Training with a QuestUAV 200 In Finland

At the beginning of this year I entered an interesting email discussion with Romi Rancken, a senior lecturer in Finland. He was thinking about leaving his current work at Novia University in Finland  and studying the forests and archipelago of Southern Finland using UAVs (rotary and fixed wing) as part of his set of research tools. 

We threw a few ideas around, but they all seemed quite challenging, either involving lots of trees or lots of water… neither of which are particularly healthy for UAVs.  Trees can generate huge turbulence and wind shear, whilst large bodies of water or sea leave any UAV pilot feeling slightly uncomfortable unless … um … whatever … there is so much not to desire flying over the sea! Living on the North Sea coast in Northumberland UK I’ve done a lot of UAV flying by/over the sea and lost a couple of  UAVs there. Total write-off; the UAV may survive but the salt kills everything pretty soon afterwards.

Anyway, as I write this blog I’ve just come back from a great few days in Finland, living with Romi and his wife Tuula, and flying around the forests of Southern Finland. We all agree that at times it was tough and challenging!  I’m sure Romi occasionally have wondered what he had let himself in for. I wondered too, but at least I had great faith in the QuestUAV 200 and it’s ability to handle a breathtaking amount of violence, and still live.  My faith was well placed.

Romi, Jon and George with the 1990 Novia UAV
and the QuestUAV 200

Prior to our arrival, Romi, an experienced forester, wanted to operate from a clearing that is in the heart of his University research area. My instincts put me decidedly on edge operating so close to trees and in a relatively small clearing.  To the inexperienced eye, a 200m x 250m clearing would seem fine as a flyng site, especially with the protection at ground level from the winds. As it turned out, the winds and turbulence would be considerably worse than expected.

Our arrival to Helsinki was in torrential rain and the forecast wasn’t brilliant with more rain and winds expected throughout our short stay. The first night of our arrival I introduced Romi to the simulator and we spent a couple of hours working on flying, perspective, take-offs and landing. 

A quick 3d ortho created from100+
images on a very windy flight

The simulator is a key piece of training, complete with a QuestUAV aircraft to fly around and get to know the transmitter. Whilst the QuestUAV is an autonomous beast and will do all the flying itself, like many other sUAV’s, its unwise to leave every situation for the UAV flight systems to sort out.  Some things will happen (bad planning, weather change, unexpected winds or turbulence etc) that the intervention of a pilot, if only for a few seconds, might save the day.  Of Romi’s crew (including a Russian businessman/researcher called George Rybakov) I thought that one of them would have some previous RC experience.  I didnt pick up that the these skills didnt exist before arriving in Finland and our normal process in UK of sending a simulator and transmitter a few weeks ahead of the main delivery didnt happen.  Little did we know we were going to step into such a tough training scenario with  a difficult flying field, rough weather and inexperienced flyers,

Day one was classroom training and an introduction to the equipment with flying in the afternoon.  Classroom training was standard – lots of information to take in, lots of questions to answer and lots of interest in the whole process.  When we arrived at the flying site, it was agreed that the wind was too strong for Romi/George to operate in, but just within limits for us to operate in. I saw trees in the distance being affected by winds in a way that I didn’t like and gave that unease in my stomach. However the forecast wasn’t expecting to get any better and it was the “well we are here now” decision that won.  We launched. Thankfully the QuestUAV is such a tough beast because it got pounded and thrown around in the turbulence and winds like nothing on earth.  Jon and I thought at one stage the wings would break off the turbulence was so strong. It was almost impossible at times for us to operate in manual mode and assisted mode was essential to provide the rapid responses required for stable flight. Unbelievably the flight returned a load of good photographs and a test DTM was produced from the hundred or so images taken.

A typical single image.  Lots of trees!

Day two brought the opposite weather – calm winds brought about through heavy rain. A publicity shoot with local news and papers was forecast around lunchtime.  The interviews brought a lot of interest in the new technology and Romi's intentions for them, and our fame was soon to hit the front pages of the local newspapers, despite not a single take off or landing happening. It also gave us time to try out a new concept in turbulence assessment.  The recovery poles became very tall windsocks. This allowed a much clearer assessment of how the wind was moving and twisting through the clearing. Even in the lights winds the tapes would be completely in opposition to each other showing strong windshear and rotors.  

Windpoles showing varying winds,
even in light winds

With the press through and a clearance coming we finally got onto proper training with Romi and George practising launches, take offs, autonomous flight and landings. The site was a former landfill that had been compacted to a hard, wet surface. So landings were muddy and hard too and the UAV got covered in a layer of wet mud and grit, but the new crew got in sufficient practice to be able to autonomously take off and land the QuestUAV.  It would have been good to spend more time on the hands-on flying, but the conditions and circumstances were all pretty much working against us. The concept we have of the QuestUAV being akin to a Landrover rather than a Lamborghini held true. Any damage from the day was repairable with the minimum of tools and experience. Tape and superglue repairs ended up with a UAV even stronger than it arrived and ready for more action.

Romi launching

We all learnt so much on this trip that would take pages to write about. The two and a half days packed in a huge amount and there was an agreement with all of us on a number of things. Firstly, the turbulence, wind and weather conditions were something we had no control over but choosing a better flying site would have had much better results.  Secondly, our recurring theme of “training needs to take more than two days, and it needs to be done on our own turf”, held true.  This was our first foreign training and a lack of familiar training facilities made quality training hard to achieve. And thirdly …  and most importantly, the QuestUAV is incredibly tough, easily repairable, and produces excellent imagery under the harshest of conditions.

We look forward to Romi and George's results in the weeks ahead...

And from Romi...

"I read your blog - it's a nice description of your experiences here and lessons learnt! 


Being a QuestUAV customer means that you don't only get the ordinary, correct and smooth treatment you could expect from a seller, but you also get a feeling of being a co-developer of a novel technology in a team of enthusiastic people.


We had two successful flights in Västankvarn on Tuesday with very good landings. We flew at 80 meters and the imagery was good. Next time we will fly at 110m for better mosaics."

           

Mars, The Atacama Desert and Digital Terrain Models

As space missions become more enterprising, the technology driving the systems becomes more demanding. The next round of planetary Lunar and Mars rovers must be able to survey more, quicker, and with little input from their commanders on earth. In effect they need to become autonomous; thinking for themsleves, driving without human command, sensing the landscape, assessing dangers and navigating safely in order to uncover the fundamental questions they have been sent to answer.


This technical wizardry cant be be left to just happen on an alien planets surface and has to be thorougly tried and tested long before launch. Ruatherford Appleton Labs in Harwell have been developing just this technology driven through a multinational project funded primarily by NASA but with a significant input bfrom the European Space Agency (ESA) and companies such as British Aeropsace Engineering (BAE) and Mcdonnel Douglas. Their testing includes settings tasks on real terrestial surfaces that simulate the martian landscape and this is where QuestUAV systems have been empolyed. For every task, a detailed three dimensional map, a Digital Terrain Model (DTM) is required of the lanscape.

The partnership for us started after staff at RAL systematicallly looked at every UAV system available on the market, in both the rotary and fixed wing classes. After a detailed analysis they chose to ask QuestUAV to provide a sample DTM of one of theri test sites. Despite the DTM being conducted in high wind situations (est 60 mph gusts) the resuts were successful and and DTM was produced. RAL were very happy with this - their very first successful DTM -and so the relationship developed between RAL and QuestUAV. Training took place over Jan, Feb and March and in May they departed for the pre test sites in the Attacama desert where NASA had set the gruelling tasks that the navigation systems would have to prove themselves on.

The Atacama desert is a hostile place; high, dry, rocky and gritty. Landings are harsh and the air is thin. At almost 10000 ft with little graphical information to bind overlapping images, it provides a real challenge for an unmanned aircraft to operate successfully. However the crew, Prof Brian Maddison, Dr Aron Kisdi and Dr Wayne Tubby performed an excellent job, learnt quickly and brought images back that can only be described as stunning. There were issues though. The harsh surface and thin air made for rough landings and the foam nose of the aircraft became increasingly pitted. Though the nose is designed to be sacrificial the concern was that the continuous impacts with rocks would eventually break something off. In hindsight it was just fine, but subsequent modifications added a stronger "Skid Pan" to the base of the nose that worked well. Also the mixture of sharp and grity sand got everywhere. Each landing threw up a mini duststorm that got into anything that was exposed (both internally and externally). Landings also took a toll on propellors. Propellors are changed in a few seconds so its not a problem, as long as it is noticed. On the last flight damage probably wasn't picked up and when the rate of climb was significantly reduced after takeoff and the UAV seemed to be performing poorly, the pilot elected to abort the flight and conduct an emergency landing. The UAV landed fine, though probably fast and directly into a rock field, suffering mild damage to the body. Wisely the crew chose to cease flying (though the damage could have been reapired in field) and with suitable testing and imaging complete in this phase of the project the crew decided to conduct final repairs back in the UK.

QuestUAV got the aircraft body back for repairs a few days later, tidied it up, beefed up the nose and sent it back ready for round two of testing back in the Atacama desert in early June.

Over a three week period it performed faultlessly on all twelve flights, even manging days with high winds (over 40 mph). The spare UAV sent out to accompany it was never needed. The cameras got very dusty, but performed flawlessly. The motor had the harshest of lives but never missed a heartbeat. The UAV got pumellled and even accidentally flown into the ground at cruising speed, but flew again without a hiccup. (The crew, looking down on the UAV from what they thought was high ground, mistook sloping ground for flat gound and impacted the UAV on the upslope of the desert floor). Damage? A cracked £6 propellor.

It was all the news that we hoped for but didn't dare to ask ... that the QuestUAV 200 had not only survived, but passed wiyh flying colours. We even got the message that Gianfranco, the head of ESA Robotics, was pretty impressed with the UAV ... he was convinced it would not manage the harsh tasks chosen for it. But it did.

RAL will be composing the DTM's over the next months and the results will eventually form a research paper earlt next year. We are now working with RAL on a project to integrate the very capable (and very expensive) Tetracam Mini MCA into a QuestUAV 300 for a number of spectrometry tasks set aside for it in the coming months.

2 – 5 June 2012 Training with Hannes and Hans from Austria

 Image taken by Grid-IT 1 week after UAS delivery

We have been discussing high accuracy surveying through the use of UAVs for a few months now with a company in Innsbruck named GRID-IT. The company is a reseller of ERDAS, has won some interesting contracts in the Alpine areas and wanted to expand their services into the new market of low level aerial survey using small UAVs. GRID-IT aren’t without UAV experience and have been experimenting with various own-builds over the past years, including the Bormatec Maja.  However they found that the route to a successful survey using a sUAV isn’t an easy one, takes up a lot of time and money and often results in poor data.  Hence they started to make enquiries about turnkey (complete systems such as Mavinchi, Cropcam, Sensefly, SmartPlanes and QuestUAV).  They decided that QuestUAV was the right one for them and are now, at the time of writing this (one week after their training in UK), successfully flying their QuestUAV 200 system in the Alpine foothills (see image above).

After a series of emails and Skype calls throughout May with Hannes (Director of GRID-IT) both companies felt comfortable to proceed with the sale and training.  Hannes in accompaniment with an associate, Hans, (yes remembering names on this particular training was very easy!!!) came over in early June.  They found some incredibly cheap flights to Edinburgh and travelled down on the train to Alnmouth (a good train booking website is here), where we picked them up, took them to the workshop for a welcome and discussions and then took them in the evening to their hotel at the Schooner Inn in Alnmouth.

Hannes wanted to take his UAS system back with him so we had prepared his entire system beforehand and flew it with him on training whilst he was here. The system was all designed to pack in large international travel tough/carry box.   Hannes and Hans effectively had three days in and around the QuestUAV workshop where we could transfer information, iron out problems and do flights.  We found GRID-IT staff to be exceptionally skilled at their work with a very good understanding of IT systems and architecture, brought about through many years in the survey industry, and they even helped in an information exchange with us over NDVI issues.  Thanks Hannes and Hans.

One of our big problems was sorting out international differences in the preparation of laptops and lat-long systems etc.  This all had to be done before we could fly. IT problems take an unknown amount of time to solve and for each hour taken up there is an hour less for training. We have since decided to promote the purchase of a laptop as part of the equipment list.  The laptop will by setup by us in the workshop prior to manufacturer training with all software installed, working and talking to the aircraft.

Hans preparing for launch. Red tips chosen for
greater snow visibility.

Interestingly we had negotiated a process of demo first then decision to buy (or not). Always before a big sale (more with commercial companies rather than universities) there are concerns on both the part of the buyer and the manufacturer about genuiness of interest.  The buyer wants to know that they aren’t going to get a system that might have been advertised well but doesn’t actually perform, and the manufacturers are always concerned about companies or people wanting to find out about QuestUAV, general sUAV manufacturing know-how and secrets as quickly and easily as possible for their own ends (and, yes, we’ve been in that situation a few times!).  We did the normal tip-toeing around each other for a while, trying to protect our individual interests, but at the end of the day it simply wasn’t needed. I wish there was a better way to get past this.

Though the weather prevented flying on the first day, we all felt comfortable enough to proceed with training and forego the demo.  Hannes felt that he had seen enough of the system and its performance to make the commitment to come over to the UK in the first place so the system was pretty much sold anyway before he arrived, and only something fairly major would make a change.  So the loss of the demo wasn't a great loss. Interestingly I don’t think we ever had a discussion as to whether it was bought or not – it just seemed accepted as time went on.

Our data loggers have only recently been released by SkyCircuits and still had minor bugs in their operation during training.  The primary issue is finding SD cards that work with the system. Most SD cards are designed to capture large files (typically large .jpg files for photos) rather than the tiny files that are written in data logging. We were right in the middle of another  “yes it’s working – oh no it’s not” during training that I find annoyingly unprofessional, but it’s the nature of the beast in something that is novel and cutting edge.  Thankfully Hannes was patient and helpful as we worked through this.  We now have the right SD cards and software that works, and of crucial importance, the written procedures that accompany success with the equipment. However, the holdup during training added unnecessary stresses because logger data was required to match imagery position and time for accurate Photoscan and ERDAS processing.

Power off autonomous landing

The weather soon improved though and the normal training with QuestUAV phases of design, mission preparation, flight checklists, deployment, launch, flights and landing all went well. With the initial bugs ironed out we had a good time in the sunshine and in our training field, watching the aircraft fly and do it’s thing whilst we relaxed and watched.

When most of the flying training was complete Hannes displayed his excellent flying skills whilst flying under Pilot control (full pilot control but with stability assistance). He landed the UAV just a few metres from the crew. So the challenge was set! Hans then took off to do better, but ended up providing an added dimension to training when he was flying it on his approach to land. The flightpath was slightly wide and the UAV suddenly disappeared with a sound of breaking branches as it hit the top section of a tree. It’s forgivable because we each thought the UAV was on a good flightpath to land but the perspective of this tree against it’s background put it closer than expected.  No panic… we were prepared.  Hans was a bit embarrassed but it gave us the chance to set in motion something we have never done before with QuestUAV aircraft – a full UAV recovery from a tree.  The recovery went flawlessly with the Aeroloops (see pics) and the aircraft was flying five minutes later after a quick checkover.  There was no damage at all, structural or cosmetic, apart from a few twigs stuck in the foam and no repairs had to be made.  The QuestUAV is designed to be pretty tough!

Recovery with two poles

Safe and sound, completely undamaged.
Wings and Pitot tube are designed to separate.

Hans launching again 5 minutes after tree recovery

The logger was working well by this stage and even showed the flightpath into the tree, the recovery to ground level and the walk back to the flying site.  Cool!

In all it was great working with Hannes and Hans.  They are a great team, professional, patient and skilled.  They also came knowing something that is invaluable, and it goes something like this: no matter what any sUAV advertiser promotes, it is never the case that any sUAV is autonomous and “does it all the work by itself”.  Hannes came knowing that and willing to learn the knowledge and systems that go alongside the system to make it as safe and productive as possible and to protect the investment he has made both for his productivity and his company integrity. We reckon that if a person, company or establishment already has RC or UAV experience then a minimum period of time from first call to full operation is expected to be at least two months.  For a company, person or establishment with no RC or UAV experience this needs to be extended to six months. This obviously isn’t full time but includes the gentle shift in skills and awareness gained by reading, training and discussion in order to display the required level of safety and competence to command an unmanned aircraft.

Happy guys at the end of the day. Look closely
and you might sees some tiny scratches from the tree.

On the Monday afternoon we packed the kit into the large explorer box. Unnecessary packaging had to be removed but, apart from batteries (which went in Hans backpack) it all fitted in.  Not having scales we estimated 25kg in prep for the check-in at the airport.  We got it wrong.  It was actually 35kg! It still went but it’s something we are prepared for now with scales to measure weights prior to departure.

Hannes is now home with his aircraft, flying and operating in his local area - in the process of becoming fully operational. We wish him the greatest success and look forward to the time when we expect to do more work together.


And from Hannes about two weeks after his training and many successful flights


GRID-IT is a small company located in Austria (Innsbruck), specialised in Geoinformatics and Remote Sensing. Since we are the Austrian reseller for ERDAS software, and since we did many remote sensing projects, many of them related to natural disasters and crisis management, we thought we could start to create our own remote sensing data using UAV's. Therefore, our focus is on monitoring of natural hazards in an alpine environment, which leads to some special requirements on autopilot navigation.

First we though, we could compile a system of our own, did some tests with the MAJA from Bormatec (a very nice UAV, but not quite the right tool for our applications in remote alpine areas). We soon discovered that finding the right components, which would smoothly work together was a very time consuming challenge.

After that, we did quite some research on different UAV providers, some of them in Germany, Sweden, Belgium, Poland, United States, even Pakistan (not seriously though). Finally, we found QuestUAV, got in touch with Nigel, had some very good discussions and decided to order a Quest200 together with a training in Amble. The reason for this decision was a common understanding of the requirements, the concept of a gimballed camera, which we were having in mind as well, and of course the price.

During the three-days training we found that the QuestUAV team are very nice people, excellent to work with and very professional. Since we want to carry out photogrammetric analysis of the imagery, the on-board logger is an essential part for our needs. We had to solve some problems, but now the logger and the whole system is working perfectly well.

What I really like about the QuestUAV systems is the launch concept. It is so easy and safe. And I think, the SkyCircuit autopilot is a very nice and powerful piece of technology - I think there is so much more potential in that component. In an alpine environment with narrow valleys, it is important to know how the autopilot is setting the aircraft's route. I hope, we'll get some more detailled documentation on the autopilot, soon.

So, to finalise this short introduction, we are really happy with our Quest200 system. We are currently running test with different flight plans to come up with a strategy on flight plan optimisation. We already wrote our own flight plan software, which allows us to load and visualise both the SPX flight plan and the logger data. We just recently added a module to georeference the images using the timestamp as link to the logger data. With that, we are ready to become operational, with a fully tested workflow.

Thanks, Nigel, for the good cooperation.

Hannes

29 May 2012 - Receipt of Fox Talbot Award

Now this was a stunning surprise!  Last year I had submitted a portfolio of work the British Institute of Profession Photographers (BIPP).  Bearing in mind that my workhorse is just a compact camera (albeit a very nice one) and not a brace of fancy DSLR’s with extensive studio equipment to back it up, I hoped simply to be accepted into the institute at the first level of Licentiate.  However at assessment I seemingly skipped past that, past Associate level and straight into the top level as a Fellow of the Institute.  Hearing that somewhat blew me away, especially as it was a unanimous decision by the board, but that was only the start.  (My head took months to shrink to its normal size).  Unbeknown to me they had gone on to award something outstanding beyond that, which was their top photographic award that isn’t even awarded annually, only when there is an outstanding achievement that deserves special recognition from the Photographic community.

I can’t really say much more – I am still surprised to wake up each morning and remember that this happened to me.  And of course I’m pretty chuffed too!

22 – 23 May 2012 Training with National Centre for Atmospheric Science NCAS

QuestUAV provided UAV flight experience and a lecture last year (2011) at the NCAS summer school in Arran under atrocious wind conditions.  Despite this they liked the system, liked the way QuestUAV worked and found the budget to purchase one for atmospheric research earlier this year. 

Barbara, Brian and James arrived from Leeds looking remarkably refreshed after a four-hour drive through heavy traffic. (We actually discovered later that they had stayed overnight in a B&B for a much more relaxing regime away from work!).  We chose to conduct the training based at our workshops this time following the feedback from Newcastle’s training. (It is a little more cramped but all the equipment, data and processes are there.)  James and Brian had been practicing for some months on the QuestUAV simulator in their office in Leeds and James had also bought a small RC trainer to help with his learning process. Brian was an old hand at radio control with previous experience in years gone by (we wont say how far though). 

The two days were a different format to the Newcastle training with a flying demo right at the start of the first day. The demo led into a very successful two days of training with the crew and an exceptionally competent Barbara emerging with brilliant Commander skills.  Barbara however was adamant that she should never, ever touch a live flight control which was fine because the rest of the crew was happy to step in and do the launching and flying. Sunshine, and light to medium winds dominated the two days.

As always the QuestUAV crew learnt also from the training session. Our particular lesson on this session was actually a relearnt lesson along the lines of “never fly with equipment that hasn’t been carefully tested before.” In this case we had a snazzy new, tiny locator beacon that had flown without any problems before, installed in the wingtips.  However on this day because we didn’t know until the last moment that NCAS wanted the equipment we popped it in the centre of the aircraft right by our twin receivers.  In that location, unbeknown to us, it caused a momentary interference that made the aircraft enter a failsafe in flight and bring the aircraft down a couple of fields away.  Ironically it was the locator that found the aircraft but the whole process gave us an unexpected opportunity to examine and demonstrate the safety designs, the robustness and flight systems employed with QuestUAV aircraft.   (The locator now sits safely in the end of each wingtip on all new aircraft where is has been extensively tested and is the joy of any UAV “retriever” who has to find the aircraft after landing in long grass!)

Barbara Brooks “Thanks for hosting us up in Amble last week. I was really impressed with how everything performed. I've been very impressed with service I have received from QuestUAV. Working with the scientific research community is not the easiest of things for a small company to do but your friendly and flexible approach has resulted in us developing a platform that best suits our needs. The training has been excellent with great patience and clarity of instruction given and I am looking forward to many happy hours of flying and future collaboration. 

Cheers team QuestUAV

14 – 15 May 2012 Training With Newcastle University (Part 1)

Rachel, Pauline, Martin and Gareth arrived with us for our very first official QuestUAV training session at our training centre in Northumberland.  Though we had completed training at client’s locations in earlier months, this was our first using local facilities. We chose a large conference room in the Amble Development Trust building for the first day of ground school, followed by training in our workshop and local flying site for day two.  Winds were high so there wasn’t going to be any flying on day one, but there were lots to learn anyway. We all packed into our favorite lunch spot at Jaspers where the flying discussions continued over lattes, hot chocolate and penne pasta. The whole day was centred on legal requirements, an introduction to equipment and a work through of the all–important flying checklists.

Day two was held at the unit, but winds were still too high to start live flying (est 35 mph plus at operating height).  The Newcastle crew had been practicing flying onscreen in the previous months with a QuestUAV simulator (Phoenix based) and one of their Spektrum DX8 transmitters but were otherwise completely new to flying apart from one member who had a little experience.  The day was still completely packed though with the equipment preparation work streams and Pilot/Commander routines.  There was still lots to learn about the UAV software and discussions around image processing, particularly with infrared and NDVI.  The Newcastle team actually left wondering whether there would have been any time for flying anyway because so much was packed into two days.  The live flying training was planned for a few weeks later when everyone else was available again.

As our first full-on in-house training session we learnt as a company just how much there is to put across to a fledgling flying team, with little RC flying experience, in terms of learning.  The major focus for us has always been on flight safety.  Capturing and understanding the safety elements from each process takes time and care though and confirmation that safety messages have sunk in is important. Our earlier training with RAL in Harwell on the Mars Rover Project had taught us about the clear need for Commander and Pilot roles, the importance of accurate checklists and the impact of human factors in maintaining the primary goal of flight safety above everything else.  The lesson is how long this can take to implement and to ensure that that the concept of the clients “in-house training schedule” continues long after they have left us.

Rachel Gaulton– “Thanks for 2 days of excellent training!”

Martin Robertson– “Thanks for the training over the last couple of days and for making it fun and interesting. The training was one of the most thorough I have attended with instructions to cover, in detail, the planning of the mission, checks on equipment and hands on experience."

The guys from Newcastle went on to receive flying training at a time when the weather was better, in late June. However the onlookers were more than just interested, it only for a while. Message to self... cows generally don't seem to understand the Rules of the Air.