New open-source autopilot

[zik]

I've been designing an autopilot for a little while and after reading some of the discussion here I thought that other people might like to use it too. I've decided to open-source and open-hardware the whole thing so anyone can get involved.

Here's the feature list as it currently stands:

* IMU with 3 axis accelerometer, 3 axis gyros and 3 axis magnetometer
* Barometric altitude sensor
* GPS - 4 updates per second
* thermopile leveling sensor (similar to the copilot)
* 8 servo outputs
* 55 MIPS ARM7 processor
* can piggyback an "easyradio" or a Maxstream XTend-OEM for telemetry
* R/C receiver input with separate manual override circuit
* airspeed sensor
* can piggyback a gumstix if more processor power is needed
* SD card slot for plug-in mission profiles
* USB for PC downloads
* battery voltage monitor
* piezo speaker for audio readiness indications
* optional infrared LED to control shutter release on a camera
* optional tiny camera for navigation and low-res shots via telemetry (not a video camera - stills only)
* optional temperature sensor
* optional ultrasonic sensor for landing
* optional battery current sensor
* expansion connectors so you can add your own extras

Is there anything I've missed? Where I need your help is to tell me if there's some brilliant feature I haven't put in there or if some of these features are a waste of space. I'd like to finalise the prototype design in the next week before I send the prototype board off for fabrication.

If people are sufficiently interested by the time it's debugged and ready to go I may get a few fabricated for people to buy. Or you can always just make it yourself of course.

[O.L. Adcock]

Zik, It sounds like a dandy and you list more features then most I've seen. I like that you've got 2 attitude references, 3 if you count the camera an operator can monitor. Let us know how the testing goes....O.L.

[zlite]

Wow! Zik, that's very impressive. How much of it is now working?

[zik]

It's based on some IMU + processor tinkerings which I've been working on. This fancy new version is currently just a schematic - I've added a lot of extra sensors for this version compared to my earlier ones.

I'm expecting to go through a couple of prototypes before I have a final version so it'll be a couple of months at least.

Incidentally my design has two parts - the main board with most of the parts and a sensor board with the GPS and the thermopile sensors. The idea is that the main board can be hidden away in the interior of the plane and the sensor board can sit on top where it has a clear view of the sky. If anyone can see any problems with this setup for their application I'd love to hear from you.

I'm also interested in what platforms people are using. The main board will probably turn out about 4cm x 10cm or maybe a little larger. Is this size a problem for anyone?

[samanpriya]

Hi Zik,
It’s very interesting project, and I am waiting for more details.

My I idea is if you can create it based on SMD size, it will be smaller and can be use in the small electric platforms

Rgds.


:D

[zik]

Hi Samanpriya,

I'm running fairly small electric platforms myself so I'm definitely planning to use it that way. I have an optional current sensor which when used with the built-in voltage sensor can be used to calculate the power output of electrics.

Basically I'm designing something which I can just drop into my planes. I don't want lots of bits and pieces drifting around, all linked together with bits of wire. I'm trying to come up with a system which you can just plug in and go, no matter what plane you have.

[dracul]

The hardware sounds fantastic but I'm wondering about the software.

What language/toolchain will you write the code in? I presume some form of C/C++ ?

Have you already developed the code to integrate the sensor data into the navigation algorithm ?

I've been researching autopilots for about 2 years now planning on building something much like your project but have always been stumped by one thing or another. I can't program in C which is a major disadvantage and I don't wish to learn assembly!

This is very exciting for me personally and I wish you the best of luck with your prototypes.

Kudos

[zik]

Thanks dracul. I plan to start with the paparazzi code and work from there. That way there should be working software from day 1 and improvements from there.

I'm hoping that work from this project can benefit the paparazzi project and vice versa. (Incidentally I really should think of a better name for this project than just "this project"...)

The code will be written in C/C++. For embedded programming C is hard to beat. It's worth learning if you're getting into UAVs.

[fly_asmara]

Hi Zik,

Great effort this one. I'd shake your hand if I was in australia. I'm not much of an expert in the electronics to make a good autopilot but I've used of-the-shelf autopilots a lot. Some are ok and some are well... terrible. I guess I will help to draw up a list of things you 'want' your autopilot to do if you like, and since you are the hardware and software expert, it probably will help the design process to some extend. One thing that comes to mind now is the GPS option you would like the autopilot to have. This is no doubt great to have. The faster the rate, the better. But I would add another navigational capability in the form of an electronic compass, at least as a backup - for 'returning home' capability perhaps incase of gps malfunction. Will go thru my old notes and see what else I can contribute and let you know soon. In the mean time, all the best in your project and look forward to see some results soon. All the best!

--
fly_asmara

[toxicmouse]

are you going to use a normal RC receiver into the MCU, then outputs to servos, or are you going to fly using the maxstream only? if the former, it might be worth looking into a 74XX157. it complicates the hardware but simplifies the software and also speeds it up- for my setup anyway. using the 74XX157 allows an easy way to overide the MCU with RC in the event of a problem with code.

the idea is that servo inputs can easily be changed from the RC to the MCU, and give precedence to the RC if available. i think that this is the safest approach.

http://www.futurlec.com/74HC/74HC157.shtml

good luck on your project

[zik]

Thanks fly_amsara - the current design includes a 3-axis magnetometer similar to the PNI MicroMag3. This is basically a digital compass except that it operates in 3 dimensions rather than just two. It can be used in the same way as a normal compass. It can be used in conjunction with the IMU and GPS or separately as you suggest. It's just a matter of how the system is programmed to make use of it.

Nice idea on the 74xx157 toxicmouse. The design has a microcontroller-based manual override switch already though your hardware solution sounds very neat. You can attach a radio control receiver to the unit on one side and servos on the other. One channel of the receiver is configured as the override channel and if that's triggered the servos are switched from autopilot control to manual R/C receiver control. This works even if the main processor has failed.

The design has two processors - one main ARM processor (an AT91SAM7X256) and a peripheral microcontroller (a PIC16F887). The peripheral microcontroller really just talks to servos and does some other bits and pieces. Basically it does nothing complicated. The idea is that its program is simple and stable and rarely changes. All the fancy (and potentially flakey) programming is in the ARM. If the ARM processor malfunctions the peripheral controller still keeps on doing its thing and can be switched to manual mode to bring your baby home. The receiver and the servos connect directly to the peripheral controller - not the ARM - so you have total control of your plane even if the ARM is dead. Flick the manual override switch on your R/C transmitter and the autopilot's peripheral microcontroller becomes a straight-through link between your receiver and the servos.

[Jack Crossfire]

Having already been through this, the dual processor configuration is a waste of time. The problem is communicating between the 2 chips and the extra weight. A single processor in a PLCC package would be ideal. Individuals can't afford to run off a $50-$100 PC board with all the components every time they change something.

Tried using an STR91 as a single chip solution. It had 8 A/D converters, 96Mhz internal clock, 512kB flash. Never could get it to run a program although I could verify the flash was being programmed. It didn't seem to accept mounting on an electro-board.

Moved on to a Gumstix + PIC18f because the idea was to use a Kalman filter. With the Kalman filter off the table again, the processing requirements are much smaller, and a single 8 bit microprocessor is more attractive. The problem with PIC is the lack of a free C compiler.

[zik]

Hi Jack,

The whole manual override feature is a safety vs. complexity tradeoff. Personally I chose the side of safety. The PIC only adds $3.48 to the cost and also provides some much needed I/O. On compilers: I'm using the free C compiler "SDCC" (which I'm also a contributor to).

[zik]

Hey, a question for you guys. How many channels do you need from the R/C receiver? My current plan has up to 7 control channels available and up to 10 servo outputs. The servo outputs can also be used as sensor inputs for things like motor temperature etc.. I can trade them off receiver inputs against servo outputs - more servos/sensors means less channels of receiver control. I can't quite decide what's best:

* 5 channel R/C control / 12 servos+sensors
* 6 channel R/C control / 11 servos+sensors
* 7 channel R/C control / 10 servos+sensors
* 8 channel R/C control / 9 servos+sensors

I figure most people don't fly using more than 5 or so channels but some people might want to trigger cameras, put gear up etc. with the extra channels.

How many channels can you see yourself using on a UAV? Your thoughts would be very much appreciated. And you get to have input on the design of this autopilot!

[_helitron_]

Hi zik,

for my UAV/FPV-Cularis glider I need for the plane alone 7 channels: throttle, rudder, elevator, 2 ailerons, 2 flaps (for butterfly mixing, mixing flaps to ailerons/ ailerons to flaps, etc.) ! Then additional I'd like to have 2 for pan/tilt camera, one for autopilot override (switch/slider) and so on. Too much I know :) but when I asked to write a Xmas wishlist :D ...

Ah, yes, why do you use the PIC instead of an AVR, Paparazzi is using an AVR (ATmega8) as the slave controller. The core of the AVR is way better usable for a high level language.

Cheers, Erwin

[AntonK]

Correction... Paparazzi used to use the Mega8 as the slave controller. We have all but abandoned the AVR architecture. Now we run a ARM 2148 on our Tiny AP, and 2 2148 on the Classix AP. On the Classix the second ARM does serve as this slave function, but it was designed to do some higher level stuff as well. My 2 cents btw, if you are trying to invent something yourself, all the power to you. If you are looking for a solution, Paparazzi would get you pretty far into your endevour faster than you could design hardware and core software yourself.

AntonK

[_helitron_]

Correction... Paparazzi used to use the Mega8 as the slave controller. We have all but abandoned the AVR architecture. Now we run a ARM 2148 on our Tiny AP, and 2 2148 on the Classix AP. On the Classix the second ARM does serve as this slave function, but it was designed to do some higher level stuff as well. My 2 cents btw, if you are trying to invent something yourself, all the power to you. If you are looking for a solution, Paparazzi would get you pretty far into your endevour faster than you could design hardware and core software yourself.

AntonK

Hi Anton,

was only a hint for zik regarding his planned 8-Bit slave controller. I know the Paparazzi hardware very well and knew the fact with the double LPC-2148 :) . Had the intention to build a Tiny13 board two month ago but was not able to get a hardware kit from David Armstrong (had exchanged several mails with him) :( . Gave then up.

Cheers, Erwin

[sergey123]

Paparazzi is a great project with only one minus for now – it is very difficult to get hardware kits (correct me if I’m wrong. I reaaly want to be wrong in this case :) ).
For people who are not hardware engineers the first step (getting hardware) basically kills all enthusiasm what is why we see so many attempts to build something like paparazzi or compatible with paparazzi software.

zik have a look at paparazzi pcb and software it will help you speedup your project and will give ability to reuse what they have built and … yes 2x2148 from Classix AP is much better PIC and arm7. Actually I would prefer to have 1 arm7 for communications (servos and etc) and second Arm9/11 for navigation/autopilot software …..

sergey

[_helitron_]

Agree fully with you Sergey.

Erwin

[clolson]

This discussion has me moderately curious. How much does a Paparazzi hardware kit cost? What would be an expected lead time? Any "realistic" estimate for how long it would take a person with medium hardware skills to assemble the kit and debug it and get it all running? The direction I'm going with this is that even though it's "open-source" hardware, it would be interesting to come up with a "realistic" cost and time frame for getting a unit in house and up and running ... for comparison purposes relative to other options.

Thanks,

Curt.

[sergey123]

As far as I know no any kits yet available for paparazzi…
Probably AntonK can give better answer about cost and time frame to build.

Cost: something like 300-500USD for all hardware (with sensors, modems and etc) and assembly time is a week or two…. Once again - I could be wrong ..... I'm not in paparazzi group - just looking for the best options for now.

zik: what do think about your kit cost and how flexible it will be in terms of adding sensors step by step, so starting point will not be very expensive

[Unterhausen]

I was looking at the CVS for the paparazzi hardware, and I'm not sure all the different platforms are in there. If you are satisfied with the boards that are there, you could have Sparkfun or Phoenix or your board house of choice build the boards, and build it yourself.

[_helitron_]

Boards are not the problem can you buy here

http://www.rctechnix.com/index.php?manufacturers_id=10&sort=2a&language=en&osCsid=7ff0b01606770454beef735bb14ac288

but all the other stuff you have to obtain for yourself.

Erwin

[AntonK]

Assembled boards are not readily available as of yet. PCBs are around a plenty though. The Tiny has the GPS receiver built in, so its a bit more pricey if you talk about just the AP. You kind of have to break down the hardware to realize what each AP does. This is explained on our wiki. The Tiny takes a bit of patience and at the least a hot air soldering station for the LPC. I believe David has PCBs, and is working towards assembled units, I havent chatted with him recently as Ive been working on my own little project. I guess as proof that Paparazzi is a great start, I took one of the older AVR boards(called Paparazzi Lite) which has 1 mega128 and 1 mega8, and built a Quadrotor heli with it. 4 motors very similar to the draganfly. The neat thing with the program is I plugged in 3 gyros and some control code, and bam I have a rate stabilized quad. Next I want to through some IR sensors on, and bam an outdoor attitude stabilized heli. Hopefully someday I can get some filtering code to work, so I can fly indoors with attitude control using some accels. Anyways my point is, if you are going to have to build hardware anyways, take a look at our system, it may allow you to have a nice base, instead of starting from scratch. If you need an IMU for your project, then create one, or better yet use the one for another quadrotor project thats on the CVS(1 ARM2148 + 3 accels + 3 gyros) which runs a kalman and feeds the AP using SPI. The Possibilities are endless. Ok now off my soapbox.

AntonK

Btw, I didnt mean to steal your thread, just thought I would suggest an option.

[zik]

Thanks _helitron_, I guess I'll go with the 8 channel approach then. If you need as many as 8 channels from the receiver I'm sure other people will too.

PICs and AVRs do pretty much the same job but the PIC was a little cheaper. Either can be programmed in C so that's no problem either way.

[zik]

Hi to all the paparazzi guys! I think the paparazzi project is great and I'm hoping to use the paparazzi software as the basis for my Skybot board's software. I'm really hoping that we can work together and boost each other's projects.

My Skybot hardware is different from the paparazzi (classix) hardware in a few ways:

* It's all on one board. I wanted something I could just plug in and go. With paparazzi you need a few boards to get a useful system. I wanted less boards dangling around the place. paparazzi's approach gives you lots of flexibility but the idea with Skybot is to have everything built in.

* You can plug in missions in the field using SD cards

* Radios just plug right in, no special wiring is required

* It has a 3-axis magnetometer to help keep track of the plane's attitude and correct gyro drift (although apparently the paparazzi guys are currently working on an add-on IMU board which will also support this)

* Optional ultrasonics for autonomous landings

* Battery voltage and current monitoring, audio status indications, etc. Lots of handy stuff.

* The control code is all on one processor which makes programming easier

* Both Skybot and paparazzi use an ~60MHz ARM7TDMI core, with optional gumstix for more power. So they're pretty similar processor-wise.

Basically Skybot's more "plug in and go", has more features and is more expensive. The plan is to make fully built-up units available. Feature-wise it's more like a MicroPilot than paparazzi. But I should point out that it's still going to be at least a couple of months before I have beta hardware so if you're looking at an autopilot don't hold off from getting something else in the meantime.

Paparazzi is a great system. I actually started my Skybot project before I'd heard of paparazzi but since I found it I've learned a lot from their experiences. I'd probably be using their "classix" for my main project except that I'm working on a complex UAV project which does some unusual stuff which "classix" isn't really set up for.

Making Skybot available to other people is really just a spinoff of my own UAV project. But hey, if Skybot can be useful to other people why not make it available?

[zik]

Sergey123, I haven't really thought about whether I'll offer multiple options with different features installed. If you want to start cheap and add features the paparazzi approach is better - you can add more boards as your budget permits. Skybot is really intended to be a "one board does everything" approach.

[sergey123]

>>But hey, if it can be useful to other people why not make it available?

I really like your comments :)), yes, please keep going :) and tell us how we can help !!!

sergey

[sergey123]

>Sergey123, I haven't really thought about whether I'll offer multiple options with >different features installed. If you want to start cheap and add features the paparazzi >approach is better - you can add more boards as your budget permits. Skybot is really >intended to be a "one board does everything" approach.

I was talking about more powerful CPU and may be some additional sensors. I like "one board solution too" but if everything for it will be too expensive not a lot of people will be ready to invest. Kits approach in addition to fully assembled boards can give some flexibility and will attract much more people

[AntonK]

Making your own boards available is awesome, and I truly believe that if more projects develop thats awesome. I do want to clarify some of the bullets you made, just so people dont get any sort of misunderstanding

"
* It's all on one board. I wanted something I could just plug in and go. With paparazzi you need a few boards to get a useful system. I wanted less boards dangling around the place. paparazzi's approach gives you lots of flexibility but the idea with Skybot is to have everything built in.
"

Be careful what you mean buy built in. To have every component attached, may or may not be a good idea. EMI from components can cause lots of trouble. In the case of the Tiny, the GPS, Processor, Servo driver, and powersupply are all on 1 board, this has proven to be difficult to work with as the stuff messes with the GPS a lot. The classix was specifically designed to try stuff with, and thus the reason its just a processing board, it really isnt an autopilot, more of just a processing board. To fly paparazzy with a Tiny you really only need 3 things. An AP, a modem, and a sensor board. which really isnt a lot.


* You can plug in missions in the field using SD cards

This is a really neat idea. However Im not sure how useful it is. Dynamic waypoint addition would be better I think, although paparazzi doesnt do it. Instead we create generic flight plans which have references based on the boot position, from there you can do just about anything within the flight plan, including circle, figure 8s, line follow, survey, hmmm I know there is more!

* Radios just plug right in, no special wiring is required

The only "special wiring" Ive ever needed was... a cable to run from the radio to the AP. Here Im assuming you are talking about the datalink, and not the RC receiver.

* It has a 3-axis magnetometer to help keep track of the plane's attitude and correct gyro drift (although apparently the paparazzi guys are currently working on an add-on IMU board which will also support this)

Indeed, the IMU being worked on has a 3 axis mag. Of course this has only been tested on a hovering platform. The filtering required to fly MAVs or Small UAVs is pretty complex, dealing with some serious dynamics. Most users fly pretty small aircraft, largest I know of is the Twinstar II

* Optional ultrasonics for autonomous landings

Already built in to paparazzi. Depending on the sensor you choose your interfacing options include SPI, I2C, PWM, or easy ADC. Im planning on using the one from sparkfun on the quadrotor.

* Battery voltage and current monitoring, audio status indications, etc. Lots of handy stuff.

Battery voltage is monitored in flight, current sensor we have, but havent played with(not sure what it would help with, other than to calculate power remaining) audio status indications. The groundstation is set up to yell at you if you change flight modes, if the battery gets to low, if the flight termination is triggered, and a host of other things(Im pretty sure this is user defined, but Ive never played with it)

* The control code is all on one processor which makes programming easier
All control code in ever architecture has been on 1 of the processors, by control code. But


* Both Skybot and paparazzi use an ~60MHz ARM7TDMI core, with optional gumstix for more power. So they're pretty similar processor-wise.

Indeed!

Basically Skybot's more "plug in and go", has more features and is more expensive. The plan is to make fully built-up units available. Feature-wise it's more like a MicroPilot than paparazzi. But I should point out that it's still going to be at least a couple of months before I have beta hardware so if you're looking at an autopilot don't hold off from getting something else in the meantime.

I wouldnt suggest you compare your product to the Micropilot. You will lose interest from a lot of people in a hurry. If your AP is like the Micropilot, better for people to send there money to a needy family(sorry personal feelings inserted here)

Paparazzi is a great system. I actually started my Skybot project before I'd heard of paparazzi but since I found it I've learned a lot from their experiences. I'd probably be using their "classix" for my main project except that I'm working on a complex UAV project which does some unusual stuff which "classix" isn't really set up for.

Making Skybot available to other people is really just a spinoff of my own UAV project. But hey, if Skybot can be useful to other people why not make it available?

More power to you seriously. I just wanted to clarify, in case people wondered. Doing somethign yourself will be extremely rewarding, and damn you will learn a lot.


AntonK

[zik]

Hi Sergey,

Yes, if you want to add more processor power the Skybot will have a connector to attach a gumstix (as does the paparazzi classix). Like the paparazzi it'll have I2C, SPI, PWM and ADC expansion available if you want to add more sensors. So you can definitely move up from the basic setup.

The most expensive bits in the board are the gyros and the magnetometer. Together they make up nearly half of the total price. Now I'm wondering if I can make them an optional extra so people can get started more cheaply without them and add them later if they like.

AntonK - you're right about the GPS. I actually plan to put the GPS module on a separate board with its antenna to avoid EMI issues. This GPS board will mount on top of the plane where it has a clear view of the sky.

On the wiring for radios etc - I'm trying to minimise that kind of cabling. I'm aiming for a setup where you can piggyback gumstix->skybot->radio by plugging them directly into each other, no cables required. I want a beginner without soldering skills to be able to buy one, connect it up and go.

If the paparazzi group is interested in sharing or joining forces with me in some way let me know! The more co-operation the better.

[zik]

On incidentally AntonK, what's so bad about the MicroPilot? I've never used one so I'm curious.

[AntonK]

Ive heard nothing but horror stories from others, but my own personal experience. We received 1 while I was at university to use. Simple equipment, RC receiver 3 servos and the micropilot. We couldnt get the GPS to lock. The documentation said something like wait 30 minutes, mayeb it was 15. Anyways its definitely not made for small UAVs for sure, maybe a ground robot or something. The second instance was at my last job Miraterre Flight Systems, who use Paparazzi. An existing project we took over used the MP2028, and our board died when during a demo flight the MP decided it needed airspeed, so it dove from 500 feet into the ground at 100MPH. I was kinda glad it did. We had it wrapped in 4 layers of foil to block its EMI from radiating on everything. We also removed the onboard GPS and installed a UBLOX receiver which helped. The one thing that we couldnt get over was the poor initiation of the sensors, during take off(hand launch) the plane would lose attitude orientation(filter would go crazy) and it would think it was flying every way but straight. This coupled with the horrible EMI on the RC led to the MP takign control thinking it had bad RC and rolling the plane over(thinking it was upside down already) and then getting RC back after it rolled the plane over inverted for us... Really we shouldnt have been flying the way it was, but at that point to keep schedule we really didnt have a choice. So ya I have a bad taste. Jeff from the paparazzi group had a similar experience and has started a club whos slogan is "friends dont let friends fly micropilot"

AntonK

[AdamKt1]

I have some experience in Programing microcontroller especially PIC microcontroller.
I would like to learn from the big guys.
Adam

[octane-link]

I would be willing to give you a hand in flight testing your system, depending on the size.

I have two aircraft that it could fly on, but one has yet to maiden, I'd like to get it some hours before I go autonomous in it.

[sergey123]

Hey zik,

How you project is going??? Do you see light in the end of tunnel :) :)
If you need any help with design/software/testing just tell us…

Sergey

[zik]

Thanks Sergey - I've finished the design of the prototype. A downside of prototyping is that it can be much physically larger than the final product - and in fact this one is large enough that I don't really have a suitably large airframe to put it in. All my current planes have pretty small fuselages. So I've been a bit stumped on whether to spend the time making a bigger plane or whether to try to cram the electronics down in size right now. Unfortunately this will take a lot of extra time which will essentially be wasted since I'm expecting to redesign before the final version anyway. So you can see the quandry I've been in for the last couple of weeks.

In the end I decided to go the middle road - I'm now making a simpler prototype which isn't so large. This will have a reduced set of features but will still be a useful autopilot in itself while still being a guide for the larger autopilot design. I'll probably offer this version as a cheaper/simpler entry-level design. Depending on what people want they can choose this small, cheap, light version or the larger fully-featured version (though the small one is my focus for now).

Here's the feature list for this small version:

* small and light. The final version will be around the same size as a large receiver.
* 6DOF IMU
* thermopile leveling
* GPS
* altitude sensor
* up to five servo channels
* connects to a standard R/C receiver
* USB for course setting and programming
* optional radio telemetry (using easyRadio)
* expansion connector for optional airspeed sensor etc.

I've been pretty busy with designing this new version - in fact both prototypes will be ready to send off to the PCB shop for fabrication in about a week.

[AntonK]

Just a suggestion, 5 servos may not be enough. If you have an airframe with a rudder, and 2 aileron servos you end up with 5 right there, and that doesnt include if you want a gimballing camera, or a nose wheel, or flaps, or whatever. Just a thought

[Unterhausen]

I wanted to put an autopilot on my Cub, but it requires 6 channels right off the bat: elevator, ailerons 2, flaps, rudder, throttle. I guess I could go with a splitter for the ailerons

[AntonK]

6-7 seems to be the right number. Currently the paparazzi Tiny has 7 I believe, numbered 0-6. if you are using a 4017 its not much more to add the extra channels I dont think, a little board space for the pins I guess.

[zik]

Ok, by popular demand I've scrounged a couple more servo outputs. Now the total is eight servo outputs and six receiver inputs. This means that the last couple of servos could be used for pre-programmed shutter releases or other programmed behavior of some sort since there's no direct way to control them from your transmitter.

If you have extra transmitter channels you want to use there's nothing stopping you from connecting servos directly to them of course and just remotely controlling them like you normally would.

If you need more servo outputs then I'd suggest my other design "SkyBot" might be the one for you - it has fourteen servo outputs! It's intended for big, serious UAVs where this simpler one "Flying Fox" is intended for small hobby UAVs.

[sergey123]

It is great progress zik!!!
I am sure lots of people will sign for your autopilot. If even “simple version” will have full 6dof it will be much better than anything else available (except very expensive commercial units).
I can not imagine what the “full version” in this case will have – auto landing and takeoff or may be obstacle avoidance :) :) ???

Question: are you planning of adding SD card support. It is pretty simple to add but it could be very useful for waypoints and saving flight logs. Yes, it cold be done by using modem link but it is just nice to have feature and not difficult to add.

[zik]

Hi Sergey,

Yes, the more complex version "SkyBot" has provision for auto-landing and SD card support. There's no SD card in the simpler "Flying Fox" version - it just takes up too much room. I considered microSD but it doesn't seem very practical for changing in the field. However it should be easy to change your mission at the field using the radio link from a laptop (or USB if you prefer).

[sergey123]

Thanks zik, both versions sounds cool !!!

[stevievb]

To anyone who may be interested,
I have been working on a fully functional auto pilot for a little under a year now and am frankly just not getting anywhere. I started with small 8 bit atmel boards (Olimex dev boards - sparkfun is a five minute bike ride away) a little before last Christmas and am now working with olimex lpc boards. I've played around with just about everything, meaning I've got all the components of the autopilot to sorta half work, but between getting my private pilots license at embry riddle this summer, and attending school full time, I just haven't been able to put it all together. I would also attribute that to the fact that I'm just frankly not a programming person. I've taken almost three years of C++ in school, and I've been messing around with all the code which is in C, but it just gets beyond my skills every time I start to dig in deep. I have an idea of everything that needs to be done and how it should be done and how it will and should work (down to a very specific level - im not just saying i know i need some software that does this and this and this ), but translating that to code with organized structure and meaning is just not my skill in life. I also should note I am witting and compiling all my code in Linux with gcc so that took a while to get the hang of but I think thats for the most part under my belt. I have all the parts and the will power so I guess what I'm looking for is a few people that want to put in the time with me and publish everything up here so everyone can have access (I'm using all off the shelf parts from sparkfun except the pressure sensors and radio modems). I'm shooting for :
- imu stabilized
- full two communication through airtronics radio modems (uploads live data to your computer throughout the flight, and is controlled fully from a joystick)
-airspeed/altitude - from pressure
-gps

Thanks,
Steven

[Aeroflot]

Hi Zik, What Electronic CAD are you using?

[zik]

Aeroflot, I'm using Cadsoft's Eagle. It's cheap and very capable - and probably the most popular one among hobbyists at the moment.

[Aeroflot]

I had a chance to read through some of your threads, it maybe worthwhile
if you have'nt already checkout the WPS-2 thread. MX might have a few helping hints in regards to parts that he is using in his new project.
helpful too.