Building an UAS. Inputs on autopilot and mission computer.
I am planning on building an UAS.
The idea is this. Have a autopilot with the required sensors connected to a mission computer. The mission computer will control other sensors and ground communication.
The first choice is the autopilot. I was thinking of using the Paparazzi for this, but I am not sure if there is cheaper/better alternatives around.
Second there is the mission computer. It should be running Linux, preferably from a memory card. USB/serial connections are needed to connect sensors and communication modems. Not sure if the ARM or x86 architecture is best suited for this.
This is a private research project testing whether a process network written in python (pyCSP) can be used to control the different sensors and the communication modems(GSM and WLAN).
What alternatives is there on mission computers? Have seen the Gumstix. I guess the easiest would be a full featured x86 processor board.
The airframe at the moment is a Funjet.
Any inputs will be appreciated.
Oerjan Pettersen
The idea is this. Have a autopilot with the required sensors connected to a mission computer. The mission computer will control other sensors and ground communication.
The first choice is the autopilot. I was thinking of using the Paparazzi for this, but I am not sure if there is cheaper/better alternatives around.
Second there is the mission computer. It should be running Linux, preferably from a memory card. USB/serial connections are needed to connect sensors and communication modems. Not sure if the ARM or x86 architecture is best suited for this.
This is a private research project testing whether a process network written in python (pyCSP) can be used to control the different sensors and the communication modems(GSM and WLAN).
What alternatives is there on mission computers? Have seen the Gumstix. I guess the easiest would be a full featured x86 processor board.
The airframe at the moment is a Funjet.
Any inputs will be appreciated.
Oerjan Pettersen
Replies to This Discussion
-
Permalink Reply by Curt Olson on -
The gumstix probably costs a bit more than the ATMega328, but it does run linux and you have access to a ton of open-source software and libraries, as well as device drivers for just about anything that linux supports. The gumstix is small and light. It's what I've been using and I feel I've been getting good results. I'm running a 15 state kalman filter on mine as well as all the round-earth navigation code and autopilot.
-
Permalink Reply by Ørjan Pettersem on -
Have you tested running Python on your Gumstix?
-
Permalink Reply by Curt Olson on
-
Sorry, I haven't done any python scripting on the gumstix. I would be shocked if there were any problems though, pretty much everything runs on the gumstix. I have been doing all my autopilot development in C/C++. At some point (time permitting) I plan to experiment with integrating the nasal scripting language into my autopilot. (Nasal is a fast, efficient, tight scripting language developed by the most talented software guy I've ever met and is used heavily by the FlightGear project.)
I do realize nasal is not python so I'm not suggesting you do anything other than python. I can't imagine python wouldn't work just fine for you ... the only possible concerns I can imagine might be memory footprint and performance.
Curt.
-
Permalink Reply by Ørjan Pettersem on
-
Thanks. I have ordered a Gumstix Overo Fire. Can't wait until I get it.
It will be used together with a Lisa/L board running Paparazzi.
-
Permalink Reply by Ørjan Pettersem on
-
Not to a great extent, no. Have read the home page for each AP in the menu, but that is about it. Do you think the ArduPilot would have been a better choice?
-
Permalink Reply by Curt Olson on
-
I'm pretty sure you won't be running python on an ardupilot. I suppose your thought process and mission requirements will direct where you go. If you want to start very simple and cheap and just get a basic autonomous plane flying, then ardupilot isn't a bad way to go. You probably can't get to where you want to go with an ardupilot, but you will gain a ton of valuable experience that could be a stepping stone to the next level. Or you could start by building a more complex system that does satisfy all your initial requirements, but depending on your experience level, this could be a lengthier process with many more pitfalls and risks.
The funjet is a pretty small airplane, so you have to be really careful about how much you plan to put in there ... boards, sensors, wires, connectors, batteries, radio modems ... it all adds up *really* quickly, and even if you get some stuff prototyped on your desk that looks nice and tidy, it always seems to double in size and weight and random wires by the time you get it installed in the airplane.
If you've never built a uav before, I would suggest starting small and cheap and simple, even if the system you end up with doesn't come close to meeting your initial objectives. You will gain a ton of experience and be in a much better position to then do what you really wanted to do. It takes more time, but UAVs are difficult and time consuming, no matter how you approach them.
If you really really really need to meet your initially stated requirements, another option would be to build a simulated UAV using a flight simulator. You could run your autopilot code on separate embedded hardware, but you wouldn't have to worry about many practical details ... like flying and crashing and destroying all your expensive gear. And in the lab debugging is a lot easier when you can directly look at things or directly look inside your code. Everything is much more difficult when something odd happens in flight and you are the sole pilot/engineer.
Contests
Season Two of the Trust Time Trial (T3) Contest has now begun. The third round was a reliablilty/aerial photography round for both planes and copters, which is now closed. Stay tuned for the next round, beginning soon.A list of all T3 contests is here.
Top Content
Groups
-
South African DIY Drones…
18 members
-
DIYDrones - CNC Project
58 members
-
UK Ardupilot Group
26 members
-
German Ardu-Group
41 members
-
San Francisco Bay Area F…
1 member
© 2012 Created by Chris Anderson.
