dissertation idea 3rd year computer science
|Benjamin Yeomans||28/04/2013 00:16:41|
6 forum posts
I am currently in my 2nd year at uni doing a computer science degree.
My idea is to get a rasberry pi (basically a mini computer about the size of a credit card) and add a few modules to it, these would include a GPS module and some others for a few other features I am yet to come up with. basically I would have this GPS module which tracks the speed, alltitude and location and flight path of an RC plane I would then create some software which would display all this information/data gathered from the flight in a program e.g. graphs for altitude and speed, I could use a google map api for the GPS side which would allow me to display the flight path in 3d using google maps. I was also thinking of adding a small camera to the nose of the plane so the user can record their flight or maybe stream it to the program so it could be used in real time.
any other ideas for this would also be greatly appreciated so if you can think of any other features you'd like to see added please don't hesitate to say so
1423 forum posts
So you are just looking to use the RPi to collect data from a camera, gps and some MEMS sensors? I'd expect that to be possible.
If you wanted a challenge try going for the Raspberry pi collecting the data and controlling the model at the same time.
|2699 forum posts|
Interesting... However a playback on Google maps would only have the pane circling around a 100yd wide circuit, unless if course you are going to use a FPV and fly it cross-country..!
The real challenge would be to relay this in real time to a gound station for interpretation and display on a laptop. So, the pilot does the flying, and nothing else. The other person (me!) studies the flight envelope on a laptop in real time with a dashboard readout of speed, angle of attack, height, RPM, motor current (if electric), attitude, battery voltage, temperature...etc. You get the idea.
You will also need an electronics person to transmit this data (simplex broadcast) to a ground station and interface to a laptop.
Good luck !!
|Pete B - Moderator||28/04/2013 18:23:07|
7639 forum posts
Welcome to the forum!
An interesting project you have there but, if you're new to flying, don't underestimate the learning to fly aspect, so be sure to get some practice in before you install the Pi
|Ian Jones||28/04/2013 20:22:58|
3220 forum posts
A few screen shots from a laptop enabled to receiver the data:
Telemetry can also bee displayed on Hitec Eclipse 7 Pro & Aurora 9 transmitters, not the maps of course.
I think it's agood idea you have and an interesting project but sorry, it's not new.
All the best, Ian
Edited By Ian Jones on 28/04/2013 20:24:34
|Benjamin Yeomans||29/04/2013 18:12:09|
6 forum posts
Thanks for all the responses ,
Stevo, I did look into FPV ,a video on youtube of a FPV first gave me this idea however am I right in thinking it is quite expensive? also I think the real time readouts is possible that was my original idea having a real time readout of speed, flight footage etc. however this could prove quite challenging I would have thought? but defiantly a possibility!
163 forum posts
This is just a thought but would the police be intrested in this project regarding drones and bombs etc, our club had a visit about this type of idea.
Sorry if it puts a spanner in the idea
|Ben B||29/04/2013 19:15:34|
1423 forum posts
And presumably when they visited they discovered nothing illegal was going on and toddled off?
|Martyn K||29/04/2013 22:41:07|
5072 forum posts
The biggest problem you may have is getting power to the pi. It runs off USB - 5v I would guess- but the current consumption is quite high.I wouldn't try and run it off my reciever battery.Otherwise a cracking project, the sensitivity of GPS is about 3-5m with a 1sec positional lag so it probably wouldn't be brilliant recording fast maneuvres. You may be better building a Super60 type model for this. Lots of space and wing loading low enough to carry the additional power packs.
Let us know how you get on
|Benjamin Yeomans||30/04/2013 00:28:34|
6 forum posts
The police shouldn't have to worry about this I would have thought, but I'll check with the Uni thanks .
I have thought about the pi being powered by a separate battery pack I have found a few good tutorials on hacking the micro sd and hijacking it up to a battery pack to make the pi portable. Thanks I am glad the project is getting lots of positive attitude towards it that's always means I've got come up with something good and something that interests me!
16 forum posts
Like the idea, Ben. RPi is a very flexible device indeed. There are lots of people out there doing similar things with RPi's, Arduino's, Snapdragon's and such. They all seem to focus around unmanned flight or telemetry feedback or FPV flying. Now, if you tie these together into something task focused... I'll give you an example, a semi automated flight that takes off and lands under normal control but when in flight, follows a preset track or search pattern controlled by GPS and autopilot. Flight data is recorded along with video or still capture for something like search an rescue or aerial surveillance or even transmitted back live.
Just my take on your idea, certainly sound interesting.
|Olly P||30/04/2013 12:14:05|
3215 forum posts
I would perhaps have a waypoint based flight with gyro stabilisation for images - perhaps using the idea as a S&R type system - tie it into google earth for programing the initial flight and then have a 'return to waypoint' control also, so it can go back to a position and circle if something is found. A good final test might be a costal cliff area with a hi vis jacket somewhere.
The next stage would then be auto ID and circle of an object via on-board computer without ground station 'interference'
|Simon Chambers||30/04/2013 13:35:56|
|789 forum posts|
Check out www.openpilot.org. Does pretty much everything you was looking to do.
HOWEVER don't be put off doing it, as undergrad projects don't have to be unique, different or innovative.
My advice I can give from a couple years experience teaching electronics based undergrads is that those that do a difficult project but which has achievable goals, always invariably do better than those who have ambitious projects but don't achieve the intended goals. I.e. what ever you do, make sure it works and meets your original goals! So choose a project that you can complete in a realistic time frame (I'd aim for it to be fully working by Christmas) - at that point you can extend it to meet new goals.
Nothing worse for a supervisor (and marker) to have what looks to be a great project idea but is only a collection of half working systems that aren't integrated.
Now is the best time to create a proof of concept with a simpler platform to test it before you start your final year.
Your actual project idea could be readily and easily integrated into a smart phone - which will probably be substantially easier to complete on than using a RPi. As all the API's will be there for you to access the hardware without much trouble (i.e. Camera/GPS/Accelerometer/etc API). Also app developers are hot property at the moment, so it could be beneficial to your future career job prospects to develop on a smartphone.
Plus an App could be published on the App store and out to market - which always looks good to an employer. It could also be used for many other things too - e.g. tracking your run/walk or even wrap one up in a package and post it to see where it goes and how your package is treated!
Developing on a RPi will require a lot more linux knowledge and experience than developing on a smartphone, likely requiring kernel recompilation to get the necessary webcam drivers working. It's likely that you'd have to develop your own APIs (or wrappers around them) to access this hardware - likely in C/C++ (rather than in Java which is likely what you've learnt already). Plus you may experience difficulties that you haven't envisaged - such as performance limitations from capturing video from a webcam to SD card, etc, etc.
Edited By Simon Chambers on 30/04/2013 13:37:43
|Benjamin Yeomans||02/05/2013 18:26:58|
6 forum posts
Sorry about the late response guys, Certainty an interesting idea that search and rescue idea may have to do something around that.
Edited By Benjamin Yeomans on 02/05/2013 18:28:27
|Max Cairns||03/07/2013 17:09:32|
|14 forum posts|
FWIW, PIC microchips run on 5V, cheap, tiny current, come in lots of sizes and capabilities, and need very few external components to run. (Most can run on just +- pins connected). Smaller and lighter, too. Several languages - JAL is very easy. Bit more of a project to get them working, though.
|Benjamin Yeomans||03/07/2013 20:08:43|
6 forum posts
2. Mosaicing aerial imagery captured from a low-level UAV.
Using a multi-rotor radio control helicopter we have acquired multiple datasets of images of fields of crops. From this data, we can calculate a number of measures which can indicate how healthy the crops are in the field. However, the multiple individual images need stitching together to form one large image representing the field (see left). Current automatic methods do not work well on such sets of images, as the content of the images is so similar. In this project, the challenge is to build a tool which will allow the user to load in such a set of images, try to automatically stitch the images together based on their content, and allow the users to refine the final results manually if necesary.
|Martin U||02/08/2013 23:17:35|
96 forum posts
Hi Ben, the GoPro cameras have great quality. Take a quick look on YouTube and you'll see loads of examples.
Please login to post a reply.
Want the latest issue of RCM&E? Use our magazine locator link to find your nearest stockist!