CN101208874A - Pilot transmission and channel estimation for communication systems employing frequency division multiplexing - Google Patents

Abstract

A transmitter generates a pilot having a constant time-domain envelope and a flat spectrum based on a polyphase sequence. To generate the pilot IFDMA symbol, a first sequence of pilot symbols is generated based on the polyphase sequence and the sequence is replicated multiple times to obtain a second sequence of pilot symbols. A phase ramp is applied to the second pilot symbol sequence to obtain a third output symbol sequence. A cyclic prefix is added to the third sequence of output symbols to obtain IFDMA symbols, which are transmitted in the time domain via the communication channel. The pilot symbols may be multiplexed with the data symbols using TDM and/or CDM. Pilot LFDMA symbols may also be generated with a polyphase sequence and multiplexed using TDM or CDM. A receiver derives a channel estimate based on received pilot symbols and using minimum mean square error, least squares, or some other channel estimation technique.

Description

Pilot Transmission and Channel Estimation for Communication Systems Using Frequency Division Multiplexing

background

Priority claims under 35 U.S.C. §119

This patent application claims to be filed on March 7, 2005 and is assigned to the assignee of the present invention and is hereby expressly incorporated by reference to the title "Estimation for Pilot Design and Channel Interleaved Frequency Division Multiple Access Communication Estimation and Frequency Division Multiple Access Communications with Channel Interleaving)" is the priority of Provisional Application No. 60/659,526.

I. Domain

The present invention relates generally to communications, and more particularly to pilot transmission and channel estimation for communications systems.

II. Background

Orthogonal Frequency Division Multiplexing (OFDM) is a multicarrier modulation technique that divides the total system bandwidth into multiple (K) orthogonal subbands. These subbands are also called tones, subcarriers, and frequency bins. In OFDM, each subband is associated with a corresponding subcarrier that can be modulated with data.

OFDM has certain desirable properties, such as high spectral efficiency and robustness to multipath effects. However, the main disadvantage of OFDM is the peak-to-average power ratio (PAPR), which means that the ratio of peak power to average power of an OFDM waveform can be high. The high PAPR of the OGDM waveform results from possible in-phase (or coherent) addition of all subcarriers when independently modulated with data. In fact, it can be shown that for OFDM, the peak power can be as high as K times the average power.

High PAPR for OFDM waveforms is undesirable and may degrade performance. For example, large peaks in the OFDM waveform may cause the power amplifier to operate in highly nonlinear regions or may clip, which in turn will cause intermodulation distortion and other artifacts that may degrade the signal quality. The reduced signal quality will adversely affect the performance of channel estimation, data detection, and the like.

There is thus a need in the art for techniques that can mitigate the detrimental effects of high PAPR in multicarrier modulation.

summary

Pilot transmission techniques and channel estimation techniques that can avoid high PAPR are described herein. The pilots may be generated based on the polyphase sequence and using single-carrier frequency division multiple access (SC-FDMA). A polyphase sequence is a sequence with good temporal properties (eg, constant time domain envelope) and good spectral properties (eg, flat spectrum). SC-FDMA consists of (1) interleaved FDMA (IFDMA) that transmits data and/or pilot on subbands that are evenly spaced across the K total Localized FDMA (LFDMA) for sending data and/or pilot.

In one embodiment using IFDMA for pilot transmission, a first sequence of pilot symbols is formed based on a polyphase sequence, and the sequence is replicated multiple times to obtain a second sequence of pilot symbols. A phase ramp may be applied to the second sequence of pilot symbols to obtain a third sequence of output symbols. A cyclic prefix is added to the third sequence of output symbols to form IFDMA symbols, which are transmitted in the time domain via a communication channel. The pilot symbols may be multiplexed with the data symbols using time division multiplexing (TDM), code division multiplexing (CDM), and/or some other multiplexing scheme.

In one embodiment using LFDMA for pilot transmission, a first sequence of pilot symbols is formed based on a polyphase sequence, and the sequence is transformed to the frequency domain to obtain a second sequence of frequency-domain symbols. A third sequence of symbols is formed by mapping the second sequence of frequency-domain symbols onto a group of subbands used for pilot transmission, and mapping zero symbols onto the remaining subbands. The third sequence of symbols is transformed to the time domain to obtain a fourth output sequence of symbols. A cyclic prefix is added to the fourth output sequence of symbols to form LFDMA symbols, which are transmitted in the time domain via a communication channel.

In one embodiment of channel estimation, at least one SC-FDMA symbol is received via a communication channel and processed (e.g., demultiplexed for TDM pilots, or de-channelized for CDM pilots) to obtain Received pilot symbols. SC-FDMA symbols may be IFDMA symbols or LFDMA symbols. The channel estimate is derived based on the received pilot symbols and using minimum mean square error (MMSE) techniques, least squares (LS) techniques, or some other channel estimation technique. Filtering, thresholding, truncation, and/or tap selection may be performed to obtain improved channel estimates. Channel estimation may also be improved by performing iterative or data-assisted channel estimation.

Various aspects and embodiments of the invention are described in further detail below.

Brief description of the drawings

The nature and nature of the present invention will become more apparent from the detailed description set forth below when read in conjunction with the accompanying drawings, throughout which like reference numerals identify correspondingly.

FIG. 1 shows an interleaved subband structure of a communication system.

Figure 2 illustrates generating IFDMA symbols for a set of N subbands.

Figure 3 shows a narrowband subband structure.

FIG. 4 illustrates generating LFDMA symbols for a group of N subbands.

5A and 5B illustrate two TDM pilot schemes in which pilot and data are multiplexed across multiple symbol periods and multiple sampling periods, respectively.

5C and 5D illustrate two CDM pilot schemes in which pilot and data are combined across multiple symbol periods and sampling periods, respectively.

Figure 6 shows wideband pilots time division multiplexed with data.

Figure 7A shows a process for generating pilot IFDMA symbols.

Figure 7B shows a process for generating pilot LFDMA symbols.

Fig. 8 shows the process of performing channel estimation.

Figure 9 shows a block diagram of a transmitter and receiver.

10A and 10B illustrate transmit (TX) data and pilot processors for the TDM pilot scheme and the CDM pilot scheme, respectively.

11A and 11B show IFDMA and LFDMA modulators, respectively.

12A and 12B show IFDMA demodulators for TDM and CDM pilots, respectively.

13A and 13B show LFDMA demodulators for TDM and CDM pilots, respectively.

Specific instructions

The term "exemplary" is used herein to mean "serving to serve as an example, instance, or illustration." Any embodiment or design described herein as "exemplary" is not necessarily to be construed as preferred or superior to other embodiments or designs.

The pilot transmission and channel estimation described herein can be used in various communication systems that employ multicarrier modulation or perform frequency division multiplexing. For example, these techniques may be used in Frequency Division Multiple Access (FDMA) systems, Orthogonal Frequency Division Multiple Access (OFDMA) systems, SC-FDMA systems, IFDMA systems, LFDMA systems, OFDM-based systems, and so forth. These techniques can also be used for the forward link (or downlink) and the reverse link (or uplink).

FIG. 1 shows an exemplary subband structure 100 that may be used in a communication system. The system has a total bandwidth of BW MHz, which is divided into K orthogonal subbands, which are given indices 1 to K. The spacing between adjacent subbands is BW/K MHz. In a spectrally shaped system, some subbands at both ends of the system bandwidth are not used for data/pilot transmission, but instead function as guard subbands that allow the system to meet the spectral mask requirements. Alternatively, the K subbands can be defined over an available portion of the system bandwidth. For simplicity, the following description assumes that all of the total K subbands are available for data/pilot transmission.

For the subband structure 100, the K total subbands are arranged into S disjoint subband groups, which are also called interleaves. The S groups are disjoint or non-overlapping because each of the K subbands belongs to only one group. Each group contains N subbands, which are evenly distributed across the K total subbands such that successive subbands in a group are separated by S subbands, where K=S·N. Thus, group u contains subbands u, S+u, 2S+u, . . . , (N−1)·S+u, where u is the group index and u∈{1, . The index u is also the subband offset indicating the first subband in the group. The N subbands in each group are interleaved with the N subbands in each of the other S-1 groups.

Fig. 1 shows a specific subband structure. In general, a subband structure may include any number of subband groups, and each group may include any number of subbands. Each group may include the same or a different number of subbands. For example, some groups may include N subbands, while other groups may include 2N, 4N, or some other number of subbands. The subbands in each group are evenly distributed (ie, evenly spaced) across the total K subbands to achieve the benefits explained below. For simplicity, the following description assumes the use of subband structure 100 in FIG. 1 .

The S subband groups can be viewed as S channels available for data and pilot transmission. For example, each user may be assigned a set of subbands, and data and pilot for each user may be sent on the assigned set of subbands. S users may simultaneously transmit data/pilot to the base station on the S subband groups via the reverse link. The base station may also transmit data/pilot to S users simultaneously on the S subband groups via the forward link. For each link, up to N modulation symbols may be sent on each group of N subbands in each symbol period (in time or frequency) without causing interference to other groups of subbands. A modulation symbol is a complex value corresponding to a point in a signal constellation (eg, M-PSK, M-QAM, etc.).

For OFDM, modulation symbols are sent in the frequency domain. For each subband group, N modulation symbols may be sent on the N subbands in each symbol period. In the following description, a symbol period is the duration of one OFDM symbol, one IFDMA symbol, or one LFDMA symbol. One modulation symbol is mapped to each of the N subbands used for transmission, and zero symbols (ie, zero signal values) are mapped to each of the K-N unused subbands. The K modulations and null symbols are transformed from the frequency domain to the time domain by performing a K-point Inverse Fast Fourier Transform (IFFT) on the K modulations and null symbols to obtain K time domain samples. These time domain samples may have high PAPR.

Figure 2 illustrates generating IFDMA symbols for a set of N subbands. The original sequence of N modulation symbols to be transmitted on the N subbands in group u in one symbol period is denoted {d ₁ , d ₂ , d ₃ , . . . , d _N } (block 210). This original sequence of N modulation symbols is replicated S times to obtain a spread sequence of K modulation symbols (block 212). The N modulation symbols are sent in the time domain and occupy a total of N subbands in the frequency domain. S copies of the original sequence result in N occupied subbands separated by S subbands, and adjacent occupied subbands are separated by S-1 zero-power subbands. This spreading sequence has a comb spectrum occupying subband group 1 in FIG. 1 .

The spreading sequence is multiplied by a phase ramp to obtain a frequency-shifted output symbol sequence (block 214). Each output symbol in the frequency-shifted sequence can be generated as follows:

x _n ＝d _n ·e ^{-j2π(n-1)(u-1)/K} , n=1,..., K, formula (1)

where _dn is the nth modulation symbol in the spreading sequence and _xn is the nth output symbol in the frequency shifted sequence. The phase slope e ^{- j2π·(n-1)·(u-1)/K} has a phase slope of 2π·(u-1)/K, which is determined by the first subband in group u. The terms "n-1" and "u-1" in the index of the phase ramp are because the indices n and u start from '1' instead of '0'. Multiplication with the phase ramp in the time domain shifts the comb spectrum of the spreading sequence towards high frequency so that the frequency shifted sequence occupies subband group u in the frequency domain.

The last C output symbols in the frequency-shifted sequence are copied to the beginning of the frequency-shifted sequence to form an IFDMA symbol comprising K+C output symbols (block 216). These C copied output symbols are often referred to as a cyclic prefix or guard interval, and C is the cyclic prefix length. The cyclic prefix is used to combat inter-symbol interference (ISI) caused by frequency selective fading—that is, a frequency response that varies across the system bandwidth. The K+C output symbols of the IFDMA symbols are transmitted in K+C sampling periods, one output symbol in each sampling period. The symbol period of IFDMA is the duration of one IFDMA symbol and is equal to K+C sampling periods. The sampling period is also often referred to as the chip period.

Since IFDMA symbols are periodic in the time domain (in addition to having a phase ramp), an IFDMA symbol occupies a set of N equally spaced subbands starting from subband u. Similar to OFDMA, users with different subband offsets occupy different subband groups and are orthogonal to each other.

FIG. 3 shows an exemplary narrowband subband structure 300 that may be used in a communication system. For the subband structure 300, the K total subbands are arranged into S non-overlapping groups. Each group contains N subbands that are contiguous to each other. In general, N>1, S>1, and K=S·N, wherein N and S of the narrowband subband structure 300 may be the same as or different from N and S of the interleaved subband structure in FIG. 1 . A group v contains subbands (v-1) N+1, (v-1) N+2, ..., v N, where v is the group index and v ∈ {1, ..., S}. In general, a subband structure may include any number of groups, each group may contain any number of subbands, and each group may contain the same or a different number of subbands.

FIG. 4 illustrates generating LFDMA symbols for a group of N subbands. The original sequence of N modulation symbols to be transmitted on the group of subbands in one symbol period is denoted {d ₁ , d ₂ , d ₃ , . . . , d _N } (block 410). The original sequence of N modulation symbols is converted to the frequency domain using an N-point Fast Fourier Transform (FFT) to obtain a sequence of N frequency-domain symbols (block 412). The N frequency domain symbols are mapped onto the N subbands for transmission and the KN zero symbols are mapped onto the remaining KN subbands to generate a sequence of K symbols (block 414). The N subbands used for transmission have indices k+1 to k+N, where 1≦k≦(KN). The sequence of K symbols is then converted to the time domain using a K-point IFFT to obtain a sequence of K time-domain output symbols (block 416). The last C output symbols of the sequence are copied to the beginning of the sequence to form an LFDMA symbol containing K+C output symbols (block 418).

The LFDMA symbols are generated such that they occupy a group of N contiguous subbands starting from subband k+1. Similar to OFDMA, users may be assigned different non-overlapping groups of subbands, thereby making them mutually orthogonal. Different subband groups can be assigned to each user in different symbol periods to achieve frequency diversity. The group of subbands for each user may be selected based on, for example, a frequency hopping pattern.

Like OFDMA, SC-FDMA has certain desirable properties, such as high spectral efficiency and robustness against multipath effects. Furthermore, SC-FDMA does not have very high PAPR because the modulation symbols are sent in the time domain. The PAPR of the SC-FDMA waveform is determined by the signal points in the chosen signal constellation (eg, M-PSK, M-QAM, etc.) used. However, due to the non-flat communication channel, the time-domain modulation symbols in SC-FDMA are susceptible to inter-symbol interference. Equalization may be performed on received modulation symbols to mitigate the deleterious effects of inter-symbol interference. Equalization requires a reasonably accurate channel estimate of the communication channel, which can be obtained using the techniques described in this paper.

A transmitter may send pilots to facilitate channel estimation by a receiver. A pilot is the transmission of symbols that are known a priori to both the transmitter and receiver. As used herein, a data symbol is a modulation symbol that corresponds to data, and a pilot symbol is a modulation symbol that corresponds to pilot. Data symbols and modulation symbols may be derived from the same or different signal constellations. As will be explained below, pilots may be transmitted in various ways.

5A shows a TDM pilot scheme 500 in which pilot and data are multiplexed across multiple symbol periods. For example, data may be sent in D ₁ symbol periods, then pilot may be sent in the next P ₁ symbol periods, then data may be sent in the next D ₁ symbol periods, and so on . In general, D ₁ ≥1 and P ₁ ≥1. For the example shown in FIG. 5A , D ₁ >1 and P ₁ =1. A sequence of N data symbols may be sent on one subband group/group in each symbol period for data transmission. A sequence of N pilot symbols may be sent on one subband group/group in each symbol period for pilot transmission. For each symbol period, the sequence of N data or pilot symbols may be converted to an IFDMA symbol or an LFDMA symbol as described above with respect to FIGS. 2 and 4, respectively. SC-FDMA symbols may be IFDMA symbols or LFDMA symbols. An SC-FDMA symbol containing only a pilot is called a pilot SC-FDMA symbol, which can be a pilot IFDMA symbol or a pilot LFDMA symbol. An SC-FDMA symbol containing only data is called a data SC-FDMA symbol, which can be a data IFDMA symbol or a data LFDMA symbol.

Figure 5B shows a TDM pilot scheme 510 in which pilot and data are multiplexed across multiple sampling periods. For this embodiment, data and pilot are multiplexed within the same SC-FDMA symbol. For example, data symbols may be sent in D ₂ sample periods, then pilot symbols in the next P ₂ sample periods, then data symbols in the next D ₂ sample periods, and so on . In general, D ₂ ≧1 and P ₂ ≧1. For the example shown in FIG. 5B , D ₂ =1 and P ₂ =1. A sequence of N data and pilot symbols can be sent on one subband set/group in each symbol period, and can be converted into an SC-FDMA symbol as described with respect to FIGS. 2 and 4 .

TDM pilot schemes can also multiplex pilot and data across both symbol periods and sample periods. For example, data and pilot symbols may be sent during some symbol periods, only data symbols may be sent during other symbol periods, and only pilot symbols may be sent during some symbol periods.

Figure 5C shows a CDM pilot scheme 530 in which pilot and data are combined across multiple symbol periods. For this embodiment, a sequence of N data symbols is multiplied by the first M-chip orthogonal sequence { _wd } to obtain a sequence of M scaled data symbols, where M>1. Each scaled data symbol sequence is obtained by multiplying the original data symbol sequence by one chip of the orthogonal sequence {w _d }. Similarly, a sequence of N pilot symbols is multiplied by a second M-chip orthogonal sequence { _wp } to obtain M sequences of scaled pilot symbols. Each sequence of scaled data symbols is then added to the corresponding sequence of scaled pilot symbols to obtain a combined sequence of symbols. The M combined symbol sequences are obtained by adding the M scaled data symbol sequences to the M scaled pilot symbol sequences. Each combined symbol sequence is converted into an SC-FDMA symbol.

These orthogonal sequences may be Walsh sequences, OVSF sequences, and the like. For the example shown in Fig. 5C, M=2, the first orthogonal sequence is { _wd }={+1+1}, and the second orthogonal sequence is { _wp }={+1-1}. The N data symbols are multiplied by +1 over the symbol period t, and also multiplied by +1 over the symbol period t+1. The N pilot symbols are multiplied by +1 for the symbol period t and by -1 for the symbol period t+1. For each symbol period, the N scaled data symbols are added to the N scaled pilot symbols to obtain N combined symbols corresponding to that symbol period.

Figure 5D shows a CDM pilot scheme 540 in which pilot and data are combined across multiple sample periods. For this embodiment, a sequence of N/M data symbols is multiplied by the M chip orthogonal sequence { _wd } to obtain a sequence of N scaled data symbols. Specifically, the first data symbol d ₁ (t) in the original sequence is multiplied by the orthogonal sequence {w _d } to obtain the first M scaled data symbols, and the next data symbol d ₂ ( t} is multiplied by the orthogonal sequence {w _d } to get the next M scaled data symbols, and so on, and the last data symbol d _N/M (t) in the original sequence is multiplied by the orthogonal sequence {w _d } to get the last M scaled data symbols. Similarly, a sequence of N/M pilot symbols is multiplied by an M-chip orthogonal sequence {w _p } to obtain N A sequence of scaling data symbols. This sequence of N scaling data symbols is added to the sequence of N scaling pilot symbols to obtain a sequence of N combined symbols, which is converted into a SC-FDMA symbols.

For the example shown in Fig. 5D, M=2, the orthogonal sequence for data is {w _d }={+1+1}, and the orthogonal sequence for pilot is {w _p }={+1- 1}. A sequence of N/2 data symbols is multiplied by the orthogonal sequence {+1+1} to obtain a sequence of N scaled data symbols. Similarly, a sequence of N/2 pilot symbols is multiplied by the orthogonal sequence {+1-1} to obtain a sequence of N scaled pilot symbols. For each symbol period, the N scaled data symbols are added to the N scaled pilot symbols to obtain N combined symbols corresponding to that symbol period.

As shown in Figures 5C and 5D, a CDM pilot may be sent in each symbol period. CDM pilots may also be sent only in certain symbol periods. The pilot scheme can also use a combination of TDM and CDM. For example, a CDM pilot may be sent on some symbol periods and a TDM pilot may be sent on other symbol periods. Frequency division multiplexed (FDM) pilots may also be sent on a designated set of subbands, eg, downlink.

For the embodiments shown in Figures 5A through 5D, TDM or CDM pilots are sent on the N subbands used for data transmission. In general, the subbands used for pilot transmission (or simply pilot subbands) may be the same as or different from the subbands used for data transmission (or simply data subbands). There may also be fewer or more subbands used to transmit pilot than subbands used to transmit data. Data and pilot subbands may be static for the entire transmission. Alternatively, the data and pilot subbands may be hopped across frequency in different time slots to achieve frequency diversity. For example, a physical channel may be associated with a frequency hopping (FH) pattern that indicates one or more specific groups or groups of subbands that the physical channel uses at each time slot. A slot may span one or more symbol periods.

Figure 6 shows a broadband pilot scheme 600 that would be more suitable for the reverse link. For this embodiment, each user sends a broadband pilot, which is a pilot sent on all or most of the total K subbands, eg, all subbands available for transmission. Wideband pilots can be generated in the time domain (eg, with a pseudorandom number (PN) sequence) or in the frequency domain (eg, using OFDM). A broadband pilot for each user, which may be generated using LFDMA (as shown in Figure 6) or IFDMA (not shown in Figure 6), may be time multiplexed with the data transmission from that user. Wideband pilots from all users can be transmitted in the same symbol period, which avoids interference of data on pilots used for channel estimation. The wideband pilots from each user may be code multiplexed (eg, pseudo-randomly) relative to the wideband pilots from other users. This can be achieved by assigning each user a different PN sequence. The wideband pilot for each user has a low peak-to-average power ratio (PAPR) and spans the entire system bandwidth, which enables the receiver to derive a wideband channel estimate for that user. For the embodiment shown in Figure 6, each data subband is hopped across frequencies at different time slots. For each slot, a channel estimate may be derived for each data subband based on wideband pilots.

5A to 6 illustrate exemplary pilot and data transmission schemes. Pilot and data may also be transmitted in other ways using any combination of TDM, CDM, and/or some other multiplexing scheme.

TDM and CDM pilots can be generated in various ways. In one embodiment, the pilot symbols used to generate the TDM and CDM are modulation symbols from a known signal constellation, such as QPSK. The TDM pilot scheme shown in FIG. 5A and the CDM pilot scheme shown in FIG. 5C may use a sequence of N modulation symbols. The TDM pilot scheme shown in Figure 5B and the CDM pilot scheme in Figure 5D may use a sequence of N/M modulation symbols. The sequence of N modulation symbols and the sequence of N/M modulation symbols may each be chosen to have (1) a frequency spectrum that is as flat as possible, and (2) a temporal envelope that varies as little as possible. A flat spectrum ensures that all subbands used for pilot transmission have sufficient power to allow the receiver to correctly estimate the channel gain for these subbands. The constant envelope prevents distortion from circuit blocks such as power amplifiers.

In another embodiment, the pilot symbols used to generate the TDM and CDM pilots are formed based on polyphase sequences with good temporal and spectral properties. For example, pilot symbols can be generated as follows:

n＝1，…，N，                                        式(2)

n=1,...,N, formula (2)

where the phase  _n can be derived based on any of the following formulas:

 _n = π·(n-1)·n, formula (3)

 _n ＝π·(n-1) ² , formula (4)

 _n = π·[(n-1)·(nN-1)], formula (5)

In formula (6), Q and N are relatively prime. Formula (3) corresponds to the Golomb sequence, formula (4) corresponds to the P3 sequence, formula (5) corresponds to the P4 sequence, and formula (6) corresponds to the Chu sequence. The P3, P4 and Chu sequences can be of any length.

Pilot symbols can also be generated as follows:

l＝1，…，T且m＝1，…，T，                           式(7)

l=1,...,T and m=1,...,T, formula (7)

Among them, the phase  _{l, m} can be derived based on any of the following formulas:

 _{l, m} = 2π·(l-1)·(m-1)/T, formula (8)

 _{l, m} = -(π/T)·(T-2l+1)·[(l-1)·T+(m-1)], formula (9)

Formula (8) corresponds to the Frank sequence, formula (9) corresponds to the P1 sequence, and formula (10) corresponds to the Px sequence. The lengths of the Frank, P1 and Px sequences are constrained to satisfy N= ^T2 , where T is a positive integer.

The pilot symbol sequence generated based on any one of the above polyphase sequences has both a flat frequency spectrum and a constant time-domain envelope. Other polyphase sequences with good spectral properties (eg, flat or known spectrum) and good temporal properties (eg, constant or known time domain envelope) can also be used. A TDM or CDM pilot generated with this sequence of pilot symbols will thus have (1) low PAPR, which avoids distortions caused by circuit elements such as power amplifiers, and (2) a flat spectrum, which allows the receiver The channel gain can be accurately estimated for all subbands used for pilot transmission.

FIG. 7A shows a process 700 for generating pilot IFDMA symbols. A first sequence of pilot symbols is formed based on a polyphase sequence, which may be any of the polyphase sequences described above or some other polyphase sequence (block 710). The first sequence of pilot symbols is replicated multiple times to obtain a second sequence of pilot symbols (block 712). A phase ramp is applied to the second sequence of pilot symbols to obtain a third sequence of output symbols (block 714). The phase ramp can be applied digitally to the pilot symbols, or it can be achieved by an up-conversion process. A cyclic prefix is added to the third output symbol sequence to obtain a fourth output symbol sequence, which is a pilot IFDMA symbol (block 716). The pilot IFDMA symbols are transmitted in the time domain via a communication channel (block 718). Although not shown in FIG. 7A for simplicity, pilot symbols may be multiplexed with data symbols using TDM and/or CDM, such as described above with respect to FIGS. 5A-5D.

7B is a process 750 for generating pilot LFDMA symbols. A first sequence of pilot symbols is formed based on a polyphase sequence, which may be any of the polyphase sequences described above or some other polyphase sequence (block 760). The first sequence of N pilot symbols is transformed to the frequency domain with an N-point FFT to obtain a second sequence of N frequency-domain symbols (block 762). These N frequency-domain symbols are then mapped onto N subbands for pilot transmission, and zero symbols are mapped onto the remaining K-N subbands to obtain a third sequence of K symbols (block 764) . The third sequence of K symbols is transformed to the time domain using a K-point IFFT to obtain a fourth sequence of K time-domain output symbols (block 766). A cyclic prefix is added to the fourth output symbol sequence to obtain a fifth sequence of K+C output symbols, which is a pilot LFDMA symbol (block 768). The pilot LFDMA symbols are transmitted in the time domain via a communication channel (block 770). Although not shown in FIG. 7B for simplicity, pilot symbols may be multiplexed with data symbols using TDM and/or CDM, such as described above with respect to FIGS. 5A-5D.

For both IFDMA and LFDMA, the number of subbands used for pilot transmission may be the same or different than the number of subbands used for data transmission. For example, a user may be allocated 16 subbands for data transmission and 8 subbands for pilot transmission. Another 8 subbands can be allocated to another user for data/pilot transmission. For the interleaved subband structure in FIG. 1 , multiple users can share the same subband group, or for the narrowband subband structure in FIG. 3 , they can share the same subband group.

For the staggered subband structure in FIG. 1, an FDM pilot may be transmitted on one or more subband groups to enable the receiver to perform various functions such as channel estimation, frequency tracking, time tracking, and so on. In the first type of staggered FDM pilot, pilot IFDMA symbols are transmitted on subband group p in some symbol periods and pilots are transmitted on subband group p+S/2 in other symbol periods IFDMA symbols. For example, if S=8, the pilot IFDMA symbols can be transmitted using a staggered pattern of {3,7}, whereby the pilot IFDMA symbols are sent on subband set 3, then on subband set 7, then On subband group 3, and so on. In the second type of staggered FDM pilot, pilot IFDMA symbols are transmitted over subband group p(t)=[p(t-1)+Δp] modulo S+1 in symbol period t, where Δp is the difference between the subband group indices of two consecutive symbol periods, and +1 corresponds to an indexing scheme that starts from 1 instead of 0. For example, if S=8 and Δp=3, a staggered pattern of {1, 4, 7, 2, 5, 8, 3, 6} may be used to transmit pilot IFDMA symbols, whereby on subband group 1 Pilot IFDMA symbols are sent, then on subband group 4, then on subband group 7, and so on. Other stagger patterns may also be used. Staggered FDM pilots enable the receiver to obtain channel gain estimates corresponding to more subbands, which will improve channel estimation and detection performance.

8 illustrates a process 800 performed by a receiver to estimate a response of a communication channel based on a TDM pilot or a CDM pilot sent by a transmitter. The receiver obtains one SC-FDMA symbol for each symbol period and removes the cyclic prefix from the received SC-FDMA symbols (block 810). For IFDMA, the receiver removes the phase slope in the received SC-FDMA symbols. For both IFDMA and LFDMA, the receiver acquires K received data/pilot symbols from the SC-FDMA symbol.

The receiver then reverses the TDM or CDM performed on the pilot (block 812). For the TDM pilot scheme shown in Figure 5A, K received pilot symbols _rp (n), n=1,...K are obtained from each pilot SC-FDMA symbol. For the TDM pilot scheme shown in Figure 5B, multiple receive pilot symbols are obtained from each SC-FDMA symbol that contains the TDM symbol.

For the CDM pilot scheme shown in Figure 5C, the M received SC-FDMA symbols containing the CDM pilot are processed to recover the pilot symbols as follows:

$r_p\left(\mathrm{no}\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{w}_{p,i}\·r(t_i,\mathrm{no}),\mathrm{no}=1,\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;,K,$ Formula (11)

Where r(t _i , n) is the received sample corresponding to sampling period n in symbol period ti;

w _p,i is the ith chip of the orthogonal sequence corresponding to the pilot; and

r _p (n) is the received pilot symbol corresponding to sampling period n.

Equation (11) assumes that the CDM pilot is transmitted at symbol periods _ti to _tM , where M is the length of the orthogonal sequence. The K received pilot symbols corresponding to the CDM symbol are obtained from equation (11).

For the CDM pilot scheme shown in Figure 5D, each received SC-FDMA symbol containing the CDM pilot is processed to recover the pilot symbol as follows:

$r_p\left(\mathrm{no}\right)=\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{w}_{p,i}\&Center\; Dot;r((\mathrm{no}-1)\&Center\; Dot;m+i),\mathrm{no}=1,\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;,K/m,$ Formula (12)

Where r((n-1)·M+i) is the received sample corresponding to the sampling period (n-1)·M+i in the received SC-FDMA symbol with the CDM pilot. The K/M received pilot symbols corresponding to the CDM pilot are obtained from equation (12).

Frequency selective communication channels result in inter-symbol interference (ISI). However, because of the cyclic prefix, this ISI is confined within a single SC-FDMA symbol. Furthermore, because of the cyclic prefix, the linear convolution operation due to the channel impulse response actually becomes a circular convolution, similar to OFDMA. Accordingly, channel estimation, equalization, and other operations may be performed in the frequency domain when pilot symbols and data symbols are not transmitted in the same SC-FDMA symbol.

For the TDM scheme shown in Figure 5A and the CDM scheme shown in Figure 5C, the receiver obtains K received pilot symbols from each pilot transmission. A K-point FFT may be performed on the K received pilot symbols _rp (n) for n=1,...,K to obtain the K received pilot values for k=1,...,K in the frequency domain ( block 814). The received pilot values can be given as follows:

R _p (k) = H (k) · P (k) + N (k), k = 1, ..., K, formula (13)

where P(k) is the transmitted pilot value corresponding to subband k;

H(k) is the complex gain of the communication channel corresponding to subband k;

_Rp (k) is the received pilot value corresponding to subband k; and

N(k) is the noise corresponding to subband k.

The K-point FFT provides K received pilot values corresponding to a total of K subbands. Only N received pilot values corresponding to the N subbands used for pilot transmission (referred to as pilot subbands) are retained, while the remaining KN received pilot values are discarded (block 816). IFDMA and LFDMA use different pilot subbands, so IFDMA and LFDMA reserve different received pilot values. The reserved pilot values are denoted as R _p (k), k=1,...,N. For simplicity, the noise can be assumed to be additive white Gaussian noise (AWGN) with mean zero and variance N ₀ .

The receiver can estimate the channel frequency response using various channel estimation techniques such as MMSE technique, least squares (LS) technique, and so on. The receiver derives channel gain estimates for the N pilot subbands based on the N received pilot values using MMSE or LS techniques (block 818). For the MMSE technique, an initial frequency response estimate of the communication channel can be derived based on the received pilot values as follows:

${\hat{h}}_{\mathrm{mmse}}\left(k\right)=\frac{R_p\left(k\right)\&CenterDot;P^{*}\left(k\right)}{{\left|P\left(k\right)\right|}^2+{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="S2006800121090E00041.png"\; state="\{\{state\}\}">N</figure-callout>}}_0},k=1,\&CenterDot;\&CenterDot;\&Center\; Dot;,N,$ Formula (14)

(k)是对应于子带k的信道增益估计，并且“*”表示复共轭。该初始频率响应估计包含对应于N个导频子带的N个信道增益。该导频码元序列可基于具有平坦频率响应的多相序列来生成。在此情形中，对于k的所有值皆有|P(k)|＝1，并且式(14)可被表达为：in

(k) is a channel gain estimate corresponding to subband k, and "*" indicates a complex conjugate. The initial frequency response estimate contains N channel gains corresponding to the N pilot subbands. The pilot symbol sequence may be generated based on a polyphase sequence with a flat frequency response. In this case, |P(k)| = 1 for all values of k, and equation (14) can be expressed as:

${\hat{h}}_{\mathrm{mmse}}\left(k\right)=\frac{R_p\left(k\right)\&Center\; Dot;P^{*}\left(k\right)}{1+{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="S2006800121090E00041.png"\; state="\{\{state\}\}">N</figure-callout>}}_0},k=1,\&CenterDot;\&CenterDot;\&Center\; Dot;,N,$ Formula (15)

The constant factor 1/(1+N ₀ ) can be removed to provide an unbiased MMSE frequency response estimate, which can be expressed as:

${\hat{h}}_{\mathrm{mmse}}\left(k\right)=R_p\left(k\right)\&Center\; Dot;P^{*}\left(k\right),k=1,\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;,N,$ Formula (16)

For the LS technique, an initial frequency response estimate can be derived based on the received pilot values as follows:

${\hat{h}}_{\mathrm{ls}}\left(k\right)=\frac{R_p\left(k\right)}{P\left(k\right)},,k=1,\&Center\; Dot;\&CenterDot;\&Center\; Dot;,N,$ Formula (17)

The impulse response of a communication channel may be characterized by L taps, where L may be much smaller than N. That is, if the transmitter applies an impulse to the communication channel, then L time-domain samples (at a sampling rate of BS MHz) will be sufficient to characterize the response of the communication channel based on this impulse stimulus. The number of taps (L) of the channel impulse response depends on the delay spread of the system, ie the time difference between the earliest and latest signal instances arriving at the receiver with sufficient energy. Longer delayed splays correspond to larger L values, and vice versa.

A channel impulse response estimate may be derived based on the N channel gain estimates and using LS or MMSE techniques (block 820). L taps with n=1,...,L can be derived based on initial frequency response estimates The least squares channel impulse response of (k) is estimated as follows:

${\underset{\&OverBar;}{\overset{^}{h}}}_{L\&times;1}^{\mathrm{ls}}={\left({\underset{\&OverBar;}{W}}_{N\&times;L}^h{\underset{\&OverBar;}{W}}_{N\&times;L}\right)}^{-1}{\underset{\&OverBar;}{W}}_{N\&times;L}^h{\underset{\&OverBar;}{\overset{^}{h}}}_{N\&times;1}^{\mathrm{init}},$ Formula (18)

是包含k＝1，…，N的

(k)或

(k)的N×1矢量；in

is the one that contains k=1,...,N

(k) or

N×1 vector of (k);

W _N×L is a sub-matrix of the Fourier matrix W _K×K ;

(k)的L×1矢量；以及 is the one containing n=1,...,L

an L x 1 vector of (k); and

" ^H " means conjugate transpose.

A Fourier matrix W _K×K is defined such that the (u,v)th element f _u,v is given as:

$f_{u,v}=e^{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S2006800121090E00011.png,S2006800121090E00021.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;}\frac{(u-1)(v-1)}{K}},$ u=1,...,K and v=1,...,K, formula (19)

包含最小二乘信道冲激响应估计的L个抽头。where u is the row index and v is the column index. W _NxL contains N rows corresponding to N pilot subbands in W _KxK . Each row of W _N×L contains the first L elements of the corresponding row of W _K×K .

Contains L taps of the least squares channel impulse response estimate.

(n)的MMSE信道冲激响应估计可基于初始频率响应估计推导如下：L taps with n=1,...,L

The MMSE channel impulse response estimate of (n) can be derived based on the initial frequency response estimate as follows:

${\underset{\&OverBar;}{\overset{^}{h}}}_{L\&times;1}^{\mathrm{mmse}}={({\underset{\&OverBar;}{W}}_{N\&times;L}^h{\underset{\&OverBar;}{W}}_{N\&times;L}+{\underset{\&OverBar;}{N}}_{L\&times;L})}^{-1}{\underset{\&OverBar;}{W}}_{N\&times;L}^h{\underset{\&OverBar;}{\overset{^}{h}}}_{N\&times;1}^{\mathrm{init}},$ Formula (20)

where N _{L × L} is the L × L noise and interference autocovariance matrix. For additive white Gaussian noise (AWGN), the autocovariance matrix can be given as ${\underset{\&OverBar;}{N}}_{L\&times;L}={\mathrm{\&sigma;}}_{\mathrm{no}}^2\&Center\; Dot;\underset{\&OverBar;}{I},$ where σ _n ² is the noise variance. An N-point IFFT may also be performed on the initial frequency response estimate to obtain a channel impulse response estimate with N taps.

Filtering and/or post-processing may be performed on the initial frequency response estimate and/or channel impulse response estimate to improve the quality of the channel estimate as described below (block 822). The final frequency response corresponding to all K subbands can be obtained by (1) zero padding the L-tap or N-tap channel impulse response estimate to length K, and (2) performing a K-point FFT on the expanded impulse response estimate Estimate (block 824). It can also be obtained by (1) interpolating the N channel gain estimates, (2) performing least squares approximation on the N channel gain estimates, or (3) using other approximation techniques to obtain the corresponding to all K subbands Final frequency response estimate.

The receiver can derive a longer channel impulse response estimate based on the interleaved FDM pilots. In general, a channel impulse response estimate with L _T taps may be obtained based on pilot IFDMA symbols transmitted on L _T different subbands in one or more symbol periods. For example, if L _T =2N, an impulse response with 2N taps can be obtained based on two or more pilot IFDMA symbols transmitted on two or more subband groups in two or more symbol periods estimate. If the pilot is sent on all S subband groups using a complete staggered pattern, then a full-length impulse response estimate with K taps can be obtained.

The receiver can derive a longer impulse response estimate of length _LT by filtering the initial impulse response estimate of length N for a sufficient number of different subband groups. Each initial impulse response estimate may be derived based on pilot IFDMA symbols corresponding to one subband group. If the pilot is transmitted on a different set of subbands in each symbol period, filtering can be performed over a sufficient number of symbol periods to obtain a longer impulse response estimate.

For SC-FDMA, filtering may be performed on initial frequency response estimates, least squares or MMSE channel impulse response estimates, and/or final frequency response estimates obtained for different symbol periods to improve the quality of the channel estimate. Filtering may be based on finite impulse response (FIR) filters, infinite impulse response (IIR) filters, or some other type of filter. The filter coefficients can be selected to achieve the desired amount of filtering, and the filter coefficients can be selected based on a trade-off between various factors such as desired quality of channel estimation, ability to track rapid changes in the channel, filtering complexity, etc. .

Other channel estimation techniques may also be used to obtain frequency response estimates and/or channel impulse response estimates for the communication channel.

Various post-processing can be performed to improve the quality of the channel estimate. In certain operating environments, such as multipath fading environments, communication channels often have only a small number of taps in the time domain. The channel estimation described above may provide a channel impulse response with a large number of taps due to noise. Post-processing attempts to remove taps due to noise and preserve taps due to the real channel.

In a post-processing scheme called truncation, only the first L taps of the channel impulse response estimate are kept and the remaining taps are replaced by zeros. In another post-processing scheme called thresholding, low energy taps are replaced with zeros. In one embodiment, thresholding is performed as follows:

n＝1，…，K，    式(21)

n=1,..., K, formula (21)

(n)是信道冲激响应估计的第n个抽头，它可以等于(n)或

(n)；并且in

(n) is the nth tap of the channel impulse response estimate, which can be equal to (n) or

(n); and

h _th is the threshold used to zero the low energy taps.

The threshold h _th can be calculated based on the energy of all K taps or only the energy of the first L taps of the channel impulse response estimate. The same threshold can be used for all taps. Alternatively, different thresholds can be used for different taps. For example, a first threshold may be used for the first L taps, and a second threshold (which may be lower than the first threshold) may be used for the remaining taps.

In yet another post-processing scheme called tap selection, the B best taps of the channel impulse response estimate are retained, where B > 1, and the remaining taps are set to zero. The number of taps to be reserved (denoted B) can be a fixed or variable value. B may be selected based on the received signal-to-noise-interference ratio (SNR) of the pilot/data transmission, the spectral efficiency of the data packets using channel estimation, and/or some other parameter. For example, if the received SNR falls within a first range (e.g., from 0 to 5 decibels (dB)), the two best taps may be reserved, and if the received SNR falls within a second range (e.g., from 5 to 10 dB) , then the three best taps can be kept, if the received SNR falls in the third range (eg, from 10 to 15 dB), then the four best taps can be kept, and so on.

For the TDM pilot scheme shown in Figure 5B, the CDM pilot scheme shown in Figure 5D, and other pilot schemes that transmit data and pilot symbols in the same SC-FDMA symbol, it is possible to Perform channel estimation. A rake estimator can be used to identify strong signal paths by, for example, (1) correlating received symbols with sequences of pilot symbols transmitted at different time offsets, and (2) identifying the time offset that provides the highest correlation result. Time-domain channel estimation provides a set of taps for an estimate of the channel impulse response of the communication channel.

For all pilot schemes, the channel estimate provides an equalized channel impulse response estimate and/or frequency response estimate that can be used for received data symbols. For the TDM pilot scheme shown in Figure 5A, a sequence of K received data symbols is obtained from each data SC-FDMA symbol, while for the CDM pilot scheme shown in Figure 5C, a sequence of M received data symbols is obtained from each group The received SC-FDMA symbols yield a sequence of K received data symbols. The sequence of K received data symbols may be equalized in the time or frequency domain.

Frequency domain equalization may be performed as follows. First, K-point FFT is performed on K received data symbols rd ₍ n) of n=1, . . . , K to obtain K frequency-domain received data values R _d (k) of k=1, . Only N received data values corresponding to the N subbands used for data transmission are kept, and the remaining KN received data values are discarded. The reserved data values are denoted as R _d (n), k=1, . . . , N.

Equalization may be performed on the N received data values in the frequency domain using MMSE techniques as follows:

$Z_d\left(k\right)=\frac{R_d\left(k\right)\&Center\; Dot;{\hat{h}}^{*}\left(k\right)}{{\left|\hat{h}\left(k\right)\right|}^2+{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="S2006800121090E00041.png"\; state="\{\{state\}\}">N</figure-callout>}}_0},k=1,\&Center\; Dot;\&Center\; Dot;\&CenterDot;,N$ Formula (22)

where _Rd (k) is the received data value corresponding to subband k;

(k)是对应于子带k的信道增益估计，它可以等于

(k)或

(k)；并且

(k) is the channel gain estimate corresponding to subband k, which can be equal to

(k) or

(k); and

Z _d (k) is the equalized data value for subband k.

Equalization can also be performed on the N received data values in the frequency domain using a zero-forcing technique as follows:

$Z_d\left(k\right)=\frac{R_d\left(k\right)}{\hat{h}\left(k\right)},k=1,\&Center\; Dot;\&CenterDot;\&Center\; Dot;,N,$ Formula (23)

(n)的序列，这些估计是对原始序列中的N个数据码元的估计。For both MMSE and zero-forcing equalization, the N equalized data values Z _d (k) for k=1,...,N can be transformed back to the time domain to obtain N Data symbol estimation

(n), these estimates are estimates for the N data symbols in the original sequence.

Equalization can also be performed in the time domain on a sequence of K received data symbols as follows:

Z _d (n) = r _d (n)g (n), formula (24)

where r _d (n) represents a sequence of K received data symbols;

g(n) represents the impulse response of the time domain equalizer;

z _d (n) represents a sequence of K equalized data symbols; and

 represents a circular convolution operation.

The frequency response of the equalizer can be derived based on the MMSE technique as follows: $G\left(k\right)=h^{*}\left(k\right)/({\left|\hat{h}\left(k\right)\right|}^2+N_0),$ k=1,...,N. The frequency response of the equalizer can also be derived based on the zero-forcing technique as follows: $G_k=1/\hat{h}\left(k\right),$ k=1,...,N. The equalizer frequency response can be transformed to the time domain to obtain the equalizer impulse response g(n), n=1, . . . , N, which is used for time domain equalization in equation (24).

The sequence of K equalized data symbols from equation (24) contains S copies of the transmitted data symbols. These S copies can be accumulated on a data-symbol-by-data-symbol basis to obtain the N data-symbol estimate as follows:

$\hat{d}\left(\mathrm{no}\right)=\underset{i=0}{\overset{S-1}{\mathrm{\&Sigma;}}}{z}_d(i\&Center\; Dot;N+\mathrm{no}),\mathrm{no}=1,\&CenterDot;\&Center\; Dot;\&CenterDot;,N.$ Formula (25)

Alternatively, no accumulation may be performed and only the N equalized data symbols corresponding to one copy of the transmitted data are provided as N data symbol estimates.

The receiver can also estimate interference based on received pilot values and channel estimates. For example, the interference corresponding to each subband can be estimated as follows:

$I\left(k\right)={|\hat{h}\left(k\right)\&CenterDot;P\left(k\right)-R_p\left(k\right)|}^2,k=1,\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;,N$ Formula (26)

where I(k) is the interference estimate corresponding to subband k. The interference estimate I(k) may be averaged over all N subbands for each SC-FDMA symbol to obtain a short-term interference estimate, which may be used for data demodulation and/or other purposes. Short-term interference estimates can be averaged over multiple SC-FDMA symbols to obtain long-term interference estimates, which can be used to estimate operating conditions and/or for other purposes.

Other techniques may also be used to improve the quality of channel estimates derived from TDM pilots or CDM pilots. These techniques include iterative channel estimation techniques and data-assisted channel estimation techniques.

(n)和初始信道估计

(n)来估计数据码元对导频码元的干扰为例如 $\hat{i}\left(n\right)=\hat{d}\left(n\right)\&CircleTimes;\hat{h}\left(n\right),$ 其中

(n)表示干扰估计。在另一个实施例中，处理数据码元估计以获得经解码的数据。然后以与在发射机处执行的相同方式处理经解码的数据以获得经重新调制的数据码元，将这些数据码元与初始信道估计卷积以获得干扰估计。对于这两个实施例，皆从接收的导频码元减去干扰估计以获得已消去干扰的导频码元 $r_p^{\mathrm{iC}}\left(n\right)=r_p\left(n\right)-\hat{i}\left(n\right),$ 该导频码元然后被用来推导改善的信道估计。然后重复此过程任意迭代次数以获得逐步优化的信道估计。此迭代信道估计技术更适合图5B中所示的TDM导频方案、图5C和5D中所示的CDM导频方案、以及数据码元可能对导频码元造成码元间干扰的其它导频方案。For iterative channel estimation techniques, an initial estimate of the communication channel is first derived based on received pilot symbols using, for example, MMSE or least squares techniques. This initial channel estimate is used to derive data symbol estimates as described above. In one embodiment, based on data symbol estimation

(n) and the initial channel estimate

(n) to estimate the data symbol to pilot symbol interference as e.g. $\hat{i}\left(\mathrm{no}\right)=\hat{d}\left(\mathrm{no}\right)\&CircleTimes;\hat{h}\left(\mathrm{no}\right),$ in

(n) represents interference estimation. In another embodiment, the data symbol estimates are processed to obtain decoded data. The decoded data is then processed in the same manner as performed at the transmitter to obtain remodulated data symbols, which are convolved with the initial channel estimate to obtain an interference estimate. For both embodiments, the interference estimate is subtracted from the received pilot symbols to obtain interference-canceled pilot symbols $r_p^{\mathrm{iC}}\left(\mathrm{no}\right)=r_p\left(\mathrm{no}\right)-\hat{i}\left(\mathrm{no}\right),$ The pilot symbols are then used to derive improved channel estimates. This process is then repeated for an arbitrary number of iterations to obtain progressively optimized channel estimates. This iterative channel estimation technique is more suitable for the TDM pilot scheme shown in Figure 5B, the CDM pilot scheme shown in Figures 5C and 5D, and other pilots where data symbols may cause intersymbol interference to pilot symbols plan.

(k)代入P(k)来获得。在另一个实施例中，处理该数据码元估计以获得经解码的数据，并处理该经解码的数据以获得经重新调制的数据码元D_rm(k)。该第二信道估计可通过在式(14)到(18)中将R_d(k)代入R_p(k)并将D_rm(k)代入P(k)来获得。For data-assisted channel estimation techniques, received data symbols are used for channel estimation along with received pilot symbols. A first channel estimate is derived based on the received pilot symbols and used to obtain data symbol estimates. A second channel estimate and a second symbol estimate are then derived based on the received data symbols. In one embodiment, the received data symbols r _d (n) are converted into received data values R _d (k) in the frequency domain, and the data symbols are estimated Convert to frequency domain data values (k). This second channel estimate can be obtained by substituting R _d (k) into R _p (k) in equations (14) to (18) and

(k) is substituted into P(k) to obtain. In another embodiment, the data symbol estimates are processed to obtain decoded data, and the decoded data are processed to obtain remodulated data symbols D _rm (k). The second channel estimate may be obtained by substituting _Rd (k) into _Rp (k) and _Drm (k) into P(k) in equations (14) to (18).

The two channel estimates obtained with the received pilot symbols and the received data symbols are combined to obtain an improved overall channel estimate. This combination can be performed, for example, as follows:

${\hat{h}}_{\mathrm{overall}}\left(k\right)={\hat{h}}_{\mathrm{pilot}}\left(k\right)\&Center\; Dot;C_p\left(k\right)+{\hat{h}}_{\mathrm{data}}\left(k\right)\&CenterDot;C_d\left(k\right),k=1,\&Center\; Dot;\&Center\; Dot;\&CenterDot;,N,$ Formula (27)

(k)是基于所获得的导频码元而获得的信道估计；in

(k) is a channel estimate obtained based on the obtained pilot symbols;

(k)是基于接收的数据码元而获得的信道估计；

(k) is a channel estimate obtained based on received data symbols;

C _p (k) and C _d (k) are weighting factors corresponding to pilot and data, respectively; and

(k)是总信道估计。

(k) is the total channel estimate.

(k)可基于

(k)、(k)、数据码元估计可靠性的置信度、和/或任何其它因数的任何函数来推导。以上所描述的处理可用迭代方式执行。对于每次迭代，基于从数据码元估计获得的信道估计更新

(k)，并使用更新后的

(k)来推导新的数据码元估计。数据辅助信道估计技术可用于所有导频方案，包括图5A到5D中所示的TDM和CDM导频方案。Generally speaking,

(k) can be based on

(k), (k), confidence in the reliability of the data symbol estimate, and/or any other factors are derived from any function. The processing described above may be performed in an iterative manner. For each iteration, the channel estimate is updated based on the channel estimate obtained from the data symbol estimates

(k), and use the updated

(k) to derive new data symbol estimates. The data-assisted channel estimation technique can be used for all pilot schemes, including the TDM and CDM pilot schemes shown in Figures 5A through 5D.

FIG. 9 shows a block diagram of a transmitter 910 and a receiver 950 . For the forward link, transmitter 910 is part of the base station and receiver 950 is part of the wireless device. For the reverse link, transmitter 910 is part of the wireless device and receiver 950 is part of the base station. A base station is typically a fixed station and may also be called a base transceiver system (BTS), access point, or some other terminology. A wireless device may be fixed or mobile and may also be called a user terminal, mobile station, or some other term.

At transmitter 910, a TX data and pilot processor 920 processes the traffic data to obtain data symbols, generates pilot symbols, and provides these data symbols and pilot symbols. An SC-FDMA modulator 930 multiplexes data symbols with pilot symbols using TDM and/or CDM and performs SC-FDMA modulation (eg, for IFDMA, LFDMA, etc.) to generate SC-FDMA symbols. A transmitter unit (TMTR) 932 processes (eg, converts to analog, amplifies, filters, and frequency upconverts) the SC-FDMA symbols and generates a radio frequency (RF) modulated signal, which is transmitted via antenna 934 .

At receiver 950, an antenna 952 receives the transmitted signal and provides a received signal. A receiver unit (RCVR) 954 conditions (eg, filters, amplifies, downconverts, and digitizes) the received signal to generate a stream of received samples. An SC-FDMA demodulator 960 processes the received samples and obtains received data symbols and received pilot symbols. A channel estimator/processor 980 derives a channel estimate based on the received pilot symbols. SC-FDMA demodulator 960 performs equalization on received data symbols with the channel estimates and provides data symbol estimates. A receive (RX) data processor 970 performs demapping, deinterleaving and decoding on the data symbol estimates and provides decoded data. In general, the processing by SC-FDMA demodulator 960 and RX data processor 970 is complementary to that performed at transmitter 910 by SC-FDMA modulator 930 and TX data and pilot processor 920, respectively.

Controllers 940 and 990 direct the operation of various processing units at transmitter 910 and receiver 950, respectively. Memory units 942 and 992 store program codes and data used by controllers 940 and 990, respectively.

Figure 10A shows a block diagram of TX data and pilot processor 920a, which is an embodiment of processor 920 in Figure 9 and may be used for the TDM pilot scheme. Within processor 920a, traffic data is encoded by encoder 1012, interleaved by interleaver 1014, and mapped into data symbols by symbol mapper 1016. Pilot generator 1020 generates pilot symbols based on, for example, a polyphase sequence. A multiplexer (Mux) 1022 receives the data symbols and pilot symbols and multiplexes the data symbols with pilot symbols based on TDM control and provides a multiplexed stream of data and pilot symbols.

FIG. 10B shows a block diagram of TX data and pilot processor 920b, which is another embodiment of processor 920 in FIG. 9 and may be used for the CDM pilot scheme. Within processor 920b, traffic data is encoded by encoder 1012, interleaved by interleaver 1014, and mapped into data symbols by symbol mapper 1016. A multiplier 1024a multiplies each data symbol by the M chips corresponding to the orthogonal sequence of data { _wd } and provides M scaled data symbols. Similarly, a multiplier 1024b multiplies each pilot symbol by the M chips corresponding to the pilot's orthogonal sequence { _wp } and provides M scaled pilot symbols. An adder 1026 adds the scaled data symbols and the scaled pilot symbols and provides combined symbols, eg, as shown in FIG. 5C or 5D.

FIG. 11A shows a SC-FDMA modulator 930a corresponding to IFDMA, which is an embodiment of the SC-FDMA modulator 930 in FIG. 9 . Within modulator 930a, a repetition unit 1112 repeats the original sequence of data/pilot symbols S times to obtain a spreading sequence of K symbols. A phase ramp unit 1114 applies a phase ramp to the spreading symbol sequence to generate a frequency shifted output symbol sequence. The phase slope is determined by the subband u used for transmission. A cyclic prefix generator 1116 adds a cyclic prefix to the frequency-shifted sequence of symbols to generate an IFDMA symbol.

FIG. 11B shows a SC-FDMA modulator corresponding to LFDMA, which is another embodiment of the SC-FDMA modulator 930 in FIG. 9 . Within modulator 930b, an FFT unit 1122 performs an N-point FFT on the original sequence of data/pilot symbols to obtain a sequence of N frequency-domain symbols. A symbol-to-subband mapper 1124 maps the N frequency-domain symbols onto N subbands for transmission and maps the K-N zero symbols onto the remaining K-N subbands. An IFFT unit 1126 performs a K-point IFFT on the K symbols from mapper 1124 and is provided with a sequence of K time-domain output symbols. A cyclic prefix generator 1128 adds a cyclic prefix to the output sequence of symbols to generate an LFDMA symbol.

FIG. 12A shows a block diagram of SC-FDMA demodulator 960a, which is an embodiment of demodulator 960 in FIG. 9 and can be used for the TDM IFDMA pilot scheme. Within SC-FDMA demodulator 960a, a cyclic prefix removal unit 1212 removes the cyclic prefix for each received IFDMA symbol. Phase slope removal unit 1214 removes the phase slope in each received IFDMA symbol. Phase slope removal can also be performed by down-conversion from RF to baseband. A demultiplexer (Demux) 1220 receives the output of unit 1214 and provides received data symbols to equalizer 1230 and received pilot symbols to channel estimator 980 . A channel estimator 980 derives a channel estimate based on the received pilot symbols using, for example, MMSE or least squares techniques. An equalizer 1230 performs equalization on received data symbols with the channel estimate in the time or frequency domain and provides equalized data symbols. An accumulator 1232 accumulates equalized data symbols corresponding to multiple copies of the same transmitted data symbol and provides a data symbol estimate.

FIG. 12B shows a block diagram of SC-FDMA demodulator 960b, which is another embodiment of demodulator 960 in FIG. 9 and can be used for the CDM IFDMA pilot scheme. SC-FDMA demodulator 960b includes a data channelizer for recovering transmitted data symbols, and a pilot channelizer for recovering transmitted pilot symbols. For the data channelizer, a multiplier 1224a multiplies the output of unit 1214 by the M chips of the data orthogonal sequence { _wd } and provides scaled data symbols. An accumulator 1226 accumulates the M scaled data symbols for each transmitted data symbol and provides received data symbols. For the pilot channelizer, multiplier 1224b multiplies the output of unit 1214 by the M chips of the pilot orthogonal sequence {w _p } and provides M defined chips corresponding to each transmitted pilot symbol The target pilot symbols are accumulated by the accumulator 1226b to obtain the received pilot symbols corresponding to the transmitted pilot symbols. The processing performed by subsequent units within SC-FDMA demodulator 960b is the same as described above for SC-FDMA demodulator 960.

FIG. 13A shows a block diagram of SC-FDMA demodulator 960c, which is yet another embodiment of demodulator 960 in FIG. 9 and can be used for the TDM LFDMA pilot scheme. Within SC-FDMA demodulator 960c, a cyclic prefix removal unit 1312 removes the cyclic prefix for each received LFDMA symbol. The FFT unit 1314 performs a K-point FFT on the LFDMA symbols after removing the cyclic prefix, and provides K frequency domain values. A subband-to-symbol demapper 1316 receives the K frequency domain values, provides N frequency domain values corresponding to the N subbands used for transmission, and discards the rest. An IFFT unit 1318 performs an N-point FFT on the N frequency-domain values from demapper 1316 and provides N received symbols. Demultiplexer 1320 receives the output of element 1318 , provides received data symbols to equalizer 1330 and provides received pilot symbols to channel estimator 980 . An equalizer 1330 performs equalization on received data symbols in the time or frequency domain using the channel estimates from channel estimator 980 and provides data symbol estimates.

FIG. 13B shows a block diagram of SC-FDMA demodulator 96Od, which is yet another embodiment of demodulator 960 in FIG. 9 and can be used for the CDM LFDMA pilot scheme. SC-FDMA demodulator 960d includes a data channelizer for recovering transmitted data symbols and a pilot channelizer for recovering transmitted pilot symbols. For the data channelizer, a multiplier 1324a multiplies the output of IFFT unit 1318 by the M chips of the data orthogonal sequence { _wd } and provides scaled data symbols. An accumulator 1326a accumulates the M scaled data symbols for each transmitted data symbol and provides received data symbols. For the pilot channelizer, multiplier 1324b multiplies the output of IFFT unit 1318 by the M chips of the pilot orthogonal sequence {w _p } and provides M chips corresponding to each transmitted pilot symbol The scaled pilot symbols are accumulated by accumulator 1326b to obtain received pilot symbols corresponding to the transmitted pilot symbols. The processing by subsequent units within SC-FDMA demodulator 960d is the same as described above for SC-FDMA demodulator 960c.

The pilot transmission and channel estimation techniques described herein may be implemented by various means. For example, these techniques may be implemented in hardware, software, or a combination thereof. For a hardware implementation, each processing unit at the transmitter for generating and sending the pilot (e.g., each of the processing units shown in FIGS. (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), processor, controller, microcontroller, microprocessor , electronic devices, other electronic units designed to perform the functions described herein, or a combination thereof. The various processing units used at the receiver to perform channel estimation may also be implemented within one or more ASICs, DSPs, electronics, or the like.

For a software implementation, the techniques can be implemented with modules (eg, procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory unit (eg, memory unit 942 or 992 in FIG. 9 ) and executed by a processor (eg, controller 940 or 990 ). The memory unit can be implemented within the processor or external to the processor.

The above description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (59)

1. A device comprising: a processor for forming a first sequence of pilot symbols based on a polyphase sequence; and A modulator for replicating the first sequence of pilot symbols a plurality of times to obtain a second sequence of pilot symbols for transmission via a communication channel. 2. The apparatus of claim 1, wherein the modulator is configured to apply a phase ramp to the second pilot symbols to obtain a third sequence of output symbols. 3. The apparatus according to claim 2, wherein the modulator is further configured to use at least two different phase slopes for the phase slopes in at least two different symbol periods to obtain different The first pilot symbol sequence is transmitted on the frequency subband of . 4. The apparatus of claim 1, wherein the modulator is configured to add a cyclic prefix to the second sequence of pilot symbols to obtain a third sequence suitable for transmission in the time domain via the communication channel. Output a sequence of symbols. 5. The apparatus of claim 1, wherein the polyphase sequence has a constant envelope in the time domain and a flat spectral response in the frequency domain. 6. The apparatus of claim 1, wherein the processor is configured to form a first data symbol sequence, multiplex the first data symbol sequence in a first symbol period, and The first pilot symbol sequence is multiplexed in a symbol period. 7. The apparatus of claim 1, wherein the processor is configured to form a first data symbol sequence, multiplex the first pilot symbol sequence with the first data symbol sequence, And a multiplexed sequence of data and pilot symbols is provided. 8. The apparatus of claim 1, wherein the processor is configured to form a first sequence of data symbols, multiply the first sequence of data symbols with a first orthogonal sequence to obtain a plurality of A scaled data symbol sequence, multiplying the first pilot symbol sequence by a second orthogonal sequence to obtain a plurality of scaled pilot symbol sequences, and multiplying the plurality of scaled A data symbol sequence is combined with the plurality of scaled pilot symbol sequences to obtain a plurality of combined symbol sequences. 9. The apparatus of claim 1, wherein the processor is configured to form a first sequence of data symbols, multiply the first sequence of data symbols with a first orthogonal sequence to obtain a scaled data symbol sequence, multiply the first pilot symbol sequence with the second orthogonal sequence to obtain a scaled pilot symbol sequence, and combine the scaled data symbol sequence with the The scaled pilot symbol sequences are combined to obtain a combined symbol sequence. 10. The apparatus of claim 1, wherein the first pilot symbol sequence is sent on a first set of frequency subbands, and data symbols are sent on a second set of frequency subbands, the second The set of frequency subbands contains more frequency subbands than the first set of frequency subbands. 11. A method of generating a pilot in a communication system, comprising: forming a first sequence of pilot symbols based on a polyphase sequence; and The first sequence of pilot symbols is replicated a plurality of times to obtain a second sequence of pilot symbols for transmission via a communication channel. 12. The method of claim 11, further comprising: A phase ramp is applied to the second sequence of pilot symbols to obtain a third sequence of output symbols. 13. The method of claim 11, further comprising: adding a cyclic prefix to the second sequence of pilot symbols to obtain a third sequence of output symbols; and The third sequence of output symbols is transmitted in the time domain via the communication channel. 14. A device comprising: means for generating a first sequence of pilot symbols based on a polyphase sequence; and Means for replicating said first sequence of pilot symbols a plurality of times to obtain a second sequence of pilot symbols suitable for transmission via a communication channel. 15. The apparatus of claim 14, further comprising: Means for applying a phase ramp to said second sequence of pilot symbols to obtain a third sequence of output symbols. 16. The apparatus of claim 14, further comprising: means for adding a cyclic prefix to said second sequence of pilot symbols to obtain a third sequence of output symbols; and Means for transmitting the third sequence of output symbols in the time domain via the communication channel. 17. A device comprising: a processor for forming a first sequence of pilot symbols based on a polyphase sequence; and a modulator for transforming the first sequence of pilot symbols into the frequency domain to obtain a second sequence of frequency-domain symbols by mapping the second sequence of frequency-domain symbols to a frequency for pilot transmission The subband groups are used to form a third sequence of symbols, and the third sequence of symbols is transformed to the time domain to obtain a fourth output sequence of symbols for transmission via the communication channel. 18. The apparatus of claim 17, wherein the modulator is configured to add a cyclic prefix to the fourth sequence of pilot symbols to obtain a fifth output suitable for transmission in the time domain via the communication channel code element sequence. 19. The apparatus of claim 17, wherein the polyphase sequence has a constant envelope in the time domain and a flat spectral response in the frequency domain. 20. The apparatus of claim 17 , wherein data symbols are transmitted on a second frequency subband group, the second frequency subband group comprising more frequency subband groups than the frequency subband group used for pilot transmission frequency subband. 21. A device comprising: a processor configured to form a sequence of pilot symbols, form a sequence of data symbols, and time-multiplex the sequence of data symbols with the sequence of pilot symbols; and a modulator for generating at least one single carrier frequency division multiple access (SC-FDMA) symbol based on the time division multiplexed data symbols and pilot symbols. 22. The apparatus of claim 21 , wherein the processor is configured to multiplex the sequence of data symbols in a first symbol period and multiplex the pilot sequence in a second symbol period symbol sequence, and the modulator generates a first SC_FDMA for a data symbol sequence in the first symbol period, and generates a first SC_FDMA for a pilot symbol sequence in the second symbol period Two SC-FDMA symbols. 23. The apparatus of claim 21, wherein the processor is configured to multiplex the data symbol sequence with the pilot symbol sequence in different sampling periods of a symbol period, and the The modulator is used to generate SC-FDMA symbols of multiplexed pilot and data symbols corresponding to the symbol period. 24. A device comprising: means for forming a sequence of pilot symbols; means for forming a sequence of data symbols; means for time division multiplexing the sequence of data symbols with the sequence of pilot symbols; and Means for generating at least one single carrier frequency division multiple access (SC-FDMA) symbol based on the time division multiplexed data symbols and pilot symbols. 25. The device according to claim 24, wherein the device for time-division multiplexing the data symbol sequence and the pilot symbol sequence comprises: means for multiplexing said sequence of data symbols in a first symbol period, and means for multiplexing the sequence of pilot symbols in a second symbol period. 26. The device according to claim 24, wherein the device for time-division multiplexing the data symbol sequence and the pilot symbol sequence comprises: Means for multiplexing said sequence of data symbols with said sequence of pilot symbols in different sampling periods of a symbol period. 27. A device comprising: a processor for forming a sequence of pilot symbols and forming a sequence of data symbols; and a modulator for generating a broadband pilot based on the pilot symbols, generating at least one single-carrier frequency division multiple access (SC-FDMA) symbol based on the sequence of data symbols, and combining the broadband pilot with The at least one SC-FDMA symbol is time division multiplexed. 28. The apparatus of claim 27, wherein the processor is to form the sequence of pilot symbols based on a pseudorandom number (PN) sequence. 29. The apparatus of claim 27, wherein the modulator is configured to generate at least one interleaved FDMA (IFDMA) symbol or at least one localized FDMA (LFDMA) symbol corresponding to the sequence of data symbols . 30. The apparatus of claim 27, wherein the wideband pilot is pseudorandom with respect to at least one other wideband pilot from at least one other transmitter. 31. The apparatus of claim 27, wherein the wideband pilot is time aligned with at least one other wideband pilot from at least one other transmitter. 32. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via a communication channel, and processing the at least one SC-FDMA symbol to obtain time-domain received pilot symbols; and a processor for transforming the received pilot symbols to obtain frequency-domain pilot values, and based on the frequency-domain pilot values and using a minimum mean square error (MMSE) technique or a least squares (LS) technique to determine A frequency response estimate corresponding to the communication channel is derived. 33. The apparatus of claim 32, wherein the processor is to derive a channel impulse response estimate corresponding to the communication channel based on the frequency response estimate. 34. The apparatus of claim 32, wherein the processor is configured to filter the frequency response estimate. 35. The apparatus of claim 33, wherein the processor is configured to filter the channel impulse response estimate. 36. The apparatus of claim 32, wherein the processor is configured to derive frequency response estimates for SC-FDMA symbols transmitted on at least two groups of frequency subbands, and to derive channel and filtering said respective channel impulse response estimates to obtain extended channel impulse response estimates having more taps than each of said channel impulse response estimates. 37. The apparatus of claim 33, wherein the processor is configured to reserve a predetermined number of taps in the channel impulse response estimate and set the remaining taps in the channel impulse response estimate to zero . 38. The apparatus of claim 37, wherein the processor is to select the predetermined number of taps based on a signal-to-noise-interference ratio (SNR) or spectral efficiency of data transmission over the communication channel. 39. The apparatus according to claim 33, wherein the processor is configured to retain taps exceeding a predetermined threshold in the channel impulse response estimate, and set the rest of the taps in the channel impulse response estimate to zero. 40. The apparatus of claim 33, wherein the processor is configured to retain the first L taps in the channel impulse response estimate, and set the remaining taps in the channel impulse response estimate to Zero, where L is an integer equal to or greater than 1. 41. The apparatus of claim 32, wherein the demodulator is configured to demultiplex received symbols in the at least one SC-FDMA symbol into received data symbols and received pilots symbol. 42. The apparatus of claim 32, wherein the demodulator is configured to process the at least one SC-FDMA symbol with an orthogonal sequence corresponding to a pilot to obtain the received pilot code Yuan. 43. The apparatus of claim 32, further comprising: an equalizer for equalizing received data symbols based on the frequency response estimate. 44. A device comprising: means for processing at least one single carrier frequency division multiple access (SC-FDMA) symbol received via a communication channel to obtain received pilot symbols; means for transforming said received pilot symbols to obtain frequency domain pilot values; and Means for deriving a frequency response estimate corresponding to the communication channel based on the frequency domain pilot values and using a minimum mean square error (MMSE) technique or a least squares (LS) technique. 45. The apparatus of claim 44, further comprising: means for deriving a channel impulse response estimate corresponding to the communication channel based on the frequency response estimate, and means for setting at least one tap of said channel impulse response estimate to zero. 46. The apparatus of claim 44, further comprising: Means for filtering at least two frequency response estimates derived from at least two SC-FDMA symbols corresponding to at least two symbol periods. 47. A device comprising: a demodulator for receiving a pilot and processing the received pilot to obtain received pilot symbols, the pilot consisting of multiple copies of a sequence of pilot symbols generated using a polyphase sequence; as well as A processor for accumulating received pilot symbols corresponding to the plurality of copies of the sequence of pilot symbols. 48. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via the communication channel and time-division demultiplexing the received symbols in the at least one SC-FDMA symbol into received data symbols and received pilot symbols; and A processor for deriving a channel estimate corresponding to the communication channel based on the received pilot symbols. 49. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via a communication channel and processing the at least one SC-FDMA symbol to obtain received pilot symbols; and A processor for deriving a channel estimate corresponding to the communication channel based on the received pilot symbols and using least squares (LS) techniques. 50. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via a communication channel, and processing the at least one SC-FDMA symbol to obtain time-domain received pilot symbols; and A processor for identifying at least one tap corresponding to a channel impulse response estimate of the communication channel by correlating the received pilot symbols with transmitted pilot symbols at different time offsets. 51. The apparatus of claim 50, wherein the at least one SC-FDMA symbol comprises pilot symbols and data symbols multiplexed over multiple sampling periods, and the demodulator is configured to The received pilot symbols and received data symbols in the at least one SC-FDMA symbol are demultiplexed. 52. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via the communication channel, processing the at least one SC-FDMA symbol to obtain received pilot symbols and received data symbols , and process the received data symbols with a first channel estimate corresponding to the communication channel to obtain data symbol estimates; and a first processor for deriving the first channel estimate based on the received pilot symbols, estimating a signal due to the received data symbols based on the first channel estimate and the data symbol estimate interference, deriving interference-canceled pilot symbols based on the received pilot symbols and the estimated interference, and deriving a second channel estimate based on the interference-canceled pilot symbols. 53. The apparatus of claim 52, further comprising: a second processor for processing the data symbol estimates to obtain decoded data and for processing the decoded data to obtain remodulated data symbols, and the first processor for processing the data symbols based on the The remodulated data symbols are used to estimate the interference. 54. The apparatus of claim 52, wherein the demodulator and the first processor are configured to iteratively: derive the data symbol estimates, estimate the interference, derive the interference-canceled pilot symbols, and deriving the second channel estimate. 55. A device comprising: a demodulator for receiving at least one single carrier frequency division multiple access (SC-FDMA) symbol via the communication channel and processing the at least one SC-FDMA symbol to obtain received pilot symbols and received data symbols yuan; and A first processor configured to derive a first channel estimate corresponding to the communication channel based on the received pilot symbols, derive a second channel estimate based on the received data symbols, and based on the first and The second channel estimate derives a third channel estimate. 56. The apparatus of claim 55, wherein the demodulator is to process the received data symbols with the first channel estimate to obtain data symbol estimates. 57. The apparatus of claim 56, wherein the first processor is to derive the second channel estimate based on the received data symbols and the data symbol estimates. 58. The apparatus of claim 56, further comprising: a second processor for processing the data symbol estimates to obtain decoded data and for processing the decoded data to obtain remodulated data symbols, and the first processor for processing the data symbols based on the The second channel estimate is derived from the received data symbols and the remodulated data symbols. 59. The apparatus of claim 55, wherein the first processor is configured to base a confidence level on the reliability of the first channel estimate, the second channel estimate, and the data symbol estimate The third channel estimate is derived as a function of the indication of .

Images