Graphical Traffic Scenario Editor

Title Graphical Traffic Scenario Editor
Summary Develop an interactive graphical application to draw vehicle paths and their surrounding environment, for rapid prototyping of traffic scenarios in intelligent vehicle research.
Author Iulian Carpatorea
Supervisor Roland Philippsen
Level Master
Status Finished


The aim of this Master's project is to develop a graphical application that allows to draw vehicle paths and their surrounding environment in order to rapidly design and evaluate scenarios that may arise in traffic situations. This will provide a tool for developers of driver assistance systems to human-machine interface designers. The tool will be kept as simple as possible (e.g. planar geometries) while providing an effective editing and playback GUI (e.g. using spline handles to draw vehicles paths).

Project Description

Advanced driver assistance systems include warning and intervention functions that are aimed at increasing the safety and comfort of vehicles in everyday traffic situations. Apart from the technical challenge of integrating action-perception systems on board the vehicle, it is important to understand the specific factors that influence the outcome of dangerous situations, such as avoiding unexpected static obstacles or oncoming vehicles. This Masters project will develop an interactive graphical tool to rapidly prototype dynamic traffic scenarios, to allow exploration of different warning system ideas and help the human-machine interaction (HMI) designers to develop appropriate warning systems.

It is important to distinguish this project from a simulation approach. As opposed to a simulator, we are not concerned with closing perception-action loops, but to speed up and simplify the exploration process of a human expert. Explicitly excluded from the project goals are: physical realism, simulation of sensors and actuators, and the incorporation of dynamic constraints.

Work Description

The idea is to make this software as simple as possible and as effective as possible. Vehicle and environment models will be purely planar, and shapes will be limited to relatively fundamental types such as circles, rectangles, and convex polygons. The focus is on usability and visualization of information relevant for the human-guided assessment of risksin given traffic situations.

Several valuable avenues for exploring this application idea exist, and they will be refined into a working prototype based on feedback from interested parties at Volvo Technology Corporation in Göteborg. At a minimum, the software will allow an easy manual design, playback and stepwise exploration with the following:

  • static environment information, such as roads, intersections, buildings, trees
  • vehicle paths, based on splines and/or line sigments in combination with Clothoids and arcs of circles
  • velocity profiles of vehicles along the available paths
  • visualization of anticipated relative positions between vehicles and obstacles, with appropriate graphical methods to highlight areas of particular danger in terms of collision risk or deterioration of lane tracking

Further ideas that are up for exploration:

  • interactive control, via joystick, of one of the vehicles to ease the exploration of ``what if\ldots? scenarios
  • playback of trajectory and sensory logs taken from real vehicles
  • exportation of movies and images to help HMI developers illustrate the scenarios that underlay their designs
  • possibility of integrating the scenario playback system with (separate) path generator modules for the semi-automated evaluation of path planning and control functionality