Authors:

Basil Khwaja

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Vehicles autonomously navigate dynamic environments using advanced sensory and cognitive systems but fall short of tackling new and unfamiliar environments. The objectives of this work are twofold. First, it aims to deploy a real testbed to prototype an autonomous vehicle (AV) and evaluate its performance across various application scenarios. Second, by intentionally pushing the cars to failure through localized environmental updates, we aim to test the system's limits and demonstrate the enhanced benefits of edge computing on the path-planning process. Specifically, the software and hardware stacks of this AV prototype comprise an autonomous navigation stack on the ROS platform, various global and local path-planning algorithms, and physical racing car installations. To assess autonomous navigation performance, we conduct both visual simulations and real-world experiments on the path-planning process using Dijkstra, A*, and RRT algorithms. Utilizing the move base architecture on the ROS platform, we manipulate different parameters to compare changes in the motion control of racing cars, focusing on aspects such as RRT algorithm trajectories and environmental map information. This approach allows us to explore the advantages of edge computing concerning computation offloading. Furthermore, we integrate the MuSHR racing cars with Cisco Wi-Fi 6E networks to investigate the benefits of edge computing by receiving assistance from edge servers for autonomous navigation Keywords: Autonomous Driving; Path Planning; ROS/MuSHR; Real Experiments; Edge Computing

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Basil Khwaja