Programming

[archiver]

Posted by Nick at 1/15/2001 9:28 PM EST


Student on team #240, Mach V, from Jefferson Monroe High School and Visteon.



Joe,
Our team is attempting to mimic the drive system on your robot in 2000. How did you run through the logic of the programming. I'm having trouble figuring it out.

[archiver]

Posted by Joe Johnson at 1/15/2001 10:15 PM EST


Engineer on team #47, Chief Delphi, from Pontiac Central High School and Delphi Automotive Systems.


In Reply to: Programming
Posted by Nick on 1/15/2001 9:28 PM EST:



How our code works is a trade secret that our team is
not willing to share at this time. Sorry.

I will say this much, it is a very complicated coding
problem to make a robot that behaves as the driver wants
and to make the operator interface such that the driver
can clearly communicate what it is he/she is wanting
the machine to do.

I recommend that teams spend a lot of time thinking
about how to do both of these things.

I will wish you good luck, but I am not going to share
our solution just yet.

No hard feelings I hope.

Joe J.

P.S. It took me 40 hours to get the code right last
year. The result was a reasonably compact 9 lines of
code. Thinking is the key. The coding is easy once
you get your mind straight.

[archiver]

Posted by Nick at 1/15/2001 10:24 PM EST


Student on team #240, Mach V, from Jefferson Monroe High School and Visteon.


In Reply to: I would tell you but I would have to kill you ;-)
Posted by Joe Johnson on 1/15/2001 10:15 PM EST:



I'm already insane with programming the gyro sensor. By any chance what so ever is there PBASIC code to change the joysticks inputs to parametric or am I at least on the right track?

[archiver]

Posted by Joe Johnson at 1/15/2001 10:32 PM EST


Engineer on team #47, Chief Delphi, from Pontiac Central High School and Delphi Automotive Systems.


In Reply to: Well, no sleep for a while
Posted by Nick on 1/15/2001 10:24 PM EST:



What do you mean "parametric"?

Joe J.

[archiver]

Posted by Matt Berube at 1/16/2001 7:48 AM EST


Engineer on team #49, Delphi Knights, from Buena Vista High School and Delphi Automotive.


In Reply to: Well, no sleep for a while
Posted by Nick on 1/15/2001 10:24 PM EST:



: I'm already insane with programming the gyro sensor. By any chance what so ever is there PBASIC code to change the joysticks inputs to parametric or am I at least on the right track?

You might want to take some time to try to understand this then try to plot it in something like Excel or Matlab. I don't understand why it works but it will give you an output that roughly corresponds to the angle of the joystick. I wish I understood it better myself. Feel free to e-mail me at matthew.berube@delphiato.com if you have specific questions.

' Steer
Rear_Steer_Output = 127
'set dead band in center only
if (p1_y>124 and p1_y124 and p1_x127 and p1_x127) then mode9steer3:
Rear_Steer_Output = ((1127-(((127*(P1_x-127))/SQR((P1_x-127)*(P1_x-127)+((127-P1_y)*(127-P1_y)*2))))) MIN 1000 MAX 1254) - 1000
mode9steer3:
if not(p1_y>127 and p1_x>127) then mode9steer4:
Rear_Steer_Output = ((1127+(((127*(P1_x-127))/SQR( (P1_x-127)*(P1_x-127)+((P1_y-127)*(P1_y-127)*2) ) ))) MIN 1000 MAX 1254) - 1000
mode9steer4:

(ps)
this is kinda hard to read. It basically is breaking the joystick up into 4 quadrants and the equation for the output is : X / ( X^2 + Y^2)^0.5

Matt B.
T49

[archiver]

Posted by Nick at 1/16/2001 7:43 PM EST


Student on team #240, Mach V, from Jefferson Monroe High School and Visteon.


In Reply to: Try this.
Posted by Matt Berube on 1/16/2001 7:48 AM EST:



Thanks, maybe I can sleep tonight. I think I have found another way. Basically I changed the joystick grid to polar coordinates (cos y/sin x).

[archiver]

Posted by Matt Berube at 1/19/2001 7:54 AM EST


Engineer on team #49, Delphi Knights, from Buena Vista High School and Delphi Automotive.


In Reply to: Re: Try this.
Posted by Nick on 1/16/2001 7:43 PM EST:



: Thanks, maybe I can sleep tonight. I think I have found another way. Basically I changed the joystick grid to polar coordinates (cos y/sin x).

There is NO WAY to do trig functions.
I even tried a Taylor series aproximation. It gets way too complicated. Try messing with the equation I gave you. It is not exactly liniarly (sp) proportional to the joystick angle, but it is pretty close.

my $0.02

Matt B.
T49

[archiver]

Posted by Joe Johnson at 1/19/2001 10:11 PM EST


Engineer on team #47, Chief Delphi, from Pontiac Central High School and Delphi Automotive Systems.


In Reply to: Yah, I tried that too :=)
Posted by Matt Berube on 1/19/2001 7:54 AM EST:



Yes you can do trig in PBASIC.

The problem is a little messy if you need to do an
inverse trig function, but this can be arranged too.

The most confusing part when you first try to use them
is that the measurement of angle is in a very funky
unit. Basically the circle is divided up from 0 to
255, so the angles are measured in increments of 1.4
degrees -- not the easiest unit of measure but it works
in a pinch.

By the way, when we need to convert the X-Y coords of
the joystick into "polar" coords of Radius-Theta, we
found that it was close enough for us to just use the
following function as the "radius" rather than go
through all the bother of computing the "true" radius

Radius = abs(Y1 - $80) MIN abs(X1 - $80)

This function returns the "square" that the joystick is
on rather than the "cirle" that is is on, but again it
is close enough for robot control AND it is a LOT
simpler to code than all the nonsense needed to get the
"true" answer.

There are similar simplifications that can get you the
angle of the joystick. I leave those as an exercise
for the reader.

Joe J.

[archiver]

Posted by Matt Berube at 1/20/2001 12:14 AM EST


Engineer on team #49, Delphi Knights, from Buena Vista High School and Delphi Automotive.


In Reply to: don't be silly...
Posted by Joe Johnson on 1/19/2001 10:11 PM EST:



Hex makes my brain hurt! (EOM)

[archiver]

Posted by Joe Johnson at 1/21/2001 12:44 PM EST


Engineer on team #47, Chief Delphi, from Pontiac Central High School and Delphi Automotive Systems.


In Reply to: Hex makes my brain hurt! (EOM)
Posted by Matt Berube on 1/20/2001 12:14 AM EST:



I belive that in the long run, it is easier to use hex
in your PBASIC Code.

The numbers work out easier. The overflows and
underflows are easier to spot. The positions of various
bits in the byte are easier to work with.

Once you get the Zen of PBASIC, you will definitely
perfer to use hexidecimal numbers over decimal.

One man's opinion.

Joe J.

[archiver]

Posted by Nick at 1/15/2001 10:26 PM EST


Student on team #240, Mach V, from Jefferson Monroe High School and Visteon.


In Reply to: I would tell you but I would have to kill you ;-)
Posted by Joe Johnson on 1/15/2001 10:15 PM EST:



PS I understand. Thanks anyway. EOM