New Developments with Paparazzi

[AntonK]

Hey guys after reading the thread on the gumstix support idea I thought I would give everyone a heads up on where Paparazzi is heading. This information wont be on the site for at least another week as we work to revise the manual with all current information. At present we are working on 2 solutions that you guys may be interested in.

One is using the Robostix hardware available at www.gumstix.org This board costs $40.00 and can run the paparazzi software via its mega128. Additional components needed would be a standard digikey ISP cable for In Systems Programming( http://rocky.digikey.com/scripts/ProductInfo.dll?Site=US&V=313&M=ATAVRISP) , GPS reciever(u-blox is what we use) IR sensors(FMA), a set of radio modems(possibly maxstream 900mhz units), a video transmitter, and standard RC components. This solution should be fully tested within a week or so. The robostix also allows for docking with a gumstix daugherboard for high level processing.

The other solution is a board created by the Paparazzi team. It is named the Classix and features 2 Arm7 controllers. This board is a great step up from the orginal paparazzi hardware. This boards testing should also be done in a week or two with the board schematics posted on the website. Soon however the boards should be available for purchase from the team if a ready built solution is something your looking for.

Much of the site info is outdated so I hope this update helps, please feel free to forward any information to myself at Novaxtremeua@yahoo.com

Anton Kochevar

[shedao]

Hey guys after reading the thread on the gumstix support idea I thought I would give everyone a heads up on where Paparazzi is heading. This information wont be on the site for at least another week as we work to revise the manual with all current information. At present we are working on 2 solutions that you guys may be interested in.

One is using the Robostix hardware available at www.gumstix.org This board costs $40.00 and can run the paparazzi software via its mega128. Additional components needed would be a standard digikey ISP cable for In Systems Programming( http://rocky.digikey.com/scripts/ProductInfo.dll?Site=US&V=313&M=ATAVRISP) , GPS reciever(u-blox is what we use) IR sensors(FMA) a video transmitter with an audio channel, and standard RC components. This solution should be fully tested within a week or so. The robostix also allows for docking with a gumstix daugherboard for high level processing.

The other solution is a board created by the Paparazzi team. It is named the Classix and features 2 Arm7 controllers. This board is a great step up from the orginal paparazzi hardware. This boards testing should also be done in a week or two with the board schematics posted on the website. Soon however the boards should be available for purchase from the team if a ready built solution is something your looking for.

Much of the site info is outdated so I hope this update helps, please feel free to forward any information to myself at Novaxtremeua@yahoo.com

Anton Kochevar


Anton, this is very good news indeed, I have the robostix at home. I look forward to using paparazzi to guide my first UAV.

[LukeZ]

Anton, I appreciate very much your posting here. A lot of us check out the Paparazzi site but since it hasn't been updated in so long I was starting to wonder if the project had kind of petered off...

This may be a bit off-topic from what you've posted, but has the Paparazzi group thought any more about developing an autopilot not based on IR sensors (i.e., using gyros/accelerometers)? The thought that you might comes up often on these fora... it sure would be great if there was an effort to achieve that and I think you'd get a lot of help from the guys around here.

Just curious...


Luke

[AntonK]

I present we are working on intergating an IMU to replace the IR sensors. The IR sensors are great for what they are, but we would like to expand the capability of the autopilot. Unfortunately there are about 3-4 people actively working on the project, and most of that time has been consumed with the new hardware versions. Understand however that the new versions were a necessary step in integrating an IMU because the old versions did not have the required processing power to handle the kalman filtering. Please feel free to stop by at #paparazzi on the dalnet IRC server, normally there is at least 1 of us in there, it depends on the day and time however.

Anton

[shedao]

I present we are working on intergating an IMU to replace the IR sensors. The IR sensors are great for what they are, but we would like to expand the capability of the autopilot. Unfortunately there are about 3-4 people actively working on the project, and most of that time has been consumed with the new hardware versions. Understand however that the new versions were a necessary step in integrating an IMU because the old versions did not have the required processing power to handle the kalman filtering. Please feel free to stop by at #paparazzi on the dalnet IRC server, normally there is at least 1 of us in there, it depends on the day and time however.

Anton


From what I see in the latest source code tree it looks like you guys are using Microstrain's 3DM IMU?

[poine]

Hi Shedao

You're correct. On one of our quadrotors we are using a 3DMG commercial IMU/AHRS. On the other one, we have a homemade unit using analog device ADXR300 gyros and ADXL 320 accelerometers - We are trying our hands at AHRS on the homemade unit and use the 3DMG as a reference.

Here you have some pictures of the beast

http://www.nongnu.org/paparazzi/airplane_gorazoptere.html

A+

Poine

[LukeZ]

Poine -

Hey, I saw you over on the IRC earlier today...

I checked out the 3DM-GX1 on their website - you're saying it's a complete AHRS in addition to being an IMU? I couldn't quite tell from what I saw on their page. Perhaps I'm not even entirely clear on the difference between the two - attitude and heading reference system would seem to do kind of the same thing as an inertial measurement unit. :confused:

At any rate, does this mean their unit already utilizes Kalman filtering? They mention "filtering" taking place on the on-board processor but I didn't see what "kind" of filtering. I guess I'm wondering how much firmware it would save to buy the 3DM over, say, an IMU unit such as those sold by SparkFun.

Luke

[poine]

Hi Luke

IMU (Inertial Measurment Unit) is the name for the sensors ( gyros, accel)
AHRS ( Attitude and Heading Reference System ) is a filter that uses the measurments of an IMU and fuses them / tracks the gyros biases using the assumption that the acceleration is vertical (gravity) on the long term - this asumption is mostly true on a hovering platform, but really false on an aircraft performing coordinated turn

The 3DMG performs a simple AHRS computation ( it seems that axis are treated independantly, as we witness single axis crash by moment ).

You can get alot of informations and even sample code in the autopilot CVS and mailing list archive ( autopilot.sf.net )

There was also a very interesting paper at the following URL
http://www.ece.byu.edu/grad1/users/beard/www_docs/papers/preprints/KingstonBeard04.pdf
but it seems to have disapeared, even from google cache - it describes a way of using GPS measurments to correct the accels and thus make the AHRS work on a fixed wing aircraft.

Regarding processing power, a "normal" ahrs consists in a 7 states kalman filter. Running in a 16Mhz mega128, we get about 20hz update, which is a bit low to control an heli. The 3DMG outputs rates and angle at about 80Hz, using a pic18F452

A+

Poine

[AntonK]

Thank you poine for jumping in here. For those who dont know poine is the original creator of the paparazzi :) you will see him in the IRC channel most of the time. We are working on some formal documentation that should be available soon, maybe as soon as a week from now.

[DavidB.]

There is a huge need for UAV helicopters. I want to start taking aerial photographs from a helicopter, but flying the heli while trying to compose a shot is not really feasible. At least not without a lot of wasted time and bad shots. Something that can keep the helicopter motionless would be a HUGE help. Is that something you're looking into? Would it be hard to adapt the IMU system you're working for use on a helicopter instead of the quadrotor? I'm really impressed with what you guys are doing. Bringing something together of this scale is a huge undertaking and gives us all access to something that used to costs many $1000's. Thanks for the hard work :)

David B.

[wadiprawita]

LukeZ,

an AHRS usually use tri axial magnetometer as absolute heading reference to zeroing the gyro bias over time (so maybe the name is attitude HEADING ref system)

-doni-

[LukeZ]

Aha, yes doni, I can see that would be the difference. The Spark Fun IMUs, and others I have seen, only have gyros/accelerometers but not magnetometers. I guess an IMU could be an AHRS without the 'H'.



Luke

[poine]

Hi David

Rotary wings are in Paparazzi roadmap but are not a priority. First thing we would like to achieve is consolidate what we have at the moment ( autonomous navigation for fixed wings - attitude control on rotary wings ) to make it accessible to most people ( ie non computer/electronics experts ) - This involves a huge work on documentation, graphical interfaces, packaging - as well as solving some design flaws and finding a way to make the hardware available ready build.

Adapting the quadrirotor stuff to a single rotor helicopter is just a mater of configuration - We have an extensive configuration system that allows you to describe just any kind of commands ( ie mixers in the rc terminology) - So be it CCPM or whatever, it only takes a couple of xml lines to describe.

Now, you should know that vibrations are the worst enemy of IMUs and that you should expect some trouble in going from a relatively vibration free electric quadrirotor to a single rotor combustion engine.

To add on what wadiprawita said, it is true that most AHRS use a magnetometer to enhance the accelerometers/vertical acceleration hypothese couple - The former couple on its own is sufficient to track pitch and roll gyros biases ( but it is more robust with a magnetometer ) but gives no info for tracking the yaw gyro bias - this requires the use of a magnetometer - There are some AHRs that use only gyros and magnetometers ( no accel ) . The 3DMG has all 3 types of sensors ( Honeywell HMC1053 for the magnetometer ) - We are playing with the PNI micromag3 wich is cheaper than the Honeywell.

Now, for keeping the helicopter motionless, an AHRS is not sufficient - AHRS will allow you to control the attitude of the heli (the way the FMA copilot would allow). To go one step further ( speed control and even further position control), you'll need a motion sensor ( it is generally not sufficent to integrate acceleration measurments to obtain speed ) - GPS come to mind first, but there are other means (image processing etc...)
You'll also need a more complicated filter ( filter = algorithm ) which is called INS ( Inertial Navigation System ) - This kind of filter requires much more processing power than the AHRS - Trammel Hudson ( autopilot.sf.net ) used to run one in a PXA255 ( 400MHz arm9 )

A+

Antoine

[AntonK]

And that is why poine is the project lead. Besides working on all of these projects poine and I are tackling the large task of writing full documentation for the entire system. Our hope is to have a users guide which will outline the specific hardware configurations to help give people a better idea of what solution might work best for them. Also a complete users manual for all systems will be written. Although its a similar system, each hardware configuration will require differences in its manual. We have a very good start on this information, and will continue to write more and more in hopes that it will help people here and elsewhere feel more comfortable with trying paparazzi. These manuals will be on the web shortly and will help supliment the lack of information thats on the website regarding the new additions. Thats all I can come up with for an update for now. I hope for more soon, maybe after I type more tomorrow.

Anton

[DavidB.]

All I can say is WOW. I have a little programming experience, and a little hardware experience. However good documentation and GUI's make life SOO much easier for everyone, and make it possible for the less experienced among us to uitilize this amazing new technology. I CAN'T WAIT to see what you come up with in the future :) :) OH, and a kit of parts would be a HUGE help as well. It doesn't get much better than this :eek:

[StrangeRanger]

Hi guys, a friend of mine pointed this thread out to me. Just thought I'd chime in and say that I'm very interested in the project and I'm also an employee of MicroStrain. I'm one of the sales / applications engineers.
The GX1 does not use "Kalman" filters per say. Our engineers have developed our own proprietary filtering system for the GX1. It really is a great little unit. All the processing is done on board and it's really not that heavy when you remove the case. To give you an idea, I've been flying a little Blade CX around the office with a GX1 strapped to it! And the Blade CX is a tiny heli.
Also, you may be interested in our new product coming hopefully by the end of this month. It's called the Inertia Link (I-Link). It's about half the size of the GX1, with faster gyros, better math and an overall better design. There will be a wireless version too. The biggest difference is that it does not have mags on it. But it will still give you pitch, roll and yaw, but the yaw is not a compass heading. And of course you can still get any kind of raw sensor data out of it at up to 350Hz. If you have questions or if I can help please let me know. Very cool project. Thanks!
j

[Hovertime]

I am very happy to see you guys here, and will be following these new developments closely!, Thanks :)

[DavidB.]

Are there any plans to offer a dumb 6 or 9 axis unit aimed at the hobby market (for less $$$)?? Right now there aren't too many options so something more affordable would be awesome.

[Tuner]

Thanks for a basic explanations.

I look forward to your documentation as you chat in this forum has been enlightening.

I need a good Reference for Flight computers such as different filters and the likes.

My problem is I know enought to be dangerous but certain basics I do not have and would like as I need an INU that can interface with a machine vision system in the long term (see analogic-computers.com ace16k chip "brains in planes")

Scott

[RajeshRaman]

The IR sensors are great for what they are, but we would like to expand the capability of the autopilot.

Could you share your experiences with the IR system? How well does this
system work If I wanted to only obtain pitch and roll angles? How accurate
is it, and does it get confused at low altitudes (e.g., during takeoff/landing)
by trees or buildings?

I'm guessing that a 3-axis magnetometer could give me pitch, roll and heading.
However, would some kind of magnetic shielding be necessary if I were to use
a magnetometer in an electric r/c aircraft? What material would work best
for shielding (weight!), and would it be sufficient to 'shroud' the motor?

Thanks!

[AntonK]

By itself the 2 axis IR system has a hard time remaining accurate. This is because its linear approximation is dependent on the weather. Paparazzi has 2 ways of working around this. The first of these is a program onboard that helps adjust the contrast based on the turn rate and bank angle. This program works really well for larger aircraft and does not require any additional hardware. The second is the use of a third axis sensor called a Z sensor. Code has been written and tested to make this setup work properly and with very good results. bank and pitch angles are very accurate.

Anton

[Vindication]

So how are the IR sensors affected by mountains and such? I mean I anticipate these would work well for folk in the (relatively) flat east, but what about the west?

Vind

[AntonK]

The Competition in Garmish was conducted in the worst mountain conditions you could think of. The idea of a mountain has been thought of and researched by a school using hte paparazzi software. Antoine can explain more about that but let me explain what I think. The plane will see a mountain similar to cross wind. With wind the plane will be blown laterally, to which the plane must bank inward to compensate. With a mountain the plane will move laterally due to the plane actually being banked when it believes to be level, so it will have to bank in the opposing direction to compensate for the lateral movement away from the desired course. I hope that makes sense, I see now it might not!

Anton

PS, to the magnetometer question I dont really know, maybe Antoine will check this and answer.

[wadiprawita]

Dear RajeshRaman,

If you prefer to use absolute sensor rather than rate sensor (for instance : gyro to get attitude) try search about wahba problem, one of the solution is QUEST, but you need 2 sets of 3D vector to solve the wahba problem using quest, you can use magnetometer and accelerometer (as tilt sensor, but this introduce another problem for high dynamics system, because the tilt sensor function will be disturbed by centripetal and dynamic acceleration cause by motion).

About thermopile sensor, it works very well, I've tried in my own autopilot system (using 2 sets of co-pilot sensor, 1 is set vertically).

regards

-doni-

[Tuner]

I had given up on this good to see some actions.

If you guys need any programming or GUI work let me know.
I have a Quadrotor and would love to have complete control over the system rather than using somones proprietary system.

Scott

Keep up the great work if their is anything I can offer let me know.

[Vindication]

So what's the status on the robostix + paparazzi?

I still haven't seen anything on the web page yet.

-Vind

[AntonK]

Vind,
Yes sorry, we have been running around working on a lot of little projects and in doing so have slowed down the process of updating the documentation. However we have gotten a good majority of the Wiki complete and Ill go ahead and post the URL to it. http://www.recherche.enac.fr/paparazzi/doc2/ Most of the documentation is PW protected at this time, but we will be adding true Wiki sections soon. The gettings started section is completed to say 90% and the introduction has a good start. Reading it will hopefully help a lot of you. We have a University from Canada who is starting to use Paparazzi and they plan on helping a lot with the documentation on how to use it. As far as robostix + paparazzi goes. This is being done in conjuction with the other 3 hardware versions, avr 2 MCU, 2LPC and 1LPC, Also the robostix needs a radio modem to transmit telemetry data. I would guess a couple weeks till we have a flying proto. Right now the manual control code is done, meaning you "could" fly a plane RC through the robostix, but the AP code is only partly integrated. About the same time we will be testing the other 2 hardware revision(Classix and Tiny) so it should all kind of fall out at once! Hope thats a good enough update for now.

Anton

[dracul]

Am i the only person who is incredibly impressed?

Nice work to all concerned. This is really going to help everyone out tremendously.

I'm personally looking forward to your prototype of the Robostix board, the project I'm interested in needs to be low cost!

Kudos

[DavidB.]

Am i the only person who is incredibly impressed?


Are you kidding? These guys blow me away. Their work will make UAV's available for anyone who's handy with a soldering iron, once the kits are available then even grandma will be flying RC. I think this has applications ranging from surveillance, AP, all the way to helping those with bad eyes get into RC. The possibilities are endless :D

[Vindication]

Anton,

Yes thanks very much for the update. It does sound like what you guys are doing is very good and exciting! Keep it up! Thanks for sharing with us.

Vind

[Tuner]

I just went to the www.gumstix.org and could not find a Robostix board wtih a mega 128 cpu???

I love their new boards but I am unsure as to compatibility with them
Am I missing something?
Scott

:) Never Mind I see the robostix add on board.

[AntonK]

I should link the wiki to the actual robostix instead of the gumstix main page, thanks tuner Ill do that.

Anton

[treehog]

I shall monitor this closely and when I can contribute any usefull inputs will do so
This project seems to the most doable without costing mad money that exists so far that I know off
They are lucky to be in France to do thier stuff as there is plenty of spce for thier activity around Toulose(great town lived there for a few months pre RC days )
me stuck in north europe for this year cold and wet and much less airspace to do things and hard to get decent camabert cheese:(
Better get my 3 metre scale Discus slope glider I bought in France 7 years ago and redo its mattrix to recieve a new onboard pilot and mayby a pylon electric motor mount

Ralf

[AntonK]

I guess you all deserve an update! Work continues as usual, although it feels as thought we are a bit behind. We are doing some work this week to clean up the CVS, it has been broken for over a month with all the code changes. This way we hope to have at least a working version available on the CVS while the new revision is completed. Secondly we have moved from Dalnet over to FreeNodes, where you can find us still at #paparazzi I will try to keep in touch, if you read this and its been more than a week yell at me to make an update!

Anton

[Hovertime]

Thanks for update, we ARE reading this for sure! :)

[AntonK]

Wow, all the way at the bottom of the second page. I guess its time for an update. We have a good member returning who is helping to get the CVS tree back together from antoine and I's hacking. New additions have included color selection of the airplane :) and use of radio modems. We have done numerous flights with radio modems onboard with great success. Now we are working to update the GUI to help interface the uplink and make it a little easier to do things :) I have gotten a couple inquiries about hardware. The team has considered doing a hardware run, probably once hte new boards are tested, would there be any interest here on the boards for hardware? Hmm thats about it for now, back to work!

Anton

[Tuner]

I might be.
I have been following the paparazzi project for 2 years now from the side lines and I love IT!

My issue has been since I have an unusual design most hardware does not support 8-16 servos.
I also need to be able to interface this with machine vision system Im sure that is already obtainable.

I will be honest I am not quite sure what is correct hardware for me by looking at you site it is hard to tell the specification of each hardware unit and witch is the best let alone if it will work for me I can be a little dumb with the obvious.

I basiclly need the best possible hardware with the lowest latencies between servos.

Hear is my project maybe you can understand my needs a little better.
http://www.rcgroups.com/forums/showthread.php?t=436841

My Budget is $5-7,000 for GPS + INU + Flight Computer and Radio Modem. Camera system is an other ballpark for me since I plan on using Cellular Neural Network Based Imaging system ACE16K.
I have considered commercial Autopilot units but they seem like they are not flexible enough for my application.

Any Help From anyone Would ROCK!

Thanks

Soctt

[clolson]

I might be.
Hear is my project maybe you can understand my needs a little better.
http://www.rcgroups.com/forums/showthread.php?t=436841

My Budget is $5-7,000 for GPS + INU + Flight Computer and Radio Modem. Camera system is an other ballpark for me since I plan on using Cellular Neural Network Based Imaging system ACE16K.
I have considered commercial Autopilot units but they seem like they are not flexible enough for my application.

Any Help From anyone Would ROCK!


I don't know about driving 8-16 servos independently (maybe you could get some serial based controllers and rig up a couple of them?)

If you need flexibility, you might take a peak at the Xbow micronav:

http://www.xbow.com/Products/productsdetails.aspx?sid=133

This is a really slick little unit, costs about $2000 with flight computer and you get GPS, INS, flight computer, and servo control (although not 16 servos) all in one tight little package.

The computer runs Linux and all the sensor integration, autopilot, and communication code is open-source so you can get in there and make it do just about anything you would want. The potential down side (if you aren't a software person) is that you probably will need to dig in and do coding to make it do what you want.

But it is a very capable unit with a huge amount of flexibility and programability. We have one here at the University of MN, but the problem is it keeps getting moved to someone else's desk so I don't get to play with it as much as I'd like.

Curt.

[Tuner]

I love to code and tinker so this is down my alley just to low tech.

This is a nice setup for a average plane.
It usese 802.11 and Im not sure what the range is for that?
It has low rate sensor and they wont work for me.
This is basicly what I am looking for.

This project is really what close to what I want Hmm I will look more into them.

I am looking into Atairs INU and the GX-1 INU.
It looks like the only way I will be happy is to put it all together myself seperate INU GPS CPU Radio Modem & Servo Controller. Man this will add a lot of weight the nice think about an integrated solution is that it is light.

[Tuner]

Im sticking with Paparazzi unless Atair can wow me and I am impressed with thier specs they beat most anyone else on the market that I can compare them to. They have mentioned having algorythms that help deal with the INU being offset fromthe CG like in a paraplane I think this will be handy for my Quadra-Plane.

The more I read the more I am finding that the paparazzi project is what I need.
It is just a matter of getting the proper hardware at the proper weight.

I am leaning that its all about the hardware.

So In that spirit I noticed mention of building an INU based hardware system. I have $5-7,000 ready for everything GSP INU Flight Computer Servo Control.

I would be gladly donate some money to the project if somone can give me some of their time helping me configure the best possible hardware setup for me???

Just thought I would through that out thier.
Thanks
Scott

[clolson]

I love to code and tinker so this is down my alley just to low tech.

This is a nice setup for a average plane.
It usese 802.11 and Im not sure what the range is for that?
It has low rate sensor and they wont work for me.


Well, don't knock the device just because it has a pcmcia slot and a compact flash slot and you can plug in a wireless card and it supports it. :-)

It also has a serial port so you can connect it up to any kind of radio modem or cell phone modem or whatever you want that speaks serial. It has a usb port too and will accept a simple web camera. Run apache and a web cam on the airplane and I'm not sure what you would have that's useful, but it would definitely earn you some geek points somewhere. :-)

By default the 'processed' data rates aren't spectactular. 25hz for orientation data and 10hz for position data. However, the raw sensor output can be configured for 100hz and you have the source, so it is possible to bump up these processed rates.

The unav/stargate forms a nice tight package, but the stargate's xscale cpu doesn't have any floating point, so that limits how fast you can crunch the kahlman filtering loop. However, by my estimates, the current code with the 25/10 hz data rates is only burning about 25% of the cpu time. So I think you could at least double the rates (maybe 50hz for attitude and 20hz for location) and still have plenty of cpu left over for navigation and ground station communication.

Also, the unav (the sensor piece) simply interfaces to the stargate through a serial port. So you could plug it in to just about any other board if you want more crunching power to get your data rates up to 100hz.

Some guys here are working on interfacing it to an MPC555 which supposedly has hardware floating point. Oh and I know a guy who compiled the code on his linux PC and happily ran with that instead of the stargate ... so you might be able to consider some sort of PC104 stack.

In my opinion, it's a very capable device, and it's potential is still largely untapped. I hope to start tapping in the next month or two though.

I've got a dynamics model of our UAV built in FlightGear (www.flightgear.org.) I also have a nice 3d visual model of our UAV with animated control surfaces. I've begun to setup and tune a variety of autopilot modes in the simulation. The next thing I'm working on is to port the FlightGear autopilot code to the stargate and add a network communication layer to talk to FlightGear in 'real-time'. So FlightGear can still act as the sensor inputs, but those get sent the stargate, all the autopilot code is crunched there, and then the control commands are passed back to FlightGear. FlightGear runs the flight dynamics and uses whatever control inputs it gets from the stargate. So I should be able to excercise my code quite well on the desktop before I go flying. Then the final step would be to rig it back up to read the real sensors and drive the real servos, mount it all in the airplane and go fly. And cross my fingers and hope it doesn't shred itself in the first 100 milleseconds of attempted autonomous flight! If it actually flies in any sort of semi-stable manner, I'm on my way to fame and fortune. :-)

If things become really quiet on my end you'll have some clue what happened. :-)

Curt.

[Tuner]

I broke down and purchased a RobotStix, a Gumstix 400mhz connex with bluetooth and I got a CF card adapter.

I am going to either buy the microstrain AHRS and a GPS or the Atair INU.

Right now the Atair is the one I want not sure on their GPS??

I like to tinker. I make rigs to keep my model safe and still test though at some point you have to take a plunge.

Good luck!

Wish me luck on my first Flight computer using this project :)

[LukeZ]

You all keep us updated on your progress. It seems to me the issue with both the attitude packages you mentioned Tuner, and with most all other products available right now, is you still have to write a mountain of code and figure out this Kalman Filtering stuff to get your state estimation data to do something useful. I should correct myself and say that's not really an "issue," as that isn't the point of these INUs, but it is a real obstacle for the small-scale UAVist to overcome. I'm still looking forward to the day when we have an open-source project that is actively pursuing such a project - at that point, you could plug in either the Microstrain or the Atair unit and it should work (they both output data in compatible formats).

I'm keeping tabs on the Paparazzi guys - they seem like perhaps the first that will get there. We'll see.

Anyhow, I know plenty of people will be interested to hear about your experiences with the Robostix, so keep us posted Tuner.

I'll wish you luck, and you too Curt!

Luke

[Tuner]

Im hoping my experience in 3D animation I can levearge my understanding of taking complex movement captured from devices just like these only try 24-36 at a time and then making the data look smooth.

I have been reading about these kalman filters but know very little details on them other than what they do. I think you are underestimating these devices though they do a considerable amount of preprocessing of the information so you dont have to that is why I chose them otherwise I would do what Paparazzi did and make my own AHRS using MEMS devides.

I have a few tricks up my sleave I hope will contribute to the community.

I would love it if anyone who has done this before could chime in at some point I am dying to pick somones brains on this project. I have like 20 questions that could all be answered in 10-15 min conversation via Online chat or the phone.

I will say when I am successfull with this I will create a Kit forsale because I see no high quality turnkey kits out their. Heck even Cloud Clap though well developed is low quality.

-=PEACE=-
Scott

I can help out with GUI or Coding or Documentation if you guys need any help????

[AntonK]

Update time. The team has been busy doing specific tasks, and unfortunately the new hardware and software has been put somewhat on hold. However slowly we are working on finishing it all. Consider its a lot of work to make the software compatible with 4-5 different hardware platforms at the same time. Not to mention we are making all of the control laws configurable. During this time we have worked more on the datalink code, as well as the groundstation interface. The groundstation now links up with google maps to make a map that is perfectly calibrated. Antoine has been working on code dealing with the IMU. He plans to try to make his quad rotor craft fly autonomously in 2 months time. At this point he has it rate augmented with great success. Next step is going to be getting it to navigate. Pascal has been busy working on updating the groundstation and the code for the new hardware. Now you may ask what I have been doing(Taking midterms). Actually Ive been working on navigation algorithms because they interest me :) and also so attitude control code, thats been evolving for hmmm 7 months :) Alright theres the update. Come visit us sometime at #paparazzi on freenodes IRC server.

Anton

[Tuner]

ANTON
Thanks I will get IRC up and running again and look for you all.

--EDIT--

IRC is a great resource for paparazzi as Im finding.

Scott