You can view the full text of this article for free using the link below.

Title: System identification of rover dynamics: a comparison of three model structures

Authors: Christina M. Ivler; Norma Gowans; Cole Marfise

Addresses: University of Portland, Portland, OR, 97217, USA ' University of Portland, Portland, OR, 97217, USA ' University of Portland, Portland, OR, 97217, USA

Abstract: Dynamic testing of a small instrumented ground vehicle was conducted in order to identify an accurate state-space model for simulation of autonomous vehicles. This paper describes the application of frequency domain system identification to model the yaw/steering response of a small-scale rover. Several model structures of varying levels of complexity were adapted and applied to a small-scale rover: dynamic bicycle model, roll-yaw model and a lumped (based on Taylor series expansion) model. In comparing these three model structures, it was found that the dynamic bicycle model provided a simple model structure with good performance but cannot model roll dynamics. The roll-yaw model gave the most accurate model and better prediction for a range of vehicle speeds but is significantly more complex. Finally, the lumped model gave a highly accurate model at the identified speed condition; however, it cannot be accurately extrapolated to other speeds.

Keywords: system identification; rover; steering dynamics; bicycle model; frequency domain methods; frequency sweeps; roll-yaw model; ground vehicle; coherence weighted; composite windowing; identification of inertia.

DOI: 10.1504/IJMIC.2022.10048800

International Journal of Modelling, Identification and Control, 2022 Vol.40 No.1, pp.70 - 83

Received: 10 Jun 2021
Accepted: 26 Aug 2021

Published online: 12 Jul 2022 *

Full-text access for editors Access for subscribers Free access Comment on this article