6.888 Lab 1: OFDM Transceiver

Introduction

In this lab, you will learn how to design a simple OFDM physical layer chain. You will need to design and write the code for each block of the RX and TX chains shown below. The code for this lab will be written in matlab. We will provide you with a skeleton code. Ideally, you will only need to modify the places where you see : "... add your code here ..." . We will provide you with wireless signals on which you can test your code. The lab will be divided into several tasks. Each task will require you to run some test script which will output a file or plot a figure. You will have to keep the figures and the output files and submit them along with your src code.

For a detialed description of each block's functionality, you can refer to chapter 3 of this THESIS, and chapter 2 of TEXTBOOK 1.

You can download a copy of the lab code here : 6.888_lab1.zip

• SRC_CODE: You need to implement these functions.
• Test_Scripts: Code to test the functions you implement.
• Mat_Files: Signals to test your code.
• Results: This folder is initially empty. The test scripts will output files in this folder

Lab 1 : Part A : OFDM Basic RX/TX

In the first part of the lab, we will assume that the transmitter and receiver oscillators are synchronized. You do not need to estimate and correct for the carrier frequency offset (CFO) or perform phase tracking. Hence, you do not need to implement the 2nd and 5th blocks of the RX Chain. For the purpose of this lab, we will only use BPSK modulation. For both parts of this lab, you will use the OFDM paramters shown in the table below. The wariables are stored in the matlab file: Parameters.mat. The code loads these variables into the workspace so you can immediately use them.

Task 1: TX Chain

• Run the code test_tx_ofdm_chain.m. It will output the file Result_TX_Chain.mat and will plot Result_TX_Chain.fig
  
  

Task 2: Packet Detection

• Run the code test_packet_detection.m. It will output the file Result_Pkt_Start.mat
  

Task 3: Channel Estimation

• Run the code test_estimate_channel.m. It will output a file Result_Channel_Estimation.mat and a figure Result_Channel.fig.
  
• Run the code test_correct_channel.m. It will output the figure Result_Cons_Channel_Corr.fig.
  
  

Task 4: RX Chain

• Run the code test_rx_ofdm_chain_synced.m. It will output the file Result_RX_Chain_Synced.mat
  
  

Hints:

• There are 8 preamble symbols. The first 4 are for packet detection. The next 2 are for CFO estimation and the last 2 are for channel estimation.


• Make sure you add cyclic prefix for each symbols starting with the last preamble symbol.


• Make sure to use the proper mapping between bits and symbols in your modulation and demodulation. 0 maps to 1, 1 maps to -1.


• Make sure to use the proper subcarrier indicies. Because of fftshit, the index of the subcarrier in the fft (1 to 128) is not the same as its normal index (-64 to 63). For example, the subcarrier with index -1 becomes 128 after the fftshift. Use the supplied functions convert_bin_index_fft_to_normal.m and convert_bin_index_normal_to_fft.m to move between the two.


• The preamble does not need pilots and thus the variable bits_preamble has hundred bits


• For the variabls bits_preamble and bits_data, the first bit should map to the first subcarrier in the fft subcarrier indexing and so on (e.g. bits_preamble(1) is assigned to fft subcarrier index 2 and bits_preamble(51) is assigned to fft subcarrier index 75). However, for the variable bits_pilots, the first bit should map to the first subcarrier in the normal indexing (i.e. [1 0 1 0 0 0 1 0] are assigned to normal subcarrier indicies [-39, -28, -17, -6, 6, 17, 28, 39] respectively).


• For the double sliding window packet detection, you can use a detection thershold of 6. You can also try different thresholds and pick the one with the least error.


• Idealy, you need only 2 preamble symbols for packet detection use double sliding window. You can use 4 preambles to improve the accuracy by using a larger window or implementing the delay and correlate packet detection on the next 2 preamble symbols.


• In channel estimation and correction, make sure to ignore the unused subcarriers.


• In this part of the lab, you can ignore the pilot subcarriers while decoding the signal in the RX chain but not while encode the signal in the TX chain.


• Make sure to compenstate fo the cyclic prefix bofore you run the FFT in the RX chain.


• Make sure to skip enough samples of the cyclic prefix (3/4 of cp) to avoid inter symbol interference.

Submit:

$ curl -F file=@lab1a-handin.zip \
-F key=XXXXXXXX \
http://haithamh.scripts.mit.edu/6.888/handin.py/upload

Lab 1 : Part B : OFDM Phase Tracking

Task 1: CFO Estimation and Correction

• Run the code test_estimate_cfo.m. It will output the file Result_CFO_estimation.mat.
  
• Run the code test_correct_cfo.m. It will output the file Result_CFO_Correction.mat and the figure Result_Cons_CFO_Corr.fig which shows the constellation before and after CFO correction but before channel estimation.
  
  

Task 2: Phase Tracking

• Run the code test_phase_tracking.m. It will output the file Result_Phase_Tracking.mat and the figure Result_Cons_Phase_Track.fig which shows the constellation with and without phase tracking.
  
  .
  

Task 3: RX Chain

• Run the code test_rx_ofdm_chain.m. It will output the file Result_RX_Chain.mat
  
  

Hints:

• Incorporate the phase tracking correction in the channel value i.e. update the channel h from symbol to symbol so that the phase does not accumulate so much that it starts wraping around 2pi. This also will allows us to automatically correct for the initial channel as well as all the accumulated phase when we run correct_channel.m.


• In phase tracking, when estimating the slope make sure that the indicies of the pilot subcarriers are in the normal -64...63


• In phase tracking, you can use the builtin matlab functions regress( ) or robustfit( ) for estimating the slope of the phase.

Submit:

$ curl -F file=@lab1b-handin.zip \
-F key=XXXXXXXX \
http://haithamh.scripts.mit.edu/6.888/handin.py/upload