Open AccessStability Analysis of Semi-active Suspension Systems Using a Data-driven Approach
Author(s) -
Dániel Fényes,
Balázs Németh,
Péter Gáspár
Publication year - 2021
Publication title -
periodica polytechnica. transportation engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.388
H-Index - 15
eISSN - 1587-3811
pISSN - 0303-7800
DOI - 10.3311/pptr.18597
Subject(s) - carsim , reachability , stability (learning theory) , computer science , decision tree , fidelity , suspension (topology) , software , tree (set theory) , high fidelity , categorization , machine learning , artificial intelligence , data mining , control engineering , control (management) , engineering , algorithm , mathematics , telecommunications , mathematical analysis , homotopy , pure mathematics , electrical engineering , programming language
The modern vehicles are getting equipped with more and more sensors, which allows the engineers to collect more information about the states of the vehicle and its environment during its operation. This information can be used to increase the capacity and the performances of the control systems. In this paper, a novel data-driven approach is presented to compute the reachability sets of the vehicles, which are equipped with a semi-active suspension system. The dataset, which is used in this paper, is provided by the high fidelity vehicle simulation software, CarSim. Firstly, the dataset is categorized using a stability criterion. Then, a machine-learning algorithm (C4.5 decision tree) is trained, which can categorize a given instance using only the onboard signals of the vehicle. Finally, a possible application of the reachability sets is presented to show the use of the computed sets.