Citation

Iwamoto M, Min K, Kimpara H. Trans. Soc. Automot. Eng. Jpn. 2012; 43(4): 905-910.

Copyright

(Copyright © 2012, Society of Automotive Engineers of Japan)

DOI

10.11351/jsaeronbun.43.905

PMID

unavailable

Abstract

Muscular control algorithm using reinforcement learning, which is supposed to be a mathematical model of learning process in the basal ganglia associated with human postural control was developed to predict occupant postures in pre-crash which could affect occupant injury outcomes in a crash. The algorithm was applied to a musculo-skeletal finite element (FE) model of human whole body. Simulations using the FE model under a condition with a brake deceleration predicted occupant postures in pre-crash.

Language: ja

Keywords

anthropomorphic dummy; muscle activation; pre-crash; reinforcement learning; safety