Experiments of RAL-19-307

This repo contains code and dataset for the first experiment described in our submission to RAL-19-0307. This repository allows to re-run the experiments described in the Section 4A. The access to the additional experiments and data will be released after the reviews will be concluded or on request.

Prerequisities

The software was tested on Ubuntu 16.04 with openCV2.7 and Python 2.7.12 python-pandas 0.17.1, python-numpy 1.11.0 and python-opencv 2.4.9.1.

Compiling and running

To run it, do the following

1. Install the required packages:

• sudo apt install g++ make libopencv-dev python-opencv python-dev python-numpy python-pandas libalglib-dev transfig gnuplot imagemagick graphviz

2. Compile the test codes:

• cd Hypertime-RAL-19-0307/door_state/src
• make

3. Run the tests **.cd ../eval_scripts/ **./process_dataset.sh greg_door_2016_min

4. Check the results

• make
• mkdir tmp
• ./summarize_results.sh greg_door_2016_min
• display summary.png

Additional, custom models

The temporal models are stored in the src/models folder and by implementing the methods of CTemporal.cpp, you can make and test your own method. To re-run the experiments described in Section 4A, with your new model, modify the test models file accordingly. Then go to the src folder and type make to compile the predictive framework. After that, go to the eval_scripts folder and type make to compile the t-test utility.

Then, type:

./process_dataset.sh greg_door_2016_min

if it runs well, type

./summarise_results.sh greg_door_2016_min

and then check the summary.png file, which should look similarly to Fig 2 in the paper.