Master thesis  - 
 Automatic semantic tagging
A number of techniques for automatic fault detection and other system analyses are being adopted in the Building Management Systems communities. One big obstacle for a wider usage of these technologies is to map the numerous signals in a system to equipment, rooms, variables etc.
Using prior semantic tagging project as platform with more focus on feature engineering and choice of ontology/classification. Also consider not manually chosen features for classification and to use time as part of the input.
 Cognitive BMS
Development packages like IBM’s cloud platform BlueMix offer cognitive functionality that can be used when a person interacts with the Building Management System. We would like to explore possibilities around voice commands, e.g.
- Acknowledge alarms by voice
- Control system settings (like room temperature) by voice interaction
 Crash analysis
There are about 100 000 building automation servers deployed globally. Occasionally one crashes or experience other problems and if the connection is enabled a crash dump is sent to a central repository managed by BMS R&D. One could apply machine learning techniques to
- Cluster crashes with common root cause
- Make initial analyses of where in code base there is a problem
- Do pre-crash analyses of trace log information
Contact: Oskar.Nilsson@schneider-electric.com, Jorgen.Malmborg@schneider-electric.com
 Building knowledge graph
Information stored in Building Management Systems is typically not machine readable. One remedy to this is to model a building using a knowledge graph, i.e. an ontology. The model can contain BMS devices, HVAC equipment, physical locations, data points and their interrelations. The project could explore ontologies such as Bricks, RealEstateCore or the tagging scheme Haystack. A mockup GUI for BMS modeling using the OLGA SDK would be interesting. Also, useful queries based on the ontology data and what customer value they bring can be explored.
 Distributed energy management
Energy production is becoming more decentralized and energy consumers are becoming so called prosumers that also feed energy back into the grid and store energy locally. There is flexibility in such a system that can be optimized to reduce energy cost, CO2 footprint and or resiliency to grid disconnection. Energy components can be PV panels, thermal energy storage, flexible energy loads together with both electrical and thermal grid connections. This is the case of current (2018) project of the Lidl distribution center in Finland.
 Low energy room units based on e-ink and Bluetooth
Wireless, battery-powered products have many advantages and are becoming more and more common in commercial buildings. However, a challenge is the life of batteries, as an exchange will be very costly for large scale installations. We would like to investigate whether we can make a battery powered room unit with E-ink display and BLE communication that can meet our requirements for very long battery operation, user experience, and quality. In the scope of this project we will make a proof-of-concept including hardware and firmware and evaluate performance. To be successful in this project a strong interest in digital electronic design, embedded system, firmware development and display technologies is necessary.
 Impulse radio ultra-wideband radar sensor for human detection
Ultra Wideband (UWB) is an emerging technology in the field of indoor positioning and motion detection which has shown proof of better performance compared to other technologies. The motivation for increased performance is surveillance and intruder detection, demand-driven HVAC and efficient lighting control systems. UWB could also be suitable for battery powered applications and offer lower cost compared to e.g. camera technologies. To set the stage for this work, we will perform a study of state-of-the-art technologies in human detection, followed by a detailed comparative analysis. The focus of the project will then be on the practical part where strong interest in RF-design, analog electronics, embedded systems and signal processing is needed.
 Agile Development methodology for hardware Product developments projects
The Agile Manifesto has revolutionized software development. The principles, tools, and practices which have stemmed from it have been refined and spread through many companies. In hardware product development, Agile principles, tools, and practices have been slow to take hold. In this project we would like to do a case study on an ongoing hardware development project, make a comprehensive review of Agile and how to apply Agile to the hardware product development. Look at which Agile Principles work in hardware, which don't and why, and explore best Agile practices for hardware development projects. An engineering background with a strong interest in Agile methodologies, project management and business development is necessary to be successful in this project.
 Optimize building efficiency based on occupancy
Indoor location is getting increasing focus. The technology opens possibilities for tenants to locate themselves and navigate within buildings. It also opens the possibilities to track people and assets to ta analyze patterns.
Building automations has always been a very static segment which has controlled buildings based on fixed sensors and schedules.
New kinds of data are being generated by the indoor location system and this master thesis is about exploring the possibilities for the building management system to use this data to optimize the building.
 Cybersecurity for embedded systems
Cyber security considerations are becoming increasingly important as more and more things get connected to the Internet, and attacks on local IT infrastructure becomes easier and more prevalent. This is an especially great challenge when working with low cost embedded systems as security is often a low priority during development.
The objective is to investigate what we can do to harden and/or improve auditing of data storage on our existing server product.