Enabling decentralized wireless index coding in practice
Index coding is a problem in theoretical computer science and network information theory that studies the optimal coding scheme for transmitting multiple messages across a network to receivers with different side information. The ultimate goal of index coding is to reduce transmission time in a communication network by minimizing the number of messages based on shared information. Index coding theory extends to several key engineering problems in network communication including peer to peer communication, distributed broadcast networks, and interference alignment. Although the theoretical connection between index coding and wireless networks is valuable, we focus on finding index coding strategies for a realistic wireless network. More specifically, we investigate how index coding can be applied to an OFDMA downlink network during the retransmission phase. An orthogonal frequency-division multiple access (OFDMA) downlink network is a network where data is sent downward from a designated higher-level transmitter to a group of receiving nodes. In addition, receivers can often decode the other receivers' physical layer signals on the other sub-channels that can be exploited as side information. If this side information is sent back to the transmitter, it can then be coded to cancel the interference in subsequent retransmission phases resulting in fewer retransmission messages. In this report, we explain the coding model and characterize the benefits of index coding for retransmissions within an OFDMA downlink network. In addition, we demonstrate the results of applying this index coding scheme in such network in both simulation and in an active wireless mesh network.