Patterned Search Planning and Testing for the Robotic Antarctic Meteorite Search

Abstract

The goal of the Robotic Antarctic Meteorite Search is to enable discovery of meteorites in Antarctica by a mobile robot. The extreme environment makes it one of the best places to find meteorites, but one of the worst places for humans to work. The meteorite-finding robot will traverse an ice field in a pattern designed to cover the area completely, stopping to investigate potential meteorites with an array of sensors. High level autonomy is needed for this project in many areas: scientific sensing, scientific analysis, navigational sensing, navigational planning, and mission planning. Navigational planning for this project primarily involves generating coverage patterns for traversing the ice fields as completely as possible. Multiple types of patterns are considered, and some of their benefits and problems are discussed. Several of the coverage patterns have been tested in a gravel slag heap in Pittsburgh and on the ice in Antarctica, and the results of these tests are described. The affect of different environments and robot locomotion configurations on control issues, such as the implementation of the pure pursuit algorithm, and maintaining and regaining a path, are discussed.