IRLab Talk Schedule

From irlab
Jump to: navigation, search

IRLab talks for Autumn 2012:

Mondays 3pm Room 245 Except where otherwise stated

Responsibilities of the speaker:

  1. Announce the talk a day or two in advance.
  2. Remind people that the meeting is on ~1 hour before the talk.
  3. Check if the room needs an additional projector (LG52 does!)
  4. Collect the projector before the talk from support.
  5. Return the projector after the talk to support.


  • October 1st: Claudio, Two-Level RRT Planning for Robotic Pushing Manipulation. [Practice for IROS talk]

Please, get the paper here: http://www.cs.bham.ac.uk/~cxz004/download/paper/2012/Zito_etal_IROS2012.pdf


  • October 8th: Aaron Sloman, An introduction to toddler theorems: What makes it possible for humans to perceive, play with, think about, and learn from triangles?

Extended abstract here: http://tinyurl.com/CogMisc/irlab-lang-cog-abstract.html (Given to Language and Cognition Seminar in Psychology on Friday 5th Oct.)


  • October 15th: Nick Hawes and Mike Mistry, Talking about the new EU project


  • October 22nd: AleŇ° and Jeremy, Talking about their new grant


  • October 29th: Jeremy Baxter, Multi - UAV task scheduling. How a simple Team Orienteering problem solver evolved over a series of user trials into something much more complex.

In this talk I'll describe the development and use of a multi-uav surveillance scheduling algorithm. This forms part of a uav autonomy testbed which QinetiQ have been putting through a series of trials with potential users.

I'll describe the core scheduler - An anytime A* variant for team orienteering problems and cover the changes that were made to it in response to feedback from trials. The initial implimentation used a simple set of tasks with known locations, durations and rewards with re-planning occuring as tasks were completed, removed or added. This relatively simple model was updated several times to include variable rewards depending on uav type, the handing over of tasks from one uav to another and accounting for partial completion during re-plans.

I'll also describe the impact of having users involved and the changes we had to make to improve user acceptance (proposing multiple options and allowing the user to specify parts of the schedule as fixed.

This is based on a paper I'm writing for the innovative applications track of ICAPS.


  • November 5th: Cancelled


  • November 12th: Cancelled - Lab Lunch


  • November 19th: Min Lue Wang, TBC


  • November 26th: Vero, TBC


  • December 3rd: Cancelled - Lab Lunch


  • December 10th:


Past IRLab talks



Wiki formatting help: http://www.mediawiki.org/wiki/Help:Formatting

For up-to-date news about this schedule, please join the mailing list: https://mailman.cs.bham.ac.uk/mailman/listinfo/irlab

Personal tools