Welcome to InspirePilots.com

Join the leading DJI Inspire community for free!

Building an autonomous drone for academia, ideas on senor, req, etc

Discussion in 'General Discussion' started by Roy Hufe, Oct 7, 2016.

  1. Roy Hufe

    Joined:
    Oct 7, 2016
    Messages:
    1
    Likes Received:
    0
    Greetings,

    I am part of a thirty member team of mechanical, computer, and electrical engineers collaborating with Northup Gruman, NASA, Aerial Alchemy, and a few select other sponsors to build an autonomous drone to assist emergency responders. Currently we are scouring the industry and drone enthusiasts for ideas on requirements for a system that could record/transfer data in real time via video, camera, infared, flir, etc. Our primary clients are fire, law enforcement, hazmat, and ems. Any advice on sensors in the field already in use or prototypes coming within five years would be most helpful. I seek to learn anything from anywhere possible in this field.

    Respectfully,
    R. Hufe
     
  2. The Editor

    The Editor Moderator
    Staff Member

    Joined:
    Aug 7, 2013
    Messages:
    6,722
    Likes Received:
    3,878
    Good morning,

    I think the most important thing you should look at before anything else (sensors, imaging etc) is the safety aspect since these machine will (presumably) be deployed in potentially heavily built up/populated areas etc.

    You could look at designing a UAV with no single point of failure which currently doesn't exist. Currently there are platforms out there that have redundant IMU's (Accelerometers and gyroscopes) together with multiple magnetometers (Compasses). anything over a quad will give you redundant propulsion so a motor or esc failure will not result in a crash etc and there are many UAV's that offer multiple bus power systems to protect against battery problems whilst airborne.

    However, currently, there are no systems that give redundant flight controller options. This would mean two (or more) flight controllers processing in parallel with the same data feed from all sensors.

    This would then truely give you a 'No single point of failure' machine and would greatly ease the way for FAA and other aviation authority approvals around the world.

    The sensors and imaging are the easy bits.... :p
     
  3. RaptorMan

    Joined:
    Dec 18, 2015
    Messages:
    770
    Likes Received:
    214
    I've talked about this before and for me there are a number of issues that need to be addressed. First, there are regulatory issues related to maintaining visual sight of the drone which would severely limit its use for the applications you listed. But, if you can convince the FAA to permit it they would likely require sensors and transponders that might not now exist.

    The sensors I'm referring to are things like radar that can be used to prevent collisions and would also serve, potentially, to warn others of your presence. I think a transponder that reports the drones lat/lon/alt as well as velocity (speed and direction) and perhaps climb rate. Looking at this from the standpoint of defining a new system with expandability to other AC types I might also include in the transponder data stream AC identifiers that include AC type and tail number.

    You are charting new waters here and the task of getting a bird to fly is likely only 10% of the job and could be much less than that. For me, the biggest challenge is to develop the system infrastructure to integrate autonomous drones into the airspace and there's no getting around the need to have the FAA involved and later, other nations aviation authorities. There are clearly many applications that benefit from or outright require autonomous flight and beyond visual range. Other's looking at this and perhaps someone that you can partner with is Google, Amazon etc.


    Brian
     
  4. licensed pilot

    Joined:
    Dec 11, 2015
    Messages:
    654
    Likes Received:
    278
    Location:
    CASA GRANDE, AZ
    A long endurance aircraft is a necessity for public safety operations (endurance should be the prime objective), and that generally requires a fixed wing (or large batteries, reducing payload). Of course fixed wings require a landing/t.o. area and/or launch systems, which may not be available in a disaster zone.I would explore a VTOL approach. As the editor said, the sensors are the easy part. As a retired chief of police I'd like to see something that can be modular and carried in the trunk of a squad car or an SUV, an assembled in minutes and simple to fly. Take a look at the FireFly 6 as a starting place.
     
    The Editor likes this.
  5. The Editor

    The Editor Moderator
    Staff Member

    Joined:
    Aug 7, 2013
    Messages:
    6,722
    Likes Received:
    3,878