Deep neural network generating wireless channel heat map

The work is based on the idea similar to neural network(NN) generating van Gogh style images. However, this time the result could be used for localization and channel sounding. The objective of this project is to create a neural network which as an input receives a simplified map of the room, including walls, types of walls and access point(s), and outputs the heat-map of the given environment.

Students would go through the process of creating data to train the DNN, train a variation auto encoder that generate desired output.

Students are required to have basic knowledge in neural network, as well as programming skills in python.

Dates

The initial meeting for my group take place in ICT cubes, Room 333 on Wednesday, 10.04.2019 at 3 p.m..

Materials

Contact Person: Saeed Shojaee M.Sc. upon agreement

zurück zur Liste der Vorlesungen