Demo 1: Link throughput optimization in Wi-Fi networks by monitoring and characterizing interference

Speaker: Wout Joseph, Mostafa Pakparvar & David Plets

Company/ Institute: iMinds – UGent – WiCa

Event: 15th work meeting

Date: 25 February 2014

Abstract:

This PoC demonstrates link throughput estimation models that characterize the throughput in a wide area using predictive physical models. The model estimates the throughput based on interference characterization parameters including interference transmission rate (TxRateintf.) and interference channel occupancy degree (CODintf.). The PoC demonstration comprises 3 phases. In the first phase, the throughput of the intended link is recorded while interference is present. During the second phase, interference is characterized by sniffing packets on different channels. During the third phase, the frequency channel of the intended link will be adjusted to optimize the throughput. It will be shown that by utilizing intelligent decision algorithms, for strongly interfered channels the throughput will increase from 2.45 Mbps to 10.9 Mbps, which corresponds to a 344% improvement.