MSc thesis project proposal

[Almende] Distributed AI-based localization in IoT systems

Almende (almende.com) is a company based in Rotterdam which specializes in developing ICT solutions based around self-organizing networks in the domains of healthcare, energy, mobility, manufacturing and security. Almende has several subsidiary companies, one of which is Crownstone (crownstone.rocks) which specializes in IoT smart building systems.

One of the our project goals is to develop asset localization algorithms which run within the Crownstone network. In a Crownstone-equipped building, outlets have a Crownstone installed behind them. These Crownstones communicate with each other over Bluetooth Mesh. An asset (e.g. a person equipped with a Bluetooth beacon) transmits Bluetooth messages at a regular interval. Based on the message’s signal strength (RSSI) received at each Crownstone, in-network asset localization aims to make a prediction for the room that the asset is in in real-time. A specific focus lies on distributed machine learning algorithms to fulfill this task.

Assignment

The main project goal is to develop methods combining machine learning algorithms with signal processing techniques. These methods should be suitable for a decentralized implementation, taking into account the strict hardware constraints of the Crownstone and the available communication resources. Proof of concepts can be implemented and tested on the Crownstone platform.

A particular focus of the project relates to the translation of a static localization problem into a time-dependent one. Bayesian techniques may be used to combine knowledge of the location and movement of assets with observations by Crownstones. The developed method should be compared against the current centralized methods employed for asset localization by the Crownstone system and will be evaluated based on classification performance and use of resources.

Requirements

  • High level of proficiency in English, both oral and written. Dutch language proficiency is a plus.
  • Studying for a master degree in engineering (Electrical Engineering, Computer Science, Applied Mathematics or related fields).
  • Experience with embedded software development (C/C++).
  • Interest in distributed artificial intelligence systems.

Contact

dr. Raj Thilak Rajan

Signal Processing Systems Group

Department of Microelectronics

Last modified: 2022-09-28