Rule-based highway maneuver intention recognition

Paper in proceeding, 2015

Future advanced driver assistance systems or fully automated driving systems require an increased ability to classify and interpret traffic situations in order to appropriately account for, and react to, the behavior of surroundings vehicles. When driving on a highway, humans are able to recognize the maneuver intentions of surrounding vehicles by observing lateral and longitudinal motion cues. The main idea presented in this paper is to adapt this ability to technical systems by formulating simple logic rules for intention recognition of highway maneuvers. As such, the presented algorithm is able to recognize the intention of left and right lane change maneuvers with an accuracy of approximately 89% while maintaining the false positive rate low at approximately 3%. Further, due to the algorithm’s low computational complexity, flexibility, and straight-forward design based on easily comprehensible logic rules, the proposed method fulfills the requirements of future advanced driver assistance systems or fully automated driving systems.