Blogs

To optimize videophotography results, vibrations need to be reduced as much as possible. The objective is to disturb as little as possible a brushless gimbal holding a camera. (we all know too well how hard it is to get a brushless gimbal well tuned).

As most people, brushless gimbal tuning is done on the bench using accelerometers & gyros of the brushless gimbal's IMU. And like most, we do not get as good results when flying. Why?: frame & air vibrations. Frame vibrations are created by motors, props and wind. Direct air vibrations (external wind) also impact the brushless gimbal.

This is a blog post showing anti-vibrations objective measurements of different isolation systems, obtained on my Mr. Grey X8 workhorse shown on the introductory picture. All vibrations measurements are made during periods of min 30 sec of hover in stabilize mode. Take-off and landing periods are excluded of measurements.

Configuration of the test craft:

-Pixhawk controlled -Stabilize mode

-T-Motors MT3515-400Kv

-15x5 XOAR props

-6S batteries

-2 axis brushless gimbal - Alexmos

-Sony NEX5 camera

-Weight of the gimbal+camera+fpv+3S 1ah battery : about 1250 g

-AUW : about 6.8Kg

Vibrations measurement setup:

-All tests were made with the craft configured fully ready-to-film : except the gimbal which remained off not to measure its own generated vibrations. The camera was thus simply screwed on the gimbal with no power to the brushless motors.

-X-Cam vibration measurement tester

X-Cam is installed on the different areas of the craft to be tested. Its mesaurements are triggered remotely by a PWM pulse (on/off switch allocated on Taranis). It is powered directly through the PWM servo cable connected on the receiver.

Data are stored on board its flash memory. It is made for vibrations measurements in the 0-200Hz range and can store up to 3 min of continuous measurements.

Through the X-Cam GUI , data is downloaded from the flash memory to produce a time graph (basically useless as it shows the X-Y-Z vibrations measurements along the time axis), and a frequency graph showing X,Y,Z vibrations amplitudes (in number of g) versus frequency (in Hz). Raw data or weighted averaged data can be calculated.

Two anti-vibration systems were measured,

First a silicon ring system as shown below,

Four of these were used in a X configuration between Mr Grey's two plates (see introductory picture). Each plate is made of plywood; square shaped with 20 cm long sides.

Silicon rings come in two stiffnesses: red for softter silicon, black for stiffer silicon. Here, only the red softer ones were tested. Each Secraft system costs about 9 euros each (www.secraft.net) , which is relatively cheap.

Second tested system is a "Bell" shaped  precompressed silicon system. Each "Bell" is made for a 250g payload. Each one of them costs about 39 euros (www.altigator.com) which is quite expensive. It is mostly used in high end professional aerial photography rigs :

Four of those were used, in each corner in between Mr Grey's test plates as shown below:

Reference measurement:

In order to compare results, a reference measurement is made on the frame itself, with the X-Cam placed at the center of Mr Grey's top plate, under its protective "flower pot".

Frequency domain picture shows two frequencies were vibrations are significant : 85Hz and 175Hz; The measurements show about 0,3g of vibrations on all three axis at 85Hz; and 0,4-0,5 g vibrations at 175Hz.

First test result with Silicon rings system:

X-Cam is now placed on the second bottom plywood plate on which the gimbal is fixed,

The results show that frame vibrations are continuing to pass through to the gimbal at the same frequencies (85Hz and 175 Hz) but amplitudes are divided by a factor of 10 at 85 Hz and almost by a factor of 20 at 175Hz, versus non damped frame vibrations.

Second test with the "Bell" system:

X-Cam is placed at the exact same location as for the first test,

The results show that frame vibrations are continuing to pass through to the gimbal at the same frequencies (85Hz and 175 Hz) but amplitudes are divided by a factor of 2 at 85 Hz and by a factor of about 2 at 175Hz, versus non damped frame vibrations.

If we show the silicon ring's system on a graph at the same scale as Bell's system:

The silicon ring anti-vibes system literally erase vibrations in comparison with the "Bell" anti-vibe system....

Comparison of both anti-vibrations systems:

It appears clearly that the "silicon ring system" is much more efficient at reducing vibrations versus the "Bell" system. Further, this silicon ring system is much cheaper than the Bell system.

For once, more expensive, supposedly professionnal parts were not up to the task verus cheaper diy parts.

Cheers,

Hugues

Netflix joked about it, I thought I'd give it a try since my brother in law wanted to borrow one of my DVD's.  Here are the results.

Repost from MIT Technology Review:

How do you keep small drone aircraft safe in the world’s busiest national airspace? One idea is to have them use cellphone networks to feed data back to an air traffic control system made just for drones.

A startup called Airware is working with NASA on a project exploring how to manage the swarms of commercial drones expected to start appearing in U.S. skies. The four-year program will create a series of prototype air traffic management systems and could shape how widely commercial drones can be used. Airware’s main business is selling control software and hardware to drone manufacturers and operators.

The U.S. Federal Aviation Administration has yet to propose rules to govern the use of commercial robotic aircraft in U.S. skies. But it predicts that 7,500 unmanned craft weighing 55 pounds (25 kilograms) or less will be operating in the U.S. by 2018. There is strong interest from agriculture, mining, and infrastructure companies in using drones for tasks like inspecting crops or gathering geospatial data (see “10 Breakthrough Technologies 2014: Agricultural Drones”).

That could mean gridlock in the skies, or at least increasingly unsafe traffic patterns. “You will have competing interests trying to use the same space,” says Jesse Kallman, head of business development and regulatory affairs at Airware. “Imagine Amazon trying to deliver packages in an area that an energy company is trying to survey their power lines.”

The first prototype to be developed under NASA’s project will be an Internet-based system. Drone operators will file flight plans for approval. The system will use what it knows about other drone flights, weather forecasts, and physical obstacles such as radio masts to give the go-ahead.

Read more here: http://www.technologyreview.com/news/531811/air-traffic-control-for-drones/

Due to the safety aspects of the man drone we "invented" a combination of airbag and parachute with the purpose of opening the parachute in a spit second. We made a small testing sample. and we think it can work. So the idea is that a drone in an emergency situation would ignite a rocket parachute and that this one would open by airbag at the the top. Think that for normal drones (_tumble bee_ ) this might also be interesting. Anyone interested in such development, or sees a need for such parachute that should substantially reduce the the rescue height?

Best Winfried

PS in the blog of my www.thehummingmanproject.com I have a small document for download. 

This is a quick compilation video of a few FPV flights in Alt Hold, Sport and Drift flight modes earlier this week while on the beach. The wind was quite a factor and it shown in some of the bobbles, then again it is probably my poor piloting skills. 

Dear All,

As I announced on this forum we are aiming to build a man carrying drone; it is a multicopter. we have made quite some progress. Have a look: 

www.thehummingmanproject.com/blog

best regards Winfried

This question gets asked a lot on the forums and Facebook so I thought I'd make a video with some of the answers that I give to new pilots.  Feel free to add your own advice too.

We have been selected as one of 20 projects to compete in Engadget's competition. Vote for Easy Drone! http://insertcoin.engadget.com/

Remarkable photos and videos within this informative 11 minute interview with John Durban, a biologist with NOAA's Southwest Fisheries Science Center in La Jolla, California.

"For the first time, scientists have used an unmanned aerial vehicle to study killer whales from above. The device they're using is a remote-controlled hexacopter with a high-resolution camera mounted in its belly, and the photos it produces are beautiful and full of detail. The images offer an entirely new view of this species.

But scientists aren't taking pictures just because they look nice. The images contain detailed information that scientists can use to monitor the health of individual killer whales and of their population as a whole."

I would like to introduce droneCFD.com to the DiyDrones community. The purpose of droneCFD is to reduce the complexity of running Openfoam CFD simulations on small unmanned aircraft geometries. To do this, droneCFD uses the OpenFoam toolkit for meshing and flow solving, with a little extra automation to handle case setup, simulation domain configuration and parallel execution.

To get started with DroneCFD, you need to have OpenFoam 2.3.0 and PyFoam installed on your machine. You can install droneCFD using 'pip install droneCFD'. You can double check that OpenFoam is installed correctly by running 'dcCheck' in a terminal. Your first simulation only requires you to type 'dcRun', which will run a simulation based on a reference geometry packaged with droneCFD. More details can be found at dronecfd.com/gettingstarted.

This project was the result of many frustrating hours trying to simulate flow around a novel small unmanned aircraft geometry. It is far from perfect, but I'm hoping feedback and lots of testing will help improve it.

The code is located on Github, and I've setup a webpage with a details and a forum for collecting feedback. Please take a look, try it out and let me know what you think!

Danish Aviation Systems recently made a head impact test with eBee drone and upload the results today. 

Interesting video, from the description:

DAS have today conducted a human crash test to see what the effects of an eBee hitting a human direct in the head.
please see the link for result.

eBee during landing = 8.75J 
Frisbee = 15,8J
American football = 45J

Jordi Muñoz (left), Brandon Basso (center), and Chris Anderson (right) at 3D Robotics’ office in Berkeley, California, with one of the company’s Iris quadcopters, which sell for $750 and are aimed at recreational videographers. (All photos by Chris McPherson). Meet the inventors, tinkerers, and entrepreneurs at the forefront of the flying-robot revolution.

Congrats to Randy Mackay and the APM:Copter team, whose collaboration with Tohoku University was discussed in IEEE Spectrum today:

When Mount Ontake erupted in Japan a few weeks ago, it was completely unexpected. No significant earthquakes, no steam or gas releases, nothing. Usually, some warning does exist, and the best that we can to is to monitor active volcanoes as carefully as we can to try and spot whatever warning signs that are there. This is especially problematic with volcanoes that are undergoing frequent periods of activity, where it’s not safe to get close to them to determine when a minor eruption might turn into a major one. Not safe, you say? There’s a solution for that: send in the robots.

Keiji Nagatani, a professor at Tohoku University in Japan, has spent the last half decade developing robotic systems for volcano exploration. For the last few years, he’s been working on ways of exploring remote, potentially dangerous volcanic areas using UAVs in collaboration with ground robots. Here’s a video from a test last month on Japan’s Mount Asama:

The ground robot is called Clover, and the UAV is called Zion. The robots were controlled over a 3G connection from about 3 kilometers away. Since 3G can be unreliable and is vulnerable to terrain-induced blind spots, the researchers have developed a system to use more conveniently placed robots as signal relays. 

Clover robots could be equipped with small sensor packages, like a gas sensor, but it’s also important to be able to analyze rock samples directly. Strawberry is a robotic claw that hangs from the bottom of Zion and can be used to collect rocks or soil or stuffed alien toys or whatever else a claw dangling from the sky is good for picking up. If you watch closely, in addition to a claw device, there’s also an actuated roller that helps to scoop up smaller rocks and dust:

As far as I’ve been able to tell, none of these robots have yet been tested on an active volcano, but hopefully the Japanese researchers are getting close to doing that. Because one thing we’ve learned from previous disastersis that robots can be a huge help, so the more we test and improve them under real-world conditions, the better they’ll be when we really need them. 

[ Tohoku University ]

Someone flew on the football field with a DJI Phantom, with a national flag attached to it. One of the footballers managed to catch the flag. Take a look at the situation. What do you think about it? Will it have an impact on our community?

I did not plan on testing "RTL" this flight however when I swamped my RC TX it worked flawlessly, now time switch up the video TX. A lopass filter was used however I assume it is the culprit. The wind picked up to 15-20 knots right before this flight and I called it a day. Looking forward to getting this back up in the air after the weather system moves through. To cover all the bases I will test a UHF RC Tx also.

Dr Kabuta builds smart mechanical birds.  You can buy one of his frame design and follow him on twitter 

@KazuKaku  May be It is time to start a new line of autopilot firmware : APM:Bird or FlappyPlane ? 

3DR’s Aero-M and X8-M are fully automated and intelligent tools: drone mapping platforms for easy, fast and accurate aerial data acquisition that will enable you to take informed and targeted action. They provide a level of insight that’s invaluable to industries like agriculture, construction, mining, and land and resource management, or for gathering data for any area that needs to be looked at closely and often. They allow you to get this data cheaper, easier, quicker, with less risk and more frequency and detail than manned flight or satellite imagery.

Our mapping drones come bundled with a high-resolution Canon SX260 visible-spectrum camera, but the payload system is interchangeable so you can swap your own cameras or sensors for shooting IR, hyperspectral or other imagery. The package also includes Pix4D professional image-processing software, a version developed specifically for the 3DR platform by Pix4D, the leading image post-processing company.

Our three-step workflow is fully automated: Plan, fly, process. Select the area you want to map, and the drone computes the flight path that will cover it. While in flight, onboard software automatically captures all the right photos for you and geo-tags where each one was taken; takeoff and landing are also both automated, with built-in flight protections like Return-to-Land. Finally, our custom Pix4D post-processing software stitches these photos into a crisp, high-resolution map, down to an amazing one cm/pixel, depending on the height of your flight. Zoom in from the sky down to the grape.

Additionally, you can save and repeat any flight with as much frequency as you need, and you can compare and overlay that data across time.

Imagine the advancements you might make if you could get accurate and immediately actionable information about your land as frequently as you want. You’ll see more, see better, and save money, and once you see your operations from the air you’ll never approach work the same way again. There are two options for included Pix4D software, designed specifically for 3DR mapping platforms. Capabilities include creating 2D maps in the Standard version, and the Pro version can also create index maps (e.g., NDVI) in a single workflow, perfect for crop stress differentiation. In addition, you can “fly through your map” in a software-generated 3D movie, complete with waypoints. And we’ve made all of this available at a price up to five times less than that of our nearest competitor, which will revolutionize the commercial UAV industry.

Image processing for our mapping platforms is powered by Pix4D, the leading provider for professional UAV processing software. The Pix4Dmapper software included in your package is a complete and fully automatic mapping and modeling solution that can convert and join thousands of images into highly accurate geo-referenced 2D mosaics. The software can also generate fly-through animations from original images in a single workflow. The software is suited to beginners as well as professional photogrammetrists: Create the project and get results with one click.

A Pix4Dmapper Pro upgrade is also available for 3DR mapping platform customers. With Pix4Dmapper Pro 3DR Edition users are enabled to exploit full 3D capabilities as well as agriculture specific tools. The Pro version can generate point clouds, digital surface and terrain models, create vector objects (polylines, surfaces, stockpiles) and create index maps (e.g. NDVI).

Two wings are ready for flying with APM 2.5. Below, some more pictures with different views:

Three Part 2M Flying Wing easy to assemble fins are also removable:

Top View with folding propeller and motor installed  

Bottom View Bottom pod is composite

Composite Canopy

I'm sure we would benefit from this technology, specially in the lifespan of our batteries.

"Scientists have developed a new Li Ion battery that can be recharged up to 70 per cent in only 2 minutes. The battery will also have a longer lifespan of over 20 years. Expected to be the next big thing in battery technology."

http://www.sciencedaily.com/releases/2014/10/141013090449.htm

I’m thrilled to announce our partnership with the Linux Foundation to create the new Dronecode Foundation, a non-profit organization that brings together the leading open source UAV software projects under the professional management and leadership of one of the world’s foremost open innovation organizations.

For me personally, this is the culmination of a half-lifetime of work on open innovation and in particular a seven-year journey in drones that started with my children and me experimenting with Lego Mindstorms robotics parts and RC airplanes back in early 2007 (I love that first video we shot, shaky as it was; it was a life-changing moment):

 I was amazed by what was possible with sensors, embedded microprocessors and GPS, but knew nothing about drones. So I set up a community called DIY Drones, mostly to “ask dumb questions in public”.  Two wonderful things happened as a result. First, lots of people generously answered my dumb questions and helped me learn about all the disciplines that come together in aerial robotics, from software and electronics to aerodynamics and data analysis.  Second, the existence of this community created a place for other people to ask their own questions, and in those early days of amateur UAVs, everyone was learning as we went. Even the experts in one domain had to learn about the others that had to come together to take robots into the air.

The DIY Drones site quickly grew, and is now one of the largest robotics communities in the world (nearly 60,000 active members and as much as 2 million page views per month). But it wasn’t just talk: members were collaborating on electronics, software and aircraft designs, first informally and then increasingly in organized open source projects, including the APM/ArduPilot platform (including ArduCopter, ArduPlane, ArduRover), Mission Planner, DroidPlanner and countless others, from radio/telemetry to video.  Many of these stood on the shoulders of other open source innovation projects, especially the Arduino open computing platform, which gave many of these projects the “Ardu” prefix.

At the same time, leaders in the academic robotics world were also embracing open innovation, and we soon found a kindred spirit in Lorenz Meier from ETH in Zurich, one of the best robotics research institutions, and the PX4 and MAVlink projects that he leads. Soon we joined forces, and today the Pixhawk autopilot platform represents the fruits of that collaboration, a joint effort of an open source community, a premier academic institution and a Silicon Valley company (3D Robotics, which I co-founded in 2009 with Jordi Munoz, who started the APM project on DIY Drones).

Today, I’m proud that this platform has been adopted by more than 100,000 users and is helping bring advanced UAV technology to regular people, allowing them to do extraordinary things from Hollywood-quality aerial video to crop mapping and 3D-scanning of buildings. It’s a classic example of the power of democratizing a technology; we are entering the consumer and commercial drone age and I’m delighted that an open source platform is helping lead the way.

Now that we have reached this level of adoption and maturity, it’s time to embrace the best practices of other highly successful open source projects, including professional management and governance structures, to ensure the continued growth and independence of these efforts.  There is no better organization to lead this than the Linux Foundation. Not just because of the extraordinary success of Linux itself, but also because of all the other collaborative projects that it helps run, allowing each to reach the next level of participation, performance and innovation.  The combination of independence and a clear path for corporate participation and adoption while protecting open source ideals is the aim of all these projects. The experience and reputation of the Linux Foundation ensures that those values will be embraced and preserved in our own projects as the industry around them grows.

The creation of a formal non-profit organization to organize, lead and coordinate these projects also makes it easier for companies who want to embrace open innovation in their own UAV and robotics efforts. This allows them to participate in a more formal way that helps ensure that they can contribute back to the community, in everything from code to people power to financial resources. So I’m particularly delighted to welcome our company launch partners, including such giants as Intel, Qualcomm, Box and Baidu in addition to UAV leaders such as Yuneec, Walkera and my own 3D Robotics.

Why now? Not just because drones are hot and open innovation is hotter ;-) Most of all, it’s because as our platforms have matured, they are increasingly heading in the same direction as Linux itself. The extraordinary improvements in mobile hardware technology (thanks Intel and Qualcomm!) and cloud technology (thanks Box and Baidu!), means that our community and software is increasingly running on gigahertz-class Linux computers and with constant broadband connections to the cloud.  That means that the opportunity for us to join forces with the broader Linux and Embedded Linux communities is now, and the Linux Foundation is the perfect place to lead that convergence.

Welcome to Dronecode. I can't wait to see what this community creates next.