Blogs

I posted this earlier but did some tweaking on it to compare the Extended Kalman to my Fixed Gain Observer filters. I haven’t finished the FGO position filter but am really starting to lean more on EKF because the performance really is significantly better.

The idea for this filter is to fill in the gaps of the GPS position between the 1-4 Hz range for something like a quadrotor or just overall improved lag free navigation. You can clearly see how the fusion of more sensors can really increase your UAV's perception of reality!

**Edit, I improperly emulated GPS in the first pic so I added the correct pic. This shows how GPS comes in at 2Hz with noise and using a little logic and deadreckoning you can not only fill in the gaps but be more accurate than GPS is actually reporting. The GPS acts kinda like the accelerometers in the DCM algorithm, they keep the solution constrained.

Enjoy!

-Beall

Some testing

This is my first attempt at building an ArduCopter. I started with franticly searching the web for parts. I had no idea what I was looking for, but I was looking. Between DIY Drones and AeroQuad, I was able to come up with a list. I have a little background with the arduino. But nothing like this. Mostly would work with LEDs, LCD screens, and Temp sensors.

Ok on with the build. Here are some pictures of what I have so far.

I have a thousand questions, that I don’t even know I need to ask yet.

Thanks to all on this site....

I've finally perfected the X-Plane HIL for Ardupilot. This will be a feature on 2.6.3 which will be in beta today.

Flying Waypoints:

Here is missed waypoint detection: (I nudged the plane at the last second with the radio to miss the WPs.)

Here is fully auto take off and landing:

Here is the video of auto takeoff and landing:

Xplane and Ardupilot from Jason Short on Vimeo.

If you want to try this out you have to use some software to glue this all together.

Serproxy - used to talk serial to Ardupilot

Perl - to link Serproxy and X-Plane using "X-Plane.pl"

X-Plane 9 (the free version works fine, but it time limited)

I've only tested this on a Mac, so if anyone can try this on a PC, please let me know of changes to make, if any.

In the config header you will see a new ground station called GCS_Xplane (3). You must also use GPS_protocol 3 for ArduIMU.

Here is the Perl file:X-Plane.pl

Here is Windows Serproxy - Windows

Here is Mac Serproxy - Mac

in the config set:

newlines_to_nils=true

comm_baud=38400

Then set the com port accordingly

Go ahead and download 2.6.3 from the SVN. I'll post a beta soon as a download.

Moving forward I'd like to replace the Perl/Serproxy combination with a single C or Python based solution. If anyone can help, that would make it so much easier to set up!

Bonus:

Here you are some photographs of my flying robot.

It's a chinese replica of the t rex 450 equipped with brushless motor and lipo battery.

As you can see, there's a CMUcam mounted in the front. It points the helicopter in the direction (yaw control) of a coloured object, and follows it when this ascends or descends, lowering and raising height.

The new arduPilot mega with IMU shield is mounted on the tail boom, near the body. Currently it works just as a stabilization unit with little code modifications. In the future, when I'll have some spare time, I'll adapt it as an autopilot with gps guide and indoor navigation techniques such as the Polly algorithm (CMUcam software) which recognizes wall edges.

The goal of this project is to demonstrate (or not!) that cheap electronics and models can lead to a "successful" and reliable robot, to use as a platform for sensors.

in particular, the major goals to achieve are:

-long flight time

-payload weight

-almost absolutely fail-safe hardware and software, reliability

-autonomous navigation

upgrades can be a controller such as the RoBoard, the roboRealm vision software and a camera

I'll post more as soon (long) as i have time to work

cheeeeeers

"Environment:

• New file extension to replace the .pde borrowed from Processing (issue #13).
• Redesigned toolbar icons (issue #291).
• Ability to upload sketches using an in-system programmer (ISP) from the IDE (issue #260).
• Simplifying the process of selecting your board and serial port (issues #223 and #257).
• Command-line compilation and uploading of sketches (issue #124).

Language (most of these are possibilities and still open for discussion):

• Creating events that can be called automatically, e.g. the serialEvent() as in Processing (issue #263).
• Adding specific functions for enabling / disabling the internal pullup resistors (issue #246).
• Modifying the behavior of print() on bytes (issue #284).
• Functions for accessing more of the low-level functionality of the hardware timers and other peripherals (issues #169 and #248).
• Optimizing the digitalWrite() function (issue #140)."

I just read a great post by Bill Sweetman on Aviation Week's Ares Blog about a publication of The Mitchell Institute over the history of UAVs in the US armed forces.

The author is Dr. Thomas P. Ehrhard and although I haven't finished reading it yet it certainly is very interesting.

You can download the full document (88 page pdf) here: Air Force UAVs: The Secret History

(in the photo one of the concepts covered on the document, the D-21B)

What's so special about 3D printers? They make it possible to create complex structures, such as wings that are warped to improve performance, like the manually curved wings of a paper aeroplane, says Richter. Their printer is capable of producing features just 16 micrometres wide, and thin films just 40 micrometres thick.

The other advantage of printing is speed, says Lipson. Once they have arrived at a new wing design, printing a set takes under an hour. With the exception of its motor and battery, their latest four-winged creation is almost entirely printed from polyester films stretched over carbon fibre rods, and weighs in at just 3.89 grams – a six-fold weight reduction on their previous version – and capable of hovering untethered for up to 85 seconds.

I haven't heard much talk about Jasons sim which is strange because the more I use it the more I realise how good it is.

It is certainly the best tool to use when building your own gcs. Auto mode works brilliantly and "fly by wire" is a bit of fun. The sim automatically returns to base and then loiters once all the way points have been reached. Switching back to "auto" mode starts the sim again.

Below you can see the turns which the sim produced. The telemetry included the correct lat, lng, pitch, roll, climb rate, and speed

Another great run by Brian Wolfe with his custom PicPilot autopilot, and he wins Round 7 of the T3 competition--the autonomous takeoff and landing round! Second place goes to Mark Griffin (Paparazzi), Third to William Premerlani (UAVDevBoard), followed by Riccardo Kuebler (UAVDevBoard) and Krzysztof Bosak (EasyUAV).

This was a tough round--not all autopilots can do autonomous takeoff and landing--so hats off to the winners. They're all real UAV pros, as shown by their consistent high performance in multiple T3 Rounds.

I'll announce the next round this weekend, and it should be a bit easier (and fun!) to encourage lots of entries. Plus it's summer--what else do you have to do??

Full list of winners and standing after 7 rounds below. Huge thanks to our judge, Gary Mortimer, who has handled the difficult task of reviewing and grading the entries (requiring the smarts of Einstein and wisdom of Solomon) with charm, good humor and constant encouragement. We couldn't do it without you, Gary!