IPython Notebook and supporting Python files for my Bachelor Thesis.
The Thesis presents a system that endows a social robot with the capacity of actively want to learn about novel stimuli presented to it, expanding its knowledge base. The data provided for the evaluation of the system comes from another experiment developed in the Department of Automation and Systems Engineering at Universidad Carlos III de Madrid in the field of gesture and pose recognition (https://github.com/VGonPa/datasets-poses2012).
The notebook used for this Thesis can be viewed onlive here: http://nbviewer.ipython.org/github/UC3MSocialRobots/novelty-detection-in-hri/blob/master/Novelty%20Detection%20system%20for%20a%20Social%20Robot.ipynb
- Download, and save in the same folder:
lsanomaly.py form http://cit.mak.ac.ug/staff/jquinn/software/lsanomaly.html
data-torso-relative from https://github.com/VGonPa/datasets-poses2012 and save in a folder as 'data'
all files from folder 'helper files'
-
Open 'Novelty Detection system for a Social Robot.ipynb' to see the system working
-
Open EXPERIMENTS to see the automated tests for the novelty detection system
'excels' include spreadsheets used to present how many pose entries were needed to pass the noise filter. They were used for the Results section in my Thesis.