In the DISCO project, we build a decentralised system for discovering other Wifi users and then use machine learning models to optimise the radio spectrum use. To be able to coordinate with others, each access point needs to know its location. This is not straight forward as access points generally do not support accurate location mechanisms, such as GPS- and may also be unable to use them, due to how they are often placed in buildings.
In this master thesis we will explore ways to record the position of an access point. The primary mechanism to investigate is whether it would be possible to take a picture with an iPhone or Android phone and use image recognition techniques (such as SIFT) to recognise the router. Many modern phones include depth information in the image (obtained with front-facing LIDAR or ML techniques) that can be used as input to the estimation. Other sensor input can also be used, such as compass and gyroscope.
The goal of this thesis is to design and implement a positioning system for routers on Android or iPhone and evaluate the accuracy and feasibility of the solution in experiments.
Learning outcome
Experience
- in formulating, investigating and answering research questions
- experience in using computer vision techniques to identify objects and their location in 3D space
- experience in developing applications and reading sensor input for mobile phones
- conducting, evaluating and interpreting research results
Conditions
We expect that you:
- have been admitted to a master's program in MatNat@UiO
- take this as a long thesis
- are interested in and have some knowledge of Swift/C/C++/Java programming and development of iPhone or Android applications
- include the course IN5060 in the study plan, unless you have already completed a course on classical (non-ML) data analysis