(Originally posted on NOVAerial Robotics Blog)

NASA is now planning on using a UAV in addition to the Rover in order to explore Mars.  The reasons are very simple, and parallel the desire to use UAV’s on earth.  Having a camera in the air, instead of being stuck on the ground, allows for better perspective to see details of the terrain.  And also, avoids the problem that rovers have with negotiating difficult terrain.

But what is particularly interesting with this story, is that they have decided to use a coaxial helicopter instead of a quadcopter.  Why?  Multirotors on earth, already have enough trouble with stabilization due to the need to accelerate and decelerate the propellers to control their flight.  While small quadcopters do fly very well, as they get bigger, and the rotors get bigger, stabilization gets harder.  This is because the rotor inertia increases exponentially with diameter.  It takes much more energy to change the speed of a large rotor than it does smaller ones.  This is why you do not see very large multirotors flying very dynamically.

On Mars, the problem is even worse.  Due to the thin atmosphere, the rotor must be very large to be able to generate enough lift to fly. It would require a very large and heavy rotor to achieve lift, and the power required to change their speed quickly enough for flight control, would be much too large and heavy to fly.

Helicopters do not have this problem, because they use a swash plate to actuate variable pitch on each blade for flight control.  The motors do not need to be large enough to accelerate the rotors quickly for roll and pitch control.  So you can have a tiny motor spinning a large rotor in the thin Martian atmosphere.  Roll and Pitch control is achieved by cyclic pitch.  Yaw control of a coaxial helicopter is achieved via motor torque, and thrust can be controlled either by variable collective pitch on the rotors, or increasing the rotor speed.  I expect they used the former in this case, for the same reasons why fixed pitch props don’t work for pitch and roll control.  All of this also explains why large UAV helicopters fly so much better than large multirotor drones.

One interesting thing to point out, is that all of this also holds true for UAV’s flying at high altitudes on earth with thin air.  Multirotors will suffer much more than Helicopters will.  We have flown UAV’s in Colorado, both multirotors and helicopters.  The multirotors required installation of larger diameter propellers to achieve enough lift, in our case, changing from 10″ props to 12″. This negatively impacts the flight control, as the propellers are noticeably heavier.  The helicopter however, simply required a change to the motor controller to run the rotor slightly faster to make up for the thinner air.

We are excited to see that NASA is taking a novel approach to exploring Mars.  Here is a 37 minute video which goes into deep detail the system.

http://1080.plus/Mars_Helicopter_Scout/w3y7iJEe7uM.video

Today I am pleased to announce the formation of a new commercial support company which I have founded in combination with Bill Bonney (creator of MAVPilot and maintainer of APM Planner 2.0 GCS software) and Patrick Krekelberg who has been running an existing R&D consulting company for several years.

We decided to do this because it has become apparent that there are many companies which are attempting to use UAVs and other mobile robots, but struggle because the field is multidisciplinary, requiring skills not just in software development and electronics, but also power systems, aerodynamics and mechanical engineering.  These combined skills are difficult to find in one single person. Over the years, the Ardupilot community and DIY Drones has fostered an amazing group of people who have developed these skills, many who are working as independent contractors. But some, particularly large companies, prefer the simplicity of dealing with one single entity to simplify project management, billing, etc. and to make sourcing talent easier and more reliable.

Our goal is that as the business grows, we will be able to draw on the talent of the community of developers which has built up within this ecosystem, and help connect them with commercial companies needing assistance with their projects.

We aim to help commercial companies gain easier access to the performance, reliability and flexibility of Ardupilot, and also help find ways for developers to keep doing what they have been doing until recently.  We will be supporting the ArduPilot community with contributions to the code, support to members on discuss.ardupilot.org and its day to day operations as a commercial supporter.

ASC will offer more than just consulting on Ardupilot code, but complete end-to-end solutions. R&D, vehicle design, rapid prototyping micro-manufacturing, program management, autopilot code, payload integration, user interfaces.  ASC also offer assistance with operations, flight training, etc.  As little or as much as required to get customer’s ideas off the ground, or along it!.  Companies can focus on integrating mobile robotics into their business and developing new revenue streams, instead of having to undertake the long and difficult process of developing UAS expertise and technology in-house.

Autonomous Systems Cooperative http://autosystems.io

Today I did something a little bit unusual, and definitely more difficult than flying around with a camera. A possible commercial customer asked to see a demonstration of a helicopter flying with a 2.5kg sensor.  This sort of thing is commonly done for magnetometer surveys for mining exploration.  The sensitive sensor must be hung from a long line in order to get it away from the magnetic interference from the UAV.  This could also be done for water or gas sampling.

The big helicopter has no problem with flight control despite the hanging weight.  It just requires some special piloting to control the swinging, basic pendulum dynamics.

These sorts of surveys require many tight passes at low speed, so it takes a long time to cover a large area.  It's an ideal job for a gas powered helicopter UAV.  This one can fly for an hour easily.

It would be possible to create a payload management system which would measure the position of the sensor via the angle of the rope, and automatically control the UAV to control the swinging (basically, the reverse of a balance bot).  Something I would love to work on in the future.

I wanted to bring attention to the work that Tridge and the rest of the CUAV team have done to demonstrate the possibility of flying a Helicopter with a standard FBL controller handling the rate control duties.  They have been testing gas powered helicopter for next round of the Outback Challenge which will require a long-range VTOL aircraft.  Helicopters are a natural target for this mission of course.

Several of the CUAV team are experienced RC helicopter pilots, but not as familiar with installing a Pixhawk on a helicopter which can be difficult, especially in the case of gas engine helis.  As such, they were more comfortable having a normal Flybarless Controller handling rate control. In this case, the Skookum Robotics 540.   Tridge has made changes to the code which allow for a pure control pass-through from the RC Rx, through the Pixhawk, and straight to the FBL controller. This pass-through occurs in Acro mode, whereby the Pixhawk running ArduCopter has no effect on the flight control, so even if the Pixhawk should have a major AHRS/EKF failure, the helicopter is still controllable. 

Of course, the Pixhawk is capable of controlling a Helicopter without any FBL system, and this is the most common arrangement.  But the possibility of running an FBL controller in series with the Pixhawk helps lower the barrier to entry for many existing RC Helicopter pilots.  And also offers helicopters similar failsafe-function to airplanes, where they can survive the loss of the autopilot.

FLARES - Flying Launch and Recovery System from Insitu on Vimeo.

Recently I ran across this really interesting system being used to launch and recover the Scan Eagle, which is a 40+lb airplane. That makes this a VERY large octocopter!  It's a very interesting concept.  The launch seems much gentler on the airframe than a typical catapult launch.  And even though it's a huge octocopter, it replaces two separate launch and recovery mechanisms, which normally are trailer mounted.  They can be seen in this video to get a sense of the scale.

Paradise Fire Olympic National Park, Washington, USA from Insitu on Vimeo.

http://bigwhite.cmail2.com/t/ViewEmail/r/3059DF9ADD1D1A942540EF23F30FEDED/D84B544420A969371726EA5DA1051479

This just popped up on a Canadian Commercial Operator Facebook group and we're discussing the implications and legal possibilities.  Apparently the operator claims that this has been cleared with Transport Canada, but nobody is sure yet what the details of that are.

This does not appear to fit in with existing commercial regulations.  It seems unlikely that a commercial operator would be able to get a permit to fly in a crowded area like a ski hill.  I have taken a guess as to what is going on here:

The operator is not operating the drone.  They are renting drones to recreational users.  Not much different than selling drones to recreational users.  The user is the pilot in command, not a commercial operator.  The user captures images of himself.  Then the operator edits the footage.  It would bypass all of the commercial regulations.  Some disagree that this would be legally possible, but I don't see why it would not.

Currently, a person can rent a Uhaul truck, without being required to possess a commercial truck license.  And I think it would be over-reaching if the government tried to stop a person from paying somebody to edit footage for them, no matter how it was captured.

The key to the whole gambit, is the drone rental, and the Follow-Me device is pivotal for that to work in this case.  Some of the marketing material show a Phantom, but the video from Cape Productions clearly shows a 3DRobotics X8.

No matter how this situation is being managed, it presents a very interesting legal scenario, that could test some of the implications of the existing Canadian UAV Regulations.

On February 19th at 9PM, the Canadian Broadcast Corporation will air a new documentary on drones, following up the one they did last year called Remote Controlled War.  Last year's was more focused on the military aspects, this new one is more commercial/recreational.

I'm not sure what the availability will be for those outside Canada.  Hopefully our public broadcaster makes it widely available.

Last weekend I drove up north a bit for some real cold weather testing of my Iris+ equiped with Pixhawk.  I had to get up early in the morning to catch -30C temperatures which is about as cold as I'm interested in testing.  At -30C, my gloved fingers froze after a few minutes.  Not quite frostbite, but deep into pain territory.

Spurred on by Leonard's taunts, I also did some snow resistance testing.  It passed.  I'd say it can handle just about any amount of snow you could reasonably throw at it. :)   I simply put electrical tape over the shell vents, and that's it. Throughout the weekend, I had absolutely zero problems related to the low temperatures. -30C is a non-issue for the Pixhawk and ESCs. 

I actually haven't had this Iris for long, but so far I'm pretty impressed with it.  The out-of-the-box tune was pretty much spot on.  It flies extremely smoothly.  The only changes I've made were to allow 45° lean angle, and enabling Stab and Acro modes.  The camera I was using for on-board is a Sony AS100V on the semi-fixed front mount included with the Iris.  The results are not great, heavy jello.  I'm not sure if a real GoPro is much better?  You can see the effect of the Sony's electronic image stabilization.  If you peer through the jello, you can see that the video is actually quite smooth and watchable.  I will probably make a damped undermount for the camera.

I'm not sure if anybody who is not a member of Unmanned Systems Canada can see this report or not:

https://www.unmannedsystems.ca/media.php?mid=4633

So I will summarize:  At a meeting of the International Civil Aviation Organization in Montreal last week, Transport Canada presented a working paper entitled "Canada's Approach to Managing the Risks of Remotely Piloted Aircraft Systems".  It's only about 4 pages long so not too bad.  But one very interesting statement is made in section 3.1:

Canada intends to remove the distinction between recreational and non-recreational user categories in its accelerated rulemaking strategy.

There's really no detail beyond that.  Does it mean that recreational users will suffer the same extreme regulation the commercial industry does in Canada?  Or does it mean that they will essentially remove all regulation from the commercial side, since there is none on recreational users now?  Or will the two meet somewhere in the middle?  At this point it is not clear, but there's indication that this is intended to happen in 2016.

Ah, looks like this is a public document: (UPDATED BB: to link correct document)

http://www.icao.int/Meetings/HLSC2015/Documents/WP/wp076_en.pdf

I don't really even know what to say anymore. It has 4 motors.  And a camera. And wifi control. This one has a flash light source though, so that's new.  

https://www.kickstarter.com/projects/torquing/zano-autonomous-intelligent-swarming-nano-drone

This is just a quick note to say that the Ardupilot.com/forum/ was down temporarily due to server issues.  

Well, the moment we have all been waiting for in Canada is finally here.  Transport Canada has published the new UAV regulations.  It is important to understand that these regulations effectively come in 2 parts.  Basically, you have one set of regulations for UAV's less than 25kg and for simple operations.  Then another set for UAV's larger than 25kg and/or complicated operations.  Then, within those two groups, there are more groups.  For the first, it's broken down into <2kg class, and 2<>25kg class.

http://www.tc.gc.ca/media/documents/ca-standards/Infographic_Permission_to_fly_a_UAV_Web_English.pdf

The rules allow quite free operation of <2kg UAV's.  In a nutshell, you have to not be stupid, and fly safely.  You can't be drunk.  Know how to fly.  Have a plan.  And remain at least 30m away from people and things not involved in the operation.

For UAV's less than 25kg, there are a few additional, reasonable additions.  You must have a fire extinguisher, and remain 150m away from people and things not involved.

The only question I have is, what does it mean for a building to "be involved in the operation"?  If I am taking photo/video of a building, does that mean it's involved?  I would assume so.  But, what about neighboring buildings?  Are they involved?  Because, it is uncommon to have solitary buildings 150m away from any other buildings. If this rule is rigidly applied, then it means you can pretty much only fly larger craft in remote areas.

Now, for UAV's over 25kg and/or operations not meeting these simple rules, there is a much more complicated document which applies:

http://www.tc.gc.ca/eng/civilaviation/opssvs/managementservices-referencecentre-documents-600-623-001-972.htm

I've skimmed through this, and it appears to be similarly reasonable.  There are various levels of permissions, etc.  too much to get into in detail here.

Another recent accomplishment with the Helicopter version of Arducopter was this full auto speed run which achieved a top speed of 30 m/s or 108 km/h.  This was done using 3.2 which features improvements to the helicopter flight dynamics code, as well as Spline Waypoints which is a general Arducopter feature.  I find the Spline nav really helps helicopter flight as it avoid jerking the helicopter  which is very important at high speeds.  

108 km/h in auto mode is a good accomplishment given that the absolute maximum speed this helicopter is capable of is 140 km/h, based on Retreating Blade Stall.  This performance wasn't a fluke either, as not only do I do 2 back-to-back runs in the video, I did in fact run this mission 10 times while preparing to shoot the video.

For reference this is a relatively small helicopter, a Trex 500, which weighs only 2100g and is 90cm long.

Now we're cooking with gas!

I have just completed a project to build a helicopter with a gasoline engine, controlled by Arducopter (of course!)  This helicopter serves as a test bed for future gas heli development work for a client of mine.  The goal was to prove out that it could be done.  Gas engines pose a unique challenge for Arducopter due to the heavy vibration they produce.  But they also present a lot of promise for UAV applications as it can allow heavy payloads, long duration flight, and high speeds.

The heli is based on a Helix Heli conversion of a Trex 700.  However I stretched it to 800, used many aftermarket upgrades such as from KDE, and machined a number of custom parts myself.  The machine required quite a bit of customization in order to increase frame stiffness which is important for vibration reduction.  The engine is a Zenoah G29RC engine with pull start.  This is an increased displacement 23cc engine, which makes more power but creates more vibration than the G23.  I could have purchased an aftermarket balanced engine, but I wanted to use COTS parts and also prove the Pixhawk can survive with a worst-case engine.  

Last night I performed the first check flights of the new machine and began PID tuning.  This is one segment of that flight where I was working on the PIDs.

Today I finished the PID tuning, and it is flying very well indeed.  After increasing rotor speed to ~1600rpm and tracking in the blades, the vibrations on the Pixhawk are quite good, better than both of my quads:

I flew it around a bit in Stabilize, and hit 30 m/s without really trying.  It should easily be capable of 40 m/s flight speeds.  I'm hoping for a 10 kg payload capacity. Flight time is currently better than 20 minutes with plenty of reserve on a 630cc tank.  I will probably begin development on a larger fuel tank system on the order of about 2L which should allow for 1 hour of flight time with payload.  Currently the 2S 4400mAH avionics battery lasts for at least an hour, but will be supplemented with an on-board generator.

Later today, I will start working on auto missions and will shoot another video.  

The success of this project opens up an entirely new world of possibilities for affordable UAVs.  It will be possible to perform realistic crop spraying.  Aerial imaging with the best quality cameras available.  And extended flight ranges in a large aircraft with VTOL capability for usage in areas without a prepared landing strip.  Obviously this is not the type machine that would be used anywhere near people. But remote areas, or secured industrial lands would be suitable operational areas.

I've been working on a new helicopter platform the last few months. Based on an MSH Protos heli which I chose because it's an extremely light weight platform, weighing in at only ~1200g without battery. It has a full belt drive which I much prefer to gears as it's quieter, lower vibration and more reliable. I've had a few problems with it because the belt drive makes a really awesome Van deGraaf generator... not a good thing on a UAV. But I solved that, and am conducting test flights now.

The flight controller is a modified PX4v1. I replaced the switching regulator with a MIC29300, so that I can run it on 2S direct with the servos. Main motor power is 4S 5000, typically this heli would run on 6S 3300. Using the MSH stretch kit and 465mm Spinblade Asymmetric blades. In otherwise standard form, this heli flew for 17 minutes on an old crusty battery, in -10C temperatures.

I have now added a subframe to hold an extra battery, FPV gear with a camera in the nose, and a vibration damped NADIR camera mount to be used for aerial mapping. The idea is to develop a mapping UAV that is superior to a multirotor, offering a valid alternative to a fixed wing for short to medium range missions. The VTOL capabilities would eliminate all the nastiness of catapults, and controlled-crash landings with onboard cameras in rugged areas.  Even the price is attractive at about $400 for the basic kit with motor and ESC (no servos).

Specifications show the advantage of a heli platform. This machine has an AUW including the batteries and camera of only ~3kg. It is 80m long, and about 15cm wide not including the extended legs, and 30cm high. The blades fold for easy transport, without requiring any lose wires or vibration-prone electrical connectors as a folding multirotor does. It actually looks much bigger on the table than it really is. This seems to be very good compared to multirotors I've seen with the same performance. (payload and duration)

Vibrations are always a problem with helis, but manageable with the right design and construction techniques.

Arducopter really makes helis worthwhile. You could buy two entire heli systems including a Tx for the price of a single DJI Ace One non-waypoint controller.  Or 7 for the cost of a single Ace One waypoint enabled controller.  I strongly prefer the PX4 controller over the APM and Pixhawk, because it offers 32-bit performance in a small package that is easier to mount in a heli frame.

So does it work? I took it up for it's first photo tests yesterday, and it worked beautifully. Better than 80% photos are usable. It flies for 20 minutes in a hover with old, cold batteries (-5C). I'm hoping for closer to 30 minutes while actually moving (helis are more efficient moving than hovering), in warmer weather with new batteries.  It should have an easy cruising speed of 15 m/s with little or no reduction in flight time.  At 20 minutes, this would offer an 18km range, and 27 if it can do 30 minutes.  If you wanted to do FPV and not mapping, you could configure it with a 3rd battery in place of the SX260 and fly for... 30-45 minutes, and a range of up to 36km.  Top airspeed is still TBD, but probably 20-25 m/s.  

Wind penetration and stability is excellent compared to both multirotors and fixed-wing.  You could do a mapping mission in winds up to 40 km/h with little effect on stability or duration.

If the success continues, I'm going to consider building a large gasser heli.  This would allow flight times up to 2 hours, or payloads on the order of 10 lbs for 30 minutes.  So you could map large areas, or even perform light duty spraying operations.  I'm thinking about local application of a herbicide for things like Giant Hogweed elimination, that sort of thing. Such a large heli does pose significant danger and should only be used in industrial, agricultural or remote areas.

3.1 was released just in time for the holidays, but I felt the new Acro Mode we have developed has not been adequately demonstrated.  So I shot this video to show that Arducopter now has a competent Acro mode to go along with all the other functions.  This is kind of a big deal to me, as I am interested in learning acrobatic flying, but I really appreciate having a system which can "bail out" if I lose control.  Throughout this video I was flipping between Acro and Stabilize, with Arducopter generally doing a good job of saving the copter after I lost orientation.

This is also one of the few demonstrations of a new Pixhawk controller on something other than an Iris.  But don't be fooled, Acro works just as well on an APM2.5.  :)

While developing and testing this code, I have also done quite a bit of high speed acro flying, on both the quad and my helicopters, but it was simply too cold to bother hearing out to the flying field.  So this video is just some basic flips in one place.  I will try to get out for some better video in a few days.  If it warms up. :)

I think this is actually the second attempt at this.  Wasn't there one last year?  Good to see that they actually caught the people involved and arrested them.  Hopefully it won't result in a knee-jerk reaction, but I doubt it.  The article is horribly written, but at least they didn't call it a "drone".

http://www.walb.com/story/24047984/crooks-get-creative-to-smuggle-contraband

On Friday we experienced very high winds, 60-90km/h, or Beaufort 7-8, this caused damage in the area, actually knocking over some trees, etc.  I decided to go out and test the wind performance of my F450 quadcopter, see if it's even possible to fly a quad in such conditions, and test the inertial Loiter performance.

Overall, it went fairly well.  Flight control in Loiter was quite a bit better than flying manually.  Last year I tried flying a quad in 40 km/h winds, and it did not work at all.  As soon as I got off the ground, it was practically out of control and I spent another 2 minutes just trying to get it back on the ground safely.

I should point out that this could actually be considered a worst case setup.  The Tarot F450 frame I'm using is very flexible.  The quad can't be autotuned, because of that.  In fact, I haven't spent much time at all setting it up.  All of the settings are default, other than some of the basics like the Rate PIDs.  I haven't spent any time tuning up Loiter, or Alt Hold parameters.  I haven't even done CompassMot. (Sorry Randy!).  

Despite all of that, the copter performed fairly well, as you can see.  A properly setup quadcopter should do even better.  Unfortunately winds like these are rare so I can't go back and retest whenever I want.  So, taking this as it is, it was fun to do this test, and I think it shows off the performance potential of Arducopter fairly well. Another thing to keep in mind when watching the video, is to understand that while straight-line wind speeds were 60+ km/h, the area I was flying was in the lee of several large obstacles like houses and a forest. As such, the air was extremely turbulent, which is why you see quite a bit of violent movement. I expect if I was flying in a clear field, the copter would sit more stable, with a constant tilt angle to fight the wind.

I would have liked to do back-to-back testing of the performance of the quad vs. a helicopter, but I didn't have a heli ready for Loiter flying at the time.  Looking back at last year's video, it seems like the helicopter would have done quite well actually.  The video from last year was done on a helicopter with a flybar, using an APM1 with a worse baro, and didn't have inertial navigation at all. Even still, it was quite stable as you can see.

Less than one month ago I predicted that eventually DSLR's as we know them will disappear, and instead you will buy lenses with the imaging chip (CMOS) built right in to the back of the lens.  Then you will only attach a simple body to it that gives you IO and a viewing monitor.  This is because CMOS chips are constantly coming down in price, but the really important part, the glass, will always be expensive.

This is an interesting idea for us because it will allow us to lift only the imager section, without carrying the weight of the DLSR body, which will cut in half the weight of a quality imager.

Well, it is already starting.  Sony have just released the QX10 and the QX100.  Both are basically intended as high(er) quality imagers, for use with your existing smart phone. They sort of clip on front of the phone, and you use the smart phone as the IO system.  They can apparently operate independently as well.

The QX100 is particularly interesting, because it weighs only 179g, but gives you a 1" CMOS chip, along with some decent quality glass.  Pretty amazing.  You can imagine building a nice little brushless gimbal around this thing and have a pretty nice setup for only about twice the weight of a GoPro.

However, I must say that I am quite disappointed in some of the specs.  Sony seem to have dumbed down performance, particularly video, as it's only capable of 1440x1080 at 30 FPS whereas the RX100 which uses the same chip can do 1920x1080 at 60FPS.

But still, this is a very interesting development for aerial photography.

Otherwise, I imagine this will be a market flop as it doesn't really make a lot of sense.  It's 2/3rd's the price of an RX100, but doesn't have the same performance or features (limited manual control). Every time you want to use it, you have to pull it out of your pocket, attach it to your camera, flip to the camera app...  Meanwhile somebody with a stand alone camera could have turned it on and snapped several pics already.

Maybe the QX10 is actually a more interesting candidate for aerial photography.  The lens and imager on this is the same as Sony's mid-range point and shoots.  In reality, the image quality is quite good.  And at 105g, it's only slightly heavier than a GoPro.  You could mount this on any tiny quad, and get pretty good photos and video. (no GoPro fisheye!!!).  It also has optical image stabilization.  

I'm not sure this will be a home run product either.  But both units are signs of more interesting things to come in this market.

I did some more flights on the weekend to test out my branch of Arducopter 3.0.1 -THdev and wanted to show what it's like from on-board the heli in a reasonably high-speed run.  This run used 12 waypoints to form a rectangular pattern with rounded corners around the flying field, trying to achieve a nice smooth flight path.  I still haven't been completely successful as it's a bit jerky.  This was using WP Acceleration of 200 cm/s/s, and WP Speed of 25 m/s, but this was not achieved on this flight.  It did hit 20 m/s, and I'm happy to see that it's very very smooth.  With settings of 20 m/s and 100 cm/s/s acceleration it's much smoother but takes too much room to accelerate. This was also using the Yaw Look Ahead mode which causes the yaw to automatically look into the direction of flight.  The flight ends with an RTL which causes the yaw to return to the original orientation on the final descent.

Overall, I'm very happy with the performance, but I really anxiously await the Spline Nav feature! :)

The heli itself is still a work in progress.  Obviously the camera vibration damping is completely unacceptable and I'm redesigning that.  But other than that, it's going well.  The Direct Drive Tail is brilliant, and really simplifies the mechanics.

I am getting about 12 minutes flight time, including a 3 lb payload (yes, I'm using steel blocks as ballast), and this is with a battery load smaller than a 700 typically has.  Normally they run 12S 5000 mAh, and I've only got 8S 5000 right now.  But I plan to run a total of 8S 10,000 which should give better than 20 minute flight times.  

Once I'm comfortable with the reliability, I'll be stretching it's legs on longer runs.  I'd like to have this capable of at least 100 km/h flight with a camera on-board.  What I envisage building is a compact high-speed aerial filming platform capable of chasing race cars, etc.  Ultimately I will build a completely custom frame but for now I'm just hacking on this 550 frame stretched to 700.