Posted by Jarett Jones on January 25, 2014 at 9:39pm
I have a 2-axis brushless gimble mounted underneath a flying wing, with retractible landing gear. The goal is to install pixhawk as the autopilot.First the gimble. I have equipment that is large and heave I want to mount on a second gimble that is servo powered to handle the weight and then position it where i won't get large ammounts of drag (as apposed to with the camera gimble which is predator style under the nose and would require much larger housing).The allows the camera to keep its large feild of veiw without being restricted to the other equipments limited veiw location. The problem is I need the other equipment to look at a specific offset of where the camera is looking when it's writhin the range that the second gimble can travel. To add even more to the difficulty the offset will need to be continuously calculated based on certain conditions like airspeed. I want pixhawk to control the servo gimbles stabilization.I realize I will probably have to add code in order to do this and I have done a decent ammount of programming. So I want to know what it can already do and what will need to be added. Also possibly get assistance with programming in the additions.What I wantCamera can look around freely using inputs from the ground station. (It's an alexmos gimble controller) which i don't know if it can give feed back to pixhawk or if pixehawk will have to make assumptions.I can switch modes so both gimbles move when I look around with the camera. Both gimbles are locked to a specific max mins and the camera can possibly start at a predefined location so if it can't give feedback to pixhawk from the alexmos then we can assume where it's looking based off that.Lastly the second gimble has a continuously calculated offset based on math I have already figured out.Takeoff and landingCan the auto take off be set to use yaw to maintain heading instead of roll. My plane uses landing gear and is not hand or catapult launchable. I think I already read somewhere that with airspeed sensor it can be set to maintain 0 pitch untill above a certain speed.Auto retract landing gear and leading edge slats?Can landing be setup to approch similar to a real plane. Roughly 3 degree glide scope using pitch to change decent rate and trying to mantain a specific speed (somewhere close to stall but safely above it) with the lowest throttle setting possible; thus giving a higher pitch and higher angle of attack for the same glide scope which leads to a softer landing as apposed to noseing down to decend makeing a high speed landing?Auto deploy landing gear and leading edge slats?Can or dose it flare before touchdown? Can it side slip for touch down in wind instead of crabbing though that's only marginally important the wear on the landing gear is far less then a real plane.If it does flare what parameters will I need for it to flare correctly.Auto deploy spoilers on touchdown?Ultrasonic sensor will probably be neccesary for a lot of this so I was planning on getting one.Can it auto taxi?This took a while to type its late I'm tired and probably explained something wrong or stated incorrect information somehwere so ask for clarification if you are the slightest bit unsure.
You need to be a member of diydrones to add comments!
Luckily I have no expectations and that's why I'm asking questions in order to know what to expect. If I have to code in the extra features I would be happy to try, logic and math based codeing is right up my ally :).
I have read the wiki but it's explination is lacking compared to some of the other better documented features.
So I'm trying to get a better picture of what variables are accessible through the GUI and what would need to be changed through code. Such as can I set the max pitch for the flare or will it just attempt to glide in nose up based off minimum speed.
I've seen the autopilot crab into the wind effectively enough to mantain a fairly straight line to a waypoint it shouldn't be hard to add a decrab then flare with side slip based on altitude above touch down point. I just need a better picture of how much work is already done and how much I need to do to get it to the standard I'm looking for.
Luckly I find it less demanding on hardware and software to plan a mission that climb out and landing glide scope already a void trees rather then having the ability to see and avoid them. Though a landing glide scope looks like something I will need to add.
This autopilot isn't at that level, even the $5000 ones aren't. While this can do a huge amount it's not that good (yet:). Read the description in the wiki of the autoland, it basically just flies to a 3D point and then flares holding course, nothing more. Takeoff is similar, basically hold set pitch and course until set alt then loiter, not rocket science but therein lies the problem, it won't detect trees, head wind, cross wind, bad programming, etc, etc. It may compensate for a crosswind but only reactively.
Also it's DIY, written by very talented but non-paid guys and therefore quite a few features are not polished, like the gimbal function, it works but requires some lateral thinking and lots of fiddling to get to work and even then it may not work as you would expect or the way you would have coded it.
I'm afraid your expectations are very high and you'll be disappointed with this autopilot unless they come down many notches. (Or you join the team and make it better!)
How should the autopilot doing takeoff and landing be any different then me doing it? The airframe and components are redundant and reliable.
I fail to see what my slow human reaction times could handle that an autopilot can't. Anything that isn't a catastrophic failure is most likely going to appear during the preflight ground check.
Can you give me a scenario you would worry about and explain why the chances of a crash are increased?
Replies
I have read the wiki but it's explination is lacking compared to some of the other better documented features.
So I'm trying to get a better picture of what variables are accessible through the GUI and what would need to be changed through code. Such as can I set the max pitch for the flare or will it just attempt to glide in nose up based off minimum speed.
I've seen the autopilot crab into the wind effectively enough to mantain a fairly straight line to a waypoint it shouldn't be hard to add a decrab then flare with side slip based on altitude above touch down point. I just need a better picture of how much work is already done and how much I need to do to get it to the standard I'm looking for.
Luckly I find it less demanding on hardware and software to plan a mission that climb out and landing glide scope already a void trees rather then having the ability to see and avoid them. Though a landing glide scope looks like something I will need to add.
This autopilot isn't at that level, even the $5000 ones aren't. While this can do a huge amount it's not that good (yet:). Read the description in the wiki of the autoland, it basically just flies to a 3D point and then flares holding course, nothing more. Takeoff is similar, basically hold set pitch and course until set alt then loiter, not rocket science but therein lies the problem, it won't detect trees, head wind, cross wind, bad programming, etc, etc. It may compensate for a crosswind but only reactively.
Also it's DIY, written by very talented but non-paid guys and therefore quite a few features are not polished, like the gimbal function, it works but requires some lateral thinking and lots of fiddling to get to work and even then it may not work as you would expect or the way you would have coded it.
I'm afraid your expectations are very high and you'll be disappointed with this autopilot unless they come down many notches. (Or you join the team and make it better!)
How should the autopilot doing takeoff and landing be any different then me doing it? The airframe and components are redundant and reliable.
I fail to see what my slow human reaction times could handle that an autopilot can't. Anything that isn't a catastrophic failure is most likely going to appear during the preflight ground check.
Can you give me a scenario you would worry about and explain why the chances of a crash are increased?
Much easier to hand fly the aircraft and know that the controls and handling is fine before switching to auto.