print page

Bachelor's Thesis

Multivariate Time Series Classification using Feed Forward Networks and Transformers for Sensor Based Location Awareness


Machine learning (ML) techniques can be used to imitate basic cognitive skills of animals. An example is the sense of location awareness. It enables all higher animals to remember a discrete number of locations, to recognize where it is, and to remember how to get from one location to another. The ability to recognize its own location in a set of known locations can be replicated using ML, e.g., with artificial neural networks, in a sensor module with multiple sensors. This can be used for industrial applications such as transport boxes with integrated battery-powered sensor modules that are repeatedly transported through a factory along almost the same paths.

A constantly growing dataset is being maintained by the institute. These include sensor data (e.g., accelerometer, gyroscope, magnetometer, temperature, pressure, etc.) collected at each location, while travelling along three paths in the TUHH premises.

This thesis will investigate and evaluate the success and accuracy of training data for three paths at TUHH using Deep Neural Network (DNN) models, namely Feed Forward Networks (FFN) and Transformers. The evaluation results will be analyzed with evaluation metrics such as which parts of the path contribute to the highest percentage of errors and the pros and cons of FFN and Transformers will be discussed. The implementation will be done in Python with adequate libraries.

Student Andrea Essig
Supervisor Nisal Hemadasa Manikku Badu