Wed: Reading about traffic simulations …

Two key open source traffic sims. are mentioned: SUMO (traffic sim) and NS2 (network sim).  Trans links these 2 simulators together: http://trans.epfl.ch/.

These 2 PDFs will be worth reading: http://www.mobile.ifi.lmu.de/common/Literatur/MNMPub/Publikationen/skso05/PDF-Version/skso05.pdf and http://kom.aau.dk/~evm/dist/Vehicle_Traffic_Microsimulator.pdf

There, needless to say, are commercial versions of sims e.g. http://www.aimsun.com/site/content/section/4/72/. However, if dongles are ever mentioned in a pricing structure, it’s doubtful it’s a good ting … I have a feeling that places like this http://www.trafficware.com/simtraffic7.html won’t be ecstatic about me adapting models also.

An awful lot of the time though, I keep coming back to this link: http://www.traffic-simulation.de/. An email will probably need to be sent. It also has details about the traffic models used, which will be of use.

Thu: more checking will be required …

Looked at the graphs using the multi-loess method, and they are really inconsistent. Need an appropriate method / process to say why I’m “arbitrarily” assigning values at times …
In other news, set up a central catalogue to make finding stuff a lot easier. Also started a brief search for flock control and work exists in this area, see http://users.rsise.anu.edu.au/~bradyu/PaperCollection/IFAC05/05083.pdf / A lightweight formation control methodology for a swarm of non-holonomic vehicles GH Elkaim, RJ Kelbley – 2006 IEEE Aerospace Conference, 2006 – ieeexplore.ieee.org … and it’s quite elegant really: control the centre mass of the flock, and they’ll follow it. Clever.

There is also the robot sheep dog system, but that is kind of different: the swarm is controlled by an external party, not one of their own.  This seems more similar: Evolution of control programs for a swarm of autonomous unmanned aerial vehicles KM Milam – 2004 – dtic.mil, http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA424283&Location=U2&doc=GetTRDoc.pdf. In short, put it down, step away.

Wed: velocity plots …

Did the multi loess velocity plots tonight. Will need to perform regression on them next, plus double check that method of calc-ing SD is correct and holds water.

Tue: graphing …

Yes, it has been a while ….

Today graphed the area results with multuple loess F values. Tomorrow, velocity data.