Shop OBEX P1 Docs P2 Docs Learn Events
Dead Reckoning? — Parallax Forums

Dead Reckoning?

rwgast_logicdesignrwgast_logicdesign Posts: 1,464
edited 2012-10-18 06:10 in Robotics
Since I should have a fairly usable bot with motor control ready very soon, especially now that I scored an old nice RC chasis to strip parts from, ive been thinking about what I want my bot to do. I realized I have almost no sensors!!! But I will hopefully be using the mouse technique to make wheel encoders for motor feedback, after im all set up and everything is working.

I know dead reckoning is basically just the bot tracking where it has come from and rembering its path. But im curious to maybe hear a little more about how this is achieved Im guessing just encoders based on this http://www.ikalogic.com/wfr-a-dead-reckoning-robot/ . Is there a way to tell exactly how many inches youve gone straight and how many degrees you turned all without anything but wheel encoder?? Is this kind of navigation accurate if you dont have any real specs like your motors RPM raiting etc, id assume you may need that for some equations, idk. Im just curious to hear a little more about dead reckoning without out getting into code or math, ill save that for when I understand what all can be done using dead reckoning. Plus I like saying Dead Reckoning it sounds like the sequel to max pain or something :).

I was reading about the mars rover curosity, since im interested in autonomously navigating the dessert with my bot. What I read about how it plots courses using DR and stero vision was just amazing, it made me realize that maybe my bot will never cruse the Deseret becuase the DR stuff combined with stero vision is just wayyy over my head. But thankfully on earth we have compasses and gps :) so maybe my dream will come true.

Comments

  • Martin_HMartin_H Posts: 4,051
    edited 2012-10-16 06:16
    You can do dead reckoning open loop (no encoder feedback) or closed loop (with encoder feedback). Whit had an impressive maze navigation done on the original blue scribbler which was open loop. The problem with open loop is repeatability, small errors cause your robot to creep out of sync with where it thinks it is. Erco's figure eight challenge is done open loop which is why we all get so excited about a relatively simple task.

    I've done a few closed loop encoder projects using information I learned from Phil Pi's wheel encoder document (after reading it several times). The document and accompanying code are dense but great, and I suggest making them required reading for anyone building a robot. Don't worry if it takes a bit of work for it to sink in.

    Here's a link to the Boe bot encoder kit which has source code examples:

    http://www.parallax.com/Store/Robots/RoboticAccessories/tabid/145/ProductID/80/List/0/Default.aspx?SortField=ProductName,ProductName

    Here's a direct link to the PDF:

    http://www.parallax.com/Portals/0/Downloads/docs/prod/datast/ApplyEncoder.pdf

    Here's the sort of project you can do using encoders and dead reckoning:

    Here's Whit's Scribbler Maze navigation.
  • Tom CTom C Posts: 461
    edited 2012-10-16 06:23
    rwgast_logicdesign,

    Here is everything that you have ever wanted to know about dead reckoning complete with theory and examples: http://www.geology.smu.edu/dpa-www/robo/jbot/index.html

    Regards,
    TCIII
  • zappmanzappman Posts: 418
    edited 2012-10-16 08:20
    Tom C wrote: »
    rwgast_logicdesign,

    Here is everything that you have ever wanted to know about dead reckoning complete with theory and examples: http://www.geology.smu.edu/dpa-www/robo/jbot/index.html

    Regards,
    TCIII

    Thanks for the link Tom C.
    I am in the process of building a robot and the dead reckoning info found in your link is exactly what I was looking for.
    Information on my Robot build can be found at the Savage Circuits Fourms in this thread ZappBot 4.0 - Rover
  • ercoerco Posts: 20,256
    edited 2012-10-16 08:54
    Dead reckoning on a robot? It'll never work! :)

    http://www.youtube.com/watch?v=PX0IhUqnwrk

    my journey http://forums.parallax.com/showthread.php?118861-Homebuilt-Robotic-Platform-AKA-Retrobot
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-10-16 12:21
    Ever since seeing erco's Retrobot that could play back its commanded path, I wanted to make something similar.

    My version of a path memorizing robot is kind of a cheat though since it doesn't really memorize the robot's path, it memorizes the commands to the CR servos. Since I'm using a SD card to save the path infomation, I can afford to record each pair of servo pulse commands 50 times a second. I was surprised how well this method worked. This method of recording a path has an added advantage of being able to record curved paths.

    My main reason for purchasing a Scribbler 2 was in hopes of understanding Phil's encoder code for it. I don't claim to understand the S2 code completely but I've learned a lot from reading the S2.Spin object.
  • dmagnusdmagnus Posts: 271
    edited 2012-10-17 06:17
    [QUOTE= Plus I like saying Dead Reckoning it sounds like the sequel to max pain or something :).

    [/QUOTE]

    You should know that it's really DED Reckoning. It means [DED]uctive Reckoning. Just sayin'. Everybody gets it wrong, including some navigation textbooks...
  • Martin_HMartin_H Posts: 4,051
    edited 2012-10-17 06:46
    I've never heard of ded instead of dead. So I went to the internet and found this:

    http://www.straightdope.com/columns/read/2053/is-dead-reckoning-short-for-deduced-reckoning
  • Duane DegnDuane Degn Posts: 10,588
    edited 2012-10-17 06:56
    dmagnus wrote: »
    You should know that it's really DED Reckoning. It means [DED]uctive Reckoning. Just sayin'. Everybody gets it wrong, including some navigation textbooks...

    Wow! I'm one of those everybodies.

    I'm afraid it's a lost cause though. After reading through Wikipedia's Dead_Reckoning page it looks like "dead reckoning" has established itself in literature as well as science.
    Dead reckoning in literature In Walden, Henry David Thoreau suggests the following approach to life:
    "In the midst of this chopping sea of civilized life, such are the clouds and storms and quicksands and thousand-and-one items to be allowed for, that a man has to live, if he would not founder and go to the bottom and not make his port at all, by dead reckoning, and he must be a great calculator indeed who succeeds."
    In Moby Dick, or, The Whale, Herman Melville states on page 507: "...and in these same perilous seas, gropes he not his way by mere dead reckoning of the error-abounding log?"
  • Tom CTom C Posts: 461
    edited 2012-10-17 09:01
    erco,
    There are ways (processes) to reduce error and establish repeatability with dead reckoning. The link I posted above delves deeply into dead reckoning theory and how to make it exceptionaly repeatable.
    Just a thought.

    Regards,
    TCIII
  • dmagnusdmagnus Posts: 271
    edited 2012-10-17 11:45
    I figured I'd stir up a hornet's nest with that post. (The devil made me do it. :innocent: ). I once had a long, drawn out "argument" with a well known sailing author about the subject. He insisted that the "dead" reference meant, as in the article mentioned elsewhere here, "dead on", as in very accurate. I had to remind him that ded/dead reckoning was anything but accurate. We finally agreed to disagree. He isn't going to stop using "dead" in his writing - I'm not going to stop calling him on it. :smile:
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-10-17 12:18
    I've always assumed that, by "dead reckoning," you are pointing and moving towards a target position that is "dead ahead," without reference to other cues or sources of error. It's so natural that any other derivation seems superfluous by comparison.

    -Phil
  • dmagnusdmagnus Posts: 271
    edited 2012-10-18 06:10
    Phil Pilgrim (PhiPi) wrote: »
    I've always assumed that, by "dead reckoning," you are pointing and moving towards a target position that is "dead ahead," without reference to other cues or sources of error. It's so natural that any other derivation seems superfluous by comparison.

    -Phil

    I beg to differ, but I will never win the argument for the bastardization of the term. That is par for the course in our language.

    From Wikipedia"
    In navigation, dead reckoning (also ded (for deduced) reckoning or DR) is the process of calculating one's current position by using a previously determined position, or fix, and advancing that position based upon known or estimated speeds over elapsed time, and course.

    The "fix" may have been determined by celestial means, which is, at best, accurate to within about 1-5 miles. Or it could be just a guess or a triangulation by taking bearings on visible objects. Then you calculate a future position based upon known factors such as wind, current, etc. So that in itself adds another error factor to your [ded]uction. As you can see, it could be anything but dead ahead.
    Of course, now we hava all kinds of navigational aids to make DR practically unnecessary and will predict where you will be constantly - until the batteries go dead...
Sign In or Register to comment.