AI Enabled Aerial-Ground Coordination for Off-Road Vehicles
Date:
Mentor: Dr. Yiqiang Han, Mechanical Engineering
Co-Author: Ishan Sharma
Students: Raylan Dawkins, Chad Eisele, Connor Willoughby, Nicholas Kilbarger, Alexander Krolicki, Harnish Makkar
2 hour interactive poster session Program
Abstract:
In this project, through hands-on experience, undergraduate students at Clemson University studied the new frontier of autonomous control for off-road vehicles. A team of 8 students from the Mechanical Engineering, Industrial Engineering, and Computer Science departments worked as a team in designing and testing a tracked off-road vehicle. This project required students to develop a research vehicle that could autonomously traverse over a rugged off-road environment. The students learned by experimenting with different hardware configurations and coding control algorithms by themselves. In the first iteration, students used only a lidar sensor to map and localize the tank within its environment simultaneously. They worked with adapting the low-level motor controller to the path planning algorithm to accurately control the tank through adverse situations such as a cluttered environment under snow weather. In a second iteration, students combined the lidar sensor with a stereo camera to improve the accuracy of the path planning. The introduction of a camera also enabled students to use computer vision techniques to detect objects and make decisions based on the surrounding environment. Capability of aerial-ground coordination between unmanned MAV and the tracked vehicle will also be demonstrated. This project provides students the opportunity to explore several end-to-end training solutions for controlling autonomous ground vehicles in a variety of environments, which will prepare students for future industries and careers in AI.