Actuator- Based Optimization Motion Cueing Algorithm

Abstract

A Motion Cueing Algorithm (MCA) estimates driving simulator motions subject to the driver demands. An essential task consists in sticking to the simulator's workspace limits on position, velocity and acceleration level. In this paper the driving simulator comprises a hexapod which is a parallel robot with six degrees of freedom. In contrast to classical MCAs, which are mainly based on filtering and scaling techniques, this paper introduces a new optimization approach which is designed to estimate the simulator motion, subject to the simulator's limitations. Previous optimization based MCAs use a workspace-based prediction model to estimate the resulting reference motions at the driver position over a time horizon. Unlike these approaches, this work applies an actuator-based approach, using the direct kinematics to estimate the reference motions in the workspace. Advantages lie in the direct integration of the actuator constraints on position, velocity and acceleration level. Solving the direct kinematics of the parallel robot used in the optimization is computationally expensive. Thus, two approximations of the direct kinematics are introduced; this leads to significant reductions in the computational time, while showing only small deviations from the exact kinematics.