University of Twente Student Theses

Login

The Cupid for Signals : Sensor Data Alignment and Evaluation for Freshwater Pipeline Inspection Tools

Roelofs, Pieter Jan (2024) The Cupid for Signals : Sensor Data Alignment and Evaluation for Freshwater Pipeline Inspection Tools.

[img] PDF
5MB
Abstract:Performing inspections is crucial to maintain the reliability and integrity of freshwater pipeline networks. To solidify the findings of these inspections, multiple inspection runs are performed on the same pipeline network. In order to line up the data from these multiple inspection runs, we perform alignment on the data obtained from the IMU sensors attached to the inspection tools. To guarantee an accurate alignment, this thesis contributed the following three things. Firstly, we created an IMU run simulator that allows us to generate a large amount of varied data similar to data obtained from real-world inspections. This simulator enables the optimization of alignment methodologies by providing valuable labelled train and test data. Secondly, we provide two new metrics, the shift error and the weighted shift error. These metrics accurately describe the alignment performance by comparing the time shift obtained from alignment and the baseline time shift. Finally, we created a new alignment method specialized for aligning IMU sensor data. Our method, PAIDTW, is an extension of the DTW algorithm, using the alignment obtained from DTW on the key features of the IMU data and interpolating in between. Results show a significant improvement on the alignment performance.
Item Type:Essay (Master)
Clients:
Rosenxt, Enschede, The Netherlands
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:https://purl.utwente.nl/essays/104844
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page