Cooperative Shared Steering Control Based on Driver’s Intent and Contextual Traffic

Abstract

Cooperative shared steering control opens up new perspectives for the driver steering assistance systems which can produce interactive feedback through the steering wheel. This paper addresses cooperative shared steering control method for lane change (LC) maneuver in highway traffic. The aim behind this concept is the challenging issue of driver machine cooperative method. In order to characterize the human driving behaviors, traffic comprehensive decision factor (TCDF) is designed by using the correlation of vehicles in contextual traffic. Based on the LC experiments on a driver-in-loop Platform, driver’s intent recognition method is proposed by a stacked-layer long short term memory (SL-LSTM) algorithm. To achieve a shared steering control during the LC maneuver, a cooperative authority allocation factor is designed to integrate the inputs from driver and shared controller. The reference input of shared controller is dynamically planned according to the driver’s intent and TCDF, and it satisfies the vehicle steering dynamics. Both of the driver’s torque and TCDF jointly determine the cooperative authority allocation factor. Finally, two traffic scenarios are designed to verify the proposed shared steering control, experimental results have shown that the proposed shared steering control algorithm can improve the LC performance with better comfort and safety.