Gene Robinson the UAV Search and Rescue Legend

Gene Robinson the UAV Search and Rescue Legend

highered-4-13-gene-robinson-wi_10885209

A couple of weeks ago I spent a day at the sUSBExpo in San Francisco, organized by Patrick Egan (Thanks Patrick great job.) Quite a few people have covered the Expo already, so I’m going to focus on it from a different angle. I had the privilege of attending the Search and Rescue Workshop presented by Gene Robinson. This was a sold out workshop, but even so Patrick managed to get in most people who wanted to attend.

Gene Robinson has been a pioneer in the use of UAV for search and rescue, starting more than a decade ago. This is truly a humanitarian focus and Gene runs all his SAR work through the non-profit RP Flight Service (http://www.rpflightsystems.com/MainPage.html) /RP Search Services (http://www.rpsearchservices.org/) and working with Texas Equusearch (http://texasequusearch.org/). He was recently awarded the Recipient of the Spectra Humanitarian UA Award. If any one person is an advocate of the peaceful community based role for UAV’s, it’s Gene.

Gene uses his own fixed-wing UAV called the Spectra, using a Robota Goose Autopilot (http://www.robota.us/Autopilot/b/8829483011?ie=UTF8&title=Autopilot). Normally he flies with a 12Mps digital camera. His philosophy is simple, keep it simple, always have a back-up, plan and practice. How do you see this in his operations? Well the Spectra is a flying wing, with only 3 moving parts, two elevon servos and an electric motor. This is a design and requirement based on experience of what works and Gene has lots of it. For example, will your autopilot work below freezing, will it work at 120’F in the desert, will you LiPo batteries have full capacity at launch when it’s cold, and should you change your propellers when flying at altitude? Gene explained all this and more, with a very straight forward and simple way, but drove the message home. Here’s another example, you’re on a search in a desert, are you going to be any good to the Incident Commander if you’ve got sunstroke because you haven’t got any cover?

SONY DSC

And here is another important point he drove home, be professional. Do your emergency management training; learn the Search and Rescue team’s methodologies and protocols. Know what radios they use, how to use them etc. And practice, practice, and practice you’re flying. As Gene put it, “You only get one chance with an Incident Commander, if you throw your UAV and it bites the dirt, you’re done.”

It was absorbing to hear of the fire monitoring, and search and rescue missions he has been involved in and made a difference. The number of searches he mentioned where the ground searches had been through an area with a tooth comb, and not find the missing person, only for Gene to find them later. It’s not that the ground search teams are doing it wrong, it’s a resource issue. To find a missing person, you need to be able to see the feet of the searcher next to you, if you don’t your too far apart, and can walk straight past the person you are trying to find. So if someone is lost in Yosemite, how many ground searchers would you need?!

Another of Genes take away point, gather as much information as possible, what were they wearing etc. Here’s another fact, about 80% of missing people are wearing Jean’s. Blue is not a significant natural color in nature. As such Gene now uses image processing software that scours the images for specific colors like blue and other discrepancies, and can identify possible sites for further investigation. On that note, Gene needs help to improve his image processing software, if there are any Python programmers out there, drop Gene an email, http://www.rpsearchservices.org/contact-form.html

SAR

The sad part is, is that Gene now has issues with the FAA. Although RP Search Services is a non-profit and Gene pays for everything himself (that’s about $4000 of equipment he throws in the sky knowing he may never get it back, people come before machines and he’s willing to lose it to save lives) they consider him flying search and rescue as a commercial operation, which according to the FAA is illegal under there regulations (that is a whole other story.) As such Gene has retained Brendan Schulman, a pioneering UAV Attorney to take the FAA to court to resolve the issue. The heartache of the matter is that until this is resolved, Gene is not flying SAR within the USA. What is worse, a lot of missing children are small children, can you imagine having a resource such as Gene on hand to find your lost child, but the FAA forbids it?

The next Saturday morning we went on a birthday party with my young daughter in a forest. The first thing I did was take a picture of my daughter. My wife having known that I’d been to Gene’s workshop said “You didn’t do that as a birthday party picture did you?” We also taught her a new rule, if you cannot see our feet your too far away. She’s still checking our feet to this day.

Thanks Gene, you are a real gentleman and humanitarian. Thanks for the workshop, the knowledge and the peace you have brought to families of missing people.

Go watch the Movie from Maha Calderon about Texas Equusearch and Gene Robinson, http://www.civiliandronesmovie.com/home.html

And if you can support Gene Robinson please support with a donation http://www.rpsearchservices.org/support.html

You can also buy his book, “First to Deploy” http://www.rpsearchservices.org/learn/index.html

firsttodeploy

Regards

@theUAVguy

http://www.kextrel.com

iain.butler@kextrel.com

Which Aerial Platform to use for Precision Agriculture?

Which Aerial Platform to use for Precision Agriculture?

eBee Ag Drone With Case

In my previous post I discussed how UAV’s can help with Precision Agriculture. In this post I’ll discuss the platforms that can be used at this time, ranging from off the shelf complete packages, to do-it-yourself airframes.

So one thing to note is that no single platform is suitable for all applications. The type of platform is dependent on a number of factors such as land topography, area of coverage, environmental conditions such as wind, direction and cloud ceilings. For example a small vineyard with elevation changes in a mountainous region is going to require a different platform, than a large artichoke field which is flat and near a windy coastline. Both these scenarios require different platforms.

Multi-Rotors v Fixed-Wing

In general multirotors are great for small aerial PA projects, 0.2 sq mile. They have endurance of approximately 20 minutes allowing coverage of smaller areas, have good wind resistance to approximately 20 to 25mph dependent on model, can cope with rapidly changing land elevations, and can fly lower for higher resolution. Also the take-off and landing area requirements are significantly smaller. The downside as mentioned is their limited range, which means that for large areas multiple flights are necessary with associated image stitching and complication.

For medium to large areas, 4 sq miles, fixed wing platforms dominate. The reason is endurance, fixed-wings as the name suggest generate lift off their fixed-wings, so as long as they have enough forward motion to generate lift they stay airborne. In a multirotor all the lift comes from the motors and propellers. If both a multi-rotor and a fixed-wing use the same battery, the more efficient fixed wing will be able to fly much longer. Normally up to 3 times as long.

So what are desired requirements for an aerial platform?

Safety and Redundancy

Well for anything aerial safety is key, and the lesson learned from manned and RCMA (RC model aircraft), is that redundancy is key. On a multirotor, the lower the number of rotors you have, the more catastrophic the event if a motor or propeller fails. Losing a propeller or motor on a tricopter (3) or quad copter (4) will normally always result in a crash. Using Y6 (6), and hexacopter (6) formats allows for a motor/propeller issue, but for the platform to be landed. The same is true of X8 (8) and octocopters (8).

Another area of redundancy is servicing and repair. Multirotors as the name suggests are made from a number of motors which operate in tandem to control the flight of the UAV. Normally the motors are the only moving parts on a multirotor except for maybe a gimbal. As such there are less moving parts to fail. The safety concern with multirotors, is that they have no lift mechanism other than the motors and propellers. The problem here is that if the motors stop, they just fall out of the sky. On a fixed-wing, if the motor/s stops, the wing still generates lift and you can glide to a landing.

Now fixed-wing moving part failures tend to be catastrophic in failure, unless there are redundant actuators. On a generic fixed-wing airframe you have actuators that control roll, pitch, yaw and throttle. Normally that translates into two independent airelons for roll, two elevators which are normally joined for pitch and a single rudder for yaw. That equates to 4 actuators/servos and a throttle control. To increase safety this can reduce to two airfoil actuators and a throttle control. Here the pitch and roll are combined in what is termed an elevon system, where the left and right ailerons are moved in combination. Both airelons up induces a climb, both aileron down is a dive, left aileron down, right up is a right roll. Now if the left aileron is moved up and the right aileron is left neutral, the plane with roll left and pitches up. This type of control is normally associated with airframes that are called flying wings. To further increase safety, multiple actuators can be also placed on a single airfoil, as such if one fails, the other takes over as a backup.

So what we are doing here is reducing the number of moving parts that can fail, and where we cannot minimize anymore, then add redundancy in the form of parallel actuators or motors.

Another key area is auto-pilots, just like you Windows PC rebooting and installing updates without warning, you need to ensure you have a safety-critical autopilot. If you autopilot reboots in mid-air, on your multirotor the motors will stop and it will just fall out of the air, on a fixed wing it may glide off into the sunset. Other safety issues are flyways, this can be due to poor autopilot firmware, GPS glitches, power supply spikes etc. Ensure you are using certified autopilots with stable code, good supply distribution, and high quality GPS and even redundant GPS units. In some cases redundant telemetry, RC receivers and autopilots are also employed.

A final safety feature is geo-fencing. Here an invisible fence is placed around the flight path by the mission planning software. If a geo-fence is breached, such as due to a failure of GPS loss of lock, GPS glitching, lost telemetry link etc., that the platform returns to the takeoff site and either loiters allowing a manual landing or auto-lands.

Safety is key, this industries growth will be defined by how safe it is.

Spares

It’s great buying that cheap UAV or airframe from overseas, but what happens when you have a rough landing? Can you get spares readily? This is an important consideration, if you ding a wing, you don’t want to wait 2 weeks for international shipping. Makes sure you either have an inventory of spares, or have a nearby dealer who has a good supply.

Stability and Image Quality

It’s true that the fun focus is on the flying, but flying is only 20% of the overall work. The other 80% is flight preparation, post flight and then data retrieval and processing of the images. However you can have poor images due to poor camera choice, no gimbal or a badly stabilized camera gimbal, badly designed autopilot or inefficient stability algorithm, a badly setup multirotor with bad gain settings, poorly balanced propellers etc. All these issues can lead to inferior image quality. And as with most processes, if you put bad images in, you are going to get poor data out. As such although flight time is only 20% of the process, it is key to getting quality data.

So you need to look for a good camera which has approx. 12Mega pixel, with the best possible dynamic range, with integrated image stabilization. The cameras are normally modified with new filters for NIR etc. I’ll discuss cameras and modifications in the next blog. At present the go to camera is a Canon S110 with NIR using filters from people like Event 38. Other cameras such as the Canon SX260 and S100 are also used. You notice the predominance of Canon cameras, this is because they are easily modified for filters and updated with control software called CHDK.

For fixed wing operations having the camera hard mounted to the airframe is the norm. Normally due to the forward speed of a fixed-wing platform and the autopilot, it is normally flying wings level during image capture. On multirotors, due to wind effects, direction of travel, the airframe can be leaning into the wind/direction of travel, and as such a 2 axis gimbal is normally used. These gimbals can be servo or brushless, as this is still pictures a quality servo gimbal can work as well as a brushless gimbal, without the associated cost. The gimbal levels the camera and removes any autopilot sudden corrections.

Autopilots are very important for image quality, a poor autopilot can wander off path giving incorrect image overlaps, jerkiness in control response and just about spoil your day.

So based on the above information, I know there was a lot compacted into a small space, here are some picks for Agriculture UAV’s with some pros and cons:

Fixed-Wing Complete (with associated GCS)

senseFly eBee Ag https://www.sensefly.com/drones/ebeeag.html

Pro – Well integrated package using S110 cameras and PIX4D software. Very small.

Con – Integration costs approx. $25,000 including PIX4D software

Event 38 E384 http://www.event38.com/ProductDetails.asp?ProductCode=E384

Pro – Well sorted package based on 3D Robotics Aero, $2,399. $3700 extra for AgiSoft software

Con – Highwing design, more susceptible to damage, more moving parts

3D Robotics Aero https://store.3drobotics.com/products/3DR-Aero

Pro – Designed by 3DR around Pixhawk autopilot $1350, supported by Pix4D +$3000 annual

Con – Highwing, more susceptible to damage, more moving parts, no defined camera area

Ritewing Zephyr II with Ruby Autopilot http://www.ritewingrc.com/Zephyr_II_ARF.html

Pro – Proven design

Con – Ruby autopilot does not support waypoint navigation at this time

Fixed-Wing Airfarmes

Range Video RVJet http://www.rangevideo.com/en/18-rvjet

Pro – Very stable design for autonomous flight

Con – Long wing easy to damage, no defined horizontal camera area

Skywalker X8 as used by Robo Flight RF 1 http://www.roboflight.com/products/

Pro – Large stable flying wing, able to carry large payloads over long distances, has camera area

Con – Parts from overseas, very large

Phantom V2 Flying Wing http://pixhawk.org/platforms/planes/phantom_fpv_flying_wing

Pro – Small airframe, flying wing that can be taken apart, very stable, has camera area

Con – Parts from overseas

Fixed-Wing Autopilots (with associated Ground Control System and Mission Planning Software)

3D Robotics APM2.6 https://store.3drobotics.com/products/apm-2-6-kit-1

Pro – Established fixed-wing autopilot

Con – Running out of processing power and memory, surpassed by 3DR Pixhawk

3D Robotics Pixhawk http://3drobotics.com/pixhawk/

Pro – Carries on from where APM2.6 left-off

Con – Newer architecture

Roberta Goose http://www.robota.us/Goose/dp/B00EGT0ZCY

Pro – Very reliable, used by Gene Robinson for Search and Rescue

Con – Expensive $3995

Airware http://www.airware.com/

Pro – Modular autopilots for who range of UAV’s

Con – Very Expensive

Complete Multirotor for Scouting (with associated GCS)

3D Robotics Iris https://store.3drobotics.com/products/IRIS

Pro – Integrated proven design with support

Con – In-demand

DJI Phantom 2 Vision+ http://www.dji.com/product/phantom-2-vision-plus

Pro – Integrated proven design with support

Con – In-demand

Complete Multirotor with Lift capability for NVDI Cameras (with associated GCS)

3D Robotics X8 https://store.3drobotics.com/products/3dr-rtf-x8-2014

Pro – Proven Pixhawk X8 copter with PA mission planning and Pix4D software support, $1350

Con – Needs more hands on experience required with mission planning

3D Robotics Y6 https://store.3drobotics.com/products/3dr-rtf-y6-2014

Pro – Proven Pixhawk Y6 copter with PA mission planning and Pix4D software support, $1000

Con – Needs more hands on experience required with mission planning

Aerialtronics Zenith http://www.aerialtronics.com/products

Pro – Very professional high quality design

Con – Expensive at $25,000

DJI S800 http://www.dji.com/product/spreading-wings-s800-evo

Pro – Very professional high quality design

Con – Expensive at $10,000

DJI S1000 http://www.dji.com/product/spreading-wings-s1000

Pro – Very professional high quality design

Con – Expensive at $25,000

Multirotor Flight Controllers (with associated Ground Control and Mission Planning Software)

DJI A2 http://www.dji.com/product/a2

Pro – Established highend multirotor autopilot

Con -Cost

DJI Wookong-M http://www.dji.com/product/wookong-m

Pro – Established multirotor autopilot

Con -Cost

DJI NAZA V2 http://www.dji.com/product/naza-m-v2

Pro – Established entry level multirotor autopilot

Con – Recorded flyaways due to GPS glitching

3D Robotics APM2.6 https://store.3drobotics.com/products/apm-2-6-kit-1

Pro – Established multirotor autopilot

Con – Running out of processing power and memory, surpassed by 3DR Pixhawk

3D Robotics Pixhawk http://3drobotics.com/pixhawk/

Pro – Carries on from where APM2.6 left-off

Con – Newer architecture

OpenPilot Revolution http://www.openpilot.org/products/openpilot-Revolution-platform/

Pro – Open source firmware, proven

Con – Not as widespread as other autopilots at this time

Airware http://www.airware.com/

Pro – Modular autopilots for who range of UAV’s

Con – Very Expensive

So the question is do you spend $25,000 on a senseFly eBee AG or do you buy a 3DR Aero for $1350? Same with the multirotors, do you spend $25,000 for an Aerialtronics Zenith or $1350 for a 3DR X8. I guess it depends on how deep your pockets are.

Thanks for reading, next posting will be about cameras for PA.

Regards

@theUAVguy

http://www.kextrel.com

iain.butler@kextrel.com