Modeling of occupant behavior under lateral vehicle motion (Proposing of control laws based on human posture control mechanism and experimental validation)

Abstract

To achieve simple and accurate modeling of human-like behavior with the aim of realizing a model-based design (MBD) methodology incorporating vehicle occupants, this research applied a simplified two-mass system for the upper body of a vehicle occupant and control laws derived based on natural human characteristics. To express the behavior of a seated occupant in response to lateral inputs, the motion of the upper body with respect to the vehicle and the motion of the head with respect to the upper body are modeled using a two-mass system projected onto a plane that includes the horizontal and vertical vehicle axes. Torques to maintain the posture calculated by the control laws were applied to each joint. This research focused on somatic and vestibular sensations as natural human characteristics. Somatic sensations are a source of information concerning the relative spatial and positional relationships between the body and its supporting surfaces (in this research, a vehicle). Seated occupant behavior to hold the upper body upright in the vehicle was defined as somatic feedback, which calculates torque based on the upper body posture angle. On the other hand, vestibular sensations are a source of information concerning the acceleration applied to the head. The vestibular system contains two types of receptor organs: the semicircular canals and the otoliths. The semicircular canals are the receptor organs for rotational acceleration of the head and are sensitive to rapid head motions. The otoliths are the receptor organs for linear acceleration of the head and are sensitive to slow head motions. In this research, we focused on the otoliths and otolithic feedback was defined as the tilting motion of the head against lateral acceleration, which generates torque based on the head posture angle and the lateral acceleration of the head. Validation experiment was performed on a test bench using a vibration generator and vehicle tests confirmed that the proposed model is capable of generally reproducing occupant behavior in response to lateral acceleration inputs.

Keywords

• vehicle control
• human model
• passenger model
• multibody dynamics
• model based design