Abstract:
Dynamic driving simulators have nowadays become common devices in the automotive field, their aim possibly being virtual prototyping, training, medical studies, R&D. A cr...Show MoreMetadata
Abstract:
Dynamic driving simulators have nowadays become common devices in the automotive field, their aim possibly being virtual prototyping, training, medical studies, R&D. A crucial point is the capability of reproducing a driving feeling as realistic as possible to the final user. This is the task of the Motion Cueing Algorithm (MCA), which calculates the motion displacements to be provided to the motion controller to achieve the better performance in terms of driver perception, while assuring not to exceed the working operational area. Model Predictive Control (MPC) has been successfully applied to MCA, being well suited for the particular task where an optimal problem is posed with the need for fulfillment of physical boundaries. Nevertheless, the predictive feature of the algorithm has been exploited only in a trivial way, due to the evident difficulties in predicting the driver behavior for a significantly long time window, and the hard real-time requirement. In this paper, we present a solution to apply real-time, realistic prediction to a MCA based MPC, where the variability introduced by the driver is taken into account. In particular, the proposed technique can effectively handle possible unexpected driver's behavior and can be adapted to the driving ability of the specific user.
Published in: 2016 American Control Conference (ACC)
Date of Conference: 06-08 July 2016
Date Added to IEEE Xplore: 01 August 2016
ISBN Information:
Electronic ISSN: 2378-5861