Ag Drones — Future and Present


No longer can farmers plant seeds from last years crop, pray for rain, and expect to feed the world.  Things have changed, mostly for the better.  My grandfather plowed with mules, killed weeds with a hoe, fertilized with manure, harvested by hand and produced about 60 bushels of corn per acre.  About average for the time.  Then came the age of mechanization.  Farmers now plow with tractors, plant GMO seeds, kill weeds with herbicides, fertilize with liquid nitrogen, harvest with combines and produce about 150 bushels of corn per acre; more than double the previous yield.  Informed sources estimate that farmers will have to double their yield again by 2050 to feed the world.  That requires precision farming, and precision farming requires data.  You can’t manage what you don’t understand, and you can’t understand what you can’t see.
That is why farmers need the UAVs to image their fields.  Now they use satellites and manned aircraft; neither is efficient.  Farmers need to image their fields often, and on their schedule, in order to manage fertilization,  irrigation and apply herbicides and pesticides to affected areas quickly before damage is done or the infestations spread.  Only UAVs can provide this data in a timely and cost effective manner.
Down on the Farm with Drones, the Vision.
After breakfast the future farmer strolls out to the garage, which replaces the barn of years past.  In the garage, with the combine, automated liquids dispenser, and self steering tractor are six Scout Drones on the charging bench and two larger Spray Drones on the garage floor.  The farmer ascertains that all the Scout Drone batteries have been fully charged before entering his air conditioned office.
The first order of business is to check todays weather.  The automated weather monitor shows current weather as light and variable wind and no significant cloud cover.  The computer forecasts show the same for the morning, however cloud buildup for the late afternoon with scattered evening showers.  We can always use more rain.
Next, the farmer reviews the field maps of his crops on the large touch screen computer display.  Three fields were previously showing slight heat stress and he decides to image them again in the 920 nm spectrum.  He has been having problems with weed infestation on another field, especially in the corner near the stream.  He programs the Drones accordingly and switches on the surface radar that will scan the surrounding lands for conflicting traffic.  In the event of a conflict with a manned aircraft or another Drone, the computer will deconflict the situation, and if necessary recall all his Drones back to their assigned landing pads.
At the charging bench, the farmer moves the three IR imaging equipped drones out to their assigned landing pads.  After each is preflighted and conducts a self check, they are sent on their way.  The weed searcher Drone’s imaging device has to be changed and programmed for the crop it will be scanning.  It will scan for any image not related to the programmed crop and provide density and location information on return.  When finished, he takes the Drone to its assigned landing pad, preflights it and completes the self tests before releasing it for flight.  Time for lunch.
After lunch, the farmer returns to the garage.  The four Drones are on their respective landing pads and shut down.  The farmer moves the Drones back onto the charging bench and begins the IR, and weed infestation data down load.  In the office, he analyses the data, determines one field is doing OK, but adjusts the irrigation schedule for the other two.  The weed mapping data confirms his suspicions that the infestation is migrating from the fallow land near the stream.  He identifies the weed species, selects the recommended herbicide and potency and prepares a program card for the Spray Drone.   He moves the Spray Drone to its assigned landing pad, mixes the herbicide and fills the hopper.  He then starts the engine on the drone and allows it to warm up while the Drone performs its self tests.  He then releases the Drone which goes to the designated area, identifies the individual weed plants and applies a squirt of herbicide to each plant before returning to the landing pad and shutting down.  The farmer then washes down the Spray Drone blow dries it and returns it to the garage.  In the office he  completes the computer entries to document his accomplishments for the day.  The job’s not done till the paperwork is finished. He turns off the radar and as he walks up to the house he thinks, “Maybe tomorrow I should take the ATV down by the stream to see if there is some way to stop the migration of weeds into that field.”
So ends, “Another day down on the farm with Drones.”
That was fantasy; or maybe not, in the not too distant future.  Now let’s look at the present reality.  In order to get to that future, AG Drone operators are going to have to prove they can operate safely and responsibly.  Below are some thoughts on how to accomplish that goal.
To diminish the chance of a fly away, or other un-commanded maneuvers there must be redundancy in certain systems.  The Aeroscout III can be used as an example.  It is a commercial off the shelf (COTS) air-vehicle  called the F 550, “Flamewheel Hexicopter” by DJI.  The control system avionics (FMS) are the NAZA M v2 with GPS, making it a 12 DOF system.  The HD camera video is recorded on board to an SD card, and combined with inputs from an OSD and transmitted to the ground station’s seven inch monitor to be used as a view finder by the pilot monitoring.  The radio control link is a Spektrum DX 8 system with frequency diversity capability and a satellite receiver to provide both horizontal and vertical antenna orientation.
The Aeroscout III is primarily constructed of plastic and is frangible on impact.  It is basically a small ( 550 centimeter diameter) plastic model hexacopter. Operational gross weight of the UAV is two kilos(4.4 lbs.) and max flight time is twelve minutes in this configuration.  It can image a forty acre field in ten minutes, flying at ten meters per second (22 MPH).  Optimum imaging height is between 20 and 30 meters. It has a stabile hovering  capability which allows the imaging to be completed within the boundaries of the land owners property.  Autonomous flight is not available.  Visual waypoints (up to 28) are available as an aid to remaining over the imaged field.
In the unlikely event of a radio control link failure, the Aeroscout III will go into the failsafe mode.  At loss of link, the Aeroscout III will come to a hover at its current altitude.  It will then wait a ten seconds for link re-establishment.  If that does not occur, the air-vehicle will climb in hover to 60 feet altitude or remain at its present altitude if that is higher than 60 feet.  It will then proceed directly to the previously established “home point” and descend to a landing and motor shutdown.
If while flying the Aeroscout III under visual conditions (VLOS) the aircraft orientation is lost, simply release pressure on the right stick and push the throttle to full open.  The air-vehicle will come to a hover, and climb to the preset altitude limit and remain stationary.  The pilot then selects the “home lock” mode.  Pulling the right stick (elevator) toward him (up elevator) causes the air-vehicle to fly toward the “home point” regardless of its orientation.  If visual orientation is re-established the mission may continue.
Why a hexa-copter vs. a quad-copter?  On a hexa-copter, if a motor, ESC, or propeller should fail in flight a similar procedure, as above, can recover the aircraft.  The loss of power on one arm will cause the aircraft to rotate about its center of gravity, and possibly descend.  Apply increased throttle as necessary to maintain altitude, and switch to “home lock” mode.  The aircraft will continue to rotate, however, it may be flown to a safe landing area by using the right stick to move it toward or away from the “home point”  and left and right relative to the course flown no matter what the aircraft orientation during the rotations.  The same loss of thrust on one arm of a quad-copter will cause a loss of control, and a crash.
The Aeroscout III has a video down link, even though intentional FPV  flight is never planned.  The video link is normally used by the pilot monitoring as a view finder to ensure the photo coverage is optimum for each swath.  However, if (VLOS) visual contact is lost for any reason the pilot monitoring can “talk” the pilot flying back to a point where visual contact is reacquired.  The backup plan is to select “fail safe” mode on the transmitter and wait for the air-vehicle to land at the “home point.”
Fly-aways are also prevented by limits set in the Master Controller’s FMS via a PC computer and USB cable.  The lateral limit is typically set at 1000 meters from the “home point” and the altitude limit is set at the planned height of the photo run, typically 20 to 40 meters AGL.  The air-vehicle will not fly past these limits.  The photo runs rarely exceed 402 Meters laterally from the “home point” as it is just too difficult to maintain accurate orientational awareness past that point even though the air-vehicle is still well in sight.  (Also, 402 Meters equals one side of a square 40 acre field.)
Multi-copters do not auto-rotate or glide.  Hexa copters and Octo copters can fly with a motor, ESC, or propeller failure, however, if they run out of power, they crash.
The battery voltage is overlaid on the OSD video transmitted to the ground station and monitored by the pilot monitoring.  In addition, alarm points are set in the Master Controller for low battery voltage.  Exceeding the first level causes the LED indicator on the rear of the air-vehicle to insert a red flash between the green flashes that indicate normal operation.  This is to alert the pilot to return for a landing.  The second level alert begins with a red flashing LED (no greens) followed by a slow descent, landing and motor shut down.  However, that landing may not be at the preferred location. This would be the multi-copter equivalent of an auto-rotation.
This redundancy provides a robust operational platform for aerial photography.  Simultaneous multiple failures would be required to cause a crash.  Barring pilot error, the Aeroscout III is an exceptionally safe system.
There is a saying about Airbus aircraft that the computers fly the aircraft and the pilot merely makes suggestions to the computers.  In the Aeroscout III this is very true; conventional pilot flying skills are not required.  The Aeroscout III will hover without pilot input, however, pilot control input is required for the air-vehicle to change position in the airspace.  Fully autonomous flight is not available.  While conventional stick and rudder skills are not required, following established procedures and “mode awareness” is mandatory.  This must be the emphasis of the training.
The Aeroscout III is unlike any certificated manned aircraft, and therefore no FAA category, class or type applies.  The Hexicopter has no wings or rotors and is neither airplane nor rotorcraft.  It has instead, six electric motors with fixed pitch propellers arranged symmetrically around the perimeter of the vehicle.  Three motors turn clockwise and three counterclockwise.  Control is by the Master Control FMS unit using inputs from three gyros, three accelerometers, three magnetometers, GPS position, and barometric altitude to vary the speed of each motor independently to maintain a stabile hover.  Lateral and vertical excursions from this hover are controlled by a radio control link from an operator on the ground. No manned aircraft is so designed, and no existing FAA regulations for air worthiness or pilot operating skills can apply.  This is new technology.
Pilot error will be avoided by following correct procedures, maintaining mode and situational awareness, and being familiar with the redundancy built into the system.  Crew resource management, coordination between the pilot monitoring and the pilot flying, also greatly enhances the safety of the operation. In the event the pilot flying is temporarily confused, centering all control sticks will cause the air-vehicle to come to a stationary hover at its present altitude. (All except the throttle are spring loaded to the center position.)  This allows time to determine the appropriate course of action and discuss it with the pilot monitoring before continuing.
Provided sufficient training emphasis is placed on using correct procedures, maintaining mode and situational awareness, employing crew resource management techniques and being familiar with the redundancy built into the system, I believe history will show this to be a very safe air-vehicle.
EQUIVALENT LEVEL OF SAFETY, Alternative Method Of Compliance (AMOC)
This is new technology and current FAA regulations for certification of aircraft and pilots cannot apply.  Therefore, the FAA must accept an Alternative Method Of Compliance (AMOC).
Airspace allocations are one method of promoting safety, and collecting data for future rule making.  Knowledgable sources predict that 70 percent of small UAVs will be used in agricultural operations, so that seems to be an appropriate venue to begin collecting data.  Also, rural areas are typically in uncontrolled airspace.  An acceptable airspace restriction would be to operate only in uncontrolled airspace and remain within the boundaries of the landowners property below 400 feet AGL.  Legal presidents for allowing these operations exist.  The drone is merely another farm implement, and historically most states do not require licensing of farm vehicles, or operators, as long as they are not operated on public roads.  This should also apply to UAV farm vehicles operated in the airspace belonging to the farms.  The Supreme Court agreed in its decision which states, “We have said that the airspace is a public highway. Yet it is obvious that if the landowner is to have full enjoyment of the land, he must have exclusive control of the immediate reaches of the enveloping atmosphere.  The landowner owns at least as much of the space above the ground as the can occupy or use in connection with the land.”
The ATO is not responsible for separating traffic in uncontrolled airspace.  Therefore, “see and avoid” would apply.  An aircraft is a noisy environment with limited visibility and for a single pilot involves a high work load.  This degrades the see and avoid capability of the manned aircraft.
The UAV requires two pilots, one pilot flying one pilot monitoring.  While the pilot flying must keep his attention on the UAV, the pilot monitoring has the freedom to scan the entire sky for conflicting traffic.  Since the UAV makes very little sound, both pilots will be able to hear a powered manned aircraft approaching; probably before they can see it.  The manned aircraft has the right of wayALWAYS, and the UAV must take evasive action.  Usually, the UAV will have landed before the pilot of the manned aircraft even knew one was operating in the area.  This makes the “See and Avoid” safer between a UAV and a manned aircraft than between two manned aircraft.  This definitely satisfies the equivalent level of safety.
Peeking drones are in the news lately, however this privacy issue is not germane to agricultural operations.  There are two reasons for this.  First, the agricultural UAVs are operated in rural areas and only over the land owners property.  Second, only orthogonal photography is used for agricultural imaging, meaning the camera is pointing straight down at the crop.  Not at the neighbors window.
Intentional Hacking and Signal Jamming
The radio control link for the Aeroscout III is the Spectrum, frequency hopping system.  When turned on, the Transmitter and receiver are bound by a discrete digital code, and an unused pair of frequencies are selected.  Should interference occur on one of the frequencies the radio automatically switches to an alternate frequency.  This system has been in use by recreational pilots for quite a few years, during which loss of the radio control link is extremely rare and usually caused by something other than frequency jamming.
The GPS link for the Aeroscout III is primarily a convenience item, the loss of which will not cause the UAV to fly away.  Since the primary control of the UAV is by visual line of sight (VLOS), loss of GPS signal or spoofing will be quickly noted by the pilot flying.  Simply switching from “GPS/ATTI mode” to “ATTI Mode” removes the GPS inputs to the FMS.  The FMS will still maintain level flight, and barometric altitude is not impacted by the loss of GPS.  All that is lost is wind drift correction at hover, return to home, and home lock functions.  The mission should be aborted however, and the UAV returned to a suitable area, landed, and shut down until the reason for the discrepancy can be determined and corrected.
Although aerial applicators share the same airspace as the UAVs, this is not a high risk situation.  Crop dusters typically “ferry” (travel between the point of loading and spraying) at or above 300 feet AGL.  This is to avoid being surprised by another crop duster pulling up out of a field in a “Duster Turn.”  This is well above the normal height of the imaging swath of 60 to 120 feet for the UAV.  In addition, the UAV operator SHALL descend to a hover and land to deconflict any encounter with a manned aircraft.  This is really necessary for the UAV pilot, as the wake turbulence from the larger manned aircraft flying overhead could upset the UAV with expensive results.  While the manned aircraft use “see and avoid” for traffic separation, the UAV operators use “hear, then see and avoid” to provide and even greater level of safety.
Coordination with the aerial applicators operating in the area is not only neighborly, it can improve scheduling to prevent having to sit out a planned flight while a manned crop duster works an adjacent field and overflies the boundaries during his duster turns.
This is a beginning, and who can predict all the wonderful things that Drones will accomplish in the future — as long as we act responsibly and fly safely now.
The technology is available now and the agricultural community is ready to begin using this technology.  (Some already have.)  Now, we the people are patiently waiting for the FAA to learn and understand this technology and act to implement its safe commercial use.  With the total value of our nation’s crop estimated at $140 billion per year, even a modest improvement in yield would have a substantial aggregate economic impact.
I wish the FAA would hurry up.
Comments and questions welcome.  Respond to: