A hybrid motion cueing algorithm

Current closed-loop, optimization-based motion cueing algorithms (MCAs) use a driver model to predict a future driving dynamics reference. These models are often inaccurate and/or computationally expensive because future driving behaviour is unknown. In some cases, the vehicle's trajectory is known in advance. In such so-called open-loop simulations, a driver sits passively in a vehicle and is being driven through a pre-recorded manoeuvre. In this case, optimization-based MCAs can compute an optimal trajectory for a pre-defined manoeuvre in a pre-processing step. This work presents the development of an MCA that uses the optimal trajectory of an open-loop, optimization-based MCA as a reference in a closed-loop simulation, resulting in a quasi-optimal pre-positioning of the motion platform. Deviations between closed-loop driver and the reference are compensated by a closed-loop, state-of-the-art MCA. By combining a closed-loop MCA with the predictions obtained by an open-loop MCA, a hybrid motion cueing algorithm is obtained. One of the challenges faced with the implementation of a hybrid MCA is how to merge the data of the driver with the reference. To this end, a preparatory experiment was performed to measure and analyse the driving behaviour of various drivers. Then, a follow-up experiment was conducted to evaluate the novel hybrid MCA using the continuous rating method in an open-loop simulation. In order to analyse deviations between open-loop and closed-loop rating, a novel rating method for closed-loop simulations was developed and applied. Here, participants gave a section-wise oral rating during a closed-loop drive. Results show correlations between the open-loop and the closed-loop rating method. Both ratings indicate an improvement in motion cueing quality for the hybrid MCA.