Safety margins in complex environments

Research Project, 2016 – 2020

Within a few years, the first unsupervised self-driving passenger vehicles will start driving on public roads. To begin with, these vehicles will only be able to drive unsupervised on a very limited set of roads for which they are specifically developed and tested. To make them available in more diverse and complex environments they must be able to drive safely while interacting closely with other road users. This research project focuses on developing methods for automatically assessing how self-driving vehicles must be able to interact with other road users in different types of traffic environments. Knowing how to interact with others is a prerequisite for driving autonomously, otherwise the vehicle may obstruct the traffic flow or even become unsafe. A self-driving vehicle will only be able to drive in a new environment if its safety margins are both sufficient for safety and are consistent with margins used by other road users. The outcome from the work in this project is expected to be used in the DriveMe project and to have additional applications within collision avoidance and supervised automotive applications