go_away

Author Topic: modifing an existing rc controller to do basic wall following?  (Read 1322 times)

0 Members and 1 Guest are viewing this topic.

Offline little-cTopic starter

  • Robot Overlord
  • ****
  • Posts: 186
  • Helpful? 1
modifing an existing rc controller to do basic wall following?
« on: December 20, 2009, 12:01:57 PM »
problem, i have

basic wall following algorithm, proved to work. well the algorithm did in simulation. and its obviously going to actually work if its coded properly and the hardware is set up correctly. mostly cause its the most basic. there is minor issues with repeatability and dealing with doorways and so on. but they are not important, its meant to follow some walls in a "maze" not much more.
small naff(i.e cheep) remote control tank. (the base to the robot)
some ultrasonic sensors, micros, motors and stuff like that.
one utterly fried sealed Pb battery unit with short protection and charging circuitry. (yeah, you know the fuse holder bracket? it does matter if it breaks off its mounting...)

so the main question is (apart from a new battery) do i modify the existing controller, or get out the wire cutters and consign it to solder practice/parts bin?

answers preferably in the form of what sort of difficulties I should expect, the sort of things needed and vague ideas about potential costs please.

having no idea how the controller is actually coded, empirical evidence suggests,

motors have three states. 1,0,-1. there are two, both mounted through gearboxes to tracks.
some other functions I don't care about, mostly noise and light shows.
a motor controlling the turret through its range of rotation. again 3 states, gearbox. (I have intresting later applications for the turret...)
a pair of motors controlling the main gun thing with gearboxes and random stuff. < don't care. they are facing wire cutters and parts bin.
there is(was, the remote died to find out what it was sending to the tank) a remote control function bit. boring.

little bit about the robot design and use.
supposed to drive around ether following walls or avoiding objects. not really much more than that so far.
its a tank. two tracks and a turret, will have an ultrasonic range finder on the hull(servos and other stuff included, so it does left right identification)
basically its a bit naff. its designed to impress little cousins/grandparents/uni grants people, with cool little thing running around by itself.

it can work happly using the forward/stop/back function and two tracks. it has nine possible paths of movement. including stopped. can work with that for above. doesn't lend much to been able to play with stuff latter on though nasty horrible paths around objects.
 
I'm a robotics student, so point me at a giant textbook and I'm happy. design and coding and so on isn't a problem. its more if its worth doing a lot of work on potently a limited use platform. (in otherwords, it looks cool, but isn't really going to cope with anything more hardcore than a lawn on a mildly damp day) given I have until mid January before I really need anything that vaguely works. and latter on in the year before I actually need to produce something remotely clever/innovative.

future use, not actually designed properly yet. cause I haven't worked out how to actually move the robot yet! and its like week 2 of starting robot project tomorrow.
additions of infrared rangefinders, more servos, bigger micro, more sensors, wires etc. and it will map its world! and do semi clever stuff. which links to the money/parts issue, which the dumb randomly driving around not hitting stuff is supposed to solve.



my normal field of intrest is data logging, and sensor arrays. and generally clever stuff with that data.  making it drive around isn't something I normally do.
Its ether data logging an object moving of its own accord or just moving sensor positions via a nice little servo or two. whilst been bolted down to something solid. unfortunately for me there is a mobile robots practical latter on in the course... and turrets are only so intresting.

intresting note from last battery creation. fuses work quite well for short protection. except when the terminal nearest the battery connects with the opposite terminal on the battery. cue flames, heat and masses of thick choking smoke. then the wire burnt through. pretty horrible stuff. and difficult to explain to the parents...
so when working with battery's bring the fire blanket up from the kitchen!

Offline little-cTopic starter

  • Robot Overlord
  • ****
  • Posts: 186
  • Helpful? 1
Re: modifing an existing rc controller to do basic wall following?
« Reply #1 on: December 23, 2009, 09:55:35 AM »
with more reading up on things like H bridges, design is sorted... well small issue of the exact board layout to minimise the amount of board required, as well as not actuly shorting anything.


Im going to build the sensors as normal, and then have output pins controling the motors. pwm if I can get it to work.

old controler has gone to the parts bin.

new battery arived. yay!

so topic closed... for now.

 


Get Your Ad Here