EMMA4Drive : a digital human model for occupant simulation in dynamic driving maneuvers

Abstract

About two-thirds of all German workers currently commute to their workplace by car without being able to engage in any meaningful activity. In the future, occupants of autonomous driving vehicles will be able to perform new activities, such as regeneration exercises, working, or consuming entertainment media [1]. This leads to many new challenges in assessing the layout of vehicle interiors, both in terms of comfort and vehicle safety. The goal of the EMMA4Drive project is to further develop the muscle-activated multi-body human model EMMA (Ergo-dynamic Moving Manikin) for use in next-generation partially or fully autonomous driving vehicles. The resulting software prototype EMMA4Drive will be able to analyze and evaluate safety and ergonomics equally during driving maneuvers under dynamic loads as a digital image of the occupant. In the automotive industry, digital human models (DHM) are widely used to simulate the human driver in the early stages of product development. Furthermore, detailed finite element (FE) models of the human body are used to simulate the highly dynamic impact and resulting injuries in the human body in crash simulations [2]. Moreover, DHM based on multibody system (MBS) kinematics are widely applied in reachability investigations and (posture- based) ergonomic assessment of the driver [3]. However, to predict active movement in dynamic driving maneuvers such as cornering, sudden braking, or lane change and pre-crash scenarios, neither FE nor simple MBS kinematic models are applicable. For a more detailed overview on DHMs in this application case, we refer to [4]. In this work, we will present an approach for the enhancement of a multibody based DHM to generate human like motions for a highly dynamic driving simulation. For motion prediction, an optimal control problem (OCP) is set up and solved. A model order reduction (MOR) approach is used to transfer driver seat interaction from detailed FE simulations to the Optimal Control (OC) framework. We use our RODOS driving simulator [5] to validate simulation results (e.g., motions, seat pressure distribution) and to identify OC parameter configurations for motion prediction.