raquib is currently certified at Journeyer level.

Name: Tayeb Karim
Member since: 2003-07-24 16:44:28
Last Login: 2007-08-14 22:00:34

FOAF RDF Share This


Well, currently, I'm a student at RIT. I've helped build 2 robots in highschool (Stuyvesant) for the FIRST competition. I bought an ER1 and am working on some projects with it. I'm really into robotics, and I know it'll be the field of the future.


Recent blog entries by raquib

Syndication: RSS 2.0
19 Dec 2006 (updated 19 Dec 2006 at 02:03 UTC) »

It's been a while. moved the senscience project over to describe my thesis. updates more regularly soon...

Just a quick update, the darpa project is moving along steadily we named our group gcart (grand challenge autonomous yada yada) website is now gcart.rit.edu

yea, i know, lots of random sentences there... link em together. In other news, i'm messing around with a new type of neural net, which i'm calling the "messynet". Yea, that's right, because it's that messy. I'll put up initial results soon... churning through test cases as i'm typing.

OK, starting with the DARPA project. Got funding (in the form of software) from evolution robotics. Sencience will have to take a back seat as this project devolopes. More coming soon...


18 Aug 2003 (updated 18 Aug 2003 at 22:40 UTC) »

OK, here's some really basic code for one of my subroutine programs. This one will be the one that "blindly" builds a map of where the robot is and where other things may be. The map itself is only used as a secondary source (more for seeing "am i headed in the right direction" then for planning out complex moves).

sensors used: 3 IR sensors that measure proximity to robot. They are placed facing foward, foward left, and foward right.Distance tracking: x,y, and degree values of the robot are measured by the robot. There may be a lot of imprecision here, which is why the map is a secondary source! (NOTE: the x and y values may be negative! the robot starts off in the state 0,0 )

The pseudocode:

/* NOTES: if you use this, email me...! raquib0@lycos.com i'd like to hear about your robot! (credit would be nice too..)

a map is made up of areas

an area is a matrix of doubles (real numbers) where the location of each element (x,y) represents a set distance and the value of each element represents the probability that is is "filled".

when I say "mark" an area, i mean that you will do this:

your position is x,y
add .1 (or some other constant) to the elements 1 element away from x,y
add .05 (or some other contant) to the elements 2 elements away from x,y
add .0025 (or some other contant) to the elements 3 elements away from x,y...
.... etc
.... etc
keep doing this for N elements away from xy

the elements in the matrix with a value above a certain threshold are likely to be filled... you can use this to plan a overall route to a point.


constant Tx; // max x size on the map
constant Ty; // max y size on the map
constant Td = 360; // max degrees
constant Tir; // threshold for IR values to signal a mark event

var doff = 0;
var xoff = 10;
var yoff = 10;

var ir1, ir2, ir3, x,y,d

begin critical loop{
read in ir1, ir2, ir3, x,y,d values from robot
// bring the x,y values to our system
x += (Tx/2);
y += (Ty/2);

// now put x,y to our degree position
var distance = sqrt(x*x + y*y);
x = distance*sin(doff+d);
y = distance*cos(doff+d);
// check bounds, create or move to new area if need be
if(! inbounds(x,y) ){
make xoff and yoff starting positions in new area
doff = (doff+d)%Td;
go back to start of loop
} // inbounds check

// check if i should update this area of the map
if(ir1 + ir2 + ir3 > Tir){
update values around me in the area
} // end of crtical loop

So here's another project i'm working on cocurrently. A music analyser using neural nets. I'm using the Maate MPEG library to decode mp3s and running them through the neural net to look for classifications. Okay, that made no sense. Here we go. Let's call the program MA. You pass in 2 lists to MA, a list of songs you want to hear and a list of songs that you don't want to hear. You then train the network (a standard FFN) with the lists that you passed in, expecting a result of 1 for a good song and 0 for a bad. I'm currently using the RMS value of each subband as the inputs to the network. Layers: variable node input layer, 20 and 10 node hidden layers, and a 1 node output layer. I want to extend the program so I have multiple nets (RMS, Centroid, Energy Movement, etc) and allow the songs that have a majority of the nets firing to enter the "good" list. I've only programmed in the RMS network so far, and that seems to be working well. I'm looking at a 98% success rate so far... i'll get back to the robot soon...

1 older entry...


raquib certified others as follows:

  • raquib certified motters as Master
  • raquib certified blueeyedpop as Master
  • raquib certified Nic as Journeyer

Others have certified raquib as follows:

  • lnxfergy certified raquib as Journeyer
  • nitro certified raquib as Journeyer
  • freqrobot certified raquib as Apprentice
  • Axle certified raquib as Journeyer

[ Certification disabled because you're not logged in. ]

Robot of the Day


Built by
Christian Vlaskowits

Recent blogs

30 Sep 2017 evilrobots (Observer)
10 Jun 2017 wedesoft (Master)
9 Jun 2017 mwaibel (Master)
25 May 2017 AI4U (Observer)
25 Feb 2017 steve (Master)
16 Aug 2016 Flanneltron (Journeyer)
27 Jun 2016 Petar.Kormushev (Master)
2 May 2016 motters (Master)
10 Sep 2015 svo (Master)
14 Nov 2014 Sergey Popov (Apprentice)
3 Jul 2014 jmhenry (Journeyer)
10 Jun 2014 robotvibes (Master)
13 Nov 2013 jlin (Master)
23 Jun 2013 Mubot (Master)
13 May 2013 JLaplace (Observer)
21 Apr 2013 Pi Robot (Master)
12 Apr 2013 Pontifier (Apprentice)
16 Mar 2013 gidesa (Journeyer)

Newest Members

Newest Robots

7 Aug 2009 Titan EOD
13 May 2009 Spacechair
6 Feb 2009 K-bot
9 Jan 2009 3 in 1 Bot
15 Dec 2008 UMEEBOT
10 Nov 2008 Robot
10 Nov 2008 SAMM
24 Oct 2008 Romulus
30 Sep 2008 CD-Bot
26 Sep 2008 Little Johnny
Share this page