Current Work

I am enrolled in a PhD at the Australian National University, under the supervision of:

• Dr. Nick Barnes   (ANU RSISE/NICTA)
• Professor Mandyam Srinivasan    (ANU RSBS, CVS),
• Professor Guilio Sandini     (Uni of Genova, LIRA-Lab)

My interest is in computer vision and robotics. I am most interested in the use of visual motion (optical flow) for biologically inspired mobile robot navigation. I am interested in implementing more robust navigational behaviours through biologically inspired techniques, but also in better understanding biological vision systems through the building and testing of artificial models (i.e. robots).

My PhD work is extending the links between the study of biological vision systems (specifically the honeybee), and the building of robust algorithms for the visual control of an autonomous vehicle. Bees possess an incredible ability to achieve high precision movements, and solve complex navigational tasks. Such abilities include navigating cluttered environments, performing graze landings with pin point precision, flight stabilisation, and an ability to accurately estimate distance travelled. Perhaps what is most remarkable about all this is that they achieve it all using a brain possessing no more than a million neurons. To put this in perspective, humans have something closer to a million million neurons! This suggests a close coupling between motor responses and a highly specialised vision system, and much has been learnt about this vision system over the last half century. From a robotics persective, there is obviously much to learn from the honeybee as well. My project looks specifically at how the honeybees panoramic view of the world may be exploited to design better algorithms for robot navigation, given a similar digital spherical projection from two cameras. To do this, we are building a ground-based mobile robot, equiped with two 180 degree FOV cameras, simulating the eye configuration of the honeybee. By modelling the honeybee in this way, we also hope to pose new questions for the biological vision community, and contribute to the understanding of biological vision systems (including our own).

Publications

Journal Papers

1. [in press] "A robust docking strategy for a mobile robot using flow field divergence", by C. McCarthy, N. Barnes and R. Mahony. To appear in IEEE Transactions on Robotics, 2008

Conference Papers

1. "Real time biologically-inspired depth maps from spherical flow", by C. McCarthy, N. Barnes and M.V. Srinivasan. Proceedings of IEEE International Conference on Robotics and Automation (ICRA 2007), Rome, Italy, 2007.
2. "Insect Inspired Robots", by J. Lim, C. McCarthy, D. Shaw, N. Barnes and L. Cole. Proceedings of the 2006 Australiasian Conference on Robotics and Automation (ACRA2006), Auckland, New Zealand, 2006.
3.  "A robust docking strategy for a mobile robot using flow field divergence", by C. McCarthy and N. Barnes. Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2006), Beijing, China, 2006.

4. "Comparison of temporal filters for optical flow estimation in continuous mobile robot navigation", by C. McCarthy and N. Barnes. Springer Tracts in Advanced Robotics: Experimental Robotics IX: The 9th International Symposium on Experimental Robotics (Singapore 2004), Volume 21, pp 481-90, 2006.
   Paper was awarded "International Foundation of Robotics Research Student Fellowship Award".

5. "Performance of optical flow techniques for indoor navigation with a mobile robot", by C. McCarthy and N. Barnes. Proceedings of IEEE International Conference on Robotics and Automation (ICRA2004), pp 5093-5098, New Orleans, USA, 2004.

6. "Performance of temporal filters for optical flow estimation in mobile robot corridor centring and visual odometry", by C. McCarthy and N. Barnes. Proceedings of the 2003 Australasian Conference on Robotics and Automation. Brisbane, Australia, 2003.

Workshop Papers

1. [in press]"Towards a Hazard Perception Assistance System using Visual Motion", by C. McCarthy, N. Barnes, K. Anstey and M. Horswill. ECCV Workshop on Computer Vision Applications for the Visually Impaired (CVAVI 08). Marseille, France, Oct. 18, 2008.

Sample Video Demonstrations

• [McCarthy et al, ICRA 2007]: The "belco kitchen" sequence (MPG file, 9.6 Mb). Video demonstrates the estimation of egomotion direction using the Nelson and Aloimonos (1988 Bio Cybernetics) de-rotation algorithm. Blue and yellow flow vectors signify direction of flow on the circle (CW or CCW).
• [ McCarthy and Barnes, ISER 2004, IROS 2006]: Docking using time-to-contact estimation from flow field divergence(WMV file). Video shows three temporal filters used with Lucas and Kanade's gradient-based flow algorithm.
• [McCarthy and Barnes, ICRA 2004]: Centring and Docking comparisons(55Mb). Temporal filter comparisons in corridor centring and flow divergence based docking trials.

Conference Presentations/Seminars

• Scheduled:
  • None scheduled

• Presented:
  • [WKS] Making flow-based navigation strategies work. To be presented at the ANU RSBS summer school on insect inspired robotics, Dec 2006, Canberra, Australia.

  • [CNF] "A robust docking strategy for a mobile robot using flow field divergence". To be presented at the IEEE/RSJ International Conference on Intelligent Robots and Systems Oct 2006, Beijing, China, 2006.

  • [WKS] "Performance of Optical Flow Techniques for continuous mobile robot navigation". To be presented at Insect Sensors and Robotics Workshop. Aug 2004, Brisbane QLD , Australia.

  • [CNF] "Comparison of temporal filters for optical flow estimation in continuous mobile robot navigation". Presented at ISER'04 International Symposium on Experimental Robotics. Jun 2004, Singapore.

  • [SEM] "Performance of Optical flow techniques for contiuous mobile robot navigation". Presented at ANU, Sys Eng dept seminar. Jun 2004, Canberra ACT, Australia.

  • [CNF] "Performance of optical flow techniques for indoor navigation with a mobile robot". Presented at IEEE ICRA04, the 2004 International Conference on Robotics and Automation. Apr 2004, New Orleans LA, USA.

  • [CNF] "Performance of temporal filters for optical flow estimation in mobile robot corridor centring and visual odometry". Presented at 2003 Australiasian Conference on Robotics and Automation. Dec 2003, Brisbane Qld, Australia.

  • [SEM] "Performance of optical flow techniques for indoor navigation with a mobile robot". Presented at Univ. Melb CSSE postgrad seminar. Nov 2003, Melbourne Vic, Australia.

  • [CNF] "The Trials and Tribulations of World Cup Robot Soccer". Presented at 2002 Australian Computer Society National Conference: Student and Young Professionals. March 2002, Lorne Vic, Australia. (Survey conducted at conference gave talk highest ranking for interest and delivery)

Past Projects

• Performance of Optical Flow Techniques for Mobile Robot Navigation: My Masters thesis compared optical flow techniques for mobile robot navigation. While comparisons have been done previously, these have not adequately addressed the specific issues associated with embedding such techniques in the control loop of a real-time system (like a robot). This needs to be addressed if this paradigm of robot navigation is to gain more attention from main stream mobile robot reseacrh. We seek to provide a systematic basis on which to choose an optical flow technique, which previously did not exist.

• Robocup: Click here to read about the RoboCup project. I have been involved with the Dog leauge (which uses the Sony Aibo robots shown below) since 2001 as a co-supervisor and general team adviser and have also assisted with software development. In 2002, our team came 4th in the world in Fukuoka, Japan.

• Web Based Telerobot: My honours project, completed in June 1999, was to develop a telerobotic control and monitoring system on the WWW. The robot had a radio link with a local PC that performed all local navigation. This system also communicated with a remotely run applet which displayed images from an overhead camera and a GUI for different modes of control (Direct or Supervisory).

Professional Society Memberships

• Robotics and Automation Society (IEEE_RAS).
• Australian Robotics and Automation Association.

Stress management in the robotics labs

Interesting/Related Links

• Melb Uni Computer Vision and Machine Intelligence Lab
• Australian National University Center for Visual Sciences
• University of Genova "LIRA" Lab
• Melb Uni RoboCup Site
• Official RoboCup Site
• Worldwide Computer Vision Homepage