4 ms·

The follow feature is killer. One day maybe when people go places drones will always follow, will really help for safety/documentation. For instance a trip in

The follow feature is killer. One day maybe when people go places drones will always follow, will really help for safety/documentation. For instance a trip into the Grand Canyon or a skiing trip if things went awry. Battery life, networking/communication distance needs to be much longer but there are so many possibilities. Maybe we'll see Russian videos with drones in addition to their dash cams that follow their drive.

robotresearcher 12y ago
> a trip into the Grand Canyon GPS needs a very good view of the sky, from lots of angles, to work well. Lots more work to do to do it reliably outdoors with patchy-or-no GPS.
- drawkbox 12y ago
  True. One thing you could do is have the drones map out areas using models and maybe even a few physical root points for them to base it on. Mapping it themselves something like this Dyson 360 vacuum: https://www.youtube.com/watch?v=oguKCHP7jNQ https://www.youtube.com/watch?v=oguKCHP7jNQ
- jjwiseman 12y ago
  The NASA Airspace Operations Challenge is trying to advance the state of the art in UAV operation in conditions where there is no GPS or GPS is degraded: http://www.nasa.gov/directorates/spacetech/centennial_challenges/uas/ http://www.nasa.gov/directorates/spacetech/centennial_challe... As examples, inertial navigation or laser or vision-based techniques (VSLAM etc.) might be ways to deal with GPS issues.
  - hueving 12y ago
    Missiles used by the military already use inertial navigation to avoid GPS jamming issues so it's definitely possible.
    - asynchronous13 12y ago
      it's definitely possible. but the price point for inertial sensors that are accurate enough to work without GPS is a lot higher than the price of the MEMS sensors that are currently used.
      - robotresearcher 12y ago
        > it's definitely possible. You can't track an external, independently moving object using only inertial sensing.
  - robotresearcher 12y ago
    How does inertial navigation help with following people for a flying camera platform? Even if the subject carries a sensor, the reference frames will drift unbounded. Visual tracking should work, but is not reliable yet. Lots of people are working on it.
    - jjwiseman 12y ago
      Right, I'm thinking of inertial navigation to help out in a case where you have spotty GPS and you lose it for 10 seconds. It's not practical for these sorts of applications to use only inertial nav, but it can help in the region between full GPS and no GPS.
- cfo123 12y ago
  there is quite a lot of research on purely camera-based localization of drones. have a look at this video for example: http://www.youtube.com/watch?v=3mNY9-DSUDk http://www.youtube.com/watch?v=3mNY9-DSUDk
  - robotresearcher 12y ago
    Yep, Scaramuzza has been putting out some great work recently. So far this year, this work from TUM has been perhaps the most impressive, and the code is published online: http://vision.in.tum.de/research/lsdslam http://vision.in.tum.de/research/lsdslam
jkestner 12y ago
I want one to tag along after my child as she plays outside, with a voice channel and geotracking (video channel only when she asks for it). True helicopter parenting. Quadcopter parenting and all the interesting applications will require major improvements in battery capacity:weight ratio, though.
- drawkbox 12y ago
  Shipping drones will up the battery time and weight they can carry. But ultimately it will probably be a team of drones when doing tasks, may not even be owned by the controller, rented eyes. New drones tagging in after 15 minutes from another location to complete a full view without battery problems. Drones will be everywhere.
- jjwiseman 12y ago
  Paul Wallich did quadcopter parenting a couple years ago, using a 3DR Arducopter to follow his son 400 meters to the bus stop: http://spectrum.ieee.org/geek-life/hands-on/the-diy-kidtracking-drone http://spectrum.ieee.org/geek-life/hands-on/the-diy-kidtrack...

robotresearcher 12y ago

> a trip into the Grand Canyon GPS needs a very good view of the sky, from lots of angles, to work well. Lots more work to do to do it reliably outdoors with patchy-or-no GPS.

drawkbox 12y ago
True. One thing you could do is have the drones map out areas using models and maybe even a few physical root points for them to base it on. Mapping it themselves something like this Dyson 360 vacuum: https://www.youtube.com/watch?v=oguKCHP7jNQ https://www.youtube.com/watch?v=oguKCHP7jNQ
jjwiseman 12y ago
The NASA Airspace Operations Challenge is trying to advance the state of the art in UAV operation in conditions where there is no GPS or GPS is degraded: http://www.nasa.gov/directorates/spacetech/centennial_challenges/uas/ http://www.nasa.gov/directorates/spacetech/centennial_challe... As examples, inertial navigation or laser or vision-based techniques (VSLAM etc.) might be ways to deal with GPS issues.
- hueving 12y ago
  Missiles used by the military already use inertial navigation to avoid GPS jamming issues so it's definitely possible.
  - asynchronous13 12y ago
    it's definitely possible. but the price point for inertial sensors that are accurate enough to work without GPS is a lot higher than the price of the MEMS sensors that are currently used.
    - robotresearcher 12y ago
      > it's definitely possible. You can't track an external, independently moving object using only inertial sensing.
- robotresearcher 12y ago
  How does inertial navigation help with following people for a flying camera platform? Even if the subject carries a sensor, the reference frames will drift unbounded. Visual tracking should work, but is not reliable yet. Lots of people are working on it.
  - jjwiseman 12y ago
    Right, I'm thinking of inertial navigation to help out in a case where you have spotty GPS and you lose it for 10 seconds. It's not practical for these sorts of applications to use only inertial nav, but it can help in the region between full GPS and no GPS.
cfo123 12y ago
there is quite a lot of research on purely camera-based localization of drones. have a look at this video for example: http://www.youtube.com/watch?v=3mNY9-DSUDk http://www.youtube.com/watch?v=3mNY9-DSUDk
- robotresearcher 12y ago
  Yep, Scaramuzza has been putting out some great work recently. So far this year, this work from TUM has been perhaps the most impressive, and the code is published online: http://vision.in.tum.de/research/lsdslam http://vision.in.tum.de/research/lsdslam

jkestner 12y ago

I want one to tag along after my child as she plays outside, with a voice channel and geotracking (video channel only when she asks for it). True helicopter parenting. Quadcopter parenting and all the interesting applications will require major improvements in battery capacity:weight ratio, though.

drawkbox 12y ago
Shipping drones will up the battery time and weight they can carry. But ultimately it will probably be a team of drones when doing tasks, may not even be owned by the controller, rented eyes. New drones tagging in after 15 minutes from another location to complete a full view without battery problems. Drones will be everywhere.
jjwiseman 12y ago
Paul Wallich did quadcopter parenting a couple years ago, using a 3DR Arducopter to follow his son 400 meters to the bus stop: http://spectrum.ieee.org/geek-life/hands-on/the-diy-kidtracking-drone http://spectrum.ieee.org/geek-life/hands-on/the-diy-kidtrack...