Safety Prompt Advanced Driver-Assistance System with Lane-Change Prediction and Free Space Detection

Abstract

With the map navigation system achieving the positioning accuracy at the lane level, the lane level navigation system began to be applied in digital navigation software. The new generation of navigation system allows human drivers to accurately identify road conditions without road experience. However, in actual use, drivers still need to complete lanechanging behavior independently. The navigation system does not provide relevant real-time traffic flow information and safety prediction. Therefore, this paper innovatively combines the free space detection and lane change prediction algorithm in actual driving, and puts forward this Safety Prompt Advanced Driver-Assistance System (SPADAS). It can predict the lanechange decision and lane-change trajectory around in a short period of time in the future according to the current state and historical trajectory of the surrounding vehicles, and combine these prediction information to provide the driver or the path planning module of the self-driving vehicle with the predicted free space in the near future. The free space provided by this Advanced Driver-Assistance System (ADAS) includes three types: continuing driving according to the current state or changing lanes to the left or right. At the same time, it also provides the required recommended driving trajectory. The evaluation test of this ADAS has been done through a simulation testbed of the NGSIM dataset built with Eclipse SUMO.