AAO Starbugs: software control and associated algorithms

TL;DR

This paper presents the software system and algorithms for controlling 150 Starbug robots in the TAIPAN instrument, focusing on automated collision-free path planning and real-time position tracking to achieve precise fibre positioning.

Contribution

It introduces a tiered, multi-stage path-finding algorithm for collision avoidance and detailed software for monitoring Starbug positions during operation.

Findings

Successful collision-free path planning for all Starbugs

Achieved fibre positioning accuracy of 0.3 arcsec

Efficient automated control system implemented

Abstract

The Australian Astronomical Observatory's TAIPAN instrument deploys 150 Starbug robots to position optical fibres to accuracies of 0.3 arcsec, on a 32 cm glass field plate on the focal plane of the 1.2 m UK-Schmidt telescope. This paper describes the software system developed to control and monitor the Starbugs, with particular emphasis on the automated path-finding algorithms, and the metrology software which keeps track of the position and motion of individual Starbugs as they independently move in a crowded field. The software employs a tiered approach to find a collision-free path for every Starbug, from its current position to its target location. This consists of three path-finding stages of increasing complexity and computational cost. For each Starbug a path is attempted using a simple method. If unsuccessful, subsequently more complex (and expensive) methods are tried until a valid path is found or the target is flagged as unreachable.