Automated driving

​The aim of the project is to develop a tactical framework for decision-making regarding how to interact with surrounding traffic. The objective of the tactical framework is thus to provide a generic model that captures how relevant cues in the traffic environment can be interpreted into a traffic situation, and subsequently determines an appropriate course of action in that situation. The situation interpretation, and corresponding decision-making process, should be fully automated in terms of not relying on the driver to provide additional context- related information or overrule the selected actions. The behaviour selection should also not be dependent on any lead vehicle. As such, the framework will provide a scheme for fully automated drive in terms of not relying on a driver or a lead vehicle in order to interpret the traffic information and perform appropriate action selection. ​An important feature of the framework is that it should be scalable in order to handle many different types of traffic scenarios and driving-styles. However, for the purpose of illustration, in this project, the target application will be low-speed driving in e.g. highway traffic jams or car park manoeuvring. Nonetheless, irrespective of the scenario, the vehicle should be able to travel in a fully automated manner while obeying the traffic rules, and handling some exceptional situations, in order to allow for an appropriate, safe, and comfortable automated driving-style, suitable for the everyday driver. The research project can be divided into two main parts, namely: Model the traffic situation - given the current state of relevant objects e.g. other traficants (vehicles and pedestrians), traffic signs, rules of the road etc. in the surrounding traffic environment, interpret their respective significance into a current traffic situation and make a probabilistic estimation regarding how the traffic situation will evolve in a future time horizon. Interact with/regulate the traffic situation - given the traffic situation, determine an appropriate behavioural strategy and realize this behaviour by controlling the vehicle. The tactical framework will function as an intermediate layer of a higher strategic layer and a lower execution layer. Per se, the input to the tactical level will be the planned route from start to destination, as well as additional road information such as curvature, number of files, type of road, speed limit, etc. The tactical layer will also be given input by means of perception e.g. through the vehicle's sensors. This information will include the state of relevant objects e.g. other traficants and traffic signs, in the surrounding environment. The output of the tactical layer will be a state trajectory defining the desired position and velocity profile of the vehicle.

Partner organizations

  • Volvo Cars (Private, Sweden)
Start date 01/12/2011
End date The project is closed: 30/11/2016

Page manager Published: Thu 31 May 2018.