Broadband service providers receive many help‐desk calls because of networking issues in their customers' homes, and ever more of those are related to wireless technologies and security issues. Current technologies for home network troubleshooting do not in the slightest fulfill the service providers’ requirements, because they are all based on the same design philosophy: collecting the available relevant information provided by the various devices in the home network in a central place, for instance the home gateway or the TR‐069 ACS, and perhaps apply some algorithms to deduce some properties of the links directly interfacing with the home gateway.
We propose to include Machine Learning (ML) techniques. By applying ML, we can learn from the history of individual home networks, from feedback provided by the end user, and from wrongly or correctly analyzed performance and security indicators in other home networks.
The goal of this project is to improve the completeness, the accuracy, as well as the discovery time of the performance and security monitoring results from “unacceptable” to “acceptable” for service providers, by applying ML. The objectives of the Home Intelligence project are therefore to
- develop novel home network performance and security analysis algorithms based on ML,
- design an intelligent home network performance and security analysis architecture,
- position the architecture in the currently deployed remote management architectures,
- build a proof‐of‐concept
Type of work
The research is a combination of simulation using e.g. NS3 and Python ML routines, and implementation and testing of a proof of concept. The research may be partly carried out in close collaboration with a large international technology vendor, and may include field testing in real apartment blocks. It is also expected that the project will become part of a larger "Industry PhD Program" (to be established), which aims to bring PhD projects in close relation to the needs of the industry, in their content as well as the skills acquired by the student.