Amazon's new Prime Air Delivery drone (Credit: Amazon)
From Gizmag
ERIC MACK JANUARY 19, 2016
You've heard of the Internet of Things – the generic name given to all the various networked sensors, machines, devices and even buildings in the world – but most of those "things" stay in one place for the most part. The world is primed for an explosion of autonomous ambulatory devices, which led a team of engineers from the University of Waterloo in Canada to draft a conceptual framework for an "Internet of Drones."
The authors of a paper on the concept (linked at the bottom of the page) lay out what is essentially a structure for how drone traffic could be managed. It combines elements of the current air traffic control system, cellular networks and the internet.
The paper proposes terminology for key components of the system, with airspace divided up into "zones," each managed by a "zone service provider" (ZSP) that operates their own section of airspace.
The zone service provider, which could be software-based rather than an actual human operator, is sort of like a combination of a cell tower and an air traffic controller for a specific airport. Drones and zone service providers communicate via the cloud to ensure that autonomous traffic flows through that zone safely, and according to whatever rules have been established for that zone. When a drone passes into a new zone, it is handed off in much the same way that a wireless device is transferred to a new cell tower as it travels.
The infrastructure can also allow for third parties outside the zone (such as administrators, retailers on either end of a delivery or possibly even consumers) to communicate with drones in flight. This would be particularly useful for proposed drone delivery services like those that Amazon, Google and others are working on.
The paper suggests that the existing cell network base stations could be used to actually deploy the system.
"Since these base stations are already deployed, the physical space is available and they are capable of running the ZSP software," it reads. "Therefore, they seem well positioned to implement ZSPs and provide wide network coverage for [the Internet of Drones]."
Within each zone, defined airways, intersections and nodes are established, which can be thought of as being similar to the system of roads, intersections and destinations that cars currently use on the ground. Even though drones could theoretically fly anywhere in the three-dimensional airspace, the idea is to establish designated airways and regulate traffic through them to avoid collisions. Drones would also be responsible for avoiding collisions with objects outside the system (such as birds) on their own, and keeping the ZSP advised of those maneuvers.
Plenty of others are also working on how the coming, drone-filled world will fly. NASA has been working with Exelis on another drone-tracking system, but it's not immediately clear if it could integrate with the newly-proposed architecture.
Full article here
Comments
Lol Rob buddy, people do the craziest things! ;-)
I can completely agree that an automated system can operate better in a closed system with known dynamics and physics, but on a openly dynamic common road, (there is no upper limit to variables) computer vision in no way can compete with an "aware and capable" human driver in cognitive capability. It can only pay attention to the variables you program it with...flying is much much easier than driving in that regard. Look at how many get through the DARPA challenges. Reaction time and physics handling maybe, but not including danger recognition and instinct. Give computers some 10-20 years then maybe.
I also completely agree that 99.9% of all car accidents are caused by humans, but that would remain the same even if a automated machine was in control and failed to avoid an accident...as every machine, including the car itself is also human made. ;-)
I did write a section on how human inattentiveness is cause for most accidents, but unlike most, i don't think automated driving cars are a solution at all. I don't have a problem in the interim to augment human ability to drive cars with computer vision or sensors, even my Prius follows the cars in front of it at a safe distance automatically and brakes in a emergency etc. but avoiding something is a completely different kettle of fish. However, if you make it a rail system like the schweeb, you can ditch most of the automation, computer vision and sensory problems cars have and replace it with a simple range sensor. I'd go for that solution in an instant.
My point is that driving cars on roads is both environmentally and engineering wise, a really stupid way to move from point a to point b. If it was up to me I'd get rid of cars in a heartbeat, there's many much better solutions. Let alone the fact that most car travel is in fact avoidable in the first place with some "intelligent" planning of how, where and why we live our lives. Commuters should be outlawed for starters along with centralization. We're creating most of the problems with transportation (and our other so called "time saving" devices) ourselves, by accommodating the whimsical notions of a population who think they know what they want, and why they want it like that!
It starts with the fact that the only real currency is time, and that one can only spend it once. Wasting one's life sitting in traffic is a tragic result of a perverted system. Also forget fake fiat money as a motivation to make decisions, or as a life pursuit to gain financial "freedom" so that one can "afford to spend time doing as one would like" (which is one of the only freedoms most desire, as "freedom" cannot be achieved by simply voting in a democratic nation of mob rule). We do only as we can "afford". That is the extent of our expression of freedom under the systems we have developed as humans. Further anything based on the faulty accounting of cost expressed in fake dollars, is always going to neglect the real cost of our actions on nature, despite what nature provides for "free". Our balance sheets are corrupt. On top of that we have this crazy idea that we can own something, yet we are born into this world only in our baby suits and can't take anything with when we leave! If we'd only understand how to best leverage nature to our desires and according to our needs, and create "intelligent" systems that benefit both us and the environment we live in, without being subjected to the ideological fallacies and desires of popular culture, consumerism and politics, that are driving us to dam-nation.
But enough of the philosophical discussion of cars... ;-)
For example, technically, a rail based system (or even the hyperloop for long distance) is a far better solution, that uses less resources to build (road paving is made from fossil too, so if we all drive electric in the future so we can ditch polluting fossil powered cars, what are we going to drive on?) and it automatically allows vehicle to "track" a course to their destination. The schweeb for example takes it to the next level, by lifting the rail out of the way of surface users way by placing the monorail just 3-4m above the ground, simplifying intersections and avoiding the need for traffic lights, pedestrian and wildlife crossing's etc altogether. It also uses a fraction of the energy of a normal car EV, to build and to run, to the point you can propel yourself along at 40kmh using pedals if you wanted too. That's already a faster average speed then you get in most metropolitan roads...and it saves time because you don't need to go to the gym as well, alleviating obesity and diabetes to boot! Congestion on the shweeb means everyone actually goes faster as each pod can help push the next along, reducing drag like every good train does, without compromising a individual pod's ability to steer to a specific destination. It's even possible to do "drone deliveries" using an automated schweeb pods, it will bring you your shopping and also take your garbage and compost away for recycling, and the rail system can even be used as a electrical supply to network each house along with internet etc too. In fact schweeb vehicles can all operate without the need for making a billion batteries for EV's, to replace the billion fossil powered cars we drive now, just so we can have our "own" "independent" transportation apparatus.
The solutions are there we just need to stop "improving" and perpetuating the stupid ideas, like cars, fiat currency, the fallacy of freedom etc, we invented over a hundred years ago! ;-)
Regards JB
JB, the discussion about situational intuition missed the fact that, the vast majority of drivers are unskilled and inattentive. They would run over the dog, the man in the fur coat, the old lady, and the pram, all at once. (And in some cases, not even notice they have done so...)
Meanwhile, the autonomous vehicle will always be paying attention, and will avoid all of them, so long as time/speed/distance physically allows this. And if physics dictates the object would be hit, a human driver will hit it anyway. They will even hit it, if physics would have allowed them to stop in time, in many cases.
The hobbyist is on notice; we will be virtually eliminated from the skies, or the cost of meeting more and more stringent regulations and other mandates will make it unaffordable for the vast majority. Either that or flying will be limited to very limited assigned flying areas.
Just go to RCG and read some of the threads.
BTW, how will drones deal with birds? Anyone with extensive experience flying their MR's has had some encounter with them. I certainly have, and in such situations I maneuver away and head for safety (usually the ground or further away) when especially larger aggressive birds find my invasion of their airspace quite objectionable. Youtube is full of bird attacks on MR's and even RC planes. I'd like to see how an Amazon drone will use AI to cope with that.
One more thing. If anyone has not tried the FAA B4UFLY app, you should. It is completely voluntary to enter a flight plan and log it at the moment, but who wants to bet within a year it will not be? Where we (myself and multiple other old timer RC guys) have been flying for over 35 years according to FAA, we have been breaking the law. What they call airports is completely laughable.
This whole "public safety" meme is nothing but a ruse. The air will be a corporatocracy
Autonomous vehicles don't have situational intuition and therefore are unlikely to be safer than human drivers, depending on the situation.
For example automated systems are not yet complex enough to distinguish between a person on all fours in a fur coat or a dog. Or find it difficult to "morally" prioritize between possible outcomes, for example missing the old lady or the lady pushing a pram. So there is a degree of complacency within an automated systems that doesn't respond as a person would or should.
But in saying that there is a way to make "automated" systems much safer by design, either by excluding human activity around vehicles altogether, or to use automation to augment human capacities to react, or react appropriately for the situation. Most vehicle accidents occur in congested areas where automated systems suffer sensory overload, as do humans, hence the high rate of accidents. Most accidents are caused by momentary loss of situational awareness...read distracted drivers. No healthy human intentionally wants to crash. So augmenting human drivers will likely result in the fastest way to reduce accidents and at the same time still leave humans morally in control.
The same rules should apply to airbourne automated systems, meaning that any airframe that is capable of human injury or serious property damage should never be operated purely autonomously, and should always have a backup, in the loop, human pilot. Accordingly I can't think of any reason or product, apart from medical deliveries, that need a "within the hour" delivery service using UAVs. SO developing UAV parcel deliveries makes no sense to me in the first place.
On another note regarding the internet of drones: I don't see why the drones can't establish there own airborne mesh networking abilities to augment the terrestrial ones, which in turn could provide internet and other information services to users on the ground whilst making one of those useless parcel deliveries. That is what I would call the internet of drones.
FAA is serious about aircraft safety and residents on the ground deserve some level of safety.
Selling low-altitude airspace to Amazon by FAA closes private drone business forever.
Ok, Amazon can go to Africa to make business but others can do either.
Self-driven or autonomous cars failed to show some standard level of road traffic safety.
You can equip your autonomous car with 100+ sensors and best technology but how can you get a-autos protected against impact by driver-driven cars ?
Ok, you can build all a-autos highways for over-head charged a-autos
like trolleybus but a-autos don't fall to the ground.
If you sell Low-Altitude Airspace to Amazon and others, you ban development of personal model aircraft at the same time and you close DIYDrones on the same day.
Ok, world is governed by bit fat lobbysts so individual modellers have already lost their chance to build small aircraft models for hobby use.
@Darius
Do you have a picture or CAD Drawing of your delivery UAS? I'm always curious how people tackle the sideslip issue.
That's ridiculous. Automated anything will be more safe and reliable than manned anything. Google has proven that with their autonomous cars. Any company investing in man these days, is on their way out.
If the FAA bans amazon drones here, they will find a country that allows it and we will be the only ones suffering from slow delivery. No one will be suffering from your imaginary death from above. (other than military/law enforcement targets)
This drone design by Amazon is still a big problem if exposed to side wind.
Frankly speaking none drone design showed fit to deliver a parcel yet.
My Pizza Delivery Drone design is reliable and fit to work since pizza is much lighter.
I am afraid, following a number of tests, FAA bans project by Amazon to protect residents.
Amazon Prime Air idea is ok to attract investors but not proven worth of increased crash and personal injuries risk.
Parcel deliveries based on blimps could work for some regions.
Drone alone is ok to fly over the unpopulated land from time to time.
Ultralights can do the job for Amazon and competitors since manned aircraft can offer higher standards of safety.
Manned VTOL ultralight is what is sought by Amazon and competitors today.
Amazon seems to be on board.
https://www.yahoo.com/tech/exclusive-amazon-reveals-details-about-1...
This sounds like a good start to helping control the coming drone-filled world.
Regards,
TCIII AVD