University of Twente Student Theses

Login

Automated Grade Classification and Route Generation with Affordances on Climbing Training Boards

Stapel, F.T.A. (2023) Automated Grade Classification and Route Generation with Affordances on Climbing Training Boards.

[img] PDF
7MB
Abstract:With the rise of training boards for climbing, research exploring the possibilities of classifying the grade of a route and generating new routes is becoming more popular. This thesis describes a framework using machine learning algorithms to generate training board routes tailored to individual climbers' action capabilities and training needs. The principles of ordinal regression are implemented in the grade prediction process to use the order in climbing grades. The grade classifier performs similarly to previous related work and human benchmarks with accuracies of respectively 46.5%, 46.7%, and 45%. However, the move-based structure of the grade classifier allows it to be used in the route generator. The route generator uses a climber's reach, finger, power, and core strength to fit a generated route. The generated routes are tested with user studies. Even though some generated routes are of bad quality, other generated routes perform well and show potential for use in training. Unfortunately, the user studies are unreliable due to their small sample size and lack of variety in the demographic of participants. To use this method in practice, improving the performance of grade classification, gathering training board route data, and receiving detailed feedback on existing routes is essential.
Item Type:Essay (Master)
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/94487
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page