Our Custom Quad Mobile Landing Platform with modified F450/APM!


Greetings DIYDrones! 

We started building this project a few months ago for our Senior Design Project at the University of Central Florida's Electrical Engineering Program. We used tons of information and products based off this community and figured we would share our project. We haven't found really much like it so we are hoping that it may help shed some light into a different area.

There are a lot of sub-systems and sections to this project and I will be more than happy to expand on any of it if anyone wants to know. But the basic idea is we have a modified F450 running an APM 2.6 with extra features that is able to land on a mobile rover. This rover is able to transport the quad and then the quad is able to fly off and it's own thing. It will then return to the rover and land on it.

The quad features a readymadeRC 200mW transmitter for the FPV, minimOSD, XBEE PRO 900 MHZ telemetry, GoProHERO3 on a pan and tilt servo setup,  Turnigy 9X with frsky tx/rx, and a 5000mAH nanotech. We also modifed the frame with 3D printed parts for all the housings and magnetic feet for the locking. For those of you who are immediately questioning the weight distributions, we actually weighed the printed parts versus using zipties and adhesives, they come within about 2 grams of eachother (the printed parts are heavier of course). We offset the frame the best we could to evenly distribute the weight. 

Now for some the more interesting parts. When the quad lands, it will actually be re-charged. The charging system utilizes what is called Bedini Motor and we are using Solar Panels for supplementary power. One thing we also realized was that the quad was going to slide around and fall off while we are driving around. So we also implemented a magnetic locking system. Simply put, we printed feet for the quad that attach to the legs that house very strong magnets. The rover then has a magnetic grid on it attached to an actuator that raises and lowers to release/lock the quad. So far we have been able to get almost a full 45 degrees sideways before it starts slipping. We also implemented Mecanum wheels into 

it so we can straffe and a few other custom modes. I can post videos of it once I figure out how to lol. 

I don't want to make the post too wordy so I will post some of the pictures. Please keep in mind that we are still developing it and I don't want to post the more final pics until it is more put together (the actual shell for the rover should be done this weekend, the plexi top right now was just for practicing landing). This project was built almost from the complete ground up from the PCB's, the software, the GUI's, the frames, and we also 3D printed a lot of our parts (oh yeah, we have  ROBO3D and a Reprap in our garages) 

 

Views: 1805

Comment by Mike T on March 20, 2014 at 12:34am

great stuff. how long does it take to charge a 5000mah battery?...


MR60
Comment by Hugues on March 20, 2014 at 12:55am

Hi, very nice project.

What I would like to see is a landing guidance system using  a combination of radio or IR beacons for mid range and then a visual recognition system for the last centimeters, so that the quad lands on the platform automatically. And if you can do this so it works on the rover platform when it moves, you would have the hell of a killer app.

I guess you land the copter manually on a the immobile rover platform?

Comment by Brandon Frazer on March 20, 2014 at 9:57pm

Mike, the charging time is kind of hard to say right now as we are still testing a lot of things. The Bedini Motor needs to run through an Acetone Bath to get all the smoothing correct and we are still testing the decrease in charging time using Solar as a supplemental power supply. We are also using nano-tech batteries to try and speed up the process. The actual numbers would be something I would have to ask one of my partners about, but I know it takes a bout a full day to charge our APC batteries that are 8AH batteries (motorcycle batteries) from a completely dead state. Again, these numbers are changing everyday. Based off some of our tests and calculations, we are fairly confident it will be able to charge a 5000mAH battery just as quickly as you would plugging it into a wall-wart based LiPo charger. It is going to be a matter of how many times we can do that before all of our reserves are drained.

Thank you Hugues. I suppose I should have elaborated more on the landing. We originally had a setup where we used laser diodes and photo-resistors. The lasers would line up with the photo-resistors and we would read the analog signal to determine if we were lined up. This was combined with our ultrasonics and performed a checking algorithm every x amount of seconds to determine it was lined up all the way during landing. Unfortunately this process failed in any situations where daylight was involved (the ambiant light combined with our lasers being too cheap/not powerful enough). We are going to be utilizing ultrasonics on both the Quad and the rover. We also have a GPS system that is kind of implemented on the rover to do a position matching. We will be using lasers as a visual guidance that we will be able to view through our go-pro and OSD system while landing to help guide us. I have already been able to catch the quad while landing using purely my own visuals. Once these last few sections are implemented, I think we will be ok for now. 

As far as how we are landing it now, the quad runs fully autonomous and when it goes to land, I am manually driving the rover and I basically "catch" the quad if our coordinates are not lined up enough. With the accuracy of GPS, we can't solely rely on it to be autonomous. The magnetic grid then helps pull the quad into it's locked state. 

Because of our dead line (which is about 3 weeks) we probably won't have enough time to get the system fully autonomous. However with the way we implemented everything, it would be a matter of tweaks and better software developers combine with some mechanicals to get everything just right. Our team is 4 Electrical Engineers. 


MR60
Comment by Hugues on March 21, 2014 at 12:56am

That's great. Keep me informed about your auto landing on platform progress because I think there is potential here to solve one of the main road block for commercial use of drones, that is how to ensure a drone can land precisely on predefined landing/take off slabs.

Comment by lukechencqu on April 16, 2015 at 6:48pm

great job!

what is the project going on currently?

i think you can use wireless tech for charging system

Comment

You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service