The PX4 team at ETH is a finalist for the IROS 2012 Best Paper award for this work, which uses the PX4 and the forthcoming PX4Flow optical flow board (coming soon from 3D Robotics) to do this:
We describe our autonomous vision-based quadrotor MAV system which maps and explores unknown environments. All algorithms necessary for autonomous mapping and exploration run on-board the MAV. Using a front-looking stereo camera as the main exteroceptive sensor, our quadrotor achieves these capabilities with both the Vector Field Histogram+ (VFH+) algorithm for local navigation, and the frontier-based exploration algorithm. In addition, we implement the Bug algorithm for autonomous wall-following which could optionally be selected as the substitute exploration algorithm in sparse environments where the frontier-based exploration under-performs.
We incrementally build a 3D global occupancy map on-board the MAV. The map is used by the VFH+ and frontier-based exploration in dense environments, and the Bug algorithm for wall-following in sparse environments. During the exploration phase, images from the front-looking camera are transmitted over Wi-Fi to the ground station. These images are input to a large-scale visual SLAM process running off-board on the ground station. SLAM is carried out with pose-graph optimization and loop closure detection using a vocabulary tree. We improve the robustness of the pose estimation by fusing optical flow and visual odometry. Optical flow data is provided by a customized downward-looking camera integrated with a microcontroller while visual odometry measurements are derived from the front-looking stereo camera. We verify our approaches with experimental results.
Here's the team and their birds. Huge and well-deserved congratulation!
Comments
Hi Graham,
Some people really like Tricopters, however, they are kind of a specialist airframe and are an unusual hybrid and they do not readily convert into hex Y copters.
Tricopters require a servo to move the one motor that spins in the opposite direction of the other two right and left.
Turnigy plush ESCs are very good and I use them regularly.
Turnigy Nano series batteries are likewise very good.
But I'm going to take a flyer here and recommend you start out with a quadcopter rather than a Tricopter. They are the easiest to workd with and even a reasonably small one can do a lot of work.
The 3DR Arducopter is an excellent choice and a lot of us like the DJI Flamewheels for their rugged simplicity, and versatility as well as excellent price.
The DJI F450 FlameWheel is avaialble as an ARF (Almost Ready to Fly) with Motors, ESCs and Props for around $180.00.
I have a F450 and a F330 and am very happy with them, the frame is very rugged and the motors and ESCs are of excellent quality and will put up with a lot and provide a long useful life.
They come with 10" props but can be upgraded to 11" or even 12" without overtaxing the motors and can run 3 or 4 cell LiPos.
All you need to add is a LiPo, receiver, a APM 2.5 and a BEC and you are ready to go.
Check out the ArduCopter and the Flamewheel assembly articles on the ArduCopter Wiki.
3 or 4 cell LiPos are as big as you should go when starting out, 5 - 8 cell is for really experienced and deep pocket guys only.
My biggest suggestion is that you start out with something to learn how to fly with it, simple, rugged and common like I have mentioned above.
You are a lot better off developing your learning curve with something affordable than building the ultimate tool and destroying it first time out.
If you were going to start racing cars you generally wouldn't start with a formula 1.
Thanks fot the Help, I was thinking that was the case, as i cant find much about the PX4 which told me it is very new , I am a very adept DIY'er, with many things but programming is not one of them.
you all seem very knowledgable in this and I am very new, would you mind looking at my parts list for me I have yet to make the purchases, and am nervous as i have no one else's direct feedback on the setup i have planned.
if you dont mind.
--
I plan on making a Tricopter that will be upgradeable to a Y6 when I need to add more survey equipment, So I want the Esc's to be over kill in the first version so I don't have to replace them (just add to them ) when i upgrade.
Arial Survey Try copter with a planned upgrade to Y6 when load requirements and funds increase
Proposed Kit
1 x ArduFlyer APM 2.5
1 x Ublox6M GPS
3 x hexTronik DT750 Brushless Outrunner 750kv
3 x Lan Yu 40A ESC /speed controller For brushless Motor 8.32 ea
or
3 x TURNIGY Plush 25amp Speed Controlle 12.48ea
1 x TR_PC – TURNIGY BESC Programming Card
1 x Turnigy 2200mAh 3S 20C Lipo
3 x OEM APC 12X3.8 props
1 x Hobbywing 3A UBEC 5V 6V Lowest RF Noise BEC
1 x 1-8S Lipo/Li-ion/Fe Battery Voltage 2IN1 Tester Low Voltage Buzzer Alarm
1 x FrSky Diy 2.4ghz Module
Some specific Questions I Have:
Have you ever herd of the Lan Yu ESC's They are pretty cheap and I am on a tight budget. the specs look good but I am not sure if they are compatible with a Tricopter? www.aliexpress.com/item/Lan-Yu-40A-ESC-speed-controller-For-brushle...
How do I know what props the DT 750's can swing? i have seen people suggest 12x 3.8 but i plan to upgrade them to 13" or 14" if i need the thrust what do you think?
and finally I have an old sony cybershot that I want to take apart and use the video as fpv? do you know any resources on this? otherwise i am flying blind… excuse the pun.
Hi Graham, Realistically the PX4 is eventually going to be a lot more powerful than the APM and do a lot more.
But when end user software is actually released in a month or 2 that is really just going to be the start of it's real development cycle. The next 3 or 4 months in the real world will inevitably result in lots of "problems" and lots of fixes.
The APM 2.5 on the other hand is very nearly a debugged working solution the near term expected release of Inertial navigation on Horizontal as well as vertical axes is going to quickly result in a very solid flight experience.
And the built in firmware is pretty well developed for the use you have outlined.
My guess is you would be having a pretty fruitful experience with the APM 2.5 in a month or 2.
Likely you would be pulling your hair out with the PX4.
If you want to help develop a new and really powerful flight controller and can work through it for 3 or 4 months the PX4 might be for you, otherwise you are likely to be much happier with the APM 2.5.
I have no choice I need to develop obstacle avoidance software for a LASER scanner; the PX4 will be my next purchase. There will be blood!
Thanks Chris, I am building a Survey UAV and have not started, do you think it is worth waiting for the PX4 or will it be stable enough for a nube like me to use. I am just hoping that the APM2.5+ will not get obsolete any time soon. especially since i plan on integrating sonar and ir for obstacle avoidance.
ps thanks for the quick reply
Graham. The PX4 UAV software is still in developer beta and is not yet released in the Mission Planner for regular consumer use, the way APM 2.5 is. It will probably be another couple months before the consumer release.
aI am debating weither to get the APM system of the PX4 I think that the higher processing is a great feature, but has the code caught up or it there a simple solution for running the AC in a shell? I am a beginer at all of this.
help would be apreciated.
Also I plan to use my UAV for arial mapping of larger sites , does the Flow board support this functionality, with a map waypoint planer for photo overlapping?
Cheers
Wow good job guys. Although when the robots finally take over, and this project is recognized as a vital step in them taking over, I'll retract my "good job".
No, we are not there yet. There are two approaches. One is to use basically start from scratch, using components of the existing PX4 system, and maybe parts of the Arducopter code. Then there's the other approach which is running the AC code in a "shell" on the PX4.
What?! This is NUTS ! :D
-
1
-
2
of 2 Next