This paper presents our approach to developing a low-cost autonomous electric vehicle research platform. The first part of the paper elaborates on our experiments across perception, planning, and control modules. We briefly explain the algorithms and also present our implementation on CARLA, an open-source SOA Urban Driving Simulator Simulator. This is followed by our proposed architecture which combines the localization and control capabilities of Pixhawk, a popular flight controller in the UAV industry, with our perception and planning modules. Our approach is highly-modular and customizable, using ROS(Robot Operating System), which allows the researcher to focus on his area of interest while using the rest of the automation stack. We also share our approach to scale this work on an actual golf cart as a further testbed for autonomous driving research.
Architecture and Algorithms for a Pixhawk-Based Autonomous Vehicle
Lect. Notes Electrical Eng.
26.05.2023
17 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch