Drone News & Drone Directory

DIY Drones Blog

Precision Landing Solution for Commercial Drones

Precision landing is a critical requirement for a large number of commercial drone applications, be
it autonomous routine patrols for security & surveillance, package delivery at multiple locations,
remote inspections using docking stations or a GPS-denied environment like a warehouse. It is
one of the key components for automating and deploying drone operations at scale.

GPS alone is not accurate, IR beacons get affected by surrounding conditions and require power
on the landing site, RTK-GPS is complex to setup, requires additional infrastructure and it still
does not give the desired results.


FlytBase, the company bringing intelligence and connectivity to the
drones, today announced the release of an automated precision-landing solution, FlytDock- the
world’s smartest visual target landing solution, compatible with the widest range of drones.

FlytDock enables the drone to precisely align and land itself on the site with a centimeter-level
accuracy. It works across conditions; whether it is landing in day or night, outdoor or indoor
(GPS-denied) environment, on a ground-level or elevated platform, or even on a moving or floating
(in water) platform. Powered by FlytOS, this intelligent plugin utilizes computer vision techniques
and dedicated landing algorithms to precisely align, approach and land the multirotor on a visual
marker on the ground. There is no infrastructure/electronics required on the landing site,
making it easy to deploy at scale. Further, the system can be remotely managed and controlled
over cloud (4G/LTE).

FlytDock is readily compatible with DJI Enterprise, Ardupilot, and PX4 based drones.

To learn more visit https://flytbase.com/flytdock

Read the full blog.

Read Full Story

Use speech recognition to control a flight

In this video I present the voice recognition feature in my GCS/Companion-app-combo FlightZoomer. The feature allows to control the flight of an RC aircraft (or multicopter) using simple voice commands. In order to get an additional safety level against wrongly recognized commands, the engine does read back whatever sentence it understood. By affirming the readback, the command is then activated. Just check out the video to see how this works.

On this image you can see the flown route. At point 1 I activated the autopilot (which b.t.w. is a copy of the Boeing 787 autopilot), then I commanded some left turns: first to a western course of 270° (point 2), then south 180° (point 3), then 50° (point 4), 270° again (point 5) followed by a right turn to 50° into the downwind (point 6). At point 7 a descend to 470 meters is initiated. Finally there is a turn to 150° into the base (point 8) and at point 9 the ILS approach is activated. The turn into the final approach (point 10) and the subsequent landing (point 11) was then performed fully automatic.

As you can see on the video, the voice engine really performs nicely and understands my commands well. You can also see how it misunderstands me in once case (and how the misunderstanding is resolved) when I pronounce a number wrongly (I am not a native English speaker).


FlightZoomer is a software solution, that runs on off-the-shelf devices and offers all you need for instrument flight and much more (e.g. a cellular network databus between gcs and companion computer, the companion computer is a smartphone, optional usage of 3rd party radio telemetry, embedding of the FPV camera feed in the cockpit app, synthetic FPV view based on telemetry alone).


The solution runs on top of the Ardupilot stack and will be released shortly. More information on https://flightzoomer.com/

Read Full Story

Archon.ai autonomous drone for industrial plant supervision and inspection.

Preview of Archon Sentinel the autonomous system for inspection of industrial plants developed by Archon.ai.

The system is equipped with a dronestation capable of supporting different types of terrestrial and flying robots.

It's based on edge computing technologies that guarantee continuity of service even in the absence of broadband Internet connection. A remote operator is able to control robots in real time, schedule and schedule inspection missions. The drone station is equipped with GPU power to support third-party apps and I.A. Technologies based on neural networks. We are looking for industrial partners to develop the project if you are interested contact Info@archon.ai 

The drone tecnology is developed around a custom vr brain hardware : arducopter and ardurover firmware. On board we use a companion computer for AI managment functionality based on Neural Network tecnology ...

The control of the drone (rover , copter , quadplane ) could be in blos mode , remotly by a web app .

We are start in europe some test with Enac blos flight. 

This is an example of neural network tecnology used for indoor localization ... a drone station use this tecnology for robot auto docking .

Wer are seraching technology and industrial partner for implement this kind of  tecnology in different market around the world. 

If you need more info contact me at rn@archon.ai

Hope that you like our last development done by virtualrobotix and archon team.

Thanks to Ardupilot team for great work done in last years around our great software :) 

Roberto Navoni

Read Full Story

Bearingless Rotor – Update

Test Video

I was finally able to perform a test of the new drive test stand.  The previous test stand used a 4-stroke engine and a less rigid test frame.  This test stand consists of a 6.5 HP, 2-stroke DuraForce engine coupled to a V-Belt centrifugal clutch with a custom made drive shaft.  Couplings between the engine/drive-shaft and drive-shaft/clutch are stainless steel flex-couplings similar to the type used on the Robinson tail rotor shaft drive.  I added a large flywheel to the engine to attenuate the torque pulses from the engine.  The flywheel also makes starting the engine much easier.

The variable pitch mechanism has the following characteristics:

I chose to NOT include a dedicated follower link which forces the upper half of the "swashplate" to follow the speed of the rotor.  Instead, the attachment of the pitch links to the control horn on the rotors has only one degree of freedom, and thus uses the pitch links to transmit torque to the upper (rotating) portion of the swash plate to make it follow the rotor.  I plan to improve this design on my next rotor build, but for proof of concept, this seemed to work reasonably well.  The final product will have 4 blades on each rotor, so I will just need to make sure of proper clearance between the control horn of one rotor, and the flex-element of the neighboring blade.

The lower half of the swashplate is connected to a giant scale servo at only one point.  I was aware that this actuation force, being applied offset from the centerline of the rotor shaft, could induce some binding or sticktion in the variable pitch collective control.  I wanted to test this approach because it so greatly simplifies the connection between the servo and the swashplate.  To combat this tendency to bind, I designed the upper portion of the swashplate as a quite long "sleeve" which slides up and down on the OD of the rotor shaft. 

The "stickiness" issues I experienced during this test I believe are due to the following effects and design choices:

  • The upper swashplate "sleeve" material is 6061 aluminum which has a quite high coefficient of friction with steel.
  • The rotor blades, in their relaxed state, have only a slight pitch angle (3 degrees) but the design pitch is around 15 degrees, so the pitch servo and linkage must overcome the torsion of the 3/8" fiberglass flex-element in order to achieve the design pitch of the rotor blades.
  • Additionally, when operating at design RPM of 1410, this blade induces approx 1600 Lbs of centrifugal force, or tension, on the fiberglass rod flex-element.  This force tends to make the blades want to straighten out, back to their relaxed position, and thereby increases the amount of force/torque required to achieve the design blade pitch.

To mitigate these issues, on the next hub design, I will attempt to:

  • Lengthen the upper sliding section of the swashplate.
  • Use lower friction brass or teflon sleeves as linear bearings.
  • Build the rotors with the blades at flight operating pitch in their relaxed state.
  • Keep the pitch servo attachment to the swashplate as close to the centerline of the rotor shaft as possible.

Another problem encountered during the test was the detection of false triggers from the engine RPM photo-interrupter.

The blue line is the engine RPM and the orange line is the clutch RPM.  You can see all of the high spikes which go off the chart for the engine RPM.  This means I was picking up false triggers on the photo-interrupter input.  My circuit does include Schmitt Triggers on the tachometer signals, which seems to be working well on the other two photo-interrupters, but was getting some noise on the engine tach.  I will have to hook up an oscilloscope to see what the issue is.

FYI, the Grey trace is the rotor RPM (approx 2.8 : 1 reduction from the clutch),

The cyan trace is the pitch servo signal in microseconds (inverted)

The yellow trace is the throttle servo signal in microseconds.

So while I discovered some issues (engine power, and pitch mechanism stickiness) I had basic success of proving that I could operate a home-made set of rotors, with an internal combustion engine, without excessive vibration.

I would still like to over-speed my rotors to ensure the safety factor ( I used a 4:1 safety factor over the static tensile strength of the flex-element fiberglass rod).  I would like to take the rotor up to 2000 RPM, which would put the tensile force at 3200 Lbs, - double the normal operating force, but still approx 1/2 of the ultimate tensile strength of the material.

Next steps:

  • New carburetor for engine to get the power/RPM I need.
  • Tuned exhaust for engine (again, for power)
  • Run rotor at 2000 RPM to verify strength / safety factor.
  • Add lower friction sleeves to swashplate to reduce binding
  • Tilt rotor at speed to observe gyroscoping effects on vibration.
  • Attempt to control/balance test stand attitude using closed loop control from a gyro. 

 For those who think this is an ill-advised project, please refrain from negative comments.  I readily acknowledge the difficulties that await me down the development road.  I am working this project for my own satisfaction, not for a commercially viable transportation vehicle of the future.  I am enjoying the process and the knowledge and experience I'm gaining at each step of the way.  If you have questions or encouragement, I welcome those comments.

Thanks!

 

Read Full Story

ZTW Seal 300A Brushless Speed control For RC Boat Test Video

ZTW SEAL Series 300 amp 14s Lipo brushless esc.

source from : http://www.ztwshop.com/product/ztw-seal-series-brushless-esc/

Specs:
Waterproof design (rates as IP 67) allows this ESC to work in wet conditions.
Continuous Current (A): 300 Amps
Burst Current(A): 350 Amps
Voltage: 18-42 NiH : 6s - 14s Lipo
Power Wire: ,160mm-Red/160mm-Black
Motor Wire ,160mm-Black
BEC output: None requires rx pack.

Optional programing with LCD programming card.

Case dimensions (not including wires):
Length: 150mm
Width: 85mm
Height: 28.6mm
Weight: 640g

Includes: Everything in picture number one

Requires: Connectors, 8mm bullet connectors or larger.

Read Full Story

Page 2 of 82712345...102030...Last »