only be based on the bits in the middle section of the OIM
(e.g., the
24
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
middle 523 bits), because terrestrial repeaters may have the option to replace
the first
and last sections of the OIM (e.g., the two short 64 bit MLSR sequences) with
local
station information.
Each identical data field section of an OIM may include any suitable type or
types of bit
portions and sequences of bit portions. In some embodiments, each identical
data field
section may include two repeating sections. For example, FIG. 6D shows an
exemplary
format for a 52 bit data field 630. The first 4 bit portion 631 of data field
630 may be a
transmission frame delay portion, and may be followed by a 5 bit portion 632
indicating
overlay degree offset. This 9 bit sequence 633 may be repeated twice, and may
be
followed by a portion 634 having 8 reserved bits. This 26 bit pattern 635 may
be repeated
to complete the 52 bit data field 630.
Each 4 bit transmission frame delay portion 631 may be sent Most Significant
Bit ("MSB")
first, and may represent the number of transmission frame delays (i.e., in a
range of 0 to
15) that may be utilized for proper Maximal Ratio Combining ("MRC") of the
data at a
receiver component. Each 5 bit overlay offset word 632 may also be sent MSB
first, and
may represent the TDM angular offset being used in degrees (i.e., in the range
of 0 to
31). The COFDM offset will calculate its corresponding amplitude offset as
explained
below. To signal the receiver that TDM1 may need to be delayed for proper time
alignment, the TDM delay pattern may be "10111101 ", and if TDM2 needs to be
delayed
at the receiver, then the compliment pattern may be used (i.e., "01000010").
The default
delay pattern may be "10111101 ". When reserved bit portions 634 are not used,
they
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
may be set to a reserved pattern, such as, for example, the "10111101"
pattern. An
exemplary default bit pattern for the 52 bit data field 630 may be, when read
from bit
location 1 to 52, "1100011111100011111011110111000111111000111110111101", for
example. This pattern represents 12 transmission frame delays, 15 degree
offset, and
that the reserved bits are not used.
Interleaving portion 220 of FEC encoding portion 210 may provide channel
interleaving
using any suitable interleaving structure. For example, in some embodiments,
interleaving portion 220 may provide channel interleaving using convolutional
interleaver
structure 700 of FIG. 7. Channel interleaver structure 700 may contain any
suitable
number of branches 702, such as, for example, 49 branches 702 (i.e., the
number of
physical frames per LDPC code block). The fundamental interleaver delay depth
D of
each branch 702 may be any suitable number, such as, for example, a delay
depth of
D=4. Each D unit of interleaver structure 700 may contain at least 1 overlay
physical
frame worth of data, which may include any suitable number of bits, such as,
for example,
755 bits.
Input branch commutator 703 coupled to data input line 701 of structure 700
and output
branch commutator 705 coupled to interleaved data output line 707 of structure
700 may
move in synchronization with each other. The branch position of either
commutator may
equal the physical frame number within each CB. The commutator may remain at
each
branch 702 for an entire overlay physical frame duration. For example, the
first physical
frame from each transmission frame may pass through the first branch 702
26
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
(i.e., branch "1" of FIG. 7), the second physical frame from each transmission
frame may
pass through the second branch 702 (i.e., branch "2" of FIG. 7), and so on for
all 49
physical frames of the first code block. This process may then be repeated for
the first
physical frame of the second code block, and then again for each of the 26
code blocks
within a transmission frame.
After all of the code blocks (e.g., 26 code blocks having 1274 total physical
frames) have
been applied to channel interleaver 700, interleaver 700 may pass the entire
OIM at the
first branch position. After the OIM physical frame is passed, no branch
increment may
take place. Interleaver 700 may effectively stall at the first branch position
to pass the
first physical frame of the next transmissions frame CB, after which normal
progression of
the branch position may continue as described above. The entire process of
passing all
code blocks, stalling the branch increment at the first branch position for
the OIM, and
then progressing, may be repeated indefinitely for all subsequent transmission
frames.
In some embodiments, each position of channel interleaver 700 may contain the
following
amount of storage capacity: (j-1) * 755 * 4 bits, where j may be the channel
interleaver
branch position. The total amount of storage required by channel interleaver
700 may be,
for example, 3.551520 Mbits (i.e., 49 / 2 * (49 - 1) * 755 * 4), and the final
output rate of
channel interleaver 700 with OIM insertion may be 2.771085 Mbits/second, for
example.
Synchronizing portion 222 may synch L1 and L2, such as, for example, at the
transmission boundary points.
27
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Transport To Uplink Modulators:
The overlay FEC processing of FEC encoding portion 210 of FIG. 2 may be
handled by
the Overlay FEC ("OFEC") elements of the broadcast and transmission
infrastructure of
the SDARS system. The output of the OFEC may be a data stream at any suitable
rate.
For example, in some embodiments, the output of the OFEC may be a data stream
at
3.7584 Mbits/second, which may be half the data rate of the legacy data stream
(e.g., a
legacy data rate of 7.5168 Mbits/second) as generated by the SPACE elements
(e.g., SPACE elements 103 of FIG. 1A and element 203 of FIG. 2). Such a 1:2
ratio of
data rates between OFEC's overlay data stream and SPACE's legacy data stream
may
be critical in maintaining bit stream alignment across the two streams.
FIG. 8 shows exemplary formats 800 of the data streams generated by the SPACE
and
OFEC elements (i.e., legacy physical frame format 810 and overlay physical
frame
format 820), as well as a composite physical frame format 830.
The outputs of the OFEC and SPACE elements may be fed to a Composite
Multiplexer
("CMUX"), such as CMUXs 108 of FIG. 1A. The CMUX may receive two overlay
streams
from the OFEC elements, each at any suitable rate, such as 3.7584
Mbits/second, and
the CMUX may also receive two legacy streams from the SPACE elements, each at
any
suitable rate, such as 7.5168 Mbits/second. Next, the CMUX may align the two
legacy
streams at transmission frame boundaries and select one of them based on
either
availability or manual-override provisioning. The selected legacy bit stream
may be
28
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
denoted as L0. The CMUX may also align the two overlay streams at transmission
frame
boundaries and select one based on either availability or manual override
provisioning.
The selected overlay bit stream may be denoted as 00. Next, the CMUX may
combine
selected bit streams LO and OO to generate a composite stream containing
sequences of
two legacy bits followed by one overlay bit. This composite bit stream may be
denoted as
CO. The CMUX may reposition a marker (e.g., the 16-second marker) in composite
bit
stream CO such that it may be aligned with the 1 pulse per 16 second ("l ppl
6s") signal
fed to the CMUX. The 1 ppl6s, or any suitable equivalent, may provide
transmission
frame alignment at each terrestrial repeater with the satellite signals.
Finally, the CMUX
may maintain physical frame and transmission frame boundaries as well as the
relative
position of the 16-second marker in the composite bit stream CO output even
when legacy
or overlay feeds are switched.
The frame format of composite stream CO is shown by frame 830 of FIG. 8. It
may be
noted that the CMUX may transmit the entire 992 overlay bits (e.g., 755 L2 +
237
additional (e.g., Layer 3 bits)) that it receives from the OFEC elements. The
output of the
CMUX may be fed to one or more Overlay Modulator ("OMOD") elements for
generating
the QPSK signal for the satellite TDM segments and the VSAT Modulator ("VMOD")
for
feeding the terrestrial sites.
FIG. 8A shows an exemplary VMOD uplink payload packet structure 850. From the
composite physical frame, the VMOD may strip out all dummy overlay CAZAC and
service channel bits, which may result in a duplicate of the legacy bit stream
header (TDM
29
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Hdr) and service channel (SC)/cluster sync(CS) bits. The TDM header may have
48 bits
and the SC/CS portion may have 16 bits. The payload portion may have 2739 bits
and
has two sections. The first section (So ... Sn_1) may contain n triplets
(e.g., n = 755
triplets), and each triplet may include two legacy bits (L2k, 1-2k+1) and one
overlay bit (Ok).
There may be a total of 755 * 3 = 2265 bits in the first section. The second
section
(L2-n...L1983) may include the remaining (1984 - 2 * 755) = 474 legacy bits.
The total
VSAT payload physical frame length may be 48 + 16 + 2739 = 2803 bits. These
bits may
subsequently be wrapped into a DVB-S frame structure for FEC encoding and
transmission to remote VSAT sites.
Both VMOD and OMOD elements may strip out dummy Constant Amplitude Zero Auto-
Correlation ("CAZAC") header bits. The OMOD elements can be configured to
utilize
both L2 and additional overlay layers or just L2 bits for the purpose of
overlay modulation.
The VMOD elements may use the L2 bits only and may strip out the additional
overlay
bits prior to DVB-S FEC processing and modulation.
Physical Layer Modulation and Overlay Framing:
To ensure robust receiver synchronization, overlay transmission frames may be
aligned
with the legacy transmission frame boundary. Each transmission frame may
contain the
same amount of interleaved code blocks (e.g., BCH and LDPC code blocks). As
shown
by overlay framing format 900 in FIG. 9, there may be 26 code blocks 904 per
each
transmission frame 902, and there may be 49 physical frames 906 per each code
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
block 904. Each transmission frame 902 may use its last physical frame 906 as
an
Overlay Identification Marker ("OIM") 908.
The distribution of CB 904 data within a physical frame 906 may be distributed
contiguously over each TDM physical frame 906. In the TDM case, the CAZAC and
SERVICE channel data may not be overlay modulated. In the COFDM case, a non-
contiguous overlay assignment pattern may used.
Legacy transmission frames may contain 1275 physical frames, for example. In
some
embodiments, the number of legacy symbols within a TDM physical frame may
contain
1024 QPSK symbols, while a COFDM physical frame may contain 1000 symbols. For
example, in one transmission frame there may be 1,305,650 TDM QPSK symbols
(i.e., 1275 * 1024 symbols) and 1,275,000 COFDM data carriers (i.e., 1275 *
1000 data
carriers). Using these exemplary values, the required overlay bandwidth may be
reconciled with respect to the legacy bandwidth on a transmission frame basis
by the
information in the following table:
TDM ICOFDM
Un-modulated CAZAC symbols: 30600 (241275) 1 N/A
Un-modulated Service channel symbols: 10200 (81275=) 110200 (81275)
Overlay bits (BCH + LDPC + zero padding): 961870 (755*49*26)1 961870
(755*49*26)
Overlay OIM: 755 1 755
Un-modulated COFDM bins (Ampl. Refs.): N/A 1302175 (2371275)
Unused TDM capacity (to rate match COFDM): 302175 (2371275) 1 N/A
-------------------------- I ------------------------
1305600 11275000
Notice the sum of each column matches exactly the number of symbols within
each
signals transmission frame. The unused TDM capacity may be required due to the
31
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
COMM pilot utilization. This unused capacity may be used, for example, for
future layer
3 services that may only be available on the TDM satellite signals.
In some embodiments, the number of bits per transmission frame out of various
stages of
the FEC encoding process of FIG. 2, for example, may be as follows: overlay L2
data 202
may be 318656 bits; after BCH encoder 212 may be 323024 bits; after LDPC
encoder 214 may be 961844 bits; after zero padding at zero pad insertion line
215 may
be 961870 bits; and after OIM insertion at OIM insertion line 219 may be
962625 bits.
Furthermore, according to the exemplary values used throughout portions of
this
disclosure (e.g., where a transmission frame rate may be 2.878676471
tx/second), the
data rate out of various stages of the FEC encoding process of FIG. 2 may be
as follows:
overlay information 202 rate may be 910.3075295 Kbits/second; BCH encoder 212
output
rate may be 929.8815884 Kbits/second; LDPC encoder 214 output rate may be
2.768837692 Mbits/second; frame shuffler 216 output rate may be 2.768912537
Mbits/second; and final overlay output rate post OIM insertion at insertion
line 219 may be
2.771085938 Mbits/second, for example. The overlay coding rate may be
0.331028178
(i.e., (12256/12424)*(12424/36995)*(1274/1275)), and the raw unused TDM
capacity may
be about 869,864 legacy symbols/second (i.e., 302,175 * 2.878676471).
As mentioned with respect to FIG. 1, the three signals to be broadcast over
SDARS
system 10 may be transmitted using two or more different modulation schemes.
For example,
the first and second signals may be transmitted via satellites 32 and 34 to
receiver 50 over
paths 42 and 44 using a TDM mode, as shown by TDM1 signal path 292 and TDM2
signal
32
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
path 294 of FIG. 2, while the third signal may be transmitted via terrestrial
repeater 36 to
receiver 50 over path 46 using a COMM mode, as shown by COMM signal path 296
of
FIG. 2.
The combination of three transmission paths using both TDM and COMM modes may
provide
for time, frequency, and space diversity within the SDARS system. For example,
COMM
signal path 296 and TDM2 signal path 294 may be delayed with respect to TDM1
signal
path 292 within processing portion 200 to provide time diversity. As shown in
FIG. 2, for
example, the time diversity incorporated into the parallel bit streams for
TDM2 signal
path 294 by delay elements 254 and 264 may match the amount of time diversity
incorporated into the parallel bit streams for COMM signal path 296 by delay
elements 256 and 266. In some embodiments, this time diversity may range from
0 to 13
transmission frames. For example, a current transmission system may utilize 12
transmission frames of delay, which may yield approximately 4 seconds of time
diversity
(i.e., 4.16868 seconds). Cluster concepts may not exist in an overlay system,
so an
overlay receiver may delay the entire TDM1 signal path 292 at the receiver-end
of the
system by 0 to 13 transmission frames. The transmission frame delay used for
the
overlay data may be conveyed to the overlay receivers by decoding the OIM
message.
To minimize degradation to hierarchically modulated signals at legacy
receivers,
complimentary keying may be used between the two TDM satellite signals. For
example,
logic 0 and logic 1 overlay data bits may be inverted between TDM1 signal path
292 and
TDM2 signal paths 294. Bit invert 274 may invert every overlay bit it receives
33
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
(e.g., including un-coded, unscrambled OIM bits) before TDM2 modulator 284 may
map
TDM2 L1 data and TDM2 L2 data to its pseudo-8PSK constellation.
Overlay information may be modulated onto a L1 legacy TDM symbol stream by
applying
a programmable angular offset to legacy QPSK symbols, forming a new
constellation
similar to an 8PSK constellation, as is shown in FIG. 10.
For example, given a complex legacy symbol L = L; + j * Lq, a transmitted
overlay symbol
O can be expressed as 0 = (L; + j * Lq) * (cos a j * sin a), where the sign
in the second
component may represent the value of the overlay bit and can thus determine,
for
example, the direction of rotation of the ultimately transmitted I,Q symbol
relative to the
original, or legacy, QPSK symbol.
As noted, this technique is illustrated in FIG. 10. FIG. 10 illustrates an
exemplary
mapping of overlay data onto a legacy QPSK symbol to form a new 8PSK-type
constellation. In FIG. 10, unit circle 130 is depicted, with real axis 110 and
imaginary axis
120. With reference to FIG. 10, the original or first modulation layer QPSK
symbols are
shown at co-ordinates (1,1), (1,-1), and (-1,1) (i.e., at angles that are
multiples of
45 degrees along the unit circle), in each of quadrants I, II , III, and IV,
respectively.
Imposing a second layer of modulation on these legacy symbols may transform
each of
these QPSK symbols to one of two possible overlay 8PSK symbols 125, and also
symbols 117 and 118, which are shown as the two points at an angle +/- cp from
each
original QPSK symbol 115, making a total of eight possible overlay 8PSK
symbols. Thus,
34
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
for example, rotating a QPSK symbol by an additional angle cp can encode an
overlay 1,
and rotating the same QPSK symbol by an additional angle -cp can encode an
overlay 0,
as is shown in quadrants I and III. Alternatively, an overlay 0 and 1 can be
mapped to the
reverse phase shifts, where rotating a QPSK symbol by an additional angle cp
can encode
an overlay 0, and rotating the same QPSK symbol by an additional angle -cp can
encode
an overlay 1, as shown in quadrants II and IV.
In FIG. 10, a Gray coding scheme is utilized. Thus, in quadrants I and III, an
overlay 1
may add an angle cp to the original QPSK symbol, and an overlay 0 may subtract
the
angle cp from the original QPSK symbol, and, in quadrants II and IV, an
overlay 0 may add
the angle cp and an overlay 1 may subtract the angle cp. A positive rotation
may provide a
counterclockwise rotation. This may be done to improve accuracy, so that if an
overlay 1
bit from a neighboring quadrant spills over into the adjacent one, it may
still be read as a
1, so all pie slices with overlay 1 may be set adjacent to each other, and all
pie slices with
overlay 0 may be set adjacent to each other. In exemplary embodiments, such a
Gray
coding scheme can be used, and in alternate exemplary embodiments, it can, if
so
desired, not be used. In general, a Gray coding scheme may reduce the error in
the
overlay bit to one-half what it otherwise would be without the adjacencies.
Thus, for example, with respect to FIG. 1, the original legacy symbol 115 in
quadrant I
can be transformed to either of two 8PSK overlay symbols 117 and 118, where
117 may
be sent if the overlay bit is a 0, and 118 may be sent if the overlay bit is a
1.
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
In exemplary embodiments, for overlaying information onto QPSK symbols to
generate a
pseudo 8PSK coding scheme, the angle cp can be set to any value from 0 to 22.5
degrees. With cp greater than 22.5 degrees there could begin to be overlap
between
received 8PSK symbols, for example, due to the fact that most real channels
can have
random phase distortions, which may not be desired. In alternate exemplary
embodiments, if such overlap is not a concern, such as in channels with very
low noise or
a known predictable (i.e., non-random) noise signature that can be reliably
removed, cp
can have any reasonable range, such as 0 to 40 degrees, and more particularly,
such as
0 to 22.5 degrees, and still avoid overlap.
The angular offset angle used for the overlay data may be conveyed to overlay
receivers
(e.g., a receiver 50 of FIG. 1 that is equipped to handle hierarchical
modulation) by
decoding the OIM message.
As the new 8PSK type overlay modulated symbols may remain in their original
quadrant,
the information from the original legacy QPSK symbols can be preserved.
However,
under an overlay modulation scheme, while the legacy decoders in legacy
receivers
(e.g., a receiver 50 of FIG. 1 that is not configured to handle hierarchical
modulation
techniques) may expect a standard QPSK signal, what they may actually see is
the
random angular offset of the overlay modulation as an unnatural noise
enhancement.
Under low Signal-to-Noise Ratio ("SNR") conditions, the angular offsets can
get lost in the
noise, but stronger signals may see an unfair bias to the internal error
calculations of the
legacy decoder. This unfair bias can hurt performance and synchronization by
allowing
36
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
adaptive algorithms, such as, for example, equalizers, digital gain control,
and carrier
recovery, to process the invalid error signal. If the legacy decoder
synchronization is
relied on for retrieval of the overlay data as well, both services can suffer
in performance.
A second degradation can also be seen in the performance of the legacy service
by
allowing the overlay modulated signal to pass through to the Forward Error
Correction
stage. Therefore, in exemplary embodiments, the angular offset may be removed
from
the signal prior to inputting a received symbol to Synchronization and Forward
Error
Correction stages of a legacy decoder.
To remove the offset created by the overlay modulation, a decoder might have
to know
exactly what was transmitted, which may not be possible. Instead, in exemplary
embodiments, a decoder can make a rough guess by hard-slicing the overlay
modulated
signal to the appropriate pie slice within the received quadrant, thereby
indicating in which
direction the overlay offset may have been added. For example, with reference
to
FIG. 10 and the overlay modulation scheme shown therein, assuming the original
QPSK
symbol was original QPSK symbol 115 with an overlay angular change, and it is
necessary to determine whether the overlay symbol that was sent was 118, with
an
angular increase of cp, or 117, with an angular decrease of cp. By slicing
quadrant I by the
line I=Q (i.e., by the line running form the origin through original QPSK
symbol 115), a
good rough guess is that if the received overlay symbol is to the right of
that line it has a 0
overlay bit, and if the received overlay symbol is to the left of that line,
it has a 1 overlay
bit. With knowledge of the angle used in transmitting the overlay signal, any
received
37
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
overlay symbol can then be de-rotated by the same amount, thus removing the
effect of
the overlay modulation. Any error vector calculated from such a de-rotated
symbol may
thus more accurately represent the true error of the received symbol.
FIG. 11 illustrates such de-rotation of a received symbol according to an
exemplary
embodiment. With reference to FIG. 11, quadrant I of the I,Q plot of FIG. 10
is shown,
with I Symbol axis 1210 and Q Symbol axis 1250. An exemplary Original QPSK
Symbol
1220 is shown, and the transmitted version of this symbol after overlay
modulation being
Transmitted Overlay Symbol 1230 is shown (e.g., by adding an angle cp to its
phase).
The angle cp between Original QPSK Symbol 1220 and Transmitted Overlay Symbol
1230
is shown with one angle sign, closest to the origin of the depicted I,Q plot.
The actually
received version of this symbol, Received Overlay Symbol 1225, is also shown
and has a
larger amplitude than, and a phase distortion relative to, Transmitted Overlay
Symbol
1230, and thus it is no longer on the unit circle. These changes to amplitude
and phase
of the transmitted symbol may be introduced by noise in the channel. After
subtracting
the known angle cp from the phase of Received Overlay Symbol 1225, a de-
rotated
symbol results, such as De-Rotated Symbol 1245. The angle cp between Received
Overlay Symbol 1225 and De-Rotated Symbol 1245 is shown with two angle signs.
Thus,
in exemplary embodiments, the error vector seen by the decoder after de-
rotation,
Corrected Error Vector 1250, may be significantly smaller than that of
Uncorrected Error
Vector 1240, which may be the difference between Original QPSK Symbol 1220 and
Received Overlay Symbol 1225.
38
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
It is noted with reference to FIG. 11 that Corrected Error Vector 1250 may
match the error
with respect to the transmitted signal (i.e., the error between Received
Overlay Symbol
1225 and Transmitted Overlay Symbol 1230, which is not shown in FIG. 11, but
is easily
discernable). Of course this method may not be absolute, and symbols received
outside
the quadrant that they were actually transmitted in may be de-rotated in the
wrong
direction. While such improperly rotated symbols may result in a more
favorable than
expected error vector, this may have negligible effects compared to the much
larger
percentage of symbols that are received within their originally transmitted
quadrant and
that are properly de-rotated.
It is noted that the actual performance gain realized due to overlay offset
compensation
may be dependent upon the actual algorithms that take advantage of the
compensation.
Thus, some algorithms may see a great improvement, while others may see no
improvement at all. In exemplary embodiments, the simplest solution to
compensating for
overlay modulation is provided without needing to modify any proven algorithms
within
legacy demodulator designs. The methods of exemplary embodiments thus allow
for
essentially any offset angle used in an overlay modulation scheme to have
minimal effect
on signal acquisition and performance.
Physical and transmission frame boundaries of the overlay system may match
that of the
legacy system. At the start of each overlay physical frame, the overlay data
may be
modulated onto the legacy QPSK symbols as outlined in FIG. 12. The starting
location of
the 1st overlay bit shall correspond to the 33rd QPSK legacy symbol. The 755th
overlay
39
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
bit shall correspond to the 787th QPSK legacy symbol. A number of reserved
unmodulated symbols (e.g., 237) at portion 1305 may be provided after the
755th overlay
bit and may be used to rate match a COMM signal.
Overlay data may be modulated onto the legacy COMM symbol stream by offsetting
an
original COMM pi/4 Differential Quadrature Phase Shift Keying ("DQPSK') symbol
amplitude. In some embodiments, the amplitude offset a, with respect to the
original
DQPSK amplitude, may be controlled by the equation a = I2.0 * sin((p), where
cp may be
the TDM angular offset in degrees. For example, a may range from +/- 0 to 0.5,
in 0.025
amplitude increments, for example. The distance from the origin to the overlay
symbol
location may have the same radial angle as the original pi/4 DQPSK symbols
with no
angular modulation.
FIG. 13 shows an exemplary DQPSK constellation. The constellation shown in
FIG. 13 may
have a nominal radial distance of 1.414, for example. DQPSK may refer to the
procedure of
generating a transmitted QPSK symbol by calculating the phase difference
between the current
and the preceding QPSK symbols. In such a modulation scheme, all information
(i.e., two
binary bits per symbol) may be conveyed by the difference in phase across
frequency bins.
Generally, a starting bin, known as a pilot bin, may be used as a reference
and all additional
bins within a group may be differentially modulated based on the starting
phase of the pilot bin
and its adjacent data bin.
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
For example, one current SDARS modulation scheme utilizes two pilot bins, one
starting at a
band edge, wherein the first 500 data bins following the pilot bin are
differentially modulated. A
null bin follows these 500 data bins, which is used to avoid carrier leakage
into active data bins.
After the null bin, a second pilot bin is used as a reference phase for a
second group of 500
active data bins following said second pilot bin. The combination of 1000
active data bins, two
pilot bins, and one null bin are used to load a Fast Fourier Transform ("FFT")
symbol. This data
can then be placed into an inverse FFT engine, appended with a guard interval,
and radio
frequency ("RF") processed for transmission.
Such a differential phase encoding technique may be viewed mathematically as
follows:
Zo = (1 / J2) + j * (1 / \2);
Z1=Yo*ZO; and
Z2=Y1 *Z1,
where Y;(n) = [+/- 1, +/- j], where i is the FFT bin number, where n is an
index for the FFT points
in a bin, and where j is the square root of -1, the basis of imaginary
numbers. Moreover, it is
noted that in the above equations ZO is a pilot symbol, which carries no
information. It may only
provide a starting phase for the modulation process. If applied to a data set
that consists of +/-1
unity symbols, such modulation may result in an exemplary symbol constellation
as is depicted
in FIG. 13. After modulation and transmission, a receiver can then accept
these symbols and
can, for example, perform an differential decoding process (e.g., similar to
the differential
encoding process described above) to demap the data symbols into normal QPSK
constellations.
41
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
In exemplary embodiments, an addition to such a first layer modulation scheme
to encode
additional overlay information onto existing symbols generated by such a first
layer of
modulation can be done in a manner that may not harm reception by existing
receivers (e.g., an
existing legacy receiver that is only designed or configured to decode legacy
layer 1 data and
not hierarchically modulated data). That is, the addition of such overlay
information may be
done such that these legacy receivers may not experience much difficulty due
to the presence
of the overlay modulation on the symbols it receives. Thus, in exemplary
embodiments, the
additional information can be carried on the amplitude of each data bin.
An overlay modulation process can start as described above, but the amplitude
of each legacy
I/Q symbol can then be changed in accordance with an overlay modulation bit.
For example, if
the additional overlay information is a logical 1, then the amplitude of the
I/Q pair for a particular
bin can be increased from its nominal value of 1 to a valuel +D. Conversely, a
logic 0 can be
transmitted in the amplitude of each I/Q pair by decreasing the amplitude to a
value 1-C. It is
noted that the nominal value of 1 for legacy amplitude is exemplary only, and
in exemplary
embodiments, nominal first layer symbol amplitude can be increased so as to
provide more
room for amplitude offsets (i.e., the range 1-C to 1 +D). In exemplary
embodiments, it may be
convenient to set C equal to D, or approximately equal to D, so as to have two
rings of received
symbols that are equidistant, or approximately equidistant, from a ring of
legacy symbols, as is
shown in FIG. 16. This can, for example, simplify the detection of the overlay
information. In
alternate exemplary embodiments, C may not necessarily be equal to D. Such an
exemplary
modulation scheme may allow for each active data bin to be modulated and,
thus, in the
example described above, allows for an additional 1000 data bits per FFT
symbol.
42
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Thus, for example, in exemplary embodiments, the FFT symbol rate can be
approximately
4 kHz per second, thus allowing for a total of 4 million additional data bits
per second.
In exemplary embodiments, a transmitted overlay modulated constellation can
appear as is
illustrated in FIG. 14, where there is now a ring of DQPSK symbols at an
amplitude of 1-C, and
a second ring of DQPSK symbols at an amplitude of 1 +D, and where the
amplitude of each of
the original legacy symbols was nominally 1.414.
In exemplary embodiments, the average power transmitted using an overlay
modulation
scheme can, for example, be essentially the same as a legacy system. In this
approach a
receiver's Automatic Gain Control ("AGC") may not see any adverse effects.
Thus, to achieve this, the average power from the new constellation can be set
to equal unity.
In exemplary embodiments, independently controlling C and D can allow for this
as well as for
possible additional system optimization procedures in the future. Thus, at a
receiver, in
exemplary embodiments, a channel equalization technique can, for example, be
used that can
be based on a unit power transmitted constellation. To maintain unit power,
the following
equation may hold: [(1 - C)2 + (1 + D)2] /(2 * 2) = 1.
In exemplary embodiments, where C may not be desired to be approximately equal
to D, one
possible candidate offset pair to implement this condition can be, for
example, C = 0.2928 and
D = 0.8708. The resulting constellation can thus average to unit power.
Alternately, in some
43
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
exemplary embodiments, as noted, setting C equal to D, or substantially equal
to D, can be
preferred, and average power can be allowed to exceed unit power.
In exemplary embodiments, an amplitude offset can be applied either before or
after the
differential modulation process. To ensure minimal impact to current legacy
receivers, optimal
performance can be obtained if overlay amplitude changes are applied after the
differential
modulation process used to generate the legacy symbols. The effect on the
received signal to
the legacy receivers and next generation receivers (i.e., those equipped to
detect both a legacy
signal and an overlay signal) may be to appear as if the signal has undergone
a multipath
distortion. This can, for example, take away some ability of such legacy
receivers to withstand
multipath distortion. Any resulting degradation cannot precisely be predicted
but is expected to
be small because the constellation will see destructive interference half of
the time and
constructive interference the remainder of the time. The overall effect can
thus be expected to
average out within, for example, a trellis decoder.
At the receiver, legacy and future generation radios that recover the
fundamental signal may
process the COMM signal with no change. As discussed above, to the legacy
decoding
process the received signal may appear to have multipath distortion induced on
the
fundamental signal. Layer 2 data (i.e., overlay data) modulation may need to
extract this
additional information. Because, in exemplary embodiments, all layer 2
modulation may be
encoded in the amplitude of the signal, an additional processing step may thus
be required.
44
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Accordingly, in exemplary embodiments, a received signal may contain multipath
distortion that
manifests itself onto a differentially modulated QPSK signal as a fan blade
type of distortion as
is illustrated, for example, in FIG. 15, with no receiver noise. This may not
generally pose a
problem to the receivers inasmuch as all legacy information may be contained
in the phase of
the signal, and not in its amplitude. Current legacy receivers may process the
signal through a
trellis decoder as is, using the fan blade effect to essentially weight the
trellis trace back metrics.
This step can, for example, remain unchanged to recover the layer one data.
In exemplary embodiments, to recover the layer 2 data (i.e., the overlay
data), channel
amplitude equalization can be used, for example, to extract the additional
data. As discussed
above, in exemplary embodiments, the data set can be transmitted with an
averaged unity
power across each active frequency bin. Thus, at the receiver, one can take
advantage of this
fact and perform channel equalization across frequency bins to isolate the
amplitude modulated
overlay signal. For example, at each FFT symbol time, an average power across
neighboring
active data bins, which may be performed, for example, by a Finite Impulse
Response filter, can
be used to determine the localized power at the corresponding FFT bins.
Channel inversion
(e.g., zero forced or Minimum Mean Square Error) can then, for example, be
performed on the
data bins to restore, as best as possible, the original transmitted symbol
amplitude. If this step
is taken, the resulting constellation can, for example, be restored as is
illustrated in FIG. 16,
which consists of two rings prior to differential demodulation.
In exemplary embodiments, overlay data can be decoded, for example, by slicing
between the
data rings. The vector distance of each point after channel equalization can
be computed and
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
compared against a center decision ring, which itself can be determined by,
for example, slicing
equally between the two received rings, or for example, by calculating the
amplitude of
reference symbols from adjacent FFT bins that have not been overlay modulated.
One of the
many advantages of this overlay technique may be the avoidance of a 3 dB loss
in detection
probability due to the multiplicative effect in differential demodulation. Yet
another advantage of
this overlay technique is that the data can be modulated in such a manner that
the overlay
modulated signal may have minimal effect on maximum ratio combining ("MRC")
that may be
done in legacy systems, inasmuch as this does not increase the Signal to Noise
Ratio ("SNR").
As noted, in exemplary embodiments, channel equalization can be used to
recover the
overlay information contained in the amplitude modulated COMM signal. This may
be
implemented, for example, as follows, with reference to FIG. 23.
It is noted that due to multi-path conditions, expected in an environment
where COMM would
be used, the lack of channel equalization can result in a received QPSK
constellation as is
illustrated in FIG. 22. Notable is the elongated constellation due to the
amplitude variations
induced by the multi-path environment. Additionally, the petal-like structure
of the received
constellation will tend to become wider as the signal-to-noise ratio ("SNR")
is decreased.
To perform channel equalization, a reference point may be needed to indicate
the
unmodified amplitude, so that it can be known what an unmodified one or zero
looks like
at the receiver end of the channel. Thus, in exemplary embodiments, every Nth
symbol
can remain unmodulated with overlay information, thus leaving it at a unity
amplitude for
46
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
both ones and zeros in the legacy data. In exemplary embodiments, N can be 4,
or 5, for
example. Or, for example, a rotating pattern can be used, whereby different
symbols in
the sequence may remain unmodulated over time.
In an exemplary embodiment where N=5, an exemplary system would only overlay 4
symbols, and then leave the 5th unmodulated, then overlay the next 4 and then
leave the
10th symbol alone, etc. The unmodulated symbols may be known as "pilots." By
this
means, a reference as to what the channel amplitude distortion is for
neighboring cells
can be obtained. If all of those references are extracted, one may obtain a
sense of what
the channel is doing in general at a given window of time (the amplitude
distortion in
general will vary). To correct this channel amplitude distortion, all of the
symbols may be
multiplied by an inverse of the distortion (i.e., a correction factor) then
seen on the pilots
(i.e., a 1/x type process). This 1/x process can be applied to all bins, thus
undoing the
amplitude channel distortion. For convenience, x, the channel state
information, will be
referred to as h, and the correction factor as 1/6.
The pilots are extracted as part of a channel estimation procedure. Because
the receiver
is already synchronized to the data stream, it "knows" which symbols are
reference pilots
and which are not, and thus which pilots to extract to estimate the channel
distortion.
Thus, once the pilots are extracted and the data is multiplied by the 1/x
correction factor
bin for bin, the data stream becomes an equalized channel.
47
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
FIG. 23 illustrates this process in an overlay receiver 2300. As noted, a
legacy receiver
will simply see the COMM overlay amplitude modulations as noise, not affecting
the
differential phase. With reference to FIG. 23, received symbols 2301 may be
sent to a
channel equalizer ("CE") 2310, which multiplies the symbols by 1/x, where x is
the then
current amplitude distortion of the pilots (described in greater detail
below). After
equalization at 2315 the data symbols may be sent to slicer 2320, which can
make a soft
decision as to the overlay bit based on the amplitude of the symbol, as
described above.
Slicer 2320 can determine which of the rings (e.g., inner or outer) the
received data
symbol is on, as shown in FIG. 16. The received constellation may look
essentially like
two rings, but if the signal is sufficiently clean, there may also be a ring
in the middle,
representing the received pilots as well, whose nominal amplitude is unity.
In exemplary embodiments the 1/x (1/6) channel correction (equalization)
factor can be
obtained by determining the distortion response of the pilots, and then using
a temporal
averaging of this channel estimate by using a sliding window of three, and
then
multiplying (equalizing) each symbol by the inverse of the then prevailing
average
channel estimate (the average is moving through time).
The overlay amplitude modulation of the COMM symbols is essentially
functionally
equivalent to the overlay phase offset used on the TDM symbols. However, it is
noted
that in exemplary embodiments a greater amplitude offset can be used than the
corresponding TDM phase offset. For example, in exemplary embodiments, a COMM
amplitude offset can be the equivalent of 15 degrees of phase offset, whereas
the TDM
48
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
overlay phase offset can be 12 degrees. As noted above, TDM phase offset can
be as
high as 22.5 degrees, and thus COMM amplitude offset can be as high as, for
example,
the equivalent of 28 degrees. Thus, in exemplary embodiments, there can be an
unequal
overlay phase and amplitude offset for TDM and COMM, respectively.
Continuing with reference to FIG. 23, the output of slicer 2320 is multiplied
by the channel
state information squared, or h2. This is done because where the channel
amplitude
distortion is highest, there will be the least confidence in the received
symbol. Thus, prior
to sending that symbol to MRC Combiner 2350, it can be scaled by
multiplication by the
square of the channel state information, or h2, so as to scale (weight) its
contribution to
the overall signal by the square of the distortion. Finally, after going
through MRC
combiner 2350, the symbol can be sent to the overlay decoder.
FIG. 24 depicts an exemplary implementation of COMM channel equalization. In
particular, Fig. 24 is a detail of blocks 2310 through 2330 of Fig. 23 in
detail. The
processing depicted in Fig. 24 extracts the I and Q values of the pilot
symbols at 2410,
then obtains the radial squared distance, or 12 + Q2, at 2415. It then takes
the square
root of those numbers at 2420. This process is performed for every one of the
pilots
received. Next, at 2425 (block at upper right of Fig. 24), the values for
missing data
points in between are interpolated using a low pass filter.
Next, at 2430 temporal averaging is performed. Here averaging is shown between
a
current channel estimate weight "ccew" and a previous channel estimate weight
"pcew"
49
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
(delayed by one time unit), each having equal weighting of 0.5, but as noted,
in exemplary
embodiments one could also perform averaging using three values, being a ccew,
a ppew
and a "pppew" ("previous previous channel estimate weight") delayed by two
time units,
each of the three values having some appropriate weighting (e.g., each at 1/3,
or each
having a different weight, where weighting is more heavily biased to the
current time, for
example).
At 2435 the inversion of the temporally averaged channel estimate 1/x is
performed (1/x is
equivalent to the 1/6 of CE 2310 of Fig. 23) and at 2440 the inverse of the
channel
estimate is multiplied by the delayed symbols (an entire FFT symbol delay so
as to match
up with the delays of the channel estimation). This multiplication at 2440 is
equivalent to
the multiplication by 1/6 at 2315 of Fig. 23.
At 2445 the ring amplitude is obtained by squaring the equalized symbol and
taking its
root, and then subtracting the reference ring (i.e., the inner dotted ring of
Fig. 16) from
said square root, to see if the result is positive or negative. Finally, at
2455, the symbol
output from the slicer is multiplied by h2 in an analogous manner to 2325 of
Fig. 23, to
appropriately scale it for input to the MRC combiner. (There is a clipping of
the
magnitude to a certain level shown as well (at "Saturate"), which may also be
implemented if desired).
At 2447, for troubleshooting or internal design use the COMM channel equalized
output
can be seen if the signals are connected to an appropriate display or scope,
and what is
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
seen is essentially the constellation depicted in Fig. 16, either two rings or
two rings and a
center ring if the signal is clean. This internal output is not generally used
by a consumer
or end user.
As noted above, FIG. 22 depicts a typical received QPSK constellation in a
multi-path
environment. As further noted, the width of the petal-structure is an
indication of SNR;
the wider the petal, the lower the SNR. This suggests a convenient way of
obtaining an
SNR estimate in such environments. One can, for example, measure across the
width of
the petal shape (i.e., perpendicular to the long axis which may essentially be
on a radial
line out from the origin), and then map the width to SNR ratio. This can be
averaged over
an entire FFT period and may be fed to an MRC combiner circuit to
appropriately weigh
the COMM signal. This method may be implemented where at every physical frame
noise levels may be adjusted.
Thus, in exemplary embodiments there are two separate noise calculations. At
the
physical frame level a physical frame wide SNR is used to calculate the MRC
noise value
for the COMM channel, and at that given PF SNR each received symbol is
appropriately
weighted by the then prevailing channel amplitude distortion.
The amplitude offset used for the overlay data may be conveyed to an overlay
receiver
(e.g., a receiver equipped to detect both a legacy signal and an overlay
signal) by decoding
the OIM message. The OIM message may contain the angular offset of the TDM
signals.
This value may be converted to amplitude offset. Amplitude offset may be
calculated by
51
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
taking the sine of the angular offset and multiplying by the nominal radial
distance of the
constellation. A total of only 21 different angular offsets may be allowed for
the TDM
signals, suggesting that the conversion to amplitude offset can be done by a
simple table
lookup.
Overlay modulation onto the legacy COMM signal may not take place on all
possible
legacy carriers. A legacy COMM system may consist of 1000 data carriers, 2
phase
reference carriers, and 1 DC null bin, totaling 1003 "active bins". After the
legacy QPSK
symbols have been frequency interleaved and differentially encoded, the
overlay data
may be modulated onto the 8-PSK differentially encoded legacy data, at
modulator 286 of
FIG. 2, for example. The first contiguous 378 L2 overlay bits from each L2
VSAT CAZAC
physical frame may be overlay modulated onto carriers in the range of 1548 to
2048 while
using a 2048 IFFT engine, for example. The remaining contiguous 377 overlay
bits from
each VSAT physical frame may be overlay modulated onto carriers in the range 1
to 502
while using a 2048 IFFT engine, for example.
Not all carriers in the range specified may be overlay modulated. After
overlay amplitude
modulation is performed, the Inverse FFT process may be performed and Guard
interval
appending may be performed. FIG. 17 outlines COMM framing 1700 and the
position of
the overlay data with respect to the legacy COMM waveform. FIG. 17 shows the
post
interleaved and differentially encoded QPSK symbols prior to the inverse FFT
operation.
Overlay data may be modulated in groups of 4 carriers, for example. FIG. 17
depicts
52
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
overlay carriers as shaded. Note that legacy carriers may load FFT bins 1 to
502 and
1548 to 2048 (e.g., when using a 2048 IFFT engine).
Terrestrial repeater sites may have the option to transmit a Station and Local
Content ID
("SLID") in the overlay data stream and allow for transmission of local
overlay content.
Each terrestrial repeater (e.g., repeater 36 of FIG. 1) may replace the first
and last bits
(e.g., first and last 64 bits) of the OIM with bit pattern 1800 shown in FIG.
18. As shown
in FIG. 18, for example, portion 1801 including the first 16 bits may be
reserved for future
usage. If not used, the original 16 bits generated by the OIM may be passed
unchanged.
The next 32 bits of portion 1802 may be a random sequence, such as the
following, if
local content is being transmitted on the overlay bit stream: "Ox7f4ead26 =
01111111010011101010110100100110". If local content is not transmitted, the
original
32 bits generated by the OIM may be passed unchanged. The last 16 bits of
portion 1803
may be used to transmit a unique station ID, for example, which may be
transmitted MSB
first. The valid range of station ID values may be from 0 to 65535, excluding
the value of
a normal 64 bit OIM pattern, for example. If station IDs are not inserted, the
original 16
bits generated by the OIM may be passed unchanged.
The SLID bits may be framed, as shown by SLID framing portion 1900 of FIG. 19,
for
example, onto the COMM carriers in the same location as the original first and
last 64
MLSR bits that provide the OIM. SLID bit 1 may map to COMM active frames 395
and
1577. SLID bit 64 may map to COMM carriers 474 and 1655. All 64 SLID bits map
in
consecutive order onto the allowable COMM carriers.
53
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
A receiver (e.g., receiver 50 of FIG. 1) may be configured to handle and
utilize
hierarchically modulated information (i.e., an "overlay receiver"). FIG. 20
shows an
exemplary embodiment of an overlay receiver 2000. Overlay receiver 2000 may
include
at least one antenna 2002, a tuner 2004, a baseband 2008, an overlay processor
or
demodulator 2010, and an application processor or decoder 2012.
The reception of an overlay element of hierarchical modulation by overlay
receiver 2000
may require the use of two or more antennas 2002 (e.g., antennas 2002A and
2002B of
FIG. 20). Each antenna 2002 of overlay receiver 2000 may have a performance
equal to
that of a single antenna in a legacy receiver, for example. The actual antenna
configuration and positioning of antennas 2002 may be dependent on the
"transport" or
environment of the receiver (e.g., a vehicle having the receiver installed
therein).
Receiver 2000 may include an active splitter element 2003 so that the signal
from one of
antennas 2002 (e.g., antenna 2002A) may be available to be sent to a "legacy
only"
receiver that may also be in the transport of the overlay receiver. This other
receiver may
be an integrated head unit. The antenna coupled to splitter 2003 should be in
an optimal
reception position on the transport, for example.
Tuner 2004 may be, for example, any satellite radio Tuner ASIC. A Discrete
Tuner may
alternatively be used due to its enhanced performance, which may mitigate the
signal
54
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
degradation due to overlay modulation. Tuner 2004 may be configured to support
two
antenna inputs.
Overlay processor 2010 may include a diversity processor and may receive the
two IF
inputs from tuner 2004. Processor 2010 may perform combining operations on the
two
signals. A full description of this combining operation is discussed in Scarpa
et al. U.S.
Patent Publication No. 20070142009, filed December 1, 2006, which is hereby
incorporated by reference herein in its entirety.
Baseband 2008 may be, for example, any satellite radio baseband processor, and
may
be configured to provide the symbol data to be used by the overlay processor
(e.g., overlay processor 2010). Baseband 2008 may be a legacy baseband, for
example.
Overlay processor 2010 may be responsible for many various functions of
overlay
receiver 2000. For example, overlay processor 2010 may de-multiplex the
incoming data,
combines the signals from the multiple sets of data provided by baseband
processor 2008, decodes the overlay data, perform error correction on the
overlay data,
close the processing loop to control the a diversity combiner, performs
content decryption,
perform subscription management tasks (e.g., key management and subscription
management), present data to external application processors, channels 12S
data from
baseband processor 2008 to application processor 2012 (e.g., 12S data can be
uncompressed audio or "data"), receive SSP+ Commands (e.g., UART, parse data
for
overlay and legacy tune commands, and navigation), controls baseband processor
2008
using SSP over a second UART (e.g., if transport sees single client or for
security key
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
communication), provide commands to application processor 2012, provide a file
system,
provide an interface to external devices including storage (e.g., USB 2.0),
provide
memory Interfaces (e.g., SDR SDRAM to support the 4 second buffer or any de-
interleaving, DDR SDRAM, FLASH for the storage of Firmware, Flash Interface
designed
to support a wide range of FLASH devices and support the industry standard
CFI, etc.),
and may provide and interface for an additional modulation layer (e.g., Layer
3).
FIG. 21 shows a physical layer functional diagram for an overlay receiver
2100.
Receiver 2100 may include complementary components and may utilize
complementary
methods of demodulation complementary to those described above with respect to
overlay
transmission. For example, as shown in FIG. 21, a symbol demultiplexer ("De-
Mux") 2102 may
receive the symbols for receiver 2100. De-Mux 2102 may demultiplex the
received symbols and
provide them to a first TDM ("TDM 1 ") Sync 2104, a second TDM ("TDM2") Sync
2106, and a
COFDM channel estimator 2108, respectively, the output of which may be
provided to a
respective Slicer 2110, 2112, 2114. Each slicer may provide a first output to
a respective OIM
detector 2116, 2118, 2120. Slicer 2110 may provide a second output to a delay
module 2122,
which may allow TDM1 signal to match the delay provided to each of TDM1 and
COFDM
signals, as shown in FIG. 2. . Slicer 2112 may provide a second output to a
negate
module 2124. Each slicer may provide a third output to an MRC combiner 2126.
Moreover,
each OIM detector, delay module 2122, and negate module 2124 may also each
provide an
output to MRC combiner 2126. Slicer 2114 may provide a third output to MRC
combiner 2126
as well. MRC combiner 2126 may provide an output to a chain of modules,
including a channel
de-interleaver 2128, a de-scrambler 2130, a de-shuffler 2132, a buffer 2134,
an LDPC
56
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
decoder 2138, and a BCH decoder 2140. LDPC decoder 2138 may be coupled to an
adaptive
noise estimator 2136.
Application processor 2012 of overlay receiver 2000 may be, for example, an
Analog
Devices BF566 Blackfin DSP running at 300 MHz. The Blackfin may run ucLinux,
for
example. A video CODEC that supports both a "Microsoft" mode and a
"Diagnostic"
mode may be used, for example. The Microsoft mode may simultaneously support
Windows Media 9 Video Professional Profile at up to 300 kbits/second 320 x 240
resolution and 30 Hz Frame rate, as well as Windows Media Audio at 32
kbits/second, for
example. The diagnostic mode may support a video receiver that may output a
standard
colorbar test pattern when put in test mode. This may be an OSD output (e.g.,
640 x 480
24 bit color 8 bits of Alpha blend) to the monitor. In some embodiments, the
receiver may
support both a H.264 (MPEG 4 Part 10) Main Profile at up to 300 kbits/second,
320 x 240
resolution and up to 30 Hz frame rate, as well as AAC Audio at 32
kbits/second, for
example.
An overlay receiver may include memory, such as, for example, 32 MB DDRAM (16M
x
16 DDR266), 16 MB FLASH, and the like. There may be a test header to enable
access
to the 656 video. Video decoder 2012 may be housed in a metal enclosure for
EMI and
environmental shielding with the connectors exiting the box for easy access.
The
package size for overlay receiver 2000 may be equal to or smaller than 145 mm
(side to
side) X 160 mm (front to back) X 38 mm (top to bottom).
57
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
Antenna 2002A input may be the primary antenna input, and may be Fakra SMB
(e.g., first part of dual configuration) and keyed to prevent cross connection
with a 50
Ohm input. Antenna 2002B may be a diversity antenna input, and may be Fakra
SMB
(e.g., second part of dual configuration) and keyed to prevent cross
connection with a 50
Ohm input. An optional antenna output to a head unit may be Fakra SMB (e.g.,
single
configuration) and may be keyed to prevent cross connection.
A default protocol for an overlay receiver may be SSP+. For OEM applications,
an
overlay receiver may use protocols specific for that OEM, and the translation
of the OEM
protocols to the internal SSP+ may be performed on the controller in the
receiver.
In exemplary embodiments, the disclosed systems and methods can be implemented
in
hardware, software, firmware, or any combination of the above, and can be
implemented in a
transmitter or transmission device. Similarly, complementary systems and
methods of
demodulation can be provided in a similar manner and implemented in a
demodulator or a
receiver. For example, a program storage device, such as a microprocessor with
memory, or for
example, a separate microprocessor memory, can store a program of instructions
sufficient to
implement exemplary methods of the invention.
While there have been described systems and methods for transmitting and
receiving
additional data, such as video, over pre-existing ("legacy") digital satellite
radio signals,
with reference to certain exemplary embodiments, it is to be understood by
those skilled
in the art that various changes may be made and equivalents may be substituted
without
58
CA 02721849 2010-10-18
WO 2009/146217 PCT/US2009/041174
departing from the scope of the invention. In addition, many modifications may
be made
to adapt a particular situation or material to the teachings of the invention
without
departing from its scope. Therefore, it is intended that the invention not be
limited to the
particular embodiments disclosed, but that the invention will include all
embodiments
falling within the scope of the appended claims.
59
