Determining vehicle pre-crash speed in frontal barrier crashes using artificial neural network for intermediate car class. (March 2020)

• Record Type: Journal Article
• Authors: Mrowicki, Adam; Krukowski, Mateusz; Turoboś, Filip; Kubiak, Przemysław
• Is Part Of: Forensic science international. Volume 308(2020)
• Publication Date: 2020-03
• Subjects: Neural networks -- Car crash reconstruction -- Car accidents; Medical jurisprudence -- Periodicals; Chemistry, Forensic -- Periodicals; Forensic Medicine -- Periodicals; Médecine légale -- Périodiques; Chimie légale -- Périodiques; Gerechtelijke geneeskunde; Gerechtelijke chemie; Gerechtelijke psychiatrie; Chemistry, Forensic; Medical jurisprudence; Electronic journals; Periodicals; Electronic journals; 614.1
• DOI: 10.1016/j.forsciint.2020.110179
• Languages: English
• ISSNs: 0379-0738

Abstract

Highlights: Authors use artificial neural network to find the precrash velocity in intermediate car class. This approach is showing improvement over linear method-9% vs up to 18%. The perceptron can be gradually improved and fine tuned producing even more accurate estimations. Abstract: This paper introduces a new, innovative approach to pre-crash velocity determination, namely the artificial neural networks. A perceptron based on a database obtained from NHTSA (National Highway Traffic Safety Administration) with numerous data concerning frontal vehicle crash tests: i.e. vehicle mass, deformation zone and deformation coefficients C1 -C6 . Part of the database entries were used to train the network to develop consistent accuracy and the remainder was used as validation and training sets.