Shop OBEX P1 Docs P2 Docs Learn Events
Arlo Platform can go down stairs . . . once. — Parallax Forums

Arlo Platform can go down stairs . . . once.

ChrisL8ChrisL8 Posts: 129
edited 2015-03-03 08:52 in Robotics
The good news is that the Arlo Robotic platform is amazingly sturdy! It can even survive tumbling down a flight of stairs with only minor damage!
The bad news is that I now know this form experience. :(

The only serious damage seems to be the front caster:
The main axle is bent, but still turns,
but it also keeps falling off now, which seems to be a result of the main swivel shaft being bent, causing it to work itself out of the set pin. So I will need to replace the front caster.


Minor damage:
All of the screws that hold the battery "pegs" in place were bent, which is understandable since the batteries are the heaviest part.
One of the big screws that holds the upright post to the bottom plate was bent (that is a big screw!).
Three of my PING/IR holders snapped, but the sensors themselves still operate fine. I think I can superglue the holders back together.

Amazingly the ASUS Xtion and the laptop survived!
The Xtion snapped its base off, so it now sports wire tie.
The laptop has a new dent, but it isn't the first (I got it used) and it still functions.

I'm experimenting with "cliff sensors" now, as well as a large red sign for the basement door . . . maybe even a buzzer. :)

Comments

  • WhitWhit Posts: 4,191
    edited 2015-03-02 12:55
    Did you witness the fall? Amazing and awful all at once!
  • mindrobotsmindrobots Posts: 6,506
    edited 2015-03-02 13:56
    Ouch! (it is good to know you have crash tested it)

    The "other" good news is that this DID NOT happen when you were testing your cliff sensors at the time! That would have been a setback in your development program.

    "big red sign" - "buzzer" ??? Oh, you're going to make your Arlo see red signs and hear warning buzzers??? Very, very ambitious!! :o)
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2015-03-02 16:19
    Many years ago (early 80s) I had a mint condition Big Trak that I was working on when my little sister pushed it down a flight of stairs. Unfortunately it wasn't nearly as durable as the Arlo. But the point is, I feel your pain. =)
  • ercoerco Posts: 20,255
    edited 2015-03-02 17:49
    Sorry to hear, Chris L8. Totally ironic that you have this fairly significant navigational "anomaly", given that you have developed and shared here your very sophisticated mapping algorithm. Hope the damage is easy to repair and that your Arlobot never experiences 32.2 feet/second/second again.

    High tech sensors are well and good, but maybe a precautionary board barrier at the top of the stairs is not a bad idea. Roomba has its battery powered "virtual walls" (IR beacons) to stop a bot from going through a doorway, but most people I know just throw down books & boxes instead. Nothing wrong with low tech if it works!
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2015-03-02 18:15
    Many years ago (early 80s) I had a mint condition Big Trak that I was working on when my little sister pushed it down a flight of stairs. Unfortunately it wasn't nearly as durable as the Arlo.

    Hope you salvaged the twin motor out of it before tossing it! Tens of thousands of homebrew robots were made from surplus Big Trak motors. The magnetic clutch was really neat.
  • NikosGNikosG Posts: 705
    edited 2015-03-03 01:15
    Maybe a distance sensor on front of the robot looking the grownd for edges and gaps is necessary.
    Very simple but no one do this.

    My school has a lot of stairs. I'm planning to do some demonstrations with my robot Artist there.
    I think that I must take a precaution to avoid a similar incident...
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2015-03-03 08:34
    Hope you salvaged the twin motor out of it before tossing it! Tens of thousands of homebrew robots were made from surplus Big Trak motors. The magnetic clutch was really neat.

    Gordon,

    Sadly that was in my younger days and while I did attempt to rebuild the robot using what was left, I can't say that it was a complete success. Let's just say I didn't really understand what an H-Bridge was back then and that 2N3904 and 2N3906 transistors weren't sufficient to control a motor that size. =/

    It was many years before I truly appreciated how transistors work.
  • ChrisL8ChrisL8 Posts: 129
    edited 2015-03-03 08:52
    Thank you all for your kind words!

    I have gotten it all back together and it seems to operate fine. I was worried about the motors/axles/wheels, but it is driving fine and the odometry seems good.
    I think that those thick plastic plates do a great job of flexing just enough to absorb impact so that the metal parts don't bend or break.
    I do have the aluminum wheels, so they are very sturdy.
    Of course, something could be tweaked that I don't know about, but I think it would affect the odometry if it was. That or make noise.

    The only real casualty seems to be the front caster. It does not swivel properly anymore, probably due to a bent shaft.
    The robot overcomes this, but eventually the entire front caster works its way out of the base plate and falls off, which is very embarrassing.

    A couple of the acrylic sensor stands are snapped off too.

    ---

    My "low tech" solution is to keep the door to the basement closed, but someone opened it. It needs a lockout. :)

    ---

    My high tech solution is two IR sensors pointing down at an angle in front and set up the robot to stop if the distance increases above a certain amount.
    I tried this yesterday.
    It works in my preliminary testing, but I need to set up a good test and test it more thoroughly before I trust it, but of course the test needs to be safe too, and I need the front caster to not fall off during testing. :)
Sign In or Register to comment.