Modelling longitudinal vehicle dynamics with neural networks

Lio, Mauro Da; Bortoluzzi, D.; Papini, Gastone Pietro Rosati

doi:10.1080/00423114.2019.1638947

Cited by 42 publications

(35 citation statements)

References 8 publications

(7 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Properly biased networks have fewer parameters than unbiased ones, which are best focused on what is necessary to approximate the system. A similar conclusion is found in [36], where neural networks with structure informed by the physical problem are more robust, accurate, and efficient than general-purpose networks.…”

Section: ) Modelingsupporting

confidence: 78%

“…In the Newtonian dynamics, the acceleration of a multibody system is the superposition of the individual forces. Each force has its distinct causes (for example, the brake force depends on the brake pedal pressure but not on the gas pedal pressure [36]). Decomposing the plant into the combination of simpler sub-plants (the force effects), each with reduced dimensionality input, is a very efficient and robust modeling approach [36].…”

Section: ) Modelingmentioning

confidence: 99%

See 1 more Smart Citation

A Mental Simulation Approach for Learning Neural-Network Predictive Control (in Self-Driving Cars)

et al. 2020

Self Cite

View full text Add to dashboard Cite

This paper presents a novel approach to learning predictive motor control via "mental simulations". The method, inspired by learning via mental imagery in natural Cognition, develops in two phases: first, the learning of predictive models based on data recorded in the interaction with the environment; then, at a deferred time, the synthesis of inverse models via offline episodic simulations. Parallelism with human-engineered control-theoretic workflow (mathematical modeling the direct dynamics followed by optimal control inversion) is established. Compared to the latter human-directed synthesis, the mental simulation approach increases autonomy: a robotic agent can learn predictive models and synthesize inverse ones with a large degree of independence. Human modeling is still needed but limited to providing efficient templates for the forward and inverse neural networks and a few other directives. One could consider these templates as the efficient brain network typologies that evolution produced to permit live beings quickly and efficiently learning. The structure of the neural networks-both forward and inverse ones-is made of interpretable "local models", which follows the cerebellar organization (and are also similar to local model approaches known in the literature). We demonstrate the learning of a first-round model (contrasted to Model Predictive Control) for lateral vehicle dynamics. Then, we demonstrate a second learning iteration, where the forward/inverse neural models are significantly improved.

show abstract

Section: ) Modelingsupporting

confidence: 78%

Section: ) Modelingmentioning

confidence: 99%

A Mental Simulation Approach for Learning Neural-Network Predictive Control (in Self-Driving Cars)

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…The network adopts a structure that is inspired by the physical model: there are 4 converging sub-networks (labelled as 1, 2, 3 & 4) that are intended to learn the effects of F p , F b , Mg sin(γ (t)) and k D v 2 + Mg k R separately. More details on the structured network approach are given in [26].…”

Section: Neural Network Modelmentioning

confidence: 99%

“…The alternative approach of data-driven learning of the vehicle forward dynamics model resolves many of the issues encountered with physical models. This may be as simple as learning a suitable linear state-space representation [20], combining local, linear models [21], [22] or more complex nonlinear approaches based on neural networks [23]- [26]. An advantage of this data-driven approach is that, if data collection continues during operation of the vehicle, multiple models for different operational conditions can be automatically generated offline (following an approach that mimics the human wake-sleep stages in the learning of motor control [27]).…”

Section: Introductionmentioning

confidence: 99%

“…based on Newton's second law of motion with forces due to engine propulsion, braking, air drag, static friction and the effect of gravity on inclined roads. For data-driven approaches a linear state-space model was identified using a prediction error method (PEM) with initialisation by subspace identification [30], and a neural network model was also identified [31], based on capturing the nonlinear impulse response due to the propulsion, braking and road gradient inputs [26]. To identify the models, we used experimental data that were collected using a typical medium-sized car and a sport utility vehicle driven along a varied urban and extra-urban route.…”

Section: Introductionmentioning

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

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

Longitudinal deep truck: Deep longitudinal model with application to sim2real deep reinforcement learning for heavy‐duty truck control in the field

Albeaik

Vurimi

et al. 2022

Journal of Field Robotics

View full text Add to dashboard Cite

To develop cooperative adaptive cruise control (CACC), the choice of control approach often influences and can limit the choice of model structure, and vice versa. For heavy-duty trucks, practical application of CACC in the field is heavily influenced by the accuracy of the used model. Deep learning and deep reinforcement learning (deep-RL) have recently been used to demonstrate improved modeling and control performance for vehicles such as cars and quadrotors compared to state-of-the-art. The literature on the application of deep learning and deep-RL for heavy-duty trucks in the field, which are significantly more complex than cars, is still sparse, however. In this article, we develop a two-layer gray-box deep learning model to capture longitudinal dynamics of heavy-duty trucks while abstracting their complexity and present an approach to properly break the nested feedback loops in the model for training. We compare this model with three other alternative models and show that it achieves x ~10 better general performance compared to a standard artificial neural network and results in x ~4 and x ~40 slower steady-state acceleration and speed error growth rates, respectively. We then present an architecture to utilize these deep learning models within the deep-RL framework and use it to develop baseline CACC controllers that can be zero-shot transferred to the field. To carry out the work, we present a setup of differently configured trucks along with their interface architecture and stochastic driving cycle generators for data collection. Numerical validation of the approach demonstrated stationary and bounded modeling error, and demonstrated transfer of CACC controllers with consistent overshoot bounds and a stable approximately-zero steady-state error. Validation from field experiments demonstrated similarly consistent results. Compared to a state-of-the-art benchmark, the deep-RL controller achieved lower speed and time-gap error variance but higher time-gap error offset.

show abstract

Modelling longitudinal vehicle dynamics with neural networks

Cited by 42 publications

References 8 publications

A Mental Simulation Approach for Learning Neural-Network Predictive Control (in Self-Driving Cars)

A Mental Simulation Approach for Learning Neural-Network Predictive Control (in Self-Driving Cars)

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

Longitudinal deep truck: Deep longitudinal model with application to sim2real deep reinforcement learning for heavy‐duty truck control in the field

Contact Info

Product

Resources

About