Acausal Approach to Motion Cueing

Abstract

Motion simulators have been used extensively by both industry and academia to train pilots, conduct psychological experiments on drivers, understand the perception of motion by humans, and cater to the burgeoning gaming industry among others. Working of a motion simulator can be summarized in following three steps: first, acquisition of motion signals; second, motion cueing: signal processing to generate motion references, and third, control: tracking the desired references. A motion cueing algorithm (MCA) acts as a bridge between the actual motions and the ones recreated by the simulator. Mathematically, MCA is constituted of the following four operations: scaling, saturation, filtering, and tilt-coordination (involves filtering too). The existing MCAs make use of causal filters to process the signals, thereby precluding the possibility of utilizing future motion signals to emulate pre-recorded scenarios. We present a new approach to generate motion cues by explicitly making use of future motion signals and causal linear filters. It is due to the usage of future motion signals (not the filter), we call the presented methodology as acausal cueing algorithm (ACA). Unlike most of the existing works on motion cueing, we choose to present the developed methodology using discrete-time models to facilitate its quick implementation by industry and other researchers in the future. The veracity of the presented methodology is examined by actuating a motion simulator (seven degrees of freedom parallel manipulator) based on the references generated by the ACA in response to test trajectories. The conducted experiments assert better performance of the ACA (over MCA) in the beginning, which eventually degrades in the last few seconds due to unavailability of future motion signals.