Chapter 1 Overview of Physical Layer (PHY) IEEE 802.16e OFDMA
1.3 Introduction to IEEE 802.16e Downlink
1.3.2 OFDMA Symbol Structure
As mentioned in [4] and [5], the OFDMA PHY defines four scalable FFT sizes:
2048, 1024, 512, and 128. Here we only take the 2048-FFT OFDMA sub-carrier
allocation for introduction. The sub-carriers are divided into three types: null (guard band and DC), pilot, and data. Subtracting the guard tones from scalable FFT size NFFT, one obtains the set of “used” sub-carriers Nused. These used sub-carriers are allocated to pilot sub-carriers and data sub-carriers.
z Preamble
The first symbol of the downlink transmission is the preamble. There are three types of preamble carrier-sets, those are defined by allocation of different sub-carriers for each one of them; those sub-carriers are modulated using a boosted BPSK modulation with a specific pseudo-noise (PN) code defined in Table 309 if [4]. The preamble carrier-sets are defined using
n 3
PreambleCarrierSet = + ⋅n k (1.1) where:
PreambleCarrierSetn specifies all sub-carriers allocated to the specific preamble, n is the number of the preamble carrier-set indexed 0...2, k is a running index 0...567.
Each segment uses a preamble composed of a carrier-set and modulates each third sub-carrier. Because the DC carrier will not be modulated at all, it shall always be zeroed and the appropriate PN will be discarded. For the preamble symbol there will be 172 guard band sub-carriers on the left side and the right side of the spectrum.
z Symbol Structure for PUSC
The symbol structure is constructed using pilots, data, and zero sub-carriers.
Active (data and pilot) sub-carriers are grouped into subsets of sub-carriers called sub-channels. The minimum frequency-time resource unit of sub-channelization is one slot, which is equal to 48 data tones (sub-carriers).
With DL-PUSC, for each pair of OFDMA symbols, the available or usable sub-carriers are grouped into clusters containing 14 contiguous sub-carriers per symbol
period, with pilot and data allocations in each cluster in the even and odd symbols as shown in Fig 1.5. A re-arranging scheme is used to form groups of clusters such that each group is made up of clusters that are distributed throughout the sub-carrier space.
A slot contains two clusters and is made up of 48 data sub-carriers and eight pilot sub-carriers. The data sub-carriers in each group are further permutated to generate sub-channels within the group. Therefore, only the pilot positions in the cluster are shown in Fig 1.5. The data sub-carriers in the cluster are distributed to multiple sub-channels.
Fig. 1.5 Cluster structure. (Source: [3])
Chapter 2
Synchronization Techniques for IEEE 802.16e Downlink
2.1 Channel Model and System Parameters
2.1.1 Modified Stanford University Interim (SUI) Channel Models
Channel models described in [6] provide the basis for specifying channels for a given scenario. It is obvious that there are many possible combinations of parameters to obtain such channel descriptions. A set of 6 typical channels was selected for the three terrain types that are typical of the continental United States. In this section we present SUI channel models that we modified to account for 30o directional antennas.
These models can be used for simulations, design, development, and testing of technologies suitable for broadband wireless applications. The parametric view of the SUI channels is summarized in Table 2.1 and Table 2.2.
Six SUI channels are constructed which are representative of the real channels, using the general structure of the SUI Channel and assuming the following scenario:
z Cell size: 7 km.
z Base station (BTS) antenna height: 30 m.
z Receive antenna height: 6 m.
z BTS antenna beamwidth: 120o.
z Receive antenna beamwidth: omnidirectional (360o) and 30o. z Vertical polarization only.
z 90% cell coverage with 99.9% reliability at each location covered.
Table 2.1 Terrain Type vs. SUI Channels
Terrain Type SUI Channels C: flat terrain, light tree SUI-1, SUI-2 B: between A and C SUI-3, SUI-4 A: hilly terrain, heavy tree SUI-5, SUI-6
Table 2.2 General Characteristic of SUI Channels
Doppler Low Delay Spread Moderate Delay Spread High Delay Spread
Low SUI-1, SUI-2, SUI-3 SUI-5
High SUI-4 SUI-6
For the above scenario, using the channel model in [6], the six specific SUI channel models are list in Tables 2.3 to 2.8 and we choose SUI-3 channel model as our simulation environment.
Table 2.3 SUI-1 Channel Model
Tap 1 Tap 2 Tap 3 Units
Delay 0 0.4 0.9 μs
Power 0 -15 -20 dB
Doppler 0.4 0.3 0.5 Hz
Table 2.4 SUI-2 Channel Model
2.1.2 System Parameters
z Primitive Parameters
The following four primitive parameters defined in [4] characterize the OFDMA symbol:
1) BW: This is the nominal channel bandwidth.
2) Nused: Number of used sub-carriers (which includes the DC sub-carrier).
3) n: Sampling factor. This parameter, in conjunction with BW and Nused determines the sub-carrier spacing, and the useful symbol time. This value is set as follows: for channel bandwidths that are a multiple of 1.75 MHz then n = 8/7 else for channel bandwidths that are a multiple of any of 1.25, 1.5, 2 or 2.75 MHz then n = 28/25 else for channel bandwidths not otherwise specified then n = 8/7.
4) G: This is the ratio of CP time to “useful” time. The following values shall be supported: 1/32, 1/16, 1/8, and 1/4.
z Derived Parameters
The following parameters are defined in terms of the primitive parameters:
1) NFFT: Smallest power of two greater than Nused
2) Sampling frequency:Fs =⎢⎣n BW⋅ / 8000⎥⎦×8000 3) Sub-carrier spacing:Δ =f Fs/NFFT
4) Useful symbol time:Tb = Δ 1/ f 5) CP time:Tg = ⋅G Tb
6) OFDMA symbol time:Ts =Tb+Tg 7) Sampling time:Tb/NFFT
z Scalable OFDMA
The IEEE 802.16e OFDMA mode is based on the concept of SOFDMA.
SOFDMA supports a wide range of bandwidths to flexibly address the need for various spectrum allocation and usage model requirements. The scalability is supported by adjusting the FFT size while fixing the sub-carrier frequency spacing at 10.94 kHz.
Since the resource unit sub-carrier bandwidth and symbol duration is fixed, the impact to higher layers is minimal when scaling the bandwidth. The SOFDMA parameters are listed in Table 2.9 from [5]. For convenience, we only take the 512 FFT size following the parameters in Table 2.9 and QPSK data modulation to simulate and implement our OFDMA PHY downlink receiver.
Table 2.9 OFDMA Scalability Parameters
Parameters Values System Channel Bandwidth (MHz) 1.25 5 10 20
Sampling Frequency (Fs in MHz) 1.4 5.6 11.2 22.4
FFT Size (NFFT) 128 512 1024 2048
Number of Sub-channels 2 8 16 32
Sub-carrier Frequency Spacing 10.94 kHz Useful Symbol Time (Tb = 1/f) 91.4 microseconds
Guard Time (Tg = Tb/8) 11.4 microseconds
OFDMA Symbol Duration (Ts = Tb + Tg) 102.9 microseconds
2.2 Synchronization Control Mechanisms
2.2.1 Network Synchronization
For TDD and FDD realizations, it is recommended (but not required) that all BSs be time synchronized to a common timing signal. In the event of the loss of the network timing signal, BSs shall continue to operate and shall automatically resynchronize to the network timing signal when it is recovered. The synchronizing reference shall be a 1 pps timing pulse and a 10 MHz frequency reference. These signals are typically provided by a global positioning system (GPS) receiver.
For both FDD and TDD realizations, frequency references derived from the timing reference may be used to control the frequency accuracy of BSs provided that they meet the frequency accuracy requirements of 2.2.3. This applies during normal operation and during loss of timing reference.
2.2.2 SS Synchronization
For any duplexing, all SSs shall acquire and adjust their timing such that all uplink OFDMA symbols arrive time coincident at the BS to an accuracy of ± 25% of the minimum guard interval or better. Ranging for time (coarse synchronization) and power is performed during two phases of operation: during (re)registration and when synchronization is lost; and second, during FDD or TDD transmission on a periodic basis.
During registration, a new subscriber registers using the random access channel,
and if successful, is entered into a ranging process under control of the BS. The ranging process is cyclic in nature where default time and power parameters are used to initiate the process followed by cycles where (re)calculated parameters are used in succession until parameters meet acceptance criteria for the new subscriber. These parameters are monitored, measured, and stored at the BS, and transmitted to the subscriber unit for use during normal exchange of data. During normal exchange of data, the stored parameters are updated in a periodic manner based on configurable update intervals to ensure changes in the channel can be accommodated. The update intervals shall vary in a controlled manner on a subscriber unit by subscriber unit basis.
Ranging on re-registration follows the same process as new registration.
2.2.3 Frequency Control Requirements
At the BS, the transmitted center frequency, receive center frequency, and the symbol clock frequency shall be derived from the same reference oscillator. At the BS, the reference frequency accuracy limited in [3] shall be better than ±2×10–6.
At the SS, both the transmitted center frequency and the sampling frequency shall be derived from the same reference oscillator. Following [4], the SS uplink transmission shall be locked to the BS, so that its center frequency shall deviate no more than 2% of the sub-carrier spacing, compared to the BS center frequency.
During the synchronization period, the SS shall acquire frequency synchronization within the specified tolerance before attempting any uplink transmission. During normal operation, the SS shall track the frequency changes by estimating the downlink frequency offset and shall defer any transmission if synchronization is lost. To determine the transmit frequency, the SS shall accumulate
the frequency offset corrections transmitted by the BS (for example in ranging response (RNG-RSP) message), and may add to the accumulated offset, an estimated UL frequency offset based on the downlink signal.
2.3 Mobile Station Synchronization Techniques
Synchronization is an essential task for any digital communication system.
Without accurate synchronization algorithms, it is not possible to reliably receive the transmitted data. In OFDM system, the received signal detection requires sub-carrier orthogonality. Variations of the carrier oscillator, sampling clock or the symbol time offset affect this orthogonality. Therefore, the synchronizer estimates and compensates any offsets in carrier, sampling time, and OFDM symbol time in the receiver in reference to the transmitter.
WLAN systems typically include a preamble in the start of the packet with reference to [1]. The length and the contents of the preamble have been carefully designed to provide enough information for good synchronization performance without any unnecessary overhead.
In this chapter, we present a novel initial synchronization algorithm for downlink of OFDMA TDD based mobile WiMAX. When the MS receiver enters the network for the first time, initial DL synchronization including timing and carrier recovery will be done. We assume that the frame synchronization is done by monitoring the power of the received signal. Upon entering the network and upon a need to handover, the MS has to identify the preamble index of the BS segment that it will communicate with.
Therefore, another important task needed to be done during initial synchronization is to find the preamble index. Fig. 2.1 from [7] depicts the overall structure of the proposed
initial DL synchronization.
Fig. 2.1 Structure of initial DL synchronization. (Source: [7])
In particular, we exploit the properties of the DL preamble described in [8] to obtain time and frequency synchronization. The proposed method does not require prior knowledge of transmitted preamble for coarse or fine time synchronization. This enables frequency domain search of the transmitted preamble and eliminates the need for computationally intensive time-domain preamble search using cross-correlation with the set of all possible preambles.
2.3.1 Symbol Timing Estimation
Symbol timing refers to the task of finding the precise moment of when individual OFDM symbols start and end. Its result defines the DFT window; i.e., the set of samples used to calculate DFT of each received OFDM symbol. In practice, it is impossible to fix the symbol timing point perfectly to the first sample of the OFDM
symbol. There will always be some variability in the symbol timing estimate around the ideal boundary. When the symbol timing point is estimated before the ideal value, the start of the DFT window will contain samples from the CP and the last samples of the symbol are not used at all. This case does not cause serious problem because the CP is equal to the last samples of the symbol. Next consider the case when the symbol timing estimate is after the ideal value. In this case, the start of the DFT window will be after the first sample of the symbol and the last samples are taken from the beginning of the CP of the next symbol. When this happens, significant ISI is created by the samples from CP of the next symbol. Additionally the circular convolution property required for the orthogonality of the sub-carriers is no longer true, hence inter-carrier interference (ICI) is generated. The end result of a late symbol timing estimate is a significant performance loss. Fortunately, there is a simple solution for this problem. Since early symbol timing does not create significant problems, the mean value of the symbol timing point can be shifted inside the CP. This means that the circular convolution is preserved and no ISI is caused by the samples from CP of the next symbol.
In IEEE 802.16e OFDMA, BS transmits a unique preamble as the first symbol in DL sub-frame. We first examine the properties of the preamble and later utilize them to develop the suitable synchronization procedure. In according to the description about the preamble in 1.3.2 the main properties of the preamble can be summarized as follows.
1) Preamble data is transmitted on every 3rd sub-carrier in the frequency domain while other two sub-carriers carrying zeros. It leads to time domain repetition, but strictly, the time domain symbol is not repetitive as IFFT size is not modulo-3, however, it does show high correlation.
2) Preamble data is real in frequency domain (i.e. BPSK modulated) so that it is
conjugate symmetry in time domain.
3) Combination of above two properties lead to repetitive conjugate symmetry in the preamble symbol that is each 3rd of the time domain preamble symbol exhibits conjugate symmetry. The time domain preamble p(n) can be written as where N is the size of FFT, conj is the conjugate operation, and flip is the operator of reversing sequence a.
In order to estimate the accurate timing offset, we propose to utilize the conjugate symmetry search for fine time acquisition. The conjugate symmetric correlation XCS(n) can be written as
where r is the received signal and the symbol timing offset estimator refer to [1] is given by We show the preamble conjugate symmetric correlation and compare with the CP delay correlation in Fig. 2.2. We find that the preamble conjugate symmetric correlation has a much sharper boundary and much larger magnitude than CP delay correlation. Thus, the preamble conjugate symmetric correlation has a better noise resistance and provides a good estimate of start of the preamble symbol as well as the tap delay profile. It also allows receiver to identify the first arriving path. We exploit this property to obtain fine time synchronization. In hardware implementations, the
correlation can be efficiently implemented using an add-subtract strategy with 2 complex multiply-and-accumulate (C-MAC) operations for each search.
0 50 100 150 200 250 300 350 400 450 500
Delay Correlation, FFT=512, SNR=0dB, SUI-3 Channel Model
Preamble Delay Correlation CP Delay Correlation
Fig. 2.2 Preamble and CP delay correlation under SUI-3 channel.
Further, as (2.2) in the preamble property (3), the conjugate symmetry search returns the peaks at roughly 1/6th of the FFT size shown in Fig. 2.2. The discrepancy of the repetitive conjugate symmetry may lead to false preamble edge detection.
However it can be resolved using the cyclic prefix search over the identified samples from the conjugate symmetry search. The CP search over the 1/6th of symbol boundary can be expressed as
1
Now, we analyze the performance of two symbol timing estimation methods in IEEE 802.16e OFDMA. One is using the preamble conjugate symmetric correlation
and another is using CP delay correlation. In addition to the estimator (2.4), we adopt the timing metric with SNR parameter proposed in [9], which is derived for a maximum likelihood (ML) estimator and is given by
( ) ( )
where r(n) is a sample of the received signal,θis the beginning of the symbol, Ncp is the length of the CP, N is the length of FFT size, andρ=SNR SNR/( +1).
The simulation environment is following FFT-512 in Table 2.9 and using SUI-3 channel model. The mobile Doppler frequency is from 0 to 300 Hz. Our simulated SNR values are in the range 0 to 20 dB.
Fig. 2.3 shows how different SNRs affect the error distributions of symbol timing during the two methods with different estimator in various Doppler frequencies. First, in fd = 0 Hz, the channel is almost fixed, we see that the symbol timing estimation is more accurate than mobile channel in fd = 150 Hz and fd = 300 Hz. Second, we compare with the two symbol timing estimation methods using preamble and CP.
According to the error probability shown in Fig. 2.3, we observe the probability density function (PDF) of a realistic symbol timing estimate has a large variance in CP correlation method. This is because the preamble conjugate symmetry correlation has a sharper boundary and stronger noise resistance than CP correlation as in Fig. 2.2. Last, we test the performance of estimator (2.6) and find that it has a better performance in high SNR when using CP correlation method. It is likely the CP correlation is sensitive to noise and the estimator (2.6) with SNR estimation can eliminate the noise effect.
When using the preamble conjugate symmetry correlation, the estimator (2.6) has almost no effect to performance. The reason is given above. The same simulation results are shown in Fig. 2.4 where the root mean square error (RMSE) is defined
as
2
E n n
⎧ ∧ ⎫
⎪ −⎨
⎪ ⎪
⎩ ⎭
⎪⎬ and we can see which one has better performance. Further, because the
preamble length is longer than CP, larger correlation values improve performance, but also increase the amount of computation required. As a whole, the symbol timing estimation method using preamble has better performance than using CP correlation.
But when we take implementation into account, the CP correlation is more suitable for the estimation. The reasons and comparison results are discussed in the next chapter (section 3.3.3).
0 2 4 6 8 10 12 14 16 18 20 10-2
10-1 100
Error sample distribution in fd=0Hz (using CP & Preamble)
SNR(dB)
Error sample distribution in fd=150Hz (using CP & Preamble)
SNR(dB)
0 2 4 6 8 10 12 14 16 18 20 10-3
10-2 10-1 100
Error sample distribution in fd=300Hz (using CP & Preamble)
SNR(dB)
Probability
CP-error>4 without SNR estimation Preamble-error>4 without SNR estimation CP-error>2 without SNR estimation Preamble-error>2 without SNR estimation CP-error>4 with SNR estimation Preamble-error>4 with SNR estimation CP-error>2 with SNR estimation Preamble-error>2 with SNR estimation
(c)
Fig. 2.3 Symbol time synchronization error distribution under SUI-3 channel with 0 Hz, 150 Hz, and 300 Hz Doppler frequency using different methods.
0 2 4 6 8 10 12 14 16 18 20
RMSE of frame offset in fd=0Hz (using CP & Preamble)
SNR(dB)
RMSE of frame offset in fd=150Hz (using CP & Preamble)
SNR(dB)
0 2 4 6 8 10 12 14 16 18 20
RMSE of frame offset in fd=300Hz (using CP & Preamble)
SNR(dB)
Fig. 2.4 RMSE of symbol timing offset synchronization using different methods under SUI-3 channel with 0 Hz, 150 Hz, and 300 Hz Doppler frequency.
2.3.2 Fractional CFO Estimation
One of the main drawbacks of OFDM is its sensitivity to carrier frequency offset (CFO). The degradation is caused by two main phenomena: reduction of amplitude of the desired sub-carrier and ICI caused by neighboring carriers described in [1] and [2].
The amplitude loss occurs because the desired sub-carrier is no longer sampled at the peak of the sinc-function of DFT. Adjacent carriers cause interference, because they are not sampled at the zero-crossings of the sinc-functions.
In the OFDMA DL, there may be large CFO in the received signal. The large
CFO can be partitioned into the fractional part of “normalized CFO” (where normalization is with respect to the sub-carrier spacing) and integral part of
CFO can be partitioned into the fractional part of “normalized CFO” (where normalization is with respect to the sub-carrier spacing) and integral part of