IMPROVING PERFORMANCE IN A TIME-FREQUENCY INTERLEAVED ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING SYSTEM

[0001] This disclosure generally relates to wireless communications and packet

based Orthogonal Frequency Division Multiplexing (OFDM) systems, and more

specifically to mitigating interference and improving performance in a time-frequency

interleaved orthogonal frequency division multiplexing (TFI-OFDM) system.

[0002] In a typical TFI-OFDM system, the frequency spectrum is divided into a

number of sub-bands each having a predetermined width. For example, a WiMedia

TFI-OFDM system uses three sub-bands each having a bandwidth of 528 megahertz

(MHz) for a total of about 1.5 gigahertz (GHz). The WiMedia system can provide

data rates from about 53.3 mega bits per second (Mb/s) to about 480 Mb/s. Spatial

capacity for the TFI-OFDM system is provided through time-frequency codes (TFC)

that each piconet uses to impart a unique frequency band-hopping sequence.

Although the TFCs enable multiple piconets to communicate at the same time, there

are instances where interference can arise when the frequency band-hopping

sequence causes the piconets to operate in the same frequency band. In-band

Interference can also arise, especially in WiMedia applications, where multiple

piconets operate in close range with respect to each; and also in the form of

narrowband interference when other devices are operating simultaneous in this band.

These types of interference can severely degrade the performance in different

channel conditions.

[0003] Therefore, there is a need for an approach that can mitigate the effects of

interference that is introduced in a TFI-OFDM system.

[0004] In one embodiment, there is a TFI-OFDM receiver system. In this

embodiment, the system comprises a receiver configured to receive data packets

transmitted in certain frequency bands corresponding to a TFC number. Each

received data packet comprises OFDM symbols segmented into a preamble section,

header section and payload section. A Fast Fourier Transform processor is

configured to transform the OFDM symbols from a time domain into a frequency

domain. A channel estimator is configured to estimate a channel impulse response

for the data packets. The channel estimator estimates the channel impulse

response from frequency domain OFDM symbols in the preamble section. A noise

variance estimator is configured to derive a noise variance estimate from the

frequency domain OFDM symbols in the preamble section.

[0005] In another embodiment, there is a TFI-OFDM receiver system. In this

embodiment, the system comprises a receiver configured to receive data packets

transmitted in a certain frequency bands corresponding to a TFC number. Each

received data packet comprises OFDM symbols segmented into a preamble section,

header section and payload section. A Fast Fourier Transform processor is

configured to transform the OFDM symbols from a time domain into a frequency

domain. A channel estimator is configured to estimate a channel impulse response

for the data packets. The channel estimator estimates the channel impulse

response from frequency domain OFDM symbols in the preamble section. A noise

variance estimator is configured to derive a noise variance estimate from the

frequency domain OFDM symbols in the preamble section. An updater configured to

update the channel impulse response estimated by the channel estimator and the

noise variance estimate derived by the noise variance estimator.

[0006] In a third embodiment, there is a method for compensating for interference

that arises from ultra-wideband devices operating in a communications link. In this

embodiment, the method comprises receiving data packets in certain frequency

bands corresponding to a TFC number. Each received data packet comprises

OFDM symbols segmented into a preamble section, header section and payload

section. The method further comprises transforming the OFDM symbols from a time

domain into a frequency domain. The method further comprises estimating a

channel impulse response from frequency domain OFDM symbols in the preamble

section. The method further comprises deriving a noise variance estimate from the

frequency domain OFDM symbols in the preamble section.

[0007] In yet another embodiment, there is a method for improving performance of

ultra-wideband devices operating in a communications link. In this embodiment, the

method comprises receiving data packets in certain frequency bands corresponding

to a TFC number. Each received data packet comprises OFDM symbols segmented

into a preamble section, header section and payload section. The method further

comprises transforming the OFDM symbols from a time domain into a frequency

domain. The method further comprises estimating a channel impulse response from

frequency domain OFDM symbols in the preamble section. The method further

comprises deriving a noise variance estimate from the frequency domain OFDM

symbols in the preamble section. The method further comprises updating the

estimated channel impulse response and the derived noise variance estimate.

[0008] FIG. 1 shows an example of a physical layer convergence protocol (PLCP)

frame format for a packet based OFDM system;

[0009] FIG. 2 shows a table listing an example of hopping patterns for different TFCs

that can be used with a TFI-OFDM system;

[0010] Fig 3 shows a block diagram of a TFI-OFDM receiver according to a first

embodiment of this disclosure;

[0011] Fig 4 shows a flow chart describing the operation of the TFI-OFDM receiver

depicted in FIG. 3;

[0012] Fig 5 shows a block diagram of a TFI-OFDM receiver according to a second

embodiment of this disclosure; and

[0013] Fig 6 shows a flow chart describing the operation of the TFI-OFDM receiver

depicted in FIG. 5.

[0014] Packet-based transmissions systems such as TFI-OFDM systems transmit

packets of data in short bursts. Each packet of data sent in a transmission includes

fields that provide information that a receiver utilizes to facilitate the transmission and

reception of the packet. FIG. 1 shows an example of a PLCP frame format 10 for a

packet of data sent in a TFI-OFDM system. The PLCP frame 10 includes a

preamble section 12, a header section 14 and a payload section 16. The preamble

section 12 includes a time domain (TD) training sequence 18 and frequency domain

(FD) training sequence 20. The TD preamble 18 has a duration that can be either

24 or 12 OFDM symbols. The duration for the TD preamble 18 will depend on the

mode of transmission (i.e., standard or streaming). A receiver in a TFI-OFDM

system uses the TD preamble 18 for packet and frame synchronization. The FD

preamble 20, which follows the TD preamble 18, has a duration of six OFDM symbols.

A receiver in a TFI-OFDM system uses the FD preamble 20 for channel estimation

(CE). The six OFDM symbols are referred to as CE symbols (i.e., CE1 , CE2, CE3,

CE4, CE5, and CE6). The header section 14, which follows the FD preamble 20,

includes 12 symbols referred to as header symbols (i.e., hdr 1 , hdr 2, hdr 3 ... hdr 12).

Generally, the header symbols are transmitted at the base rate of 53.3 Mb/s and

provide information that is specific to decoding the payload section 16. For example,

the header symbols can provide information such as the length of the payload, the

amount of bytes transmitted in the payload, the modulation scheme of the

transmission and coding rate that was used. The payload section 16, which follows

the header section 14, can have a duration that includes a variable number of

payload symbols that are transmitted at a specified rate.

[0015] FIG. 2 shows a table listing an example of hopping patterns for different TFCs

that can be used with a TFI-OFDM system. For example, in FIG. 2, TFC number

one has a hopping pattern that includes 1 , 2, 3, 1 , 2, 3. This hopping pattern in TFC

one indicates that a first symbol is transmitted in band 1 , a second symbol is

transmitted in band 2, a third symbol is transmitted in band 3 and that this pattern will

repeat itself for every three symbols. TFC number two, which has a hopping pattern

of 1 , 3, 2, 1 , 3, 2 means that a first symbol is transmitted in band 1 , a second symbol

is transmitted in band 3, a third symbol is transmitted in band 2 and that this pattern

will repeat itself for every three symbols. These hopping patterns allow one piconet

to operate in one communication link, while a second piconet can establish another

link by using a different TFC number. Note that hopping patterns are disabled in

TFC numbers 5-7 (i.e., the symbols are all transmitted in the same band for a TFC

number).

[0016] In a TFI-OFDM system, the TFC number used in the hopping pattern and the

symbol number in the PLCP frame 10 determines the carrier frequency of a particular

symbol in the frame. For example, considering the case of TFC number 1 (i.e., a

hopping pattern of 1 , 2, 3, 1 , 2, 3) and the FD preamble field, the symbols CE1 and

CE4 in the FD preamble are transmitted in band 1 , symbols CE2 and CE5 are

transmitted in band 2 and symbols CE3 and CE6 are transmitted in band 3. For the

case of TFC number 3 (i.e., a hopping pattern of 1 , 1 , 2, 2, 3, 3), the symbols CE1

and CE2 are transmitted in band 1 , symbols CE3 and CE4 are transmitted in band 2

and symbols CE5 and CE6 are transmitted in band 3. For TFC numbers 1 to 4, the

channel estimation sequence which is determined from the FD preamble symbols (i.e.,

CE1 , CE2, CE3, CE4, CE5, and CE6) is transmitted twice in each of the three bands.

For fixed frequency interleaved (FFI) modes (i.e. TFC numbers 5 to 7), where band

hopping is disabled, all the six channel estimation symbols are transmitted in the

same band. The reliability of the channel estimation for a TFI-OFDM system

improves with an increase in the number of the symbols used/transmitted for this

purpose. A TFI-OFDM system that can more reliably estimate the channel will have

a better understanding of the effects of the channel on the received symbols and

therefore be able to more accurately process the symbols in the header section 14

and the payload section 16.

[0017] FIG. 3 shows the block diagram of a TFI-OFDM receiver 22 that is able to

estimate noise variance using the FD preamble symbols. In FIG. 3, a mixer 24

receives a data packet embodied in a radio frequency (RF) signal having a certain

frequency and down-converts the signal using another signal at a different frequency

generated from an oscillator 26. The mixer 24 is one component of the RF front-end

that performs all RF processing, however, for ease of illustration, FIG. 3 does not

show the other components of the RF front-end. For this disclosure, it is assumed

that the other RF front-end processing is encapsulated in mixer 24. A

synchronization unit 28 receives a mixed-down signal generated from the mixer 24.

The synchronization unit 28 adjusts the timing and the frequency of this signal so that

it is synchronized with the transmitted signal. In addition to the timing and frequency

adjustments, the synchronization unit 28 adjusts Automatic Gain Control (AGC)

settings and detects the beginning of the frame. The synchronization unit 28

performs these functions during the TD preamble 18. Once synchronization has

occurred, the rest of the OFDM symbols starting from the symbols in the FD preamble

20 will be synchronized. Additional operations that the synchronization unit 28 can

perform include well known OFDM receiver operations that pertain to removing zero-

padded sequencing and guard intervals.

[0018] A Fast Fourier Transform (FFT) processor 30 transforms the signal from the

synchronization unit 28 which is in the time domain into the frequency domain.

Transforming the signal from the time domain to the frequency domain makes it

easier to estimate the channel impulse response from the OFDM symbols that are

encoded in the data packet. A channel and noise variance estimation unit 32 uses

the band numbers provided in the various TFCs of the frequency hopping pattern and

the OFDM symbols in the FD preamble 20 to estimate the channel and the noise

variance associated with the channel. Below is a more detailed discussion on

estimating the channel impulse response and the noise variance. Note that although

FIG. 3 shows that the channel and noise variance estimation unit 32 performs both

functions, one skilled in the art will recognize that these functions can be performed in

separate processing components.

[0019] Once the channel and noise variance estimation unit 32 has estimated the

channel and noise variation, an equalization unit 34 equalizes or compensates the

OFDM symbols in the header section 14 and the payload section 16 for the effects of

the channel. Below is a more detailed discussion of equalizing or compensating the

OFDM symbols in the header section 14 and the payload section 16 for the effects of

the channel and the noise variance. In addition, the equalization unit 34 equalizes or

compensates the OFDM symbols in the header section 14 and the payload section 16

for the effects of common phase error (CPE) that may arise because of carrier

frequency offset.

[0020] After the equalization unit 34 has equalized or compensated for the OFDM

symbols in the header section 14 and the payload section 16 for the effects of the

channel, a decoder 35 then decodes the symbols. The decoder 35 includes a

branch for processing OFDM symbols of the header section 14 and another branch

for processing OFDM symbols of the payload section 16. The branch for decoding

the OFDM header symbols includes a despreader and demapper 36. The

despreader and demapper 36 despread the symbols and then generate soft bit-

metrics from these symbols. A bit de-interleaver 38 then de-interleaves the soft bit-

metrics. A Viterbi decoder 40 receives the soft bit-metrics from the bit de-interleaver

38 and decodes the data bits. A RS decoder 42 receives the decoded bits from the

Viterbi decoder 40 and outputs the header bits.

[0021] The branch for decoding the OFDM payload symbols is similar to the branch

for processing the OFDM header symbols in that it includes a despreader and

demapper 44 and a bit de-interleaver 46. However, the branch for processing the

OFDM payload symbols differs in that there is a de-puncturer and Viterbi decoder 48

that inserts zeros in predetermined locations (defined by the puncturer in the

transmitter) and then decodes the data bits. A descrambler 50 receives the decoded

bits from the de-puncturer and Viterbi decoder 48 and de-scrambles the bits to get

back the information bits.

[0022] The decoder shown in FIG. 3 is illustrative of one embodiment that can

perform decoding operations on the OFDM symbols of the header section 14 and the

payload section 16. One of ordinary skill in the art will recognize that other

configurations are possible and that the implementation shown in FIG. 3 is not limiting

with respect to these configurations. For example, those skilled in the art will

recognize that the decoder 35 may have only one branch that performs both

processing of the OFDM header symbols and the OFDM payload symbols.

[0023] As mentioned above, the channel and noise variance estimation unit 32 uses

the band number information associated with the various TFCs of the frequency

hopping pattern and the OFDM symbols in the FD preamble 20 to estimate the

channel impulse response. The OFDM symbols received by the channel and noise

variance estimation unit 32 from the FFT processor 30 are represented as:

R _n (Jc) = H _m (k)S _n (k) + N(k) (1 )

wherein & e [0,127] is the sub-carrier index, n is the OFDM symbol number, and

m e {1,2,3} is the sub-band index, which is a function of the TFC number shown in FIG.

2 and symbol number n. H _m (k) represents the channel frequency response for

sub-carrier k on band m; S _n (k) and R _n (k) represent the transmitted and received

symbols respectively in the frequency domain; and /V(Zc) represents the additive white

noise component on sub-carrier Zc.

[0024] For the OFDM symbols in the FD preamble section 20 (i.e., CE symbols), the

transmitted symbols S _n (Zc) = A(k), where A(k) is a known training sequence. An

estimate of the channel impulse response is derived by dividing the received symbol

by the training sequence and averaging across the number of symbols in that band.

In particular, the estimate of the channel impulse response H _{CE m} (k) on sub-carrier k

in sub-band m derived from CE symbols is derived as follows:

wherein:

_{p =} 2 TFC = {1,2,3,4} 6 TFC = {5,6,7}

and n(m,p) is given as

m + 3(p - ϊ) TFC = {1,2} n(m, p) = 2m + p - 2 TFC = {3,4} (4) p TFC = {5,6,7}

[0025] Substituting for R _n {k) and S _n (Zc) in the Equation 2, results in:

1

Those skilled in the art will recognize that the quality of the channel estimates can be

improved by increasing the number of terms (i.e., P) in the summation. In addition,

the channel estimates as derived above can be further fine-tuned by using some prior

information about the channel environment. One such method is implemented by

limiting the length of the channel impulse response to equal the length of the zero-

suffix.

[0026] In addition to estimating the channel impulse response, the channel and noise

variance estimation unit 32 estimates the noise variance in each of the sub-carriers

and the sub-bands using the CE symbols. The noise variance estimate will help

improve the performance of the system 22 in the presence of interference. One type

of interference that a noise variance estimate will help mitigate is narrowband

interference that arises from narrowband devices operating in a common band. For

example, the bandwidth of an ultra-wideband (UWB) system in a WiMedia application

is about 1.5 GHz and since UWB devices do not have exclusive use of this band,

there is a high probability that there will be interference from other narrowband

devices operating in this band. Under a narrowband interference scenario, some of

the sub-carriers will be affected severely and this will degrade the overall

performance of the system. Another example of in-band interference: in a WiMedia

TFI-OFDM system, multiple access is attained by using a frequency hopping

sequence such as one shown in FIG. 2. As a result of this scheme, one piconet

might cause interference to another piconet if the piconets are in close range. Both

the piconet interference and the narrowband interference will manifest as white noise

to the desired signal at the receiver.

[0027] Usually, in a coded system, it is assumed that all the OFDM symbols have

the same signal-to-noise-ratio (SNR) or white noise, and therefore this term is

removed from the metric calculation unit, which in this disclosure may be part of the

demapper (36 or 44) or the Viterbi decoder (40 or 48). However, this is not the

case in reality due to the above mentioned in-band interference scenarios. Under

these conditions, scaling the metrics for the Viterbi decoder by a term proportional to

the noise power will help improve the receiver performance. This finding

necessitates the desire to estimate the noise variance in each of the sub-carriers for

all the CE symbols.

[0028] The channel and noise variance estimation unit 32 determines the noise

variance in each sub-carrier by using the following equation

σ _C ^∑ _E ⁽ k) = R _n (K) - A(K)H _CE ⁽ k) (6)

where n e {1,2,3,4,5,6} represents the symbol number in the FD preamble.

In order to improve the reliability of the noise variance estimate, the noise variance in

each symbol is derived by averaging OcE,n ² (k) over all the sub-carriers as shown in

the following equation:

JV-I

' CJJ, ,,n = Y _τ L ^σ C _E λ ^k ) (7)

[0029] With the channel impulse response estimate and noise variance estimates,

the equalization unit 34 equalizes and then scales the output by the noise variance

estimates for header and payload symbols as shown below:

X _n (k) = G _m (k)R _n (k) (8)

wherein X _n (Zc) is the output of the equalization unit 34, and G _m (Zc) is derived from

H _{CE m} (k) which is based on the equalization scheme adapted for the system.

Uk) = ₂ ^Xλk) ... (9)

or

Uk) = ₂ ^Xλk) (10) ^σ CE, mod(κ-l, 6)+l

where the function mod(a,b) represents the remainder of alb. X results from channel

compensation and Y is the result of scaling this output by the noise variance

estimates. Y is the input to the demapper (bit-metric calculation unit).

[0030] FIG. 4 shows a flow chart 52 describing the operation of the TFI-OFDM

communications receiver 22 depicted in FIG. 3. Operation of the TFI-OFDM

communications receiver 22 begins at 54 where the RF front-end receives a data

packet embodied in a RF signal of a certain frequency and another signal at a

different frequency generated from the oscillator 26. The synchronization unit 28

receives the mixed-down signal from the RF front-end and performs various

synchronization operations at 56. As mentioned above, these synchronization

operations include adjusting the timing and the frequency of the mixed-down signal,

adjusting the AGC settings, and removing zero-padded sequencing and guard

intervals.

[0031] The FFT processor 30 transforms the adjusted signal from the

synchronization unit 28 into the frequency domain at 58. The channel and noise

variance estimation unit 32 then uses the band numbers provided in the various TFCs

of the frequency hopping pattern along with the transformed OFDM symbols in the FD

preamble 20 to estimate the channel impulse response and the noise variance

associated with the channel at 60. Once the channel and noise variance estimation

unit 32 has estimated the channel impulse response and noise variance, the

equalization unit 34 then equalizes or compensates the OFDM symbols in the header

section 14 and the payload section 16 for the effects of the channel at 62. In

addition, the equalization unit 34 can equalize or compensate the OFDM symbols in

the header section 14 and the payload section 16 for the effects of CPE. After

equalization, the decoder 35 decodes the header OFDM symbols and the payload

OFDM symbols at 64.

[0032] FIG. 5 shows block diagram of a TFI-OFDM receiver 66 according to a

second embodiment. In this embodiment, there is an updater 68 that updates the

channel estimates and noise variance estimates on the header symbols that are

generated by the channel and noise variance estimation unit 32. Below is a more

detailed discussion on updating the channel estimates and noise variance estimates

on the header symbols. The TFI-OFDM receiver 66 of FIG. 5 further includes a

switch 70 that enables the channel and noise variance estimation unit 32 to output

the estimated channel impulse response and noise variance estimates to the

equalization unit 34 when the switch is in position 1. The equalization unit 34 will

equalize or compensate the OFDM symbols in the header section 14 and the payload

section 16 for the effects of the channel and noise variance in the manner described

above.

[0033] A decoder 72 then decodes the OFDM symbols for the header section and

the payload section. The decoder 72 is similar to the decoder 35 shown in FIG. 3 in

that there is a separate branch for decoding the OFDM symbols of the header section

and the payload section. Like FIG. 3, these branches include the same elements

(i.e., the despreader and demapper 36, bit de-interleaver 38, Viterbi decoder 40 and

RS decoder 42 for the header symbol branch and the despreader and demapper 44,

bit de-interleaver 46, de-puncturer and Viterbi decoder 48 and De-scrambler 50 for

the payload symbol branch) that perform the same functions in the manner described

above. A difference between the decoder 72 of FIG. 5 and the decoder 35 of FIG. 3

is that the decoder in FIG. 5 includes a feed back branch for further processing the

OFDM symbols of the header section.

[0034] The feed-back branch in decoder 72 includes encoding modules in order to

generate reference header symbols from the decoded header bits. It includes an RS

encoder 74 that RS re-encodes the header bits generated from the RS decoder 42.

A convolution encoder 76 receives the RS encoded transmitted header bits from the

RS encoder 74 and convolutionally encodes the bits. A bit de-interleaver 78

receives the convolutionally encoded transmitted header bits from the convolutional

encoder 76 and bit interleaves the bits. A mapper and spreader 80 receives the bit-

interleaved bits from the bit interleaver 78 and maps the bits to generate reference

symbols that are sent to the updater 68. Below is a more detailed discussion on the

processing operations performed in the feed-back branch.

[0035] The decoder 72 shown in FIG. 5 is illustrative of one embodiment that can

perform the additional processing operations on the OFDM symbols of the header

section. One of ordinary skill in the art will recognize that other configurations are

possible and that the implementation shown in FIG. 5 is not limiting with respect to

these configurations. For example, latency in the decode-encode chain (i.e., RS

decoder 42 and RS encoder blocks 74) can be reduced by removing RS decoder 42

and RS encoder blocks 74. In this case, a Header Check Sequence (HCS) can be

used to verify if the decoded header bits are correct. The estimates will then be

updated only when the header bits are decoded without errors. In another

embodiment, the despreader and demapper 36, bit de-interleaver 38, Viterbi decoder

40, RS decoder 42, RS encoder 74, convolution encoder 76 and bit interleaver 78,

may be replaced with a slicer that uses slicing, which is a simplified method that

generates reference symbols. In this embodiment, the slicer would generate the

reference symbols to the updater 68.

[0036] The updater unit 68 receives the reference OFDM header symbols from the

mapper and spreader 80. The updater unit 68 then updates the channel impulse

response estimate and the noise variance estimates by using the band numbers

provided in the various TFCs of the frequency hopping pattern, the estimated channel

impulse response and the noise variance estimates generated from the channel and

noise variance estimation unit 32 (derived from CE symbols), the symbols generated

from the FFT processor 30 which are referred to as received symbols and the

reference symbols sent to the updater 68 from the feed-back branch of the decoder

72. Below is a more detailed discussion on re-estimating of the channel impulse

response and the noise variance.

[0037] After updating the channel impulse response and the noise variance

estimates, the switch 70 is moved to position 2. When the switch is in position 2, the

updater 68 then sends the updated channel impulse response and the noise variance

estimates to the equalization unit 34, which equalizes or compensates the OFDM

symbols in the payload section 16 for the effects of the channel. The decoder 72

then processes the OFDM symbols in the payload in the upper branch in the manner

described above.

[0038] The updater 68 improves the channel impulse estimates and noise variance

estimates derived in Equations 2 and 6 by utilizing the header symbols instead of

solely the FD preamble OFDM symbols. In order to update the estimates, the

updater needs to know the information sequence. An estimate of the transmitted

data can be generated either by slicing the output of the FFT processor 30 or by

using the output of the Viterbi decoder 40. Since the channel impulse response is

assumed static for the duration of the packet, updating the estimates over payload

symbols does not provide significant performance gains (assuming that the in-band

interference is present for the complete duration of the packet). In addition, the

complexity and latency associated with the update over the payload symbols is

considerably more than the update over header symbols. In addition, the header is

usually transmitted at the lowest data rate and therefore it is more resilient to the

channel errors than the payload symbols. In addition, the receiver can process the

payload symbols only after the header is decoded. This implies that the reference

symbols for the header OFDM symbols can be generated and the estimates updated

before the payload symbols are processed. Thus, the updater 68 will update the

channel impulse response and the noise variance estimates based on the FD

preamble OFDM symbols and the header OFDM symbols.

[0039] In a WiMedia TFI-OFDM system, 12 OFDM symbols (with at least four

symbols in each band) are used to transmit the header information. The channel

estimates ( H _{HDR m} {k) ) can be derived from the header symbols as shown in the

following equation:

wherein S _n (Zc) represents the estimate of the transmitted symbol S _n (Zc) and is derived

from either the slicer output or the Viterbi decoder output. The term R _n (k) represents

the received header symbols (after FFT processor 30). Substituting Equation 1 in

Equation 11 , results in:

N λ n(m,p) ^\ k) (F) I ^ V >>( ^m _' P) ( ) (12)

" p=l S n(m,p) (k) P= ^{1 ύ} n(m,p) K ^K )

wherein

4 TFC = {1,2,3,4}

P = (13) 12 TFC = {5,6,7}

and n(m,p) is given as

m + 3(p -l) TFC = {1,2} p - \ n(m, p) = 2(m - ϊ) + p + 4 TFC = {3,4} (14)

P TFC = {5,6,7}

wherein |_xj represents the integer part of x.

[0040] The noise variance in each sub-carrier can then be calculated using the

following equation:

σ _H ² _DR,» (*) = ![|*, (k) - S _n + - S _n+6 {k)H _HDKm {k) (15)

wherein n e {1,2,3,4,5,6} .

The noise variance in each band is then derived by averaging σHDR, _m ² (k) over the

entire band as shown in the following equation:

[0041] In a WiMedia application for TFI-OFDM communications system 66, the

OFDM sub-carriers are modulated using Quadrature Phase-shift Keying (QPSK)

mapping for the lower data rate modes. This enables the TFI-OFDM receiver 66 to

make reliable decisions on the symbols by using a simple slicer which is a simplified

method to generate reference symbols. The output of the slicer is represented by

the following equation:

b = sign(X _n (k)) = sign(G _m (K)R _n (K)) (17)

The estimated bits are then mapped into symbols to form S _n (Zc), which will be used in

Equation 11 and 15 to update the estimates.

[0042] The performance of the TFI-OFDM receiver 66 is further improved by using

iterative decoding. In the first pass, channel impulse response and noise variance

are estimated on the FD preamble symbols (i.e., the CE symbols). These estimates

are then used by the equalization unit 34 to equalize and scale the received data sent

from the FFT processor 30. After the Viterbi decoder or RS decoder in the header

processing branch of the decoder 72, the bits are encoded back to the updater 68 to

generate the reference symbols (S _n ). The channel estimates and the noise variance

estimates are updated according to Equations 11 and 15, respectively. In the

second pass, after the switch has been positioned in setting 2, the updated channel

estimates (H _{HDR m} ) and noise variance estimates (σ _{HDR n} ² ) are used to equalize and

scale the received data from the FFT processor 30. Although the computational

complexity of this approach may be very high compared to a standard receiver, this

disclosure reduces the computational complexity by considering Viterbi decoder

decisions of header symbols only. Thus, the header symbols are decoded prior to

the decoding of the rest of the packet. This enables the TFI-OFDM receiver system

66 to compute reference symbols S _n corresponding to the header bits and then

update the channel estimate and the noise variance estimate accordingly.

[0043] In another embodiment, the updater 68 can perform a weighted-averaging

operation on the estimates derived from the header symbols to further improve the

results. In this embodiment, the estimates derived over the header symbols are

averaged with the estimates derived over the CE symbols as shown below:

H _m (k) = w _m H _CEtm (k) + (\ - w _m )H _HDKm (k) , (18)

< (*) = w.σ^ ² (k) + (l - w _m )σ _H ² _DR,m (k) (19)

°l = w _m σ _C ² _{E m} + (l - w _m )σ _H ² _{DR m} (20)

where 0 < w _m ≤ 1 represent the weight for band m and can be modified by the receiver

(user) dynamically.

H _{HDR m} (k) represent the channel impulse response estimates for sub-

carrier k in sub-band m derived from the CE symbols and header symbols and is

given by Equations 2 and 11 , respectively.

σ _C ² _{E m} and σl _{DR m} represent the noise variance estimate in sub-band m derived from

the CE symbols and the header symbols, respectively.

[0044] The payload symbols are equalized using:

X _n (k) = G _m (k)R _n (k) (21 )

where X _n (Zc) is the output of the equalization unit 34, and G _m (Zc) is derived from H _m (k)

based on the equalization scheme adapted for the TFI-OFDM receiver 66.

[0045] Fig 6 shows a flow chart 82 describing the operation of the TFI-OFDM

receiver 66 depicted in FIG. 5. Operation of the TFI-OFDM receiver 66 begins at 84

where the RF front-end receives a data packet embodied in a RF signal of a certain

frequency and down-converts the signal using another signal at a different frequency

generated from the oscillator 26. The synchronization unit 28 receives the mixed-

down signal and performs various synchronization operations at 86. As mentioned

above, these synchronization operations include adjusting the timing and the

frequency of this signal, adjusting the AGC settings, detecting the frame beginning

and removing zero-padded sequencing and guard intervals.

[0046] The FFT processor 30 transforms the adjusted signal from the

synchronization unit 28 into the frequency domain at 88. The channel and noise

variance estimate unit 32 then uses the band numbers provided in the various TFCs

of the frequency hopping pattern along with the transformed OFDM symbols (i.e., CE

symbols) in the FD preamble 20 to estimate the channel and the noise variance

associated with the channel at 90. Next, the switch 70 is moved to position 1 at 92,

so that the equalization unit 34 equalizes or compensates the OFDM symbols in the

header section for the effects of the channel and noise variance at 94.

[0047] The decoder 72 then decodes the OFDM symbols for the header section at

96. Transmitted header bits are generated at 98 and further processed to generate

the reference symbols that are sent to the updater at 100. The updater unit 68 then

receives the reference OFDM header symbols and updates the channel impulse

response estimate and the noise variance estimates by using the reference symbols

with the band numbers provided in the various TFCs of the frequency hopping pattern

and the estimated channel impulse response and the noise variance estimates

generated from the channel and noise variance estimation unit 32 (derived from CE

symbols), and the received symbols generated from the FFT processor 32 at 102.

[0048] After updating the channel impulse response and the noise variance

estimates, the switch 70 is moved to position 2 at 104. When the switch is in

position 2, the updater 68 then sends the updated channel impulse response and the

noise variance estimates to the equalization unit 34, which equalizes or compensates

the OFDM symbols in the payload section 16 at 106. The decoder 72 then

processes the OFDM symbols in the payload in the upper branch of the decoder at

108.

[0050] The foregoing flow charts of FIGS. 4 and 6 show some of the processing acts

associated with operating TFI-OFDM receivers 22 and 66. In this regard, each block

in the flow charts represents a process act associated with performing these functions.

It should also be noted that in some alternative implementations, the acts noted in the

blocks may occur out of the order noted in the figure or, for example, may in fact be

executed substantially concurrently or in the reverse order, depending upon the act

involved. Also, one of ordinary skill in the art will recognize that additional blocks

that describe these processing acts may be added.

[0051] The TFI-OFDM receiver 22 and 66 can take the form of an entirely hardware

embodiment, an entirely software embodiment or an embodiment containing both

hardware and software elements. In one embodiment, the operations performed by

TFI-OFDM systems 22 and 66 are implemented in software, which includes but is not

limited to firmware, resident software, microcode, etc.

[0052] Furthermore, the operations performed by TFI-OFDM receiver 22 and 66 can

take the form of a computer program product accessible from a computer-usable or

computer-readable medium providing program code for use by or in connection with a

computer or any instruction execution system. For the purposes of this description,

a computer-usable or computer readable medium can be any apparatus that can

contain, store, communicate, propagate, or transport the program for use by or in

connection with the instruction execution system, apparatus, or device.

[0052] The medium can be any apparatus that can contain, store, communicate,

propagate, or transport the program containing the instructions for performing the

image processing functions for use by or in connection with an instruction execution

system, apparatus, or device. The computer readable medium can be an electronic,

magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus

or device) or a propagation medium. Examples of a computer-readable medium

include a semiconductor or solid state memory, magnetic tape, a removable computer

diskette, a random access memory (RAM), a read-only memory (ROM), a rigid

magnetic disk and an optical disk. Current examples of optical disks include a

compact disk - read only memory (CD-ROM), a compact disk - read/write (CD-R/W)

and a digital video disc (DVD).

[0053] It is apparent that there has been provided with this disclosure, an approach

for improving performance in a TFI-OFDM. While the disclosure has been

particularly shown and described in conjunction with a preferred embodiment thereof,

it will be appreciated that a person of ordinary skill in the art can effect variations and

modifications without departing from the scope of the disclosure.