The objective of reaching net zero emissions of greenhouse gases set out by the European Climate Law requires a significant transformation of the mobility sector. Today, most emissions in this sector are caused by individual motorized trips. When compared with public transportation, these trips typically cause 3-4 times higher emissions. Interestingly, many persons tend to rely on individual motorized trips, even if more sustainable alternatives are readily available. An important reason for this is unconscious and habitual behavior (i.e., they use their car without thinking about alternatives because they are used to travel this way).
The goal of the project group SuMoC (Supporting Sustainable Mobility Choices) is to develop a set of high-quality mobile applications for devices running Android and iOS that actively incentivize users to reduce the environmental footprint caused by their mobility choices. Towards this goal, the apps shall be able to track the mobility of their users continuously during the day and to propose more sustainable alternatives for past trips, if they are available. To minimize friction, the tracking performed by the apps must be both, energy-efficient and privacy-preserving. For the latter, the apps shall rely primarily on on-device processing. For the former, the apps shall implement appropriate sensor control schemes to balance the tracking accuracy and energy usage.
From a theoretical perspective, the project group will cover the fundamentals of context recognition with mobile devices (i.e., sensing, preprocessing, ML-based classification). The focus will be on motion-related sensors (i.e., GPS, IMU) as well as techniques to increase the energy-efficiency and to protect the privacy of users. The practical part will encompass modern mobile application development using native frameworks (e.g., Jetpack Compose, Swift UI) as well as applied machine learning using Python. Students taking this course must be fluent in at least one object-oriented programming language (e.g., Python, C++, Swift, Java) and should be able to apply their knowledge to other languages quickly.