Hybrid-Learning-Based Driver Steering Intention Prediction Using Neuromuscular Dynamics

The emerging automated driving technology poses a new challenge to driver-automation collaboration, which requires mutual understanding between humans and machines through their intention identifications. In this article, oriented by human–machine understanding, driver steering prediction method is proposed better understand human driver's expectation during driver–vehicle interaction. predicted based on novel hybrid-learning-based time-series model with deep learning networks. Two different modes, namely, both hands single right-hand are studied. Different electromyography signals from the upper limb muscles collected used for prediction. relationship neuromuscular dynamics torque analyzed first. Then, developed predict continuous discrete intentions. two networks share same temporal pattern exaction layer, built bidirectional recurrent neural network long short-term memory cells. performance evaluated varied history horizon exploit capability further. experimental data 21 participants of ages experience. results show that can achieve accuracy around 95% under modes.