Answers are grounded in this paper and cite their sources.
Compared to hydraulic braking systems, estimation of tyre-road friction coefficient is more challenging for pneumatic braking systems due to their more complex dynamics. For the first time, model-free machine learning estimators have been developed to determine the peak available friction between a vehicle’s wheels and the stopping surface for a pneumatic braking system. These three model-free estimators – a support vector machine, a shallow neural network and a convolutional neural network – were compared against a model-based sliding-mode observer. In tests on unseen, real-world data gathered from a truck on a test track, the model-free estimators outperformed the sliding-mode observer by 33.3%, 38.3% and 34.6%, respectively. • Machine learning outperforms model-based methods in friction estimation for heavy vehicles. • Shallow neural network reduces error by 38.3% compared to sliding-mode observer. • Real-time friction estimation using only existing pneumatic braking system data. • Tested on real-world data from heavy vehicle emergency braking scenarios.
No takes yet. Share an insight, caveat, or question.
Midgley et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: