Motivation

Beam-steering is the backbone of millimeter-wave (mm-wave) networks and key to achieve data-rates of multiple gigabit per second. Nodes must steer their antennas so that they maximize the signal gain towards the intended communication partner. The state-of-the-art to find the best antenna configuration is to probe all possible antenna configurations. This process caused high overhead, especially in case of mobility when parameters must be adjusted continuously.

Goal

In this thesis, you apply machine learning techniques to find the antenna parameters most suitable for probing and select the optimal configuration with low overhead.

Implementation and evaluation in this thesis, should be performed by means of our mm-wave testbed platform with off-the-shelf IEEE 802.11ad devices. Experience with Linux, wireless network configuration, proper tools, and scripting languages is highly recommended.