Linear System Identification of Longitudinal Vehicle Dynamics Versus Nonlinear Physical Modelling

James, Sebastian S.; Anderson, Sean

doi:10.1109/control.2018.8516756

Cited by 14 publications

(12 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The main contribution of this paper is thus to make a comparative evaluation of longitudinal vehicle dynamics between physical and data-driven models, using experimental data collected across large-scale real-world driving (the 'ordinary domain') using two different vehicles. This work extends a recent conference paper by two of the authors into linear versus nonlinear modelling of longitudinal vehicle dynamics [28].…”

Section: Introductionsupporting

confidence: 60%

“…The first route [Lancia Delta car- Fig. 1(a)] began at Centro Ricerche Fiat in Orbassano, near Turin (Italy), then went to Pinerolo, then Piossasco before returning to Orbassano, a distance of 53 km driven in about 42 minutes, and is the same data as described in [8] and as used in our previous work [28]. The Jeep Renegade was driven along a loop of approximately 25 km similar to the first route [using some of the same roads- Fig.…”

Section: A Experimental Datamentioning

confidence: 99%

See 1 more Smart Citation

Longitudinal Vehicle Dynamics: A Comparison of Physical and Data-Driven Models Under Large-Scale Real-World Driving Conditions

2020

Self Cite

View full text Add to dashboard Cite

Mathematical models of vehicle dynamics will form essential components of future autonomous vehicles. They may be used within inverse or forward control loops, or within predictive learning systems. Often, nonlinear physical models are used in this context, which, though conceptually simple (especially for decoupled, longitudinal dynamics), may be computationally costly to parameterise and also inaccurate if they omit vehicle-specific dynamics. In this study we sought to determine the relative merits of a commonly used nonlinear physical model of vehicle dynamics versus data-driven models in large-scale real-world driving conditions. To this end, we compared the performance of a standard nonlinear physical model with a linear state-space model and a neural network model. The large-scale experimental data was obtained from two vehicles; a Lancia Delta car and a Jeep Renegade sport utility vehicle. The vehicles were driven on regular, public roads, during normal human driving, across a range of road gradients. Both data-driven models outperformed the physical model. The neural network model performed best for both vehicles; the state-space model performed almost as well as the neural network for the Lancia Delta, but fell short for the Jeep Renegade whose dynamics were more strongly nonlinear. Our results suggest that the linear data-driven model gives a good trade-off in accuracy and simplicity, whilst the neural network model is most accurate and is extensible to more nonlinear operating conditions, and finally that the widely used physical model may not be the best choice for control design.

show abstract

Section: Introductionsupporting

confidence: 60%

Section: A Experimental Datamentioning

confidence: 99%

Longitudinal Vehicle Dynamics: A Comparison of Physical and Data-Driven Models Under Large-Scale Real-World Driving Conditions

2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Of the many other scalar metrics that may be used to evaluate the quality of model predictions (e.g. model fit, variance accounted for v.a.f., Akaike's Final Prediction Error FPE [13]), we calculate here the variance accounted for metric, which evaluates the variance of the residuals with respect to the variance of the measured signal (consistently with [14]…”

Section: Convolutivementioning

confidence: 99%

Modelling longitudinal vehicle dynamics with neural networks

Lio

Bortoluzzi

Papini

2019

Vehicle System Dynamics

View full text Add to dashboard Cite

This paper studies neural network models of vehicle dynamics. We consider both models with a generic layer architecture and models with specialized topologies that hard-wire physics principles. Network pre-wiring is limited to universal laws; hence it does not limit the network modelling abilities on one side but allows more robust and interpretable models on the other side. Four different network types (with and without pre-wired structure, recursive and non-recursive) are compared for the longitudinal dynamics of a car with gears and two controls (brake and engine). Results show that pre-wiring effectively improves the performance. Non-recursive networks also look to be preferable for several reasons.

show abstract

“…It uses statistical methods to build a dynamical system's mathematical models from measured input and output data [16]. The simplest way to get started on a parametric estimation routine is to build a state-space model where the modelorder is automatically determined [17].…”

Section: Introductionmentioning

confidence: 99%

Analysis on the Performance of a Second-order and a Third-order RLC Circuit of PRBS Generator

Faudzi

Zhang

2021

JIAE

View full text Add to dashboard Cite

A pseudo-binary random signal (PRBS) has been widely utilized for system identification in complex signals to develop an experimental approach. PRBS generator is a circuit that generates pseudo-random numbers. This work aims to analyze the best fit value of the PRBS generator with second-order and third-order under-damped black-box RLC circuit of the estimated model. The procedures conducting here can be divided into three parts. First, to design two black boxes using the RLC circuit representing a critically under-damped second-order and third-order system. PRBS generated with maximum-length sequence (MLS) equals 127 bits by using seven shift registers. Second, simulate the PRBS generator using MATLAB software and validate the estimated model from the simulation using the System Identification Tool in MATLAB. Next, connecting hardware RLC circuit and reading input and output signals using an oscilloscope. Finally, 2500 samples of captured data were used for estimation. Then, analyze and compare the best fit of the simulation and experiment with second-order and third-order under-damped black-box RLC circuit. Furthermore, analyze and compare best fit using different sample time. The results showed that the best fit of the second-order model with under-damped black-box RLC circuit was autoregressive with the exogenous term (ARX) 211, where the best fit of the simulation was 99.88%, and the best fit of the experiment was 96.04%. And the results showed that the best fit of the third-order model with an under-damped black-box RLC circuit was ARX 331, where the best fit of the simulation was 99%, and the best fit of the experiment was 94.28%. It was concluded that the best fit value of the second-order was better than the third order. What’s more, the results showed that when the select range is the same, the bigger the sample time, the better the best fit.

show abstract

Linear System Identification of Longitudinal Vehicle Dynamics Versus Nonlinear Physical Modelling

Abstract: This is a repository copy of Linear system identification of longitudinal vehicle dynamics versus nonlinear physical modelling.

Cited by 14 publications

References 15 publications

Longitudinal Vehicle Dynamics: A Comparison of Physical and Data-Driven Models Under Large-Scale Real-World Driving Conditions

Longitudinal Vehicle Dynamics: A Comparison of Physical and Data-Driven Models Under Large-Scale Real-World Driving Conditions

Modelling longitudinal vehicle dynamics with neural networks

Analysis on the Performance of a Second-order and a Third-order RLC Circuit of PRBS Generator

Contact Info

Product

Resources

About