Author

Gregg Buskey

3 papers · 2000–2012

2000peak 1 in 20002012

Frequent co-authors

Papers

showing 1–3 of 3 RSS

Autonomous helicopter hover using an artificial neural network

Gregg Buskey, Gordon Wyeth, Jonathan Roberts

Computer Vision and Pattern Recognition · Queensland's institutional digital repository (The University of Queensland) · 13 Nov 2002 · cited 50× · DOI ↗ · more like this ·

Details the developments to date of an unmanned air vehicle (UAV) based on a standard size 60 model helicopter. The design goal is to have the helicopter achieve stable hover with the aid of an INS and stereo vision. The focus of the paper is on the development of an artificial neural network (ANN) that makes use of only the INS data to generate hover commands, which are used to directly manipulate the flight servos. Current results show that networks incorporating some form of recurrency (state history) offer little advantage over those without. At this stage, the ANN has partially maintained periods of hover even with misaligned sensors.

Flight control using an artificial neural network

Gordon Wyeth, Gregg Buskey, Jonathan Roberts

Computer Vision and Pattern Recognition · Queensland's institutional digital repository (The University of Queensland) · 2000 · cited 20× · OpenAlex ↗ · more like this ·

This paper details the developments to date of an unmanned air vehicle (UAV) based on a standard size 60 model helicopter. The design goal is to have the helicopter achieve stable hover with the aid of an INS and monocular vision. The focus of this paper is on the development of an artificial neural network that makes use of this data to generate hover command, which are used to directly manipulate the flight servos. Current results show that a mapping can be learnt for the training data, but generalisation to unseen data has been poor with current training sets. 1

Online Learning of Autonomous Helicopter Control

Gregg Buskey A B, Jonathan Roberts A, Gordon Wyeth B

Safety Research · 18 May 2012 · cited 0× · OpenAlex ↗ · more like this ·

This paper details the development of an online adaptive control system, designed to learn from the actions of an instructing pilot. Three learning architectures, single layer neural networks (SLNN), multi-layer neural networks (MLNN), and fuzzy associative memories (FAM) are considerd. Each method has been tested in simulation. While the SLNN and MLNN provided adequate control under some simulation conditions, the addition of pilot noise and pilot variation during simulation training caused these methods to fail. FAMs alone were found to be suitable under all simulation conditions. The performance of an FAM velocity control being used by a higher level waypoint controller in simulation is presented. As a first step toward performing real helicopter tests, an FAM attitude controller has also been developed in simulation, and its tracking performance presented. of the non-conventional…