University of Twente Student Theses
Cyclist Weight Inference using IMU Sensors on Bicycles
Angheluș, C. (2024) Cyclist Weight Inference using IMU Sensors on Bicycles.
|
PDF
34MB |
| Abstract: | Inertial Measurement Unit (IMU) sensors are widely used in various movable applications, including sports science, healthcare, and navigation. Recently, IMUs have also been used on bicycles to generate insights about road quality, fall detection, maneuver prediction, etc. IMUs are usually not known to contain sensitive data, but with the emergence of more advanced computer intelligence and machine learning techniques, we cannot be sure that certain sensitive insights, such as the weight of the cyclist, could not be inferred from the sensor data. This research investigated to what extent the cyclist’s weight can be determined using only the IMU data. This study included collecting the IMU data from a cyclist with different added weights in a controlled experiment. Next, it analyzed and preprocessed the obtained data, followed by the development of models, using machine learning techniques, which can classify the weights of cyclists based on the IMU data. Finally, it evaluated the models and explored the feasibility of classifying the weights into an increasing number of classes. It was found that we can feasibly classify the weights up to 4 classes before the accuracy drops too low. We were able to achieve models able to classify the weights with up to 84\% accuracy. The research is expected to contribute to advancing the understanding of privacy considerations when using sensor data as well as the possibilities of modern technologies to infer sensitive information from seemingly reliable and safe sensor data. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/100958 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page