Advanced Lane Finding Prototype Based on Autoware Platform

Abstract

In order to achieve functionality of autonomous driving, modern vehicles must be aware of theirs surrounding in any given moment. Complex software modules within such systems oversee vehicle’s environment and use this environmental data to pinpoint vehicles position in the world. Often these perception modules process numerous calculations and fuse data acquired from different sensors to achieve, as precise as possible, understanding of vehicles environment. Alongside these highlevel data fusion modules, many modern vehicles have redundant subsystems that handle similar functionality but in smaller scale in order to achieve shorter “Sense-Plan-Act” loop. In this paper we will present prototype for Advance Lane Finding (ALF) application, which could be utilized as an enhancement of multiple ADAS systems. Proposed solution is implemented in C+ + programming language as a part of Autoware/ROS platform. Prototype’s performances are tested on Nvidia DRIVE PX2 hardware platform.