Tutorial on 802.11n PHY layer : Part 3

The MIMO-OFDM baseband architecture with N_tTX antennas is shown in Figure11. The incoming data bits are randomized using a scrambler in order to avoid the occurrence of long zeros and ones. The output of the scrambler block ensures that the bits are equally likely to satisfy the theoretical assumptions. The scrambled bits are passed into an encoder sparser where it is demultiplexed across the N_ES forward error correction (FEC) encoders in a round robin fashion. Here, N_ES is the number of encoding streams and in the standard, N_ES = 1 for 1x1 and 2x2 systems, and N_ES = 2 for 3x3 and 4x4 systems.

The last six scrambled zero bits in each FEC input is replaced by the unscrambled zero bits. This is done to make the FEC encoder to an all zero state after the encoding is done. The FEC block encodes the data to enable channel error correction capabilities. The FEC block is made up of binary CE followed by a puncturing block. The basic block achieves the coding rate of ½ and the other coding rates like 3/4, 2/3 and 5/6 are achieved with the help of puncturing pattern defined in the standard. The output of the 2 FEC units is interleaved and is done in 3 stages. In the first stage, a stream parser is used wherein the output of the encoders are divided into block of bits where s = max{1, N_BPSC/2}is the number of bits assigned to a single axis (real of imaginary) in a constellation point. From each encoder, N_ss consecutive blocks of the sbits are taken and fed across the N_ss spatial stream. In the second stage, the bits at each spatial stream are divided into blocks of N_CBPS and interleaved using the technique given in the standard. Then, the interleaved bits from each stream are grouped into of N_CBPS bits and mapped to the constellation points. In the final stage, the output of the N_ss streams is passed through a spatial mapper. The spatial mapper distributes the complex symbols to the N_t transmit chains. In each chain, out of 64 subcarriers in 20MHz operation, data symbols are mapped on the subcarriers -28 to -1 and +1 to 28. The remaining subcarriers are loaded with guard subcarriers and pilot symbols. Then in each chain, an N-point IFFT is taken and the cyclic prefix of length N/4 is taken from the end of IFFT output is appended in front of it. The signals are then upconverted to radio frequency and transmitted through the N_t TX antennas. The total power transmitted is normalized across the N_t transmit antennas and is given as

, where s_q(n) is the transmitted signal from the ^th TX antenna.

Click here for Figure 11
Figure 11: 802.11n MIMO-OFDM baseband transmitter.

Typical 802.11n Receiver model:
At the receiver, N_r antennas are used to receive the signal. The signals in each RX antenna are down converted to baseband and sampled with a maximum sampling duration of 50ns. Assuming that the receiver is perfect time and frequency synchronized, and the exact channel knowledge is available at the receiver, the remaining processing is performed. The received signal at the p^th RX antenna is given as:

(7)

Where h_pq(n) is the impulse response of the channel between the q^thTX and the p^th RX antenna, L is the channel length and v_p(n) is the AWGN at the p^th RX antenna with zero mean and variance σ²_v. A standard OFDM receiver is used in each RX chain to obtain the frequency domain estimates.

The complex estimates corresponding to a particular subcarrier from all the RX chains are grouped together to form a vector. Then, spatial detection is done jointly on this vector of complex symbols corresponding to this subcarrier. Similar processing is done in all the subcarriers and the received signals are spatially separated into N_sssignals. Using a QAM demodulator, the complex symbols are demodulated and deinterleaved in the case of hard decision decoding. However, in soft decision decoding, the soft information from the OFDM receivers are deinterleaved and decoded. A spatial demapper collects the deinterleaved signals from the N_ss paths and multiplexes them to the N_ES viterbi decoders. After decoding, descrambling is done to rearrange the bits in a similar fashion to the input of the transmitter. Figure12 shows a typical receiver structure for the 802.11n transmitter.

Click here for Figure 12
Figure 11: 802.11n MIMO OFDM baseband receiver.

Modulation and code selection scheme
The main advantage of 802.11n standard is that the highest data could be achieved in par with Ethernet. In the 802.11 basic standards, different rates are achieved with various modulation and coding schemes. The maximum PHY layer rate is 54Mbps in 802.11a/g with 64-QAM and 3/4 convolutional coding. Based on this, formulae for calculating highest data rate could be obtained as follows;

To vary the data rate, one can consider the following options.

1. Decreasing the channel spacing or increasing the number of samples within one second
2. Decreasing the guard band overhead, i.e. increasing the number of data subcarriers out of total number of subcarriers.
3. Increasing the constellation size, i.e. choosing higher QAM
4. Decreasing the channel coding rate and
5. Decreasing the guard time, i.e. decreasing the cyclic prefix

Apart from all these, the data rate could be increased to multifold based on the number of spatial streams. Here, the number of spatial streams is the minimum between the number of transmit and the number of receive antennas. The number of encoding streams can also be used as one of the parameter to vary the data rates. Table 2 below is sample of maximum (64-QAM) and minimum (BPSK) data rates with all the above parameters.

This table is exclusively for the 802.11n standard. The data rates supported by 802.11b/g system are also supported by the new system. There are other advanced techniques like beamforming, low density parity coding (LDPC) and space time block codes (STBC) that increase the reliability of the reception of the signals in severe fading channels. The data rate calculation in all the above advanced schemes may not be straight forward as there is no linearity between the number of spatial streams and the data transmitted.

Click here for Table 2
Table 2: Data rate with different MCS and number of spatial streams.

Acknowledgements
The Author would like to thank his graduate guide Dr. Srikanth, AU-KBC Research Centre, India for providing him the knowledge for living. He also like to thank his friend N.Muthuraja for his constant encouragement and for constructive discussions.

Part 1: Introduction.

Part 2: Backward compatibility with the legacy devices (IEEE 802.11b/g).

About the author
Sathish Viswanathan is working for an MNC in the area of HSUPA and HSDPA (NodeB L1). He completed his master's in wireless communication with a specialization in PHY layer of WLAN at AU-KBC reserach centre, India. There, he worked in time and frequency synchronization aspects of WLAN systems and MIMO-OFDM.