Flexible Reconfigurable Intelligent Surfaces
HiWi – Student Assistant
Hiwi Stelle
Reconfigurable intelligent surfaces (RISs) have recently received significant interest as a solution to enable a programmable wireless signal propagation environment. They can also be mounted on non-flat sur- faces. An open question in this topic is the phase optimization of RIS when it is imple- mented on a non-plane surface based on a set of experiment data.
In this research, we are interested in approximating the non-plane surface and then designing a codebook tailored for our approximation. To do this, we first train a neural network to estimate the unknown parameters of the approximated function by collecting received power in each grid after applying a codebook specific for the plane RIS. More specifically, we aim to exploit deep learning to map the distortion in the plane surface to the known different groups of non-plane RIS with a quadratic distortion.
- Scientific skills: Basic knowledge of digital communication systems
- Programming skills: Experience in ML models and Python programming (preferable) or MATLAB
- Language skills: Fluent in English
- Duration: 3-4 months
Interested applicants are encouraged to submit their academic transcripts, and a brief statement outlining their interest in the position to mostafadarabi@ece.ubc.ca and mohamadreza.delbari@rcs.tu-darmstadt.de.
