Andrew English | PhD Candidate | School of Engineering Systems
Title: Visually Assisted Navigation of Autonomous Robots in Agricultural Fields
I am a PhD candidate in Robotics at Queensland University of Technology where I help to develop the autonomous farming robots of the near future. I hold a Bachelor of Engineering (Hons I, Electrical ) from the University of Queensland and have significant industry R&D experience in the signals intelligence, mining and construction industries.
My research projects forms part of the ARC Linkage Project "Robotics for Zero-Tillige Agriculture" (LP110200375). The goal of this project is to develop small cooperative agricultural robots to increase broad-acre crop production and reduce environmental impact. My research investigates localisation and navigation techniques suitable for navigating robots in agricultural fields. I am interested in navigation approaches that are resilient to sensor failure with a focus on using low-cost sensors such as GPS and computer vision.
- A.English, P.Ross, D.Ball, B.Upcroft, P.Corke, "Learning Crop Models for Vision-Based Guidance of Agricultural Robots" International Conference on Intelligent Robots and Systems (IROS) 2015 (under review)
- P.Ross, A.English, D.Ball, B.Upcroft, and P.Corke. Finding the ground hidden in the grass: traversability estimation in vegetation. Intelligent Robots and Systems (IROS), 2015 (under review).
- A.English, P.Ross, D.Ball, B.Upcroft, P.Corke, "TriggerSync: A Time Synchronisation Tool", in International Conference on Robotics and Automation (ICRA) 2015 [pdf]
- D.Ball, P.Ross, A.English, T.Patten, B.Upcroft, R.Fitch, S.Sukkarieh, G.Wyeth, P.Corke, "Robotics for Sustainable Broad-Acre Agriculture", Field and Service Robots, Spring Tracts in Advanced Robotics, Volume 105, 2015, pp 439-453 [pdf]
- P.Ross, A.English, D.Ball, B.Upcroft, P.Corke, "Online novelty-based visual obstacle detection for field robotics", International Conference on Robotics and Automation (ICRA) 2015
- A.English, P.Ross, D.Ball, P.Corke, “Vision Based Guidance for Robot Navigation in Agriculture”, in International Conference on Robotics and Automation (ICRA) 2014 - Best Service Robotics Paper Award Finalist [pdf]
- P. Ross, A.English, D.Ball, B.Upcroft,G.Wyeth, P.Corke, “Novelty-based visual obstacle detection in agriculture”, in International Conference on Robotics and Automation (ICRA) 2014 [pdf]
- P.Ross, A.English, D.Ball, P.Corke, "A method to quantify a descriptor's illumination variance", Australasian Conference on Robotics and Automatoin (ACRA 2014) [pdf]
- A.English, D.Ball, P.Ross, B.Upcroft, G.Wyeth, P.Corke, “Low Cost Localisation for Agricultural Robotics”, Australasian Conference on Robotics and Automation (ACRA) 2013 [pdf]
- P.Ross, A.English, D.Ball, B.Upcroft, G.Wyeth, P.Corke, “A novel method for analysing lighting variance”, Australasian Conference on Robotics and Automation (ACRA) 2013 [pdf]
- A.English, D.Ball, P.Corke, “Vision Based Guidance in Broad-Acre Agriculture” Patent Application, filed September 14 , 2013
- A.Cockerell, A.English, "Composite Structure" WIPO Patent 2013053001, issued April 19, 2013.
Robotics for Zero-Tillage Agriculture
My research projects forms part of the Australian Research Council Linkage Project LP110200375 "Robotics for Zero-Tillige Agriculture". The goal of this project is to develop small cooperative agricultural robots to increase broad-acre crop production and reduce environmental impact. I am part of the three person core team that designed and constructed the robotic research platform which can apply herbicide to weeds on broad-acre farms. The robotic platform is a John Deere Gator TE electric utility vehicle converted for autonomous operation with the addition of computers, actuators and sensors for autonomous operation. The robot has undergone extensive field trials on our industry investigator’s farm near Emerald in Queensland, Australia.
Machine Learning for Crop Row Tracking - IROS 2015 (under reivew)
This work describes a vision-based method of guiding autonomous vehicles within crop rows in agricultural fields where the crop rows are challenging to detect or their appearance is not known a-priori. The location of the crop rows is estimated with an SVM regression algorithm using colour, texture and 3D structure descriptors from a forward facing stereo camera pair. Our system rapidly learns a model online with minimal user input, and then uses this model to track crop rows.
TriggerSync: A Time Synchronisation Tool
This paper presents a framework for synchronising multiple triggered sensors with respect to a local clock using standard computing hardware. Providing sensor measurements with accurate and meaningful timestamps is important for many sensor fusion, state estimation and control applications. Accurately synchronising sensor timestamps can be performed with specialised hardware, however, performing sensor synchronisation using standard computing hardware and non-real-time operating systems is difficult due to inaccurate and temperature sensitive clocks, variable communication delays and operating system scheduling delays. Results show the ability of our framework to estimate time offsets to sub-millisecond accuracy. We also demonstrate how synchronising timestamps with our framework results in a tenfold reduction in image stabilisation error for a vehicle driving on rough terrain. The source code has been released as an open source tool for time synchronisation in ROS here: https://github.com/englishar/trigger_sync
This figure shows the importance of accurate time synchronisation between sensors.
Images show the average of 10 consecutive frames from an IMU stabilised camera
with (a) synchronised timestamps and (b) unaltered timestamps. Synchronisation
significantly improves stabilisation with features on the horizon accurately aligned in
(a) and misaligned in (b).
Monocular Crop Row Tracking - ICRA 2014
In this work I developed a novel method of visually tracking crop rows in challenging cropping fields. The new method tracks the direction lateral offset of the crop rows by estimating the dominant parallel planar texture in a synthetic overhead view of the crop. Notably, the method does not perform any binary segmentation of plants and soil, nor does it require any explicit assumptions about the appearance of plants such as row spacing, periodicity, colour, and lighting. The sensors used are a single inexpensive camera and IMU. This work was presented at the International Conference on Robotics and Automation 2014 and was a finalist for the Best Service Robotics Paper Award. The work was also the subject of an Australian Provisional Patent application.
Low-Cost Localisation for Agricultural Robots - ACRA 2013
In this work I developed a positioning system that estimates pose by fusing data from a low-cost global positioning sensor, low-cost inertial sensors and a the previously described new technique for vision-based row tracking. This approach is resilient to typical sensor failure and suitable for low cost agricultural robots. This work was presented at the Australasian Conference on Robotics and Automation (ACRA2013) in December 2013.