Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing

Ignat Georgiev, Christoforos Chatzikomis, Timo Voelkl, Joshua Smith, Michael Mistry
Proceedings of the 2020 Conference on Robot Learning, PMLR 155:552-563, 2021.

Abstract

Accurately modeling robot dynamics is crucial to safe and efficient motion control. In this paper, we develop and apply an iterative learning semi-parametric model, with a neural network, to the task of autonomous racing with a Model Predictive Controller (MPC). We present a novel non-linear semi-parametric dynamics model where we represent the known dynamics with a parametric model, and a neural network captures the unknown dynamics. We show that our model can learn more accurately than a purely parametric model and generalize better than a purely non-parametric model, making it ideal for real-world applications where collecting data from the full state space is not feasible. We present a system where the model is bootstrapped on pre-recorded data and then updated iteratively at run time. Then we apply our iterative learning approach to the simulated problem of autonomous racing and show that it can safely adapt to modified dynamics online and even achieve better performance than models trained on data from manual driving.

Cite this Paper

BibTeX
@InProceedings{pmlr-v155-georgiev21a, title = {Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing}, author = {Georgiev, Ignat and Chatzikomis, Christoforos and Voelkl, Timo and Smith, Joshua and Mistry, Michael}, booktitle = {Proceedings of the 2020 Conference on Robot Learning}, pages = {552--563}, year = {2021}, editor = {Kober, Jens and Ramos, Fabio and Tomlin, Claire}, volume = {155}, series = {Proceedings of Machine Learning Research}, month = {16--18 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v155/georgiev21a/georgiev21a.pdf}, url = {https://proceedings.mlr.press/v155/georgiev21a.html}, abstract = {Accurately modeling robot dynamics is crucial to safe and efficient motion control. In this paper, we develop and apply an iterative learning semi-parametric model, with a neural network, to the task of autonomous racing with a Model Predictive Controller (MPC). We present a novel non-linear semi-parametric dynamics model where we represent the known dynamics with a parametric model, and a neural network captures the unknown dynamics. We show that our model can learn more accurately than a purely parametric model and generalize better than a purely non-parametric model, making it ideal for real-world applications where collecting data from the full state space is not feasible. We present a system where the model is bootstrapped on pre-recorded data and then updated iteratively at run time. Then we apply our iterative learning approach to the simulated problem of autonomous racing and show that it can safely adapt to modified dynamics online and even achieve better performance than models trained on data from manual driving.} }
Endnote
%0 Conference Paper %T Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing %A Ignat Georgiev %A Christoforos Chatzikomis %A Timo Voelkl %A Joshua Smith %A Michael Mistry %B Proceedings of the 2020 Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2021 %E Jens Kober %E Fabio Ramos %E Claire Tomlin %F pmlr-v155-georgiev21a %I PMLR %P 552--563 %U https://proceedings.mlr.press/v155/georgiev21a.html %V 155 %X Accurately modeling robot dynamics is crucial to safe and efficient motion control. In this paper, we develop and apply an iterative learning semi-parametric model, with a neural network, to the task of autonomous racing with a Model Predictive Controller (MPC). We present a novel non-linear semi-parametric dynamics model where we represent the known dynamics with a parametric model, and a neural network captures the unknown dynamics. We show that our model can learn more accurately than a purely parametric model and generalize better than a purely non-parametric model, making it ideal for real-world applications where collecting data from the full state space is not feasible. We present a system where the model is bootstrapped on pre-recorded data and then updated iteratively at run time. Then we apply our iterative learning approach to the simulated problem of autonomous racing and show that it can safely adapt to modified dynamics online and even achieve better performance than models trained on data from manual driving.
APA
Georgiev, I., Chatzikomis, C., Voelkl, T., Smith, J. & Mistry, M.. (2021). Iterative Semi-parametric Dynamics Model Learning For Autonomous Racing. Proceedings of the 2020 Conference on Robot Learning, in Proceedings of Machine Learning Research 155:552-563 Available from https://proceedings.mlr.press/v155/georgiev21a.html.

Related Material