Risk Analysis Method for a Lane Change Maneuvers on Highways

Abstract

One important aspect of decision making algorithms for the autonomous driving is to behave like a human driver for two main reasons: to be understandable by the other road users, and not to cause discomfort to passengers. To this end, it is necessary to understand how the driver asses risk while driving. In this work, a risk assessment method is proposed inspired by human drivers and specific to lane changing maneuvers. For this contribution, a MOOVE dataset is used, a proprietary dataset of the VEDECOM institute. The primary assumption is that for a given lane changing maneuver, the maneuver is a less dangerous with a more samples performed in riskier situations in the dataset. The purpose of this work consists of developing a decision algorithm based on the proposed risk analysis method to decide whether a lane change maneuver should be made and how.