Welcome to PhantomPilots.com

Sign up for a weekly email of the latest drone news & information

Proximity Sensors for Phantom 2?

Discussion in 'General Discussion' started by ehpepper, Jan 15, 2014.

  1. ehpepper

    Joined:
    Jan 9, 2014
    Messages:
    13
    Likes Received:
    0
    I am interested in Beyond Visual Range (BVR) flights with my Phantom 2. There appear to be several issues still to be resolved with the firmware of the Naza M V2 flight controller embedded in the Phantom 2; it does not currently support the Bluetooth or Datalink modules... I have wasted a great deal of time figuring this out... Ultimately, I want to be able to use the GroundStation Software to upload waypoints and BVR flight paths, but i'll have to wait till a firmware update to use the Datalink & software I already bought, or swap out the Phantom's Flight system.

    Which brings me to my next question -- Collision Aviodance. In BVR flight, there exists the very real possiblity of flying into a tree, utility/cell tower, utility line, etc. Does anyone know if a proximity sensor array has ever been fitted to a multirotor/Phantom, and if so, where can I find one or the plans to build one? Ideally, a sensor array that connects to the CAN bus and fits on top of the Phantom would be an easy solution.
     
  2. Ksc

    Ksc

    Joined:
    Jul 6, 2013
    Messages:
    973
    Likes Received:
    4
    Location:
    Naples, FL
    only you can prevent forest fires. its very possible to crash into something. which is why if you plan on flying BVR you are going to want to know the area ahead of time. Should you fly BVR and hit a transformer on the pole and it explodes then falls into someones yard and kills a family of 6 who were playing in the pool (highly unlikely but just wait), I'm sure you'll be the one facing the judge for that. Sensors aren't a bad idea but the problem is that for it to work, it must connect to the naza and tell the phantom to stop or move around the obstacle. Personally I don't see this becoming something you'll see. Mainly because of liability. If you crash while using the sensor system, DJI can be held liable. There is no benefit for DJI to make a product where solving the problem may cost them much more than the problem itself.
     
  3. ehpepper

    Joined:
    Jan 9, 2014
    Messages:
    13
    Likes Received:
    0
    All good points, (which I have considered.

    The system I am proposing would be linked via CAN bus to the flight control system to intervene and negotiate around an obstacle (360 degree array plus overhead sensors)

    what about the drones Amazon was proposing for deliveries? How would they avoid dogs, people cars, etc. when delivering a package?
     
  4. Ksc

    Ksc

    Joined:
    Jul 6, 2013
    Messages:
    973
    Likes Received:
    4
    Location:
    Naples, FL

    the publicity stunt drones you mean? Its a neat concept and brought amazon a ton of publicity. in fact amazon prime sales were so high that they had a hard time fulfilling all the orders placed.
     
  5. OI Photography

    Joined:
    Sep 23, 2013
    Messages:
    5,542
    Likes Received:
    17
    Location:
    Clarksville, TN
    They wouldn't, which is why we won't be seeing them for real anytime soon if ever. That's only one of the many obstacles they would have to overcome that don't have a ready solution at this time.

    I'm sure somebody at Northrup or BAE or whatever are working on exactly this idea as we speak, but there's nothing like it ready to go for consumers right now.
     
  6. ElGuano

    Joined:
    Sep 4, 2013
    Messages:
    2,219
    Likes Received:
    4
    I'm with KSC 100% on this one. If you want to create something that controls the quad to avoid obstacles, more power to you. Maybe check on the 3DR forums, they're a lot more dedicated to autonomous flight than the low-end consumer division of DJI.
     
  7. GneeChee

    Joined:
    Jan 25, 2014
    Messages:
    164
    Likes Received:
    0
    Location:
    Oklahoma City
    Well dang!

    I'm laying in bed this morning ... half asleep/half dreaming of good flying weather ... and an idea hit me ... what I thought was a GREAT idea ... A PROXIMITY SENSOR for the Phantom! YAY!

    So, I jump online - do a quick search of the forum - and I found this thread ... thanks for bursting my bubble, guys ... you shot my idea down even before I got started.

    LOL

    Oh well ... I guess it just proves the old adage ... "there is nothing new under the sun". :D
     
  8. TechJunkieRC

    Joined:
    Jan 12, 2014
    Messages:
    87
    Likes Received:
    0
    I don't know about DJI building one but go to RCModelReviews on YouTube and shoot Bruce a message. He has stated that he has a system under development. I don't believe it is necessarily for a DJI Naza but he may know the current state of sense and avoid systems and availability if he is developing one himself. Might help answer your question on if one is available or when one might be available. He stated on is channel that his is in test. He as not provided any specific details beyond that.
     
  9. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,932
    Likes Received:
    1,804
    Location:
    Lost Angeles
    There are some inexpensive LiDAR systems rapidly coming to the market and being tested by 3DRobotics but don't expect any plug and play solution for the Phantom. You need to incorproate a whole new set of FC logic to make it work. You could hack a solution using an Arduino spoofing or even piggybacking on to ground station CAN Bus messages but that sounds pretty involved. The Phantom 4 will have full collision avoidance capability.
     
  10. mandrakecr

    Joined:
    Jun 4, 2014
    Messages:
    2
    Likes Received:
    0
  11. ianwood

    ianwood Taco Wrangler
    Staff Member

    Joined:
    Jan 7, 2014
    Messages:
    4,932
    Likes Received:
    1,804
    Location:
    Lost Angeles
    Those IR sensors have a very limited range so you'll need to crawl for the Arduino to have any chance of interceding in time. And they don't do very well in direct sunlight so you can expect a number of false alarms. And you'll need to hijack the S.BUS which puts your system on the "critical path" (i.e. if it breaks, you lose control) unless you can engineer a fail-to-passthrough (which has no guarantees of working either).

    Lots of challenges.
     
  12. syotr

    Joined:
    Jul 30, 2013
    Messages:
    896
    Likes Received:
    135