ECE 8863: Cognitive Radio Networks

Lab 1: Full-duplex Communications

Software defined radio (SDR) is a platform for developing a configurable wireless transceiver using software. The great flexibility offered by software makes SDR an ideal frequency agile radio platform for cognitive radio (CR), and a great development platform for CR testbed. The first step toward a fully-functional CR is to establish and verify the capability of digital communications using SDR.

In this lab, we aim to realize a full-duplex DQPSK digital transceiver and measure its performance on the CR testbed. After finishing this lab, you will know how to

  • Design software radio transceiver using GNU radio companion (GRC).
  • Explore the level of configurability that can be obtained by software radio.
  • Understand the major performance limitations of frequency agile transceivers.

 

Initial Environment Setup

Before starting the experiment, the following steps should be performed in order to setup the development environment appropriately:

  1. Log into your account on each host laptop and do the following steps (you only need to do this once on each host).

  2. Change the password using passwd command (the dollar sign ‘$’ is the prompt).
    $ passwd

  1. Create a directory: lab1. This is your work area <lab1>.

  2. Download lab1_dqpsk_sim.grc.

 

Lab Procedures

In this lab, we are going to develop the CR transceiver in three stages. First, we develop the CR transceiver under simulation environment (i.e., there is no actual wireless transmission). Second, we test the transceiver on one single node. Third, we test the transceiver with two nodes (full duplex mode). In each stage, perform the requested tasks and log the results for your lab report.

Table 1: Summary of CR Transceiver Specification

Parameter

Value

Comment

Transmission Frequency

900 MHz

Make sure that antenna is supporting that range of frequencies

Sample Rate

200 KSPS

The sample rate to UHD source block. The ADC sampling rate is 100MSPS. Decimation is performed in FPGA.

Modulation

DQPSK

 

Data Rate

100 Kbps

 

Symbol Rate

50K Symbols/s

 

Pulse Shaping Excessive Bandwidth

0.3

The value 0.3 is a de-facto value for pulse shape excessive bandwidth parameter.

 

A. GRC and Simulation

  1. Start with Host 1. To run GRC, type "gnuradio-companion". The GRC example lab1_dqpsk_sim.grc is a simple DQPSK communication including a transmitting path and a receiving path connected through a simulated AWGN wireless channel. Take a moment to go through each block in the flow graph.

Task 1: Identify the locations of the parameters of Table 1 inside the GRC model.

  1. Run this example in GRC. You should see a clear constellation plot and the value in the RX scope matching the value you are transmitting from the source.

Task 2: Change the modulation schemes or noise levels. Observe the difference in constellation plots.

Hw1 Q1:

  1. We know that the ADC sampling rate is 100 MSPS and the sample rate to UHD source block is 200 KSPS. Why is the decimation of samples necessary?
  2. In this lab, DQPSK is selected as our modulation scheme. What are the advantages and disadvantages of using DQPSK? What happens if QPSK is used instead? Why is the result different?
  3. What are the number of bits per symbol (bits/symbol) and the number of samples per symbol (samples/symbol) in our settings?
  4. What is the purpose of pulse shaping excessive bandwidth? How does it affect the required bandwidth?

B. Single-node Communications

  1. Change the simulation version to a UHD USRP version. You need to use and configure UHD USRP source and sink blocks in the flow graph as discussed in the class. Save it to a new file, for example, lab1_dqpsk_uhd.grc.

  2. Make sure USRP1 is in the working state. You can verify the connection between the host and the USRP by
    $ uhd_find_devices
    Run the UHD version. The results should be similar to those from the simulation.

Task 3: Estimate the bandwidth from FFT sink scope.

  1. Change both Tx and Rx frequencies to 950 MHz, and then to 140 MHz

Task 4: Check the reception for both cases.

Hw1 Q2:

  1. Calculate the bandwidth and show that the bandwidth estimated from FFT sink scope in Task 3 is what you expected.
  2. Do you still receive the signal when the Tx frequency is changed to 950 MHz? Why?
  3. Repeat part b for the 140 MHz case.

Show your results from Tasks 1-4 to a TA. This is the first check point.

C. Full-duplex Communications (Different Bands)

  1. Next, let’s extend this to full-duplex transmissions by using 2 USRPs. Copy the UHD version lab1_dqpsk_uhd.grc to Host 2. Run this on Host 2 first to make sure you obtain the same results.
  1. Configure the UHD blocks on both Host 1 and Host 2 to enable full-duplex transmissions such that one value (say 63) is constantly transmitted from Host 1 to Host 2 on 900 MHz band while a different value (say 107) is sent from Host 2 to Host 1 on 1.2 GHz band. Note that the distance between USRP1 and USRP2 needs to be approximately or at least 1m.

Task 5: Verify the reception on both hosts.

D. Full-duplex Communications (Same Band)

  1. Now set the center frequency to 900 MHz for both Tx and Rx in both directions. That is, establish the communication between both nodes when both transceivers are set at 900 MHz.

Task 6: Record the received value at each node.

  1. Change the access code of the USRP 1 packet encoder and USRP 2 packet decoder to a different value and Repeat 8.

Task 7: Record the received value at each node.

  1. Increase the center frequency of one transmitter gradually and find the minimum separation of center frequencies to achieve similar performance to that in Task 5 (full-duplex transmissions on different bands).

Task 8: Record the value of minimum separation.

Hw1 Q3: Full-duplex Communications

  1. Compare and explain the results obtained in Task 6 and Task 7.
  2. What is the theoretical minimum separation of center frequencies to achieve similar performance to that in Task 5? Explain why the theoretical value is the same or different from the empirical value obtained from Task 8.
  3. From Task 3, you know DQPSK transmissions require B Hz bandwidth in each direction. What solution do you suggest to establish full-duplex communications if you have only a white space of bandwidth B Hz available?
Show your results of Tasks 5-8 to a TA. This is the second check point.
 

Questions? E-mail: infocom@ece.gatech.edu

You are visitor:

since 04/22/1997.