Chapter 2 Overview of IEEE 802.16e OFDMA and MIMO Systems
2.4 IEEE 802.16e OFDMA Specification
2.4.4 Basic Specification in IEEE 802.16e OFDMA
IEEE 802.16e OFDMA specification defines different transmission types according to different purposes and applications. The modulation schemes of QPSK, 16-QAM, and 64-QAM are supported for data subcarrier. The transmission types are constructed by different modulation schemes with different code rates such as QPSK 1/2, QPSK 3/4, 16-QAM 1/2, 16-QAM 3/4, 64-QAM 1/2, 64-QAM 2/3, and 64-QAM 3/4. The major parameters can be derived and described in Table 2.2.
TABLE 2.2
MAJOR PARAMETERS OF IEEE802.16E OFDMA SPECIFICATION
Parameters Deriving formulas
FFT Size (N) 2048, 1024, 512, and 128
Channel Bandwidth (BW) 1.25-20MHz
Sampling Factor (n)
n=28/25 if BW is a multiple of 1.25, 1.5, 2, and 2.75 MHz
n=8/7 for the other cases Ratio of CP to Useful Symbol Time (G) 1/32, 1/16, 1/8, and 1/4 Sampling Frequency (FS) floor n BW( × 8000)´8000
Sampling Time T N b
Subcarrier Spacing (Δf) F N S
Useful Symbol Time (Tb) 1 fD
Guard Time (Tg) G T× b
OFDMA Symbol Duration (TS) Tb +Tg
Frame Duration (TF)
Number of OFDMA Symbols floor T T ( F S)
Number of Null Subcarriers (NNull)
Number of Clusters (NC) (N N- Null) 14
Number of Sub-channels (NSC) NC 2 Number of Pilot Subcarriers (NPilot) NC´ 2 DL
PUSC
Number of Data Subcarriers (NData) NC´ 12
Chapter 3
Downlink Baseband STBC-OFDM System Architecture
3.1 Introduction
In this chapter, the proposed downlink baseband STBC-OFDM system architecture will be described. This architecture can provide high transmission rate in IEEE 802.16e downlink communication as an alternative solution for WMAN in fixed and mobile wireless communication.
Recently, STBC-OFDM systems have received a lot of attention [28], [29] and are also adopted in IEEE 802.16e systems. Although STBC-OFDM systems with multiple antennas can provide diversity gains to improve transmission efficiency and quality of mobile wireless systems, accurate CSI is required for diversity combining, coherent detection, and decoding. Moreover, the system performance is also susceptible to the synchronization error. Therefore, synchronization and channel estimation are two crucial challenges for realizing a successful STBC-OFDM system.
Hence, a downlink baseband receiver scheme for STBC-OFDM systems is proposed and can be applied in IEEE 802.16e specification. In the proposed receiver, two main tasks, synchronization and channel estimation, are implemented. The synchronization includes symbol timing detection and carrier frequency recovery. A novel match filter is proposed to precisely detect symbol boundary, and a ping-pong algorithm is presented to effectively improve the performance of carrier frequency
synchronization [30]. Moreover, a refined two-stage channel estimation method with an initialization stage and a tracking stage is adopted to accurately estimate CSI over doubly selective fading channels [31], [32]. The initialization stage uses discrete Fourier transform (DFT)-based channel estimation method [33], [34] with the multipath interference cancellation (MPIC)-based decorrelation scheme to identify significant channel paths of the primary channel impulse responses. Then, the tracking stage uses decision-feedback (DF) DFT-based channel estimation with Newton’s method [35], [36] to track the path variations of these paths.
The rest of this chapter is organized as follows. In Section 3.2, we describe the system specification of the proposed STBC-OFDM system with two transmit antennas and one receive antenna. In Section 3.3, we delineate the transmitter architecture. We then describe the proposed receiver in Section 3.4. The system simulation results are presented in Section 3.5. Finally, the conclusions of this chapter are drawn in Section 3.6.
3.2 System Specification and Design Targets
IEEE 802.16e includes multiple PHY specifications such as, SC, SCa, OFDM, and OFDMA, for providing different channel conditions and applications. The OFDMA specification that supports multi-antenna technology is adopted in the proposed STBC-OFDM system. In downlink transmission, the distributed subcarrier allocation of PUSC is supported, and the major parameters and the design targets of the proposed STBC-OFDM system are summarized in Table 3.1.
The system occupies a bandwidth of 10 MHz and operates in 2.5 GHz frequency band. The sampling frequency is 11.2 MHz. FFT size (N) is set to 1024. Each OFDM symbol is composed of 1024 subcarriers among which 720 and 120 subcarriers are used to transmit data symbols and pilots, and the remaining 184 subcarriers are used as either a DC subcarrier or virtual subcarriers. In IEEE 802.16e, the modulation schemes of QPSK, 16-QAM, and 64-QAM are supported for data subcarriers, while BPSK is adopted for pilot subcarriers and preamble symbol. In the hardware design, the data subcarrier modulation schemes target to support QPSK and 16-QAM. The length of CP is 128 sampling periods, i.e., 1/8 of the useful symbol time (Tb). Fig. 3.1 depicts the frame format which starts with one preamble symbol and is followed by
40 consecutive OFDM data symbols, and a time slot is equivalent to the duration of two OFDM symbols. Based on the design targets, the proposed STBC-OFDM system is defined to support the frequency offset to ±7 ppm variation in transmitter and receiver which is equivalent to maximum CFO of 35 KHz in 2.5 GHz carrier frequency. Moreover, the proposed STBC-OFDM system is optimized to enable vehicle speed up to 120 km/hr. The maximum Doppler frequency fd is about 0.025 (normalized to subcarrier spacing). The coherence time Tc calculating by a typical way [37] is 0.423/fd=1.5 ms which is about 14.8 times of an OFDM symbol period
TABLE 3.1
MAJOR PARAMETERS AND DESIGN TARGETS OF THE PROPOSED STBC-OFDM
SYSTEM
Number of Null Subcarriers (NNULL) 184 Number of Pilot Subcarriers (NPilot) 120 PUSC
Number of Data Subcarriers (NData) 720 Design Targets
CFO ±7 ppm
Vehicle Speed Up to 120 km/hr
Modulation Scheme QSPK, 16-QAM
Uncoded Data Rate 27.32 Mbps
and is small than a frame period. Therefore, the channel can be treated as quasi-static in several symbol times but may vary quickly during one frame transmission. As the vehicle speed further increases, ICI effect caused by Doppler spread will become more significant and cannot be ignored. When fd is larger than 0.08 of the subcarrier spacing, the signal-to-ICI plus noise ratio is less than 20 dB [38]. Among various ICI cancellation methods, an important requirement is to accurately estimate channel variation for calculating ICI channel matrix [39]-[41]. The proposed system does not include the ICI cancellation block. However, a robust two-stage channel estimation method is adopted to precisely track channel gain variations. It is conducive to the future development of ICI cancellation to apply in higher mobile environment.
The proposed STBC-OFDM system with two transmit antennas and one receive antenna is shown in Fig. 3.2. The detail descriptions of the transmitter and receiver will be presented in the following sections.
3.3 Transmitter Architecture
As shown in Fig. 3.2 (a), in the transmitter, the transmitted data are first randomized to decrease the peak to average power ratio (PAPR) in OFDM transmission. The data are then channel encoded to resist channel effects and interleaved to avoid burst errors. Fig. 3.3 depicts the block diagrams of channel coder.
After channel coding, transmitted data pass through signal mapper and then go through serial-to-parallel (S/P) to form two transmitted OFDM symbols
XF = {XF[k]: 0 ≤ k ≤ N-1} (3.1) XS = {XS[k]: 0 ≤ k ≤ N-1} (3.2)
Fig. 3.1 Frame format.
Fig. 3.2 (a) Transmitter and (b) receiver of the proposed STBC-OFDMA system with two transmit antennas and one receive antenna.
within a time slot. Alamouti’s STBC encoding method is adopted to encode these two transmitted OFDM symbols [16]. An N-point inverse fast Fourier transform (IFFT) unit is used in each arm to transform the frequency domain OFDM symbols into time domain. After the parallel-to-serial (P/S) converters, the CP with time duration Tg is then inserted as a guard interval to combat ISI. A complete OFDM symbol with symbol duration Tb+Tg is converted into an analog signal with a digital-to-analog (D/A) converter. Finally, the analog signals are filtered by a low-pass filter (LPF), up converted to RF band, and transmitted in air.
3.3.1 Channel Coding
Channel coding helps the transmitted data to combat channel effects and avoid burst errors. As shown in Fig. 3.3, the procedure of channel coding in IEEE 802.16e includes randomization (Section 3.3.1.1), forward error correction (FEC) coding (Section 3.3.1.2), interleaving (Section 3.3.1.3), and repetition coding. Repetition coding can provide performance improvement for protecting the important control message transmission, e.g. FCH or DL-MAP. After FEC coding and interleaving, the data bits are segmented into slots, and a slot applied by repetition coding will be repeated R times to form R contiguous slots following the normal slot ordering that is used for data mapping, where R = 2, 4, or 6 is the repetition code factor. This repetition scheme applies only to QPSK modulation. The other functional blocks will be discussed in the following subsections.
Fig. 3.3 Block diagrams of channel coder.
Fig. 3.4 PRBS generator for data randomization.
3.3.1.1 Randomization
Data randomization shall be performed on each data burst of the downlink and uplink, except the FCH and preamble. A randomizer is used to scramble the transmitted data and consists of a PRBS generator with the polynomial, 1 x+ 14+x15, as shown in Fig. 3.4. The randomizer is initialized with initialization sequence “011 0111 0001 0101”. The bit stream sequentially enters the randomizer, started from MSB, to generate information bits.
3.3.1.2 FEC coding
There are several FEC coding methods are provided in IEEE 802.16e OFDMA specification such as convolutional codes (CC), block turbo codes (BTC), convolutional turbo codes (CTC) and low density parity check codes (LDPC). The mandatory coding scheme is the tail-biting convolutional coding, and the others are optional modes. Fig. 3.5 shows the CC encoder which has a code rate of 1/2 and a constraint length k =7 and uses the following generator polynomials to derive its two coded bits X and Y.
G1=(171)OCT for X (3.3)
TABLE 3.2
PUNCTURE CONFIGURATION
Code Rates 1/2 2/3 3/4
X 1 10 101
Y 1 11 110
Output X1Y1 X1Y1Y2 X1Y1Y2X3
Fig. 3.5 Convolutional encoder of code rate 1/2.
G2=(133)OCT for Y (3.4)
Table 3.2 defines the puncturing patterns and serialization order that shall be used to realize different code rates. In the table, “1” means a transmitted bit and “0”
denotes a removed bit.
3.3.1.3 Interleaving
Interleaving is used to combat the burst error and improve the bit error rate of FEC coding. All coded data bits shall be interleaved by a block interleaver with a block size corresponding to the number of coded bits per the encoded block size Ncbps.
A two-step permutation is adopted. The first step maps adjacent coded bits onto nonadjacent subcarriers as defined by (3.5). The second step alternately maps adjacent coded bits onto less or more significant bits of the constellation as defined by (3.6).
The value Ncpc is the number of coded bits per subcarrier, i.e., 2, 4, or 6 for QPSK, 16-QAM or 64-QAM, respectively. The value of s is Ncpc/2. Within a block of Ncbps
transmitted bits, k is the index of the coded bit before the first permutation, mk is the index of that coded bit after the first and before the second permutation, and jk is the index after the second permutation. Beside, d =16 is the modulo used for the permutation. The first permutation is defined as
( )
mod( ) ( )k cbps d
m = N d k× + floor k d (3.5)
and the second permutation is defined as
( )
( ( ) )
mod( )k k k cbps k cbps
j = ×s floor m s + m +N - floor d m N× s (3.6)
for k = 0, 1, …, Ncbps-1.
3.3.2 Signal Mapping
After interleaving, transmitted data bits serially enter to the signal mapper and then are converted into complex data symbols. IEEE 802.16e supports three constellation schemes, QPSK, 16-QAM, and 64-QAM as shown in Fig. 3.6. The constellation is Gray mapped to reduce the error bits of incorrect detected symbols.
Moreover, the constellations shall be normalized by multiplying each constellation point with the normalized factor c to achieve equal average power.
3.3.3 Subcarrier Allocation
In IEEE 802.16e OFDMA systems, interference averaging is realized by the
Fig. 3.6 QPSK, 16QAM and 64QAM constellations.
distributed subcarrier permutation which is designed to assign subcarriers pseudo randomly across the entire transmission spectrum, and users all have cross-spectrum subcarrier assignment. In-band interference and frequency selective fading affect all users evenly. This permutation method is adopted by PUSC and FUSC.
In PUSC, some of the sub-channels are allocated to one transmitter to form a segment, and remaining sub-channels are attributed to different segments. Each segment is assigned to different sector. PUSC is used both in UL and DL. FUSC is
Fig. 3.7 Subcarrier allocation of PUSC (FFT size=1024).
that all sub-channels are allocated to one transmitter, and it is used only in DL. This thesis adopts PUSC, and the subcarrier allocation is described as follows. As shown in Fig. 3.7, the subcarrier allocation to sub-channels is performed using the following procedure.
1. The subcarriers are divided into the number of clusters (Nclusters) of physical clusters, and each physical cluster contains 14 adjacent subcarriers which are 12 data subcarriers and two pilot subcarriers. The number of clusters, Nclusters, varies with FFT sizes. For example, there are total 60 physical clusters with 1024-ponit FFT.
2. Using the equation given below, the physical clusters are renumbered to form logical clusters, and this process is also called outer permutation.
( 13 _ ) mod clusters
LogicalCluster
RenumberingSequence PC DL PermBase N
=
+ × (3.7)
where PC denotes physical clusters. The renumbering sequence is given in [8].
The two adjacent logical clusters are not contiguous in frequency band.
3. The logical clusters are partitioned into six major groups depending on the FFT size. Group 0 includes clusters 0-11, group 1 includes clusters 12-19, group 2 includes clusters 20-31, group 3 includes clusters 32-39, group 4 includes clusters 40-51, and group 5 includes clusters 52-59. These groups may be allocated to segments, if a segment is being used, then at least one group shall be allocated to it (by default group 0 is allocated to sector 0, group 2 is allocated to sector 1, and group 4 to is allocated sector 2).
4. The sub-channel extracts one subcarrier from each of these groups. The exact subcarrier allocation is according to the permutation rule as
( , ) (k+13«s)modNsubcarriers, Nsubchannel is the number of sub-channels, ps[j] is the series obtained by rotating basic permutation sequence cyclically to the left s times, and DL_PermBase is an integer ranging from 0 to 31.
3.3.4 STBC Encoding
In IEEE 802.16e OFDMA specification, STBC proposed by Alamouti in 1998 can be used in DL to provide transmit diversity. In the proposed STBC-OFDM system, there are two transmit antennas on the transmitter and one reception antenna on the receiver as shown in Fig. 3.2. This scheme requires multiple-input-single-output channel estimation, and decoding is very similar to maximum ratio combining. Within a time slot (the duration of two OFDM symbols), two transmitted OFDM symbols, XF
and XS, are encoded and transmitted from different antenna as given in Table 3.3, where the superscript * stands for complex conjugate.
TABLE 3.3 STBC ENCODING
Within a Time Slot Transmit Antenna
1st Symbol Time 2nd Symbol Time
Antenna 1 XF -XS*
Antenna 2 XS XF*
Fig. 3.8 Receiver of the proposed STBC-OFDM system.
3.4 Receiver Architecture
Fig. 3.8 shows the baseband receiver architecture for the proposed downlink STBC-OFDM communication system. The proposed baseband receiver consists mainly of synchronization, channel estimation, STBC decoding, and signal detection.
The channels are assumed to be quasi-static within any two successive OFDM symbol durations. Hence, without loss of generality, the received signal processing is focused on each time slot, and the time index of symbol transmission is omitted hereafter except otherwise stated. After a RF signal is received from an antenna, it is down converted to the equivalent baseband, low-pass filtered, and digitized by an
analog-to-digital (A/D) converter. The digitized received signal is then synchronized with the symbol timing and carrier frequency to achieve the initialization in the receiver. Afterwards, when the synchronization is completed, the channel estimation estimates the channel responses, and the received signal is STBC decoded and de-mapped to retrieve the binary bits. Finally, data bits are passed to the channel decoder. The detail of the proposed receiver architecture will be described in the following subsections.
3.4.1 Synchronization
Synchronization is a critical problem in OFDM systems. The precise carrier and sampling clock frequencies are required to help the receiver to retrieve data at correct timing. However, in the transmitter and receiver, two independent crystal oscillators are used to generate the carrier frequency and sampling clock, and the oscillator mismatch results in the offsets of carrier frequency and sampling clock. Besides, the symbol timing must be derived from the received signals. In mobile environment, the Doppler effect causes the frequency shift on the received signals. Various synchronization errors in OFDM signals are defined as symbol boundary offset, sample clock offset (SCO), carrier frequency offset (CFO), and carrier phase noise.
All these unavoidable problems affect the validity of synchronization detection;
therefore, the carrier and timing recovery must be used in the receiver to solve these problems.
Based on the operations of the IEEE 802.16e system, there are two synchronization stages. First, the system acquisition stage is used to acquire the preamble symbol of the transmission when a mobile device first enters an 802.16 network. It includes the coarse symbol boundary detection, the coarse CFO estimation, and the preamble search. Then, in the transmission stage, the receiver starts to execute data reception. Our proposed system is focused on the transmission stage. The synchronization architecture includes symbol timing synchronization, sample clock and carrier frequency synchronization as shown in Fig. 3.9. The proposed synchronization scheme performs symbol boundary detection and CFO estimation as presented in the following subsections.
Fig. 3.9 Synchronization architecture.
3.4.1.1 Symbol Timing Synchronization
Symbol timing synchronization refers to the task of finding the symbol boundary of the individual transmitted OFDM symbols, and this information will be used to define the FFT window. If the detected OFDM symbol boundary is not correct, the performance of the receiver will be degraded.
Effects of Symbol Timing Offset
In wireless multipath fading channels, the transmitted signals pass through different paths to arrive at the receiver, so the receiver receives the transmitted signals which are overlapped with different delayed versions of the signals. Assume that the maximum channel excess delay is small than the length of the guard interval and a symbol timing offset nd (normalized to one sampling clock period) is existed in the symbol boundary detection. When the detected symbol boundary locates in the channel impulse response region or is later than the actual boundary, the data in the FFT window contains the components from the later or previous OFDM symbols, and the ISI effect is introduced. Also, the received signal is attenuated and has a phase rotation [42].
[ ] [ ] [ ] N nd j2 knNd [ ]
where k is the subcarriers index, N is the number of subcarriers in an OFDM symbol, R[k] is the received OFDM symbol, X[k] is the transmitted OFDM symbol, H[k] is the channel frequency response, and Z[k] is the Gaussian noise. The term of exp(j·2π·k·nd/N) causes a phase rotation to the subcarrier constellations.
Based on the requirements for symbol timing offset, an ISI free region of symbol timing detection is determined by the difference in length between the CP and the channel impulse response region. The detected symbol boundary must be located in an ISI free region, which is not affected by the previous symbol due to channel dispersion [43], [44]; besides, the orthogonality of the subcarriers is preserved, and a time offset within this region only results in a phase rotation.
[ ] [ ] [ ] j2 knNd [ ]
R k =X k H k e× × p +Z k (3.10)
Symbol Boundary Detection Scheme
Since the proposed STBC-OFDM system has two transmit antennas, the signals transmitted from different antennas may arrive at the receiver with different delays due to the multipath effect as shown in Fig. 3.10. Therefore, the detected boundary must locate in the common ISI free region to prevent the respective ISI effects from the other symbols.
Because the time-domain preamble symbol is not exactly periodic, the delay correlation method cannot be used to precisely detect the symbol boundary. Since the preamble symbol is a known sequence after the system acquisition, a match filter corresponding to the time-domain preamble sequence is applied to match the received sample sequence and find the peak of matching results to obtain an accurate symbol boundary. The complexity of the match filter is depended on the coefficient length;
Fig. 3.10 Common ISI free region of two transmit antennas.
hence, a suitable length (L) is used to take both performance and complexity into consideration. Matching the received sample sequence r[n] and a known coefficient sequence p(i)[m] which is the L-sample time-domain preamble sequence transmitted from i-th antenna, the symbol boundary can be found as
( )
1 ( ) *
0
max L [ ] i[ ]
n n
m
SymbolBounary - r n m p m
=
æ ö
= ç + × ÷
è
å
ø (3.11)where n is the sample index, and m is the coefficient index. When the sample index n corresponds to the peak of the matching results, then the symbol boundary is found.
However, the mismatch of oscillator frequency in receiver and transmitter causes
However, the mismatch of oscillator frequency in receiver and transmitter causes