Loading ...
Sorry, an error occurred while loading the content.
 

Re: [SeattleRobotics] Dead reckoning versus GPS for SRS Robo-Magellen contest

Expand Messages
  • Carbon
    Basically you build a map of the area using the waypoints and other data you may collect. Then plot a path from the start to the end that avoids the obstacles.
    Message 1 of 5 , Jul 4, 2005
      Basically you build a map of the area using the waypoints and other
      data you may collect. Then plot a path from the start to the end that
      avoids the obstacles. That path might look something like:

      (The robot is placed at the starting position pointed in a specific direction.)
      1. Travel 100 feet straight ahead.
      2. Turn 90 degrees to the left.
      3. Travel 175 feet in that direction.
      4. Turn right 75 degrees.
      5. Travel 50 feet.
      (Robot should now be within sight of the cone.)
      6. Turn on the camera and begin scanning for the orange code.

      Dead reckoning is all about being able to maintain a heading, measure
      distance travelled and make measured turns, which enables a robot to
      execute instructions like the above.

      For bonus points, put sensors on the robot to detect obstructions,
      move around any such obstructions, all the while keeping track of
      position and heading, making on-the-fly corrections to the pre-planned
      course to arrive at the original desination and heading.

      Actually, the real question is how do you use GPS to do this? Even
      with a good fix, you only know your position within a circle that's 20
      to 50 feet in diameter which is a big area compared to the entire
      Robo-Magellan course. Figuring out your heading is even worse - you
      can't figure it out from your position data: given the error in your
      position, you'd need to travel hundreds of feet to get a good heading,
      which you can't do on the course. GPS receivers use doppler shifts in
      the radio signal to infer speed and heading, which works great for a
      car or airplane, but robot speeds (in this contest) are too slow and
      bumpy to get good values from that technique.

      I think GPS is a red herring for Robo-Magellan, at least as the
      primary navigation method.

      Carbon

      On 7/4/05, Donna Smith <dsmith164@...> wrote:
      > Hello all,
      >
      > I believe I understand the basics of dead reckoning,
      > but I'd like to understand how other roboticists are
      > using this in place of GPS?
      >
      > The main question using dead reckoning is how to
      > utilize the given waypoint info using this technique,
      > or what is substituted for this to get from A to B?
      >
      > Thanks for the feedback,
      > Donna
      >
      >
      >
      > Visit the SRS Website at http://www.seattlerobotics.org
      > Yahoo! Groups Links
      >
      >
      >
      >
      >
      >
      >
    • andymich2
      The question on GPS is whether a differential ground station is within about 15 kilometers. If it is, the GPS accuracy can be as high as a 50% probability of
      Message 2 of 5 , Jul 8, 2005
        The question on GPS is whether a differential ground station is
        within about 15 kilometers. If it is, the GPS accuracy can be as
        high as a 50% probability of being within a 20 cm sphere.

        Many GPS receivers can now receive the differential signal and
        automatically correct for the errors introduced by the satellite
        signal's trip through the atmosphere. A GPS receiver receives
        position and timing signals from four satellites in orbit. It then
        calculates four equations for four unknowns: time and its position.
        Just like the twinkling of the light from a star, the GPS signals
        from the four satellites are slightly perturbed when traveling
        through the atmosphere, which is enough to vary the calculated
        position. The differential ground station is in a known position,
        and works the equations backward, to calculate the error in each
        satellite's signal. It then broadcasts this info, and differential
        GPS receivers apply this correction to their own calculations. The
        speed of the vehicle really doesn't affect the results.

        The question is whether the SeaTac airport has GPS landing
        capability, and is broadcasting the differential correction signals.
        Or if the Port of Seattle is doing the same for cargo ships.
      • Michael
        Speed doesn t effect the results but it does adjust the effects of the error in the reading. If the vehical is moving 30 mph having a +-10m reading isn t all
        Message 3 of 5 , Jul 8, 2005
          Speed doesn't effect the results but it does adjust the effects of
          the error in the reading. If the vehical is moving 30 mph having a
          +-10m reading isn't all that large; while traveling 1mph that same
          error is devastating.

          --- In SeattleRobotics@yahoogroups.com, "andymich2"
          <andrew.michalicek@h...> wrote:
          > The question on GPS is whether a differential ground station is
          > within about 15 kilometers. If it is, the GPS accuracy can be as
          > high as a 50% probability of being within a 20 cm sphere.
          >
          > Many GPS receivers can now receive the differential signal and
          > automatically correct for the errors introduced by the satellite
          > signal's trip through the atmosphere. A GPS receiver receives
          > position and timing signals from four satellites in orbit. It
          then
          > calculates four equations for four unknowns: time and its
          position.
          > Just like the twinkling of the light from a star, the GPS signals
          > from the four satellites are slightly perturbed when traveling
          > through the atmosphere, which is enough to vary the calculated
          > position. The differential ground station is in a known position,
          > and works the equations backward, to calculate the error in each
          > satellite's signal. It then broadcasts this info, and
          differential
          > GPS receivers apply this correction to their own calculations.
          The
          > speed of the vehicle really doesn't affect the results.
          >
          > The question is whether the SeaTac airport has GPS landing
          > capability, and is broadcasting the differential correction
          signals.
          > Or if the Port of Seattle is doing the same for cargo ships.
        Your message has been successfully submitted and would be delivered to recipients shortly.