Beyond Visual Line of Sight Rating (BVLOS)

[Chaosrider]

Some perspective...

Someone in the U.S. is about 5x as likely to be killed by someone driving a car as someone in Europe. Someone riding a bicycle in the U.S. is about 11x as likely to be killed as someone riding a bicycle in The Netherlands, Sweden, Germany, etc. Someone walking near a road in the U.S. is about 17x as likely to be killed as someone doing the same in Europe.

Traffic engineers in Europe are, correctly, quick to point to U.S. road designs as the primary reason for this disparity but also add that driver training along with European's higher regard for laws and law enforcement play critical roles.

While clearly defined and reasonable rules and regs are critical and I do support them, the bigger risk is likely from people who are ignorant of the rules and more so those who break so many other laws (from speeding or right-on-red without stopping to pot & prostitution) that the concept of obeying a law is watered down and they wonder why then obey a toy drone law.

Pot and prostitution are both legal in Nevada...

;-)

Something that you might want to consider is that Europeans are very substantially more risk averse than Americans, on average.

That's a choice, not a defect! A feature, not a bug!

And it's not for everybody.



TCS

 

[Ty Pilot]

There will and should be aircraft requirements to fly BVLOS, but full autonomous AI won't and shouldn't be part of the requirement.

Perhaps an enhanced RTH function that included obstacle avoidance would be useful.

TCS

What is an RTH function or obstacle avoidance if not AI? Perhaps the term 'Artificial Intelligence' suggests something more to some, but in both of those cases, the drone is using sensors and/or stored information to plot a course (or deviate from one), and then autonomously control itself to enact the solution.

BVLOS will (and does), require the UA to have the same kind of awareness and capabilities we as pilots must have within VLOS in terms of 'see and avoid' - albeit autonomously and without direct pilot control while beyond our sight.

 

[spudster]

What is an RTH function or obstacle avoidance if not AI? Perhaps the term 'Artificial Intelligence' suggests something more to some, but in both of those cases, the drone is using sensors and/or stored information to plot a course (or deviate from one), and then autonomously control itself to enact the solution.

Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts | Britannica

Artificial intelligence (AI), the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings. The term is frequently applied to the project of developing systems with the ability to reason, discover meaning, generalize, or learn from...

www.britannica.com

 

[maelstrom]

Flying a drone BVLOS is a lot like instrument flying in terms of aircraft control and navigation.

I disagree - I taught instrument flying to professional pilots for many years. A modern drone maintains attitude, heading, ground position and altitude unless you make control inputs. Real instrument flying (without using an autopilot, which is cheating) in a full sized aircraft relies on the pilot maintaining control of all of the aircraft parameters by sole reference to the instruments - a much more difficult skill.

 

[scro]

I disagree - I taught instrument flying to professional pilots for many years. A modern drone maintains attitude, heading, ground position and altitude unless you make control inputs. Real instrument flying (without using an autopilot, which is cheating) in a full sized aircraft relies on the pilot maintaining control of all of the aircraft parameters by sole reference to the instruments - a much more difficult skill.

Akin to flying a drone in "atti" mode without looking at it, nor using the video feed?

 

[Drone752H]

Aviation is based 100% on risk mitigation. There is no such thing as a 100% risk free flight.

BVLOS is one of those areas where the perception of risk is greatly overreaching in the propagation of regulation. Both 44809 and 107 require the operator to maintain the ability to see the aircraft at all time. It's a regulation that is overly biased towards safety.

My example: I routinely fly my Mavics beyond my ability to see them. And it's 100% safe. No question, or I wouldn't do it. And I do that in both controlled and uncontrolled airspace. However, in each and every instance, I have full command of the visibility of the airspace. I can see where my drone is (even if I can't see the drone), and maintain full situational awareness at all time.

BVLOS flights within reason can be 100% safe, even if they violate the letter of both 107 and/or 44809 regulations.

The issue is that some, either willingly or ignorantly, fly way beyond their ability to maintain situational awareness. That's where the issues come into play.

And you ask who are we to decide the risk? We're aviators. We use the same logic and risk mitigation process that manned aviators do. While the sheer number of risky unmanned flights severely outweigh the number of risky manned flights, the likely bad outcome is virtually nil. Numbers don't lie. And the vast majority of us (107 Pilots and 44809 flyers) have a full awareness of what the consequences of a mistake would be. Especially if that mistake involved manned aviation. To paint all 107/44809 folks by the actions of a few ignorant drone owners weakens your argument. Reason and facts are what make a viable discussion, not blanket statements.

Midair collisions between manned aviation aircraft are much more prevalent than midair collisions between manned and unmanned. It doesn't justify many reckless UAS operations (& operators), but it doesn't need to be overblown by the industry or the FAA either.

The FAA even knows that the BVLOS risk assessment in regulations needs to be taken into account for future BVLOS rules. That is why they created the BVLOS ARC this month. And I'm quite pleased that we (Drone Service Providers Alliance) are part of that rulemaking committee. Not only is our CEO on that ARC, but two of our board members are. Small scale UAS Pilots will now finally a very large say-so about changing the current overly restrictive VLOS rules. As it should have been all along. We will even be having a membership call for feedback from our members on the BVLOS white paper given to us at Wednesday's DAC meeting.

I look forward to our ability to have more open sky to legally (& safely) use.

The discussions seem to be between strict VLOS and “open the flood gates”. My original thought (posted somewhere) was just to “open the door” but just a crack. Put a reasonable cap on distance (and how do you judge distance if you can’t see it?), and altitude. The chance of encountering a manned aircraft below 500 feet is negligible, don’t fly in the path of a landing or departing aircraft - small uncontrolled field. Quite frankly, I’m not nearly as concerned about aircraft as I am about vertical obstructions and power lines. Flying the Skylane I’m “up there”. Flying the drone, I’m down low, WAY to low to encounter an aircraft (operated responsively). Drone operations in Class G airspace (for those authorized) should be the same as manned aircraft. One mile and clear of the clouds! As Martha King is so fond of saying, Class G airspace, go for it!

 

[Rich QR]

I disagree - I taught instrument flying to professional pilots for many years. A modern drone maintains attitude, heading, ground position and altitude unless you make control inputs. Real instrument flying (without using an autopilot, which is cheating) in a full sized aircraft relies on the pilot maintaining control of all of the aircraft parameters by sole reference to the instruments - a much more difficult skill.

Right. And flying in real IMC is hard because it deprives the pilot of a view out the window, which is something that the camera in modern drones continues to provide, whether or not the pilot is maintaining VLOS.

It does bring up an interesting point: Our current rules, whether Part 107 or 44809, don't make distinctions regarding the types of aircraft and the features they may have (other than overall weight, and in the special case of flying over people, shrouded propellers). The modern drones with GPS, IMU, and computer-assisted stabilization are lumped in with old-style RC airplanes with ailerons, elevator, and rudder, which rely strictly on aerodynamics, not computers, for stability. For that matter, RC helicopters with collective, cyclic, and tail rotor are also lumped in with us, from a regulatory standpoint.

If there's ever a BVLOS rating, it will likely have to come with some sort of equipment which the aircraft must have. Modern drones might finally be separated from RC airplanes in the regulations. Depending on how the rules are written, the RC airplane crowd might like that. They were largely safe and self-regulating before the much more accessible drones came on the scene.

 

[Drone752H]

I disagree - I taught instrument flying to professional pilots for many years. A modern drone maintains attitude, heading, ground position and altitude unless you make control inputs. Real instrument flying (without using an autopilot, which is cheating) in a full sized aircraft relies on the pilot maintaining control of all of the aircraft parameters by sole reference to the instruments - a much more difficult skill.

Here is a FAA link for Bird Strikes.

https://www.faa.gov/airports/airport_safety/wildlife/media/Wildlife-Strike-Report-1990-2019.pdf

B52-D

You’re absolutely right! Stand by for DCE!

 

[RoboticTundra]

Something that you might want to consider is that Europeans are very substantially more risk averse than Americans, on average.

I'm not so sure about that. I've ridden with enough German, Italian and Greek drivers to think the opposite

They generally place a higher value on human life and are more concerned about how their actions might negatively affect others. You see this in their attitudes towards making roads safer, how loud they talk in restaurants, etc.

They also tend to have higher regard for laws and to be more law abiding. People in the U.S. learn to break laws at an early age (underage drinking, a wide variety of driving offenses, pot, prostitution, etc.). We are pretty adept law-breakers by the time we enter college. People in many European countries rarely break any law ever because their laws are more aligned with their beliefs. They have many fewer laws to break. So therefore they are law-abiders and are more likely to obey the laws and rules they have than U.S. folk.

I THINK that they would be much less likely to have problems with rogue/dangerous drone flights because they are more likely to obey rules and with somewhat higher regard for others are more likely to think about how their flight might negatively affect others from a safety or annoyance standpoint. HOWEVER, while not as likely they will still have some people who will flaunt the rules.

Without that natural rule following, how do we make drone flights safer?

 

[RoboticTundra]

Do we have the technology for a controller to know when a drone is nearing the edge of VLOS? Will we be able to have a default sandbox for this and the pilot must DO SOMETHING to acknowledge it and permit BVLOS operation?

 

[Chaosrider2]

Interesting discussion!

I suspect that eventually there'll be a widespread accepted mechanism for routinely flying BVLOS. As @cgmaxed has said, there are many sUAS pilots flying BVLOS already (doesn't necessarily make it right or safe) and we are not seeing any alarming trend in sUAS/manned aircraft collisions. sUAS are relatively small target to hit, and generally tend to inhabit airspace that doesn't see usage by manned aircraft. I personally avoid flying my drone in airspace that I think has even a remote chance of being used by manned aircraft, mostly by maintaining a relatively low altitude (typically 20-50m). It is true that manned aircraft could legally fly in the same bit of airspace as me at the same time, and therefore increase the chance of hitting a still small target but then the pilot would be significantly increasing their chance of colliding with a whole lot of terrestrial objects too.

Just a thought about the presumably significantly higher numbers of drones compared to manned aircraft in the US. Do sUAS spend as high a percentage of time in the air as manned aircraft? I appreciate I'm probably at the lower end of the usage spectrum, but my drone has only seen 21hrs of use in the ~14000hrs I've owned it for - less than 0.2% of its time airborne. That's like a commercial airliner doing one 1.5hr flight a month. However, I'll not be the only low-usage drone owner bringing the average down.

Excellent point.

Other than near ariports, you just don't see manned aircraft at our elevations. Outside of take-off and landing, I have NEVER flown a manned aircraft below 400 ft AGL.

And I think the crop-dusters can handle themselves just fine...



TCS

 

[Chaosrider2]

Do we have the technology for a controller to know when a drone is nearing the edge of VLOS? Will we be able to have a default sandbox for this and the pilot must DO SOMETHING to acknowledge it and permit BVLOS operation?

That would be a good feature.

A challenge would be that VLOS varies not just from aircraft to aircraft, but from person to person.

TCS

 

[Chaosrider2]

What is an RTH function or obstacle avoidance if not AI? Perhaps the term 'Artificial Intelligence' suggests something more to some, but in both of those cases, the drone is using sensors and/or stored information to plot a course (or deviate from one), and then autonomously control itself to enact the solution.

BVLOS will (and does), require the UA to have the same kind of awareness and capabilities we as pilots must have within VLOS in terms of 'see and avoid' - albeit autonomously and without direct pilot control while beyond our sight.

RTH is definitey an AI feature. I love it, and use it all the time.

My objection was to the "fully autonomous" requirement, which as far as I'm aware, currently isn't available for civilian aircraft, and may not be for a long time.

TCS

 

[Chaosrider2]

I disagree - I taught instrument flying to professional pilots for many years. A modern drone maintains attitude, heading, ground position and altitude unless you make control inputs. Real instrument flying (without using an autopilot, which is cheating) in a full sized aircraft relies on the pilot maintaining control of all of the aircraft parameters by sole reference to the instruments - a much more difficult skill.

Not a contradiction.

BVLOS is a lot like instrument flying in terms of aircraft control and navigation, except that it's a lot easier.

TCS

 

[Chaosrider2]

The discussions seem to be between strict VLOS and “open the flood gates”. My original thought (posted somewhere) was just to “open the door” but just a crack. Put a reasonable cap on distance (and how do you judge distance if you can’t see it?), and altitude. The chance of encountering a manned aircraft below 500 feet is negligible, don’t fly in the path of a landing or departing aircraft - small uncontrolled field. Quite frankly, I’m not nearly as concerned about aircraft as I am about vertical obstructions and power lines. Flying the Skylane I’m “up there”. Flying the drone, I’m down low, WAY to low to encounter an aircraft (operated responsively). Drone operations in Class G airspace (for those authorized) should be the same as manned aircraft. One mile and clear of the clouds! As Martha King is so fond of saying, Class G airspace, go for it!

Excellent point!

"The fallacy of the excluded middle" is one of the most common argumentative fallacies.

There are very few binary choices.



TCS

 

[Rich QR]

Excellent point.

Other than near ariports, you just don't see manned aircraft at our elevations. Outside of take-off and landing, I have NEVER flown a manned aircraft below 400 ft AGL.

And I think the crop-dusters can handle themselves just fine...

You might want to consult with the crop dusters about that last point.

I definitely see manned aircraft away from airports at our altitudes. Not super frequently, but often enough. I've seen a helicopter land at my favorite drone flying practice field near a suburban schoolyard; I believe it was for a medical emergency. I often see airplanes flying offshore near the beach at below 500' AGL, and that's allowed per 91.119(c). I've twice seen helicopters land on a public beach. I've seen a glider land in a soybean field in central Illinois. I've been a passenger in a helicopter that landed in a school athletic practice field.

And yes, I've seen many crop dusters just barely above the corn tassels.

 

[Malcx]

You might want to consult with the crop dusters about that last point.

I definitely see manned aircraft away from airports at our altitudes. Not super frequently, but often enough. I've seen a helicopter land at my favorite drone flying practice field near a suburban schoolyard; I believe it was for a medical emergency. I often see airplanes flying offshore near the beach at below 500' AGL, and that's allowed per 91.119(c). I've twice seen helicopters land on a public beach. I've seen a glider land in a soybean field in central Illinois. I've been a passenger in a helicopter that landed in a school athletic practice field.

And yes, I've seen many crop dusters just barely above the corn tassels.

I'm only 40 minutes outside the centre of London. It's not super common but I've seen 2 possibly 3 times where military helicopters are at around the 400ft mark when coming over a ridge (one was definitely under), and at least 4 Air Ambulances have landed in the fields/parks near me.

These are all just in the 3 years I've been drone flying and have a better awareness of my surroundings.

So yeah it's not limited to around airports or even just remote rural crop dusters.

 

[RoboticTundra]

Two people on our lake have float planes that have to be watched out for. One of them always does one or two notification passes over the lake before landing, the other not so much.

We occasionally have national guard copters fly overhead at around 350' AGL. These are easy to hear and avoid.

My biggest concern w/ BVLOS might be other drones operated by inexperienced/uncaring pilots (or even two experienced pilots both flying BVLOS in the same area from different locations). Fortunately a collision there is not likely to result in harm to a person though it wouldn't be impossible. How much less risky will this be after 2023?

 

[maelstrom]

BVLOS is a lot like instrument flying in terms of aircraft control and navigation, except that it's a lot easier.

No amount of repeating this will make it correct! You’ve either never flown a real aircraft on instruments or you have and let the autopilot do all the work. If flying a drone BVLOS was like instrument flying, you’d continuously be making small stick inputs just to stop it going out of control and crashing. The ONLY difference between flying within VLOS and BVLOS is that you can’t physically see the drone - the mechanics of flying it are EXACTLY the same in both situations. Modern drones are designed to be very stable and easy to fly and it’s this degree of automation which has made them so accessible to the masses.

 

[RoboticTundra]

That would be a good feature.

A challenge would be that VLOS varies not just from aircraft to aircraft, but from person to person.

Thinking out loud...

I think there would be two aspects to it; 1) Distance/Vision and 2) Obstacles.

Distance/Vision would need to be set for each individual/aircraft. Though even with this there can be variation. I may find it difficult to see my M2P beyond 1100' on a clear day but I can still see the area where it is so can still have situational awareness. So perhaps ideally two settings; Absolute VLOS and Situational VLOS. In reality though I can see where it is anyway so not sure if this would be valuable. From an idiot-proofing standpoint (to protect others from idiot PIC's) maybe there are defaults (900' on a clear day for M2P, 600' for a 2s, etc.) and going beyond these require explicit action from the pilot.

I also wouldn't mind if aircraft had better lights built-in with better control. Allow me to manually or automatically turn on a brighter strobe @ 800' distance on a clear day or 400' on a misty day.

Obstacles might be more difficult. Ideally the system can know if there is a clear sight line between the controller and aircraft or if there are obstacles in the way. If approaching an area where VLOS will be reduced or eliminated due to obstacles then it creates a sandbox demarc line that requires some kind of explicit action from the PIC to cross.

I really hate bits like this that place limits on good people doing the right thing because we need to protect people from the few (or not so few?) idiots. So maybe these are defaults that can be globally overridden w/ a 107.