Timothy Wiley

Research Associate
Creative Robotics Lab | Art & Design | UNSW Australia
School of Computer Science and Engineering | Faculty of Engineering | UNSW Australia

Publications

All of Timothy's publications, sorted by category.

Journal Articles

  1. Wiley, T., Sammut, C., Hengst, B., & Bratko, I. (2016). A Planning and Learning Hierarchy using Qualitative Reasoning for the On-Line Acquisition of Robotic Behaviors. Advances in Cognitive Systems. 4, pp. 93-112
  2. Milstein, A., McGill, M., Wiley, T., Salleh, R., & Sammut, C. (2011). A Method for Fast Encoder-Free Mapping in Unstructured Environments. Journal of Fields Robotics: Special Issue on Safety, Security, and Rescue Robotics. 28(6), pp. 817-831.

Conference and Workshop Papers

  1. Wiley, T., Sammut, C., Hengst, B., & Bratko, I. (2015). A Multi-Strategy Architecture for On-Line Learning of Robotic Behaviours using Qualitative Reasoning. Proceedings of the Third Annual Conference on Advances in Cognitive Systems. Atlanta, USA, pp. 1-16.
  2. Wiley, T., Sammut, C., & Bratko, I. (2014). Qualitative Simulation with Answer Set Programming. Proceedings of the 21st European Conference on Artificial Intelligence. Prague, Czech Republic, pp. 915-920.
  3. Wiley, T., Sammut, C., & Bratko, I. (2014). Qualitative Planning with Quantitative Constraints for Online Learning of Robotic Behaviours. Proceedings of the 28th AAAI Conference on Artificial Intelligence. Quebec City, Canada, pp. 2578-2584.
  4. Wiley, T., Sammut, C., & Bratko, I. (2013). Planning with Qualitative Models for Robotic Domains. Poster Collection in the Second Annual Conference on Advances in Cognitive Systems. Baltimore, USA, pp. 251–266.
  5. Wiley, T., Sammut, C., & Bratko, I. (2013). Using Planning with Qualitative Simulation for Multistrategy Learning of Robotic Behaviours. Proceedings of the 27th International Workshop on Qualitative Reasoning. Bremen, Germany, pp. 24-31.
  6. Wiley, T. (2013). Multi-Strategy Learning of Robotic Behaviours via Qualitative Reasoning. Proceedings of the AAAI-SIGART Doctoral Consortium in the 27th AAAI Conference on Artificial Intelligence. Seattle, USA, pp. 1682-1683.
  7. Wiley, T., McGill, M., Milstein, A., Salleh, R., & Sammut, C. (2012). Spatial Correlation of Multi-Sensor Features for Autonomous Victim Identification. Lecture Notes in Computer Science. RoboCup 2011: Robot Soccer World Cup XV. ed. by T. Röfer, N. M. Mayer, J. Savage, & U. Saranli. Springer Berlin / Heidelberg, pp. 538-549.
  8. McGill, M., Salleh, R., Wiley, T., Ratter, A., Farid, R., & Sammut, C. (2012). Virtual Reconstruction Using an Autonomous Robot. Proceedings of the 2012 International Conference on Indoor Positioning and Indoor Navigation. Sydney Australia, pp. 1-8.
  9. Milstein, A., McGill, M., Wiley, T., Salleh, R., & Sammut, C. (2011). Occupancy Voxel Metric Based Iterative Closest Point for Position Tracking in 3D Environments. Proceedings of the 2011 IEEE International Conference on Robotics and Automation. pp. 4048-4053.

Tech Reports

  1. Wiley, T., Sammut, C., & Bratko, I. (2014). Qualitative Simulation with Answer Set Programming. The School of Computer Science and Engineering, The University of New South Wales. Technical Report. (UNSW-CSE-TR-201415).

Honours Thesis

Wiley, T. (2010). Autonomous Victim Identification, The University of New South Wales.
Supervisor: Dr. Claude Sammut, Assessor: Dr. Bernhard Hengst

Abstract
When a disaster (either natural or human-made) strikes in an urban environment, human casualties may become trapped within damaged buildings. Quickly finding these victims is a highly important task of Urban Search and Rescue. Robotic systems have been developed to increase the abilities of rescue teams to quickly and efficiently search large and complex building environments, however, these systems have been largely teleoperated. That is, rescue workers must operate the robot, drive it about the damaged building, and using a limited number of sensors and small set of cameras attempt to find victims. Such manual operation requires highly trained operators, and a continuous connection between the robot and operator, which is not necessarily feasible.
A more ideal solution is to develop algorithms that allow robots to autonomously explore, find and identify victims of the disaster, and then return to base with a detailed map recording the location and status of each victim. Little research has been done in the particular field. This thesis aims to extend the current state-of-the-art algorithms and propose a reliable and accurate system for the task of Autonomous Victim Identification. Two key features of this task are examined: Victim Identification methods and the Autonomous Robotic Behaviour.
For the identification of victims, the most obvious indicator of a victim is the heat signature given off by the human body. This thesis investigates methods for dynamically locating nonuniform hot features within the disaster environment using a thermal camera. Additionally, in practical situations, it is noted that victims are often trapped in small cavities. Given a 3D point cloud generated from a sequence of laser scans of the local environment, it is possible to identify features such as cavities and holes. However, current methods are unable to locate such features in real-time. This thesis proposes an efficient, real-time method for this task.
For the Autonomous Robotic Behaviour, this thesis proposes the use of a highly simple handcrafted decision tree. This is combined with a limited set of pre-defined actions (skills) the robot can perform. This leads to a simple, but predictable robotic behaviour. This thesis proposes that a simple approach can be effective even in complex environments such as a damaged buildings, as it focuses on the key attributes to finding a victim.
The performance of the Victim Identification and Autonomous Robotic Behaviour will be assessed based on the performance of the methods in the Urban Search and Rescue division during the 2010 RoboCup competition. The methods proposed in this thesis significantly contributed to CASualty (the UNSW RoboCupRescue team) winning Best-In-Class for Autonomy.