The Empathic Radio project - Clustering techniques

In this master project we will develop new features for Wi-Fi networks which make Access Points or a Wi-Fi router  (APs), to cooperate in clusters. A cluster is typically an apartment building or a multi-dwelling-unit (MDU) in which the tenants live close together and their APs disturb each other.

We will develop new features for Wi-Fi networks which make Access Points or a Wi-Fi router  (APs), cooperate in clusters. A cluster is typically an apartment building or a multi-dwelling-unit (MDU) in which the tenants live close together and their APs disturb each other.  The software will be executed by the microprocessor in the APs, without changing any hardware. The software will not replace the commonly used existing software, just complement the APs capabilities. It will make  the APs inside a cluster collaborate with other APs by choosing channels and set the transmit power of the APs. These changes will result in reduced interference and thereby improve the quality and capacity of the Wi-Fi networks in the cluster. Our system is called empathic and is completely distributed without central control. The software which will be developed, is open source.

Alone or cooperation?

Once the clustering has been performed, the APs all have enough and identical information for power and channel allocation. They all work in a Peer to Peer fashion and perform identical tasks.  The Master challenge is to find out if is smart to let each AP work alone or cooperate with other APs in the cluster. The advantage if it works alone is that there is no need to send and receive messages from other. The disadvantage is that it has to do a time consuming job alone, many times in order to find the solution. If the APs cooperate, they solve different tasks a few times, but exchange messages every time they has solved the task. To find the solution, the system must be simulated using Mininet or Opnet or ns3 in order to find the fastest solution.

Emneord: Wi-Fi networks, channel allocation, simulation-estimation
Publisert 17. okt. 2019 11:32 - Sist endret 17. okt. 2019 11:32

Omfang (studiepoeng)

60