A motion cueing algorithm (MCA) not only maintains the simulator within its physical limits but also generates such movements of the driving simulator that the necessary motion cues of drivers on the realistic vehicle are equivalently reproduced. The offline optimal MCA focuses on finding the best combination of the translational acceleration and tilt angles of the motion platform to maintain drivers’ motion perception. However, the best combination depends on the MCA’s parameters, tuned mainly by trial and error with experts in the loop. Moreover, for different amplitude input signals, the parameters are accordingly modified. This manually tuning procedure is so time-consuming that the generic optimization, named Mean-Variance mapping optimization, was proposed to search the suitable parameters for the optimal algorithm. This tuning method uses the specific cost function of constraint conditions such as workspace limits, avoiding false cues, and improving motion fidelity to achieve the best parameters for optimal MCA with the particular input signals.
