Comments for Doswa

Comment on Line Segment to Circle Collision/Intersection Detection by David

David — Sun, 07 Mar 2010 00:54:01 +0000

@Joseph Winston S
What sort of robot are you building?

Comment on Line Segment to Circle Collision/Intersection Detection by Joseph Winston S

Joseph Winston S — Sat, 06 Mar 2010 14:31:02 +0000

Hi! Well explained and i am looking for an implementation for the spatial hyper redundant robot link interference with staggered pipes . I am also implementing through python.
Thanks for the post. It really helped me to visualize and implement obstacle avoidance.
S. Joseph Winston

Comment on AVRDUDE 5.8 with FTDI bitbang patch on Linux by Mic

Mic — Fri, 26 Feb 2010 19:47:27 +0000

Great stuff, patched into avrdude 5.10 - works fine!
I burned arduino 2009 with ATmega168.

Comment on Wacom Jitter Fix (for Linux) by David

David — Wed, 17 Feb 2010 08:34:16 +0000

@M_Woody:

No, the commands will not be the same.

You might want to try looking around the preferences/control panel for your tablet driver. If you’re lucky, you’ll find something like “smoothing” or “noise suppression”. If not, there’s not much I can do to help—I’ve never used a Wacom tablet on any operating system other than Linux.

Good luck!

Comment on SparkFun’s FTDI Basic Breakout as an AVR programmer by problems uploading to mignon gamekit arduino // kaotec

problems uploading to mignon gamekit arduino // kaotec — Tue, 16 Feb 2010 21:38:13 +0000

[...] I have to try this, but for now that doesn’t really solve my what the hell is the CTS pin [...]

Comment on Wacom Jitter Fix (for Linux) by M_Woody

M_Woody — Tue, 16 Feb 2010 09:20:14 +0000

Great, too bad I don’t have Linux :-/ I’ll do some looking around on my hard-drive, the commands should be the same… Shouldn’t they?

Comment on Fourth order Runge-Kutta numerical integration by Improved RK4 Implementation | Doswa

Improved RK4 Implementation | Doswa — Sat, 23 Jan 2010 14:54:33 +0000

[...] you’re new to numerical integration or even RK4 integration, please read my other post first. It’s easier to understand because it’s a less generalized function. def [...]

Comment on Fourth order Runge-Kutta numerical integration by David

David — Sat, 23 Jan 2010 00:03:25 +0000

Hi simmuskhan:

I had never heard of the RKF method until now, but after looking into it, I feel that it overcomplicates things. For many situations, RK4 is “good enough”.

However, here’s a basic implementation of RKF. You’ll probably want to add some code to put minimum/maximum values on dt. This one is hardcoded for 1st order differential equations, but it shouldn’t be too hard to modify for higher orders.

def rkf45(t, dt, y, f, tolerance=1e-5):
	"""Runge-Kutta-Fehlberg method.
 
	t = Current time.
	dt = Timestep.
	y = Initial value.
	f = Derivative function y' = f(t, y).
	tolerance = Error tolerance.
 
	Returns a (tf, dtf, yf) tuple after time dt has passed, where:
	tf = Final time.
	yf = Final value.
	dtf = Error-corrected timestep for next step."""
 
 
	# Calculate the slopes at various points.
	# Values taken from http://en.wikipedia.org/wiki/Runge-Kutta-Fehlberg_method.
	k1 = f(t, y)
	k2 = f(t+(1/4.0)*dt, y + k1*(1/4.0)*dt)
	k3 = f(t+(3/8.0)*dt, y + k1*(3/32.0)*dt + k2*(9/32.0)*dt)
	k4 = f(t+(12/13.0)*dt, y + k1*(1932/2197.0)*dt + k2*(-7200/2197.0)*dt + k3*(7296/2197.0)*dt)
	k5 = f(t+dt, y + k1*(439/216.0)*dt + k2*(-8)*dt + k3*(3680/513.0)*dt + k4*(-845/4104.0)*dt)
	k6 = f(t+(1/2.0)*dt, y + k1*(-8/27.0)*dt + k2*(2)*dt + k3*(-3544/2565.0)*dt + k4*(1859/4104.0)*dt + k5*(-11/40.0)*dt)
 
	# 4th order approximation of the final value.
	yf4 = y + k1*(25/216.0)*dt + k3*(1408/2565.0)*dt + k4*(2197/4104.0)*dt + k5*(-1/5.0)*dt
 
	# 5th order approximation of the final value.
	yf5 = y + k1*(16/135.0)*dt + k3*(6656/12825.0)*dt + k4*(28561/56430.0)*dt + k5*(-9/50.0)*dt + k6*(2/55.0)*dt
 
	# Timestep scaling factor. From http://math.fullerton.edu/mathews/n2003/RungeKuttaFehlbergMod.html.
	err = abs(yf5-yf4)
	s = 1 if err==0 else (tolerance*dt/(2*err))**(1/4.0)
 
	return t+dt, s*dt, yf4

And here’s an example usage of that:

import math
 
t = 0
dt = 1/40.0
dt_min = 1/100.0
dt_max = 1/10.0
y = math.sin(t)
 
print "Integrating cos(t) over 0
 
while t < 100:
	t, dt, y = rkf45(t, dt, y, lambda t,y: math.cos(t))
	if dt < dt_min: dt = dt_min
	if dt > dt_max: dt = dt_max
 
print "Final value with RKF45: "+str(y)
print "Final value (exact): "+str(math.sin(100)) # sin(x) = integral(cos(x))

Comment on Fourth order Runge-Kutta numerical integration by simmuskhan

simmuskhan — Fri, 22 Jan 2010 10:43:20 +0000

Thanks soooo much for the simple way you put that RK4 code together, really helped me see how it works, as I’ve had difficulty with some of the other ways I’ve seen it written.

Just a few questions if you’ve got the time/inclination to help out!

I’m putting together an orbital code, and need to use an Runge Kutta Fehlberg routine, but can’t see how to adapt it to work. I think I’m just not sure how the time fractions fit in with the function fractions.

Thanks again for the easy to read code, worked a treat!

Comment on AVRDUDE 5.8 with FTDI bitbang patch on Linux by Tomek

Tomek — Tue, 22 Dec 2009 10:28:11 +0000

thank you very much for respond. im pretty messed up all of that ;) stk500v2 uses FT232R am i wrong? ive installed any driver, anyone claimed that is necessary but i still think, the problem is lack of right communication between the computer and ft232r. thanks for answer, ill try to search avr freaks - its impossible im the first one who has such a problem ;)

cheers