Event-based real-time sensory-motor processing

Members: Jesus Armando Garcia Franco, Bilel Belhadj, Luis Camunas, Christian Denk, Daniel Neil, Dimitra Emmanouilidou, Daniel Rasmussen, Francisco Barranco, Cornelia Fermuller, Greg Cohen, Hector Jesus Cabrera Villaseor, Mehmet Ozdas, Amir Khosrowshahi, Sergio Davies, Evangelos Stromatias, Tobi Delbruck, Will Constable, Xavier Lagorce

Leaders: Daniel Neil, Tobi Delbruck (Zurich)

This tutorial introduces the jAER software project to interested participants. In all, we had 2 tutorials, the first tutorial was centered around the organization of the project and its interface to the various AER sensors, the second tutorial was centered around the construction of a mean filter which acts as a tracker.

jAER Method Tutorial

First, the wiki page and the presentation (attached to this wiki page):

Now, to run jAER:

  • Run Netbeans.
  • If this is your first time opening the jAER project, go to File->Open Project and select "host/trunk/java" from the directory I gave you on the USB key. Otherwise the project should be open.
  • Press the hammer and broom icon to "Clean and Build" jAER. Press the green play button to run jAER. If you get errors during build, don't worry and try to run anyway. If after pressing the green play button it still doesn't work, then post to the discussion help forum.
  • You should be presented with a black screen with a blue square; go to File->Open Logged Data to open the "highway.dat" file that was in the directory I gave you. You should see the cars driving down the highway towards you. Post to the discussion help forum if this doesn't work!

The majority of the presentation was spent on the MeanEventLocationTracker which you can find if you use the search box in the upper right hand corner of Netbeans. Alternatively, it is located in the package directory here: org.capocaccia.cne.jaer.cne2011.MeanEventLocationTracker?.

Take some time to play around with how it processes events (edit the file, press the green button to test it out). Try throwing out all the events. Try shifting all the events left 10 pixels. The assignment I would like to try is to vary the square tracker size by the variance (about a 15 minute task), as you can see in the last ("Homework") slide of the presentation. We'll get to that next time!

More resources: