Developing a mapping system for an autonomous Formula Student racecar

Abstract

This thesis describes the development of a perception system for an autonomous racecar on behalf of the Driverless team at Revolve NTNU. My contribution has an extra focus on the issue of localization and mapping.

The motivation for this thesis was to develop a system that met the goals set by the organization for the first year in the Driverless class, which was to complete all race events successfully.

A simultaneous localization and mapping (SLAM) system has been implemented using an incremental smoothing and mapping backend. Through adding odometry measurements and landmarks to a factor graph, the system produces an accurate position and orientation of the vehicle, as well as the cones delimiting the racetrack. With this localization information we are able to plan a route around the racetrack which we then can follow.