Authors:

Joseph Michael McAlear

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Landing is a crucial component in robotics and control for aerial vehicles. However, it is not a trivial problem. If the robot makes contact with the target with too much velocity it could cause damage to both the machine and people. If the robot moves too slow towards the target then it would not achieve its goals in a reasonable time. In biology, bees utilize radial expansion visual cues to estimate their time to contact in order to perform landings. This phenomenon is referred to as Tau theory. Using this theory, this research developed a visual system utilizing feature detection algorithms to perform soft landings. Specifically, it uses the Oriented Fast and Rotated BRIEF (ORB) feature detector due to its efficiency and ability to find robust features in an image. To test this landing algorithm, simulations were performed using Robotic Operating System (ROS) and Gazebo. Virtual environments are utilized to mimic landing conditions, and both the estimated and ground truth Tau are collected for further examination. Both the accuracy and runtime of the algorithm are examined to ensure real- time applications. Future research will continue to optimize the algorithm and test it against an expanded set of test scenarios including different angle of approach. A different system for estimation will also be explored to improve the efficiency and accuracy of the algorithm. These improvements should help create a system that is able to be integrated with a SLAM algorithm to both map its surroundings are perform landing within it. Keywords: ORB; Feature Detection; Tau; Time-to-Contact

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Joseph Michael McAlear