DNN-Based Prediction of Standard Driving Posture for Vehicle Takeover

When drivers need to take over a vehicle during shared autonomy, the standard driving postures based on their body size are basis of non-driving posture (NDP) motion reconstruction. This study focused prediction using deep learning neural network (DNN) method. Firstly, main factors influencing were extracted through qualitative analysis, and weights analyzed an orthogonal test Based this, parameters model determined. Secondly, point cloud data typical vehicles market obtained laser scanning. After extracting key input output required for processing feature matching, dataset was established. Finally, supervised established predict different under package layouts. method allows quick evaluation corresponding activities, laying foundation risk-level assessment reconstruction in takeover. The results show that trained algorithm can with high accuracy robustness.