Re: [Flightgear-devel] Re: Real Flight PLayback ?
Sorry for joining this thread lately... As suggested by serveral people if only GPS data are available (ie: Lat/Long, Ground Speed, Ground Track and rough altitdue.) For pitch restitution : With two dated altitude data, we can compute vertical speed. From vertical speed and ground speed, we can compute the Flight Path Angle. For playback purpose we can assume the AOA is 0° and then have the pitch angle equal to the FPA (or estimate it as a linear function of ground speed - ie assuming the wind is null... - ) For roll restitution : With two dated track data, we can compute the track time derivate. From the track derivate and the ground speed, we can compute the bank angle (assuming symetric flight) For heading restitution : It could be approximated to the ground track. (ie: assuming a null wind) A couple years ago, I ran some simulations and it gave very nice results even vith 1Hz GPS output data (and software smoothed to 25Hz for display confort). Acceleros and gyros MEMs are good short term, and for quick attitude changes. The here above Pitch, Roll and Heading attitude estimation stabilizes long term the cheap MEMs IMU. An other way to stabilize cheap MEMs gyros and acceleros for attitude determination is to use 3D magnetic sensors : it's fully independant of the vehicule acceleration *but* dependant of the magnetic field environement... Olivier PS : Mat, if you still need it, I have some C code which reads GPS data from a RS232 - MNEA format, with a very low CPU usage; which let plenty of CPU to run FGFS in parallel if you see where I'm heading ;-) From: Alex Perry [EMAIL PROTECTED] While not strictly true, the assumption is close enough for playback. You can infer AOA from airspeed and add this to the flight path angle. Similarly, you can use airspeed and radius to infer roll angle. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
Warning: this is all written with not enough sleep and not enough caffiene, read and respond at your own risk. :-) Mat Churchill wrote: I can see several business ideas related to this along the lines of something that could be fun to develop but which also has a variety of commercial possibilities. Does anyone have any examples of how commercial collaborations with open source projects actually work in practical terms ? I see it kind of like a marriage, every relationship is a bit different and you have to work hard to find a happy balance and keep it there. There are certain things people have found to be helpful and work well, but in the end it's up to the happy couple to find each other, figure out how to do life together, and continue to work hard to stay ahead of issues that come up, or resurface, etc. I think it's a similar challenge to figure out how to marry an open source project with some sort of commercial venture. But I think if you look around, there are a ton of examples ... from companies effectively using open source software in their day to day lives (Linux, X11, apache, emacs, gcc, python, etc.) to companies collaborating with open source projects for some mutual benefit to companies who derive their life blood from open source. For instance if you were developing a hardware product specifically for use with Flightgear, how would you develop the software side without competitors being able to find out exactly what you are up to? Yes, unfortunately there are a lot of sleazy bastards out there who are all to happy to rip off your work as soon as they see you making a sale or two. We can try to pretend to live in a happy world where these people don't exist, but it can be hard for a business to survive this sort of thing. Unfortunately, hardware and software vendors aren't just paranoid black helicopter types ... I'm not much of an expert in this area, but you can use things like closed source drivers to hide your hardware specifics. But, that won't stop people from disassembling your compiled objects, downloading your microcode, etc. etc. The best thing you can do to combat this is never develop an interesting or useful product, because if you do, you will have someone trying to copy it before long. As an example, I'd love to see open source drivers for nvidia cards, but I certainly can't fault them for their approach. I'm no big fan of Microsoft, but I can understand why they are like they are and why they do what they do. When you are at the top of the heap you have to fight off challengers from all areas ... open source is only one of *many* challenges. But in the end, let the best company or best approach win. :-) Hopefully we can trust gov't and law enforcement to keep the playing field from getting too unlevel ... it will never be perfectly level ... and we could never all agree on exactly what is level anyway ... I guess you'd set up a software company that privately develops the software. Are there scenarios where the community helps write software and gets cash back from related profits. Could you set up a Flightgear Foundation that developed commercially useable add ons away from the gaze of potential competitors and got cash back from profits towards things that the community wanted to do ? Or can open source software be developed privately and then shared at a later date ? To be clear on this I do understand the point of open source and am a firm believer in it. I hope this is not seen as a dangerous question. You are a danger. :-) To be honest, mixing profit motives with a hobby is a dangerous thing. There are many obvious (and many very subtle) ways we could shoot ourselves in the foot. Anything we would do would need to be approached very cautiously and with *much* thought. And we need to be realistic, money has to come from somewhere. That implies that we would need to do or promise something of enough value that people would be willing to part with their money and willingly give it to us. If we have a commercial product then we need to compete with products like MSFS and X-Plane and try to match them fluff-4-fluff. Research money is an option, but that is an entirely different ball game and has it's own large set of head aches. I think the question to ask (if we are interested in finding some sort of money source) is what kind of value can we offer to individuals, companies, research institutions? Courting any of these can require up front cash, long term risk, lots of sales work, etc. I tend to think that FlightGear in and of itself is too loose of an organization to effectively go after money sources like this. We will have better luck if individuals (who are part of some company, or part of some research group) work to get FlightGear used as part of their projects and do whatever they can to contribute the results back to the project. These companies or research institutions might do
Re: [Flightgear-devel] Re: Real Flight PLayback ?
On Sun, 11 Jul 2004 22:22:23 +0100, Mat wrote in message [EMAIL PROTECTED]: I can see several business ideas related to this along the lines of something that could be fun to develop but which also has a variety of commercial possibilities. Does anyone have any examples of how commercial collaborations with open source projects actually work in practical terms ? ..ask Nazgul aka Big Blue aka IBM the SCOX whacker. ;-) For instance if you were developing a hardware product specifically for use with Flightgear, how would you develop the software side without competitors being able to find out exactly what you are up to ? ..they would probably wanna keep things secret, the GPL kicks in _on_ distribution. In-house development etc is fair use. I guess you'd set up a software company that privately develops the software. ..optional. We _are_ flightgear.org. Are there scenarios where the community helps write software and gets cash back from related profits. ..plenty. ;-) Could you set up a Flightgear Foundation that developed commercially useable add ons away from the gaze of potential competitors and got cash back from profits towards things that the community wanted to do? ..there is _no_ conflict between commercial and the GPL. You are thinking of _closed_ source software, which _is_ in conflict with the GPL, and usually sold for much more than its own value. Also google for sharewarez etc to find plenty junk for free, check out how their licenses closes their source. ..Red Hat, Novell makes tons of money _selling_ GPL software. ;-) Or can open source software be developed privately and then shared at a later date ? ..yes, assuming no distribution until release of the source. Clever people will wanna sneak their hardware etc out to retailers, then on D-day, Kaboom-sell everything and publish their source. To be clear on this I do understand the point of open source and am a firm believer in it. I hope this is not seen as a dangerous question. .. ;-) -- ..med vennlig hilsen = with Kind Regards from Arnt... ;-) ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
Mat Churchill wrote: Problem is I imagine the weight would be thrown around by centrifugal force messing up the readings. Correct, and not only by centrifugal forces. _Any_ sort of acceleration will make your data unsuable. You're looking for what is called IMU, Inertial Measurement Unit. A cheap variant is described here: http://www.cloudcaptech.com/crista_imu.htm but I don't know, how good they work. There _are_ people successfully driving autonomous helicopters with a solid-state IMU over short distances but I would not trust them for large/expensive model aircraft or 'long range' flights. For recording your own flight they should be sufficient if you add a GPS, Martin. -- Unix _IS_ user friendly - it's just selective about who its friends are ! -- ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
Mat Churchill wrote: I'd thought of drilling a hole in the ball of a serial mouse and attaching a weight on a stick to it. Mount the mouse in a cage with a spirit level on it, fix the cage in the aircraft and get a zeroed reading and then record mouse positions against time during flight. An upside down joystick with a weight on the end of the handle could do a similar job. Problem is I imagine the weight would be thrown around by centrifugal force messing up the readings. Here's a thought experiment for you -- if you strap a spirit level across your bicycle handlebars, where will the bubble be when you're turning, assuming that you don't lose your balance and tip? If you did the same thing with the mouse ball, where would the ball be during the turn? What about the ball in the turn coordinator or turn bank indicator in an airplane during a coordinated turn? In coordinated flight, or on a bicycle, the gravity vector and the centrifugal force vector should always combine to keep your mouse ball, the inclinomter ball, or the bubble in a spirit level centred -- they can tell you nothing useful about your bank angle. If that were not the case on a bicycle, the side force would tip you over. In fact, that's the problem with instrument flight. In an airplane, there is no way for your body to feel whether you're level or banking -- you can feel when you're changing your bank, past a certain threshold, but even that doesn't work if the change is gradual enough. When you're in cloud, you can be in a 30-degree bank but your body (which works a bit like the spirit level) tells you that the wings are level, and the ball in the TC or TB is perfectly centred. Non-instrument-rated pilots (or even IR pilots caught by surprise or out of currency) are often unable to believe their gyros in situations like that, with sad consequences. Attitude indicators have to operate unaffected by centrifugal force, perhaps there is one that can output data ? The attitude indicator, turn coordinator, and heading indicators all use gyroscopes, which are rigid in space and not affected by centrifugal force -- that's why they can show whether you're banking or not. There are AI's and TC's with outputs for autopilots, but you're not likely to find them in a trainer, and if you did, the owner is extremely unlikely to let you wire them into a portable computer (even if there's some legal way to do it). If you *really* want the attitude information, your best bet is to buy one of the new, portable backup gyros like this: http://www.icarusinstruments.com/microEFIS.html They're not cheap, but they'd be an order of magnitude cheaper than trying to set something up to interface with the panel gyros. All the best, David ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
On Fri, 9 Jul 2004 11:32:05 + (UTC), Martin wrote in message [EMAIL PROTECTED]: Mat Churchill wrote: Problem is I imagine the weight would be thrown around by centrifugal force messing up the readings. Correct, and not only by centrifugal forces. _Any_ sort of acceleration will make your data unsuable. You're looking for what is called IMU, Inertial Measurement Unit. A cheap variant is described here: http://www.cloudcaptech.com/crista_imu.htm but I don't know, how good they work. There _are_ people successfully driving autonomous helicopters with a solid-state IMU over short distances but I would not trust them for large/expensive model aircraft or 'long range' flights. ..2 factors: how does these handle planet rotation, and, how accurate are these IMU's. On Fri, 09 Jul 2004 08:11:44 -0400, David wrote in message [EMAIL PROTECTED]: Mat Churchill wrote: Attitude indicators have to operate unaffected by centrifugal force, perhaps there is one that can output data ? The attitude indicator, turn coordinator, and heading indicators all use gyroscopes, which are rigid in space and not affected by centrifugal force -- that's why they can show whether you're banking or not. There are AI's and TC's with outputs for autopilots, but you're not likely to find them in a trainer, and if you did, the owner is extremely unlikely to let you wire them into a portable computer (even if there's some legal way to do it). If you *really* want the attitude information, your best bet is to buy one of the new, portable backup gyros like this: http://www.icarusinstruments.com/microEFIS.html They're not cheap, but they'd be an order of magnitude cheaper than trying to set something up to interface with the panel gyros. ..this invites FG spinoff ideas; Instead of wailing for donations, we _could_ come up with something useful and profitable? ;-) ..a few gps antennas, an iMU, air pressure sensors, FG in a laptop, and you can record stuff like wing bending responses to gusts. -- ..med vennlig hilsen = with Kind Regards from Arnt... ;-) ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
Arnt Karlsen wrote: ..2 factors: how does these handle planet rotation, and, how accurate are these IMU's. Typical IMU's don't handle planet rotation at all because they never know _where_ on the planet they are located (you need to handle it different depending on which half of out earth you are sitting). Solid-state IMU's are not very accurate. As long as you have external references, like a camera picture, you are able to keep bank and attitude and navigate over short distances with a cheap solid-state IMU. For a flight over a long distance you have to readjust them like you do with the old, well known gyroscopes because the signal noise is too much to be used for accurate differentiation, Martin. -- Unix _IS_ user friendly - it's just selective about who its friends are ! -- ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
Re: [Flightgear-devel] Re: Real Flight PLayback ?
On Fri, 9 Jul 2004 14:35:17 + (UTC), Martin wrote in message [EMAIL PROTECTED]: Arnt Karlsen wrote: ..2 factors: how does these handle planet rotation, and, how accurate are these IMU's. Typical IMU's don't handle planet rotation at all because they never know _where_ on the planet they are located (you need to handle it different depending on which half of out earth you are sitting). ..makes sense, math 'n software job. Solid-state IMU's are not very accurate. As long as you have external references, like a camera picture, you are able to keep bank and attitude and navigate over short distances with a cheap solid-state IMU. For a flight over a long distance you have to readjust them like you do with the old, well known gyroscopes because the signal noise is too much to be used for accurate differentiation, ..true, and IMU's makes up only part of the navigation input. ..Mat's mouse drilled-ball-stick makes sense here, it needs a wee springloaded weight sliding on the stick, and say a sliding potmeter to report the weight's position along the stick. ..the parked plane sees 1 G, dropping vertical. Good calibration may even help it tell the latitude, by comparing it with known mass forces and the observed planet spin force offsets. ;-) ..an hacked up rodent also provides AOA, yaw etc vane circuitry, simply stick the mouse roller spoke wheel onto the vane axle and put the diodes around the wheel, and plug 'n play. -- ..med vennlig hilsen = with Kind Regards from Arnt... ;-) ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. ___ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
