Please find here supplementary code for the paper: Online validity monitor for vehicle dynamics models; Stephan Rhode and Johannes von Keler; 2019 IEEE International Conference on Connected Vehicles and Expo (ICCVE)
@InProceedings{Rhode2019,
Title = {Online validity monitor for vehicle dynamics models},
Author = {S. {Rhode} and J. {Von Keler}},
Abstract = {This paper presents a method to measure the degree of validity of vehicle
dynamics models during run time in a computational efficient way. The
method is called online validity monitor and is constructed as follows.
First, the characteristic validation error distribution of a vehicle
dynamics model is stored. Second, during run time, an error distribution is
recorded via a circular buffer. Third, the distance between stored and run
time error distribution is taken as a measure to detect if the model
operates in its validated domain.}
Booktitle = {2019 IEEE International Conference on Connected Vehicles and Expo (ICCVE)},
Year = {2019},
Month = {Nov},
Pages = {1-6},
Doi = {10.1109/ICCVE45908.2019.8965162},
ISSN = {2378-1289},
Keywords = {Model validation;online;vehicle dynamics;Jensen-Shannon divergence}
}
This software is a research prototype, solely developed for and published as part of the publication. It will neither be maintained nor monitored in any way.
You need a python environment with toolboxes defined in
requirements.txt. You can either use pip
or conda to create this environment. Then, simply run experiment.py
to reproduce result plots of the paper.
This software is licensed under AGPL-3.0 license.