Pranavi Boyalakuntla Jonathan Kelley

1. Overview


For this lab, we are implementing a receiver for a MIMO digital communication system with 2 transmitters and 2 receivers. In this lab, we will be working with data sent and recorded using USRP's. Our goal is to extract and decode the original data points from the received data by implementing various receiver schemes in the MATLAB development environment.

Figure 0: Potential set up for this data collection

Figure 0: Potential set up for this data collection

The bits were transmitted using the following order.

  1. 5000 zero samples from transmitter 1 and 2

  2. Transmitter 1 sends 128 pseudo random bits, encoded using BPSK and 40 sample long rectangular pulses, transmitter 2 sends zeros.

  3. 5000 zero samples from both transmitters

  4. Transmitter 2 sends 128 pseudo random bits, encoded using BPSK and 40 sample long rectangular pulses, transmitter 1 sends zeros.

  5. 5000 zero samples from both transmitters

6.Transmitters 1 and 2 send 1024 data bits (they are randomly generated here), using BPSK and rectangular pulses that are 40 samples each.

2. Receiver Design

2.0 Receiver Overview


In designing the receiver for our transmitter system, we first dove into the literature to gain a better understanding of what Multiple Input Multiple Output (MIMO), Space Division Multiple Access (SDMA), and channel estimation meant in the context of our system. We also made some assumptions to better constrain the final implementation. First, we assumed a flat-fading channel model as a result of ambiance white gaussian noise. Next, we assumed that the transmitted signal bandwidth with significantly less than the carrier bandwidth. We also assume that the carrier frequencies are synchronized such that the channels do not change.

Our research into the MIMO system led us to a matrix representation of the system. For any transmitted signal, we would have two corresponding received signals: one for either antenna. Let's denote the signals transmitted as $Tx_1 = x_1$ and $Tx_2 = x_2$. Let's denote the signals received by the receivers as $Rx_1 = y_1$ and $Rx_2 = y_2$.

The challenge with processing the received data is processing out the data that this receiver is not meant to capture. Both of the receivers will hear a combination of both transmitted messages weighted by the channel they were sent through and some additional noise represented by n[k].