University of Twente Student Theses

Login

Generalization capabilities of a learning from demonstration frameworkcapturing a human controller

Jaspar, S.L.J.O. (2023) Generalization capabilities of a learning from demonstration frameworkcapturing a human controller.

[img] PDF
12MB
Abstract:This exploratory study investigates whether it is possible to capture a human controller with a non-reinforcement learning from demonstration framework for closing a valve task. Based on literature, GMM/GMR was deemed to be most suitable framework to capture a human controller. The human controller for the valve task was modelled as P(D)- controller, admittance systems and corresponding feedback loops. The GMM is learned with five demonstrations with random initial conditions in a range between 0 and 2π rad. The interpolation and extrapolation capabilities were tested by reproducing the respective controller over a range of initial conditions both inside and outside the range of 0 and 2π rad. The reproduction performance of the GMM/GMR was evaluated by visual inspection of the error and the RMSE between the demonstration of the model (ground truth) and reproduction. It lead to the result that GMM/GMR is capable of learning and reproducing the PD-controller, but with a limited extrapolation capability. The extrapolation is limited by the spread of the learning data. On the contrary, the GMM/GMR was not capable to capture the admittance model due to a singularity in the input-output relation. To conclude: GMM is not suitable to capture a human controller and different methods should be used to capture a human controller such as reinforcement learning or system identification.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:50 technical science in general, 54 computer science
Programme:Biomedical Engineering MSc (66226)
Link to this item:https://purl.utwente.nl/essays/94739
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page