Robotics weekly meetings

From IridiaWiki
Jump to navigationJump to search

Purpose

  1. Community. Keep everyone updated as to what other work is going on in the lab
  2. Research. Cross germination. Get feedback. Get help on problems. Get ideas.
  3. Admin. Lab administration.


Format

  1. Presentation & Discussion (30 - 50 mins)
    • Work you have been doing
    • Reading you have been doing in the context of your work
    • Other reading on something you are passionate about
  2. Swarmanoid internal meeting (0 - 50 mins)


Try and make your session as involving as possible for everyone:

  • Send round material to be read beforehand.
  • Ask people before the meeting what aspects of your work / reading they would like to know about or discuss.
  • Think about implications of what you are presenting for the research direction of the lab as a whole - methodologies, technologies etc.


Timing & Attendance

The weekly meetings take place on

Wednesdays at 16.00 (temporary time schedule)

Unless:

  • A given Wednesday is a public holiday, or
  • Something very important (like a Ph.D. defense) is overlapping.

If either should happen the meeting is postponed to the next day.

Please indicate on this wiki page if you are not going to be able to attend a meeting (no later than the Friday before).

If you forget to note your absence in advance you will be severely punished and will be forced to buy a nice, big cake for the following meeting.

Previous Meetings

See Minutes and agendas from previous meetings (2007-2008).

See Minutes and agendas from previous meetings (older).


Wednesday 4th February 2009 - Internal Meeting - Room C2-223, 2nd floor

Agenda

  1. Presentation
    • None
  2. Internal Meeting / Discussion
    • Swarmanoid Project: Camera Segmentation Software

Reading material

People who will be absent



Wednesday 11th February 2009 - Giovanni - Room C2-223, 2nd floor

Agenda

  1. Presentation
    • Interference Reduction through Task Partitioning

The talk is about the use of task partitioning as a way to reduce interference in a spatially constrained harvesting task. Interference is one of the key problems in large cooperating groups. We present a simple method to allocate individuals of a robotic swarm to a partitioned task, and study how this task partition can increase system performance by reducing sources of interference. The method is experimentally studied, both in an environment with a constrained area and an environment without the constraint.

  1. Internal Meeting / Discussion
    • Swarmanoid Project: To be defined

Reading material

People who will be absent



Wednesday 18th February 2009 - Carlo - Meeting room in Architecture Dept.

Agenda

  1. Presentation
    • Heterogeneous Group Size Regulation (please edit)
  2. Internal Meeting / Discussion
    • Swarmanoid Project: To be defined

Reading material

People who will be absent



Wednesday 25th February 2009 - Meeting Suspended - Room C2-223, 2nd floor



Wednesday 4th March 2009 - Meeting Suspended - Room C2-223 , 2nd floor




Wednesday 11th March 2009 - Rehan - Room C2-223 , 2nd floor - Start at 16.30

Agenda

  1. Presentation
    • Morphogenesis (please edit)
  2. Internal Meeting / Discussion
    • Swarmanoid Project: To be defined

Reading material

People who will be absent



Thursday 19th March 2009 - Eliseo and Nithin - Paduart Room, 4th floor - Start at 16.30

Agenda

  1. Presentation
    • Phat-bot Assembly and Navigation
    • The behavioral toolkit
  2. Internal Meeting / Discussion
    • Swarmanoid Project: To be defined

Abstract: In this talk we will present the work we have been doing on the phat-bot assembly and navigation. All the controllers were implemented with a new architecture in mind, which made the modularity of the behaviors possible in ARGoS. We call this architecture "Behavioral Toolkit" and we will present it also during the talk.

Reading material

People who will be absent



Wednesday 25th March 2009 - Marco - Paduart Room, 4th floor - Start at 16.30

Agenda

  1. Presentation
    • Local social influence mechanisms for collective decision making: A swarm robotics experiment
  2. Internal Meeting / Discussion

Abstract: In human societies, groups are often perceived as superior decision making entities than individuals. When an important decision has to be made, groups are formed to decide on the issue. Examples include referenda, elections, and juries. When individuals are part of a decision making group, they are subject to social influences that can affect their perception and/or their opinion on the issue. Our work is inspired by this fact and we investigate whether local social influence mechanisms can make a population of robots choose the best of two competing alternatives.

In our experiment, a swarm of robots is faced with a binary decision problem: go foraging to the left, or to the right; or push or pull a family of objects, for example. Associated to each choice there is a cost which is the time it takes for a group to perform the task. I will show you preliminary results that suggest that simple local decision rules can make an uninformed population (or even a negatively biased one) choose the best strategy (in terms of task completion time).

Reading material

People who will be absent



Wednesday 1st April 2009 - Simulator staff - Room C2-223 , 2nd floor, 15:00 - 17:00

Agenda

  1. Presentation
    • None
  2. Internal Meeting / Discussion
    • Technical meeting with the simulator staff

Reading material

People who will be absent




Tuesday 7th April 2009 - Everybody - Room C2-223 , 2nd floor, 14:00 - 16:00

Agenda

  1. Presentation
    • TBA
  2. Internal Meeting / Discussion
    • The simulator
      • The Simulator Constitution
      • ARGoS BT scripting discussion follow-up
      • The BT exposed

Reading material

People who will be absent



Wednesday 15th April 2009 - Meeting suspended

Agenda

  1. Presentation
    • TBA
  2. Internal Meeting / Discussion

Reading material

People who will be absent



Wednesday 22nd April 2009 - Manuele AND Internal Meeting - Room C2-223 , 2nd floor, 15:00 - 17:00

Agenda

  1. Presentation
    • A Reliable Distributed Algorithm for Group Size Estimation with Minimal Communication Requirements
  2. Internal Meeting / Discussion
    • The ARGoS Constitution

Abstract: In this talk I will present a method that lets individual robots in a group estimate the size of the group in a distributed manner. The process is loosely based on the signaling behavior of fireflies and crickets. Melhuish et al. first devised a method based on local robot signaling to estimate the group's size. Each robot emits a signal and can perceive the signals of neighboring robots in close proximity. Following the approach of Melhuish et al., our robots count the number of emitted signals over a suitably defined period of time. Experiments show that the estimates calculated with Melhuish et al. method display a great deal of noise. We modify their method so as to sensibly stabilize the output. We assess the quality of our method through extensive simulation-based experiments.

Reading material

People who will be absent



Wednesday 29th April 2009 - Antal - Room C2-223 , 2nd floor, 15:00 - 17:00 - POSTPONED

Agenda

  1. Presentation
    • TBA
  2. Internal Meeting / Discussion

Reading material

People who will be absent



Wednesday 6th May 2009 - Ali - Room C2-223 , 2nd floor, 15:00 - 17:00

Agenda

  1. Presentation
    • TBA
  2. Internal Meeting / Discussion

Reading material

People who will be absent



Tentative Schedule for the following