This Master lecture provides an introduction to the subject area of the "Internet of Things" (IoT), in which billions of embedded systems (sensors, actuators) continuously make data about the real world available on the Internet in real time.

Topics covered are in particular: IoT system architectures (cloud vs edge vs mesh), communication protocols (IEEE 802.15.4, NbIoT, 6LoWPAN, MQTT), update protocols, data modelling, data access (stream processing, complex event processing), and data processing with machine learning for IoT devices (pruning, quantisation). In addition to theoretical knowledge, the exercise also teaches practical programming of IoT systems, e.g. with Arduino devices, sensors, Raspberry Pis and freely selectable cloud-based IoT platforms.

In previous projects, a processing pipeline for EEG data (electroencephalography) was developed, which utilises the MentaLab Explore+ EEG cap and enables a user to control a computer game avatar using thoughts.

EEG data was collected, labelled and interpreted into game controller commands using artificial intelligence. A Pacman-based video game with a client-server architecture was also implemented in Python. At the same time, work was carried out on improving the EEG recording hardware.

This pipeline is now to be evaluated and optimised. The aims of this semester include the following points:

• Software:
  • End-to-end integration from EEG data collection to the game server
  • Evaluation of prediction accuracy and latency, as well as their effects on immersion in the game
  • Optimisations regarding these points, such as improvements to data collection and processing using AI
• Hardware:
  • Acquiring the raw data from the MentaLab Explore+ with our own ElasticAI.hardware and C-runtime
  • Use of innovative "flower" electrodes from TU Ilmenau with active shielding

In the coming semesters, the AI model will be transferred to our ElasticAI.hardware, which will then function as a game controller.

Organisation:

Attendance at the kick-off meeting is mandatory for participation in this project.

This practical training is exclusive for students of the course of study M.Sc. "Cyber Physical Systems". It is offered in collaboration with the group of Prof. Weis and Prof. Schiele.