Not a surprise for anyone who has tried dealing with VTOL sporks drones, which are easy to take off but very tricky to land precisely in a wind due to the sail wing. From the WSJ:
Google has scrapped its initial drone design because it was difficult to control and is now working on a new version, according to Astro Teller, head of the Internet company’s Google X research lab.
Teller told the South by Southwest conference in Austin, Tex., that Google founders Larry Page and Sergey Brin encourage such failures, so researchers can learn from the mistakes and try different technologies and strategies.
In August, Google unveiled a drone-delivery system called Project Wing that the Internet giant was testing. The prototype had a five-foot single wing that sat vertically on the ground and then turned horizontal after take-off. The design was supposed to combine the benefits of vertical, hovering take-offs and landings with the speed of wing-based flying. In the test, the drones carried supplies including vaccines, water and radios to farmers in Queensland, Australia.
The drone was mechanically simpler than other types of drones, but it was harder to control, Teller said on Tuesday. It didn’t hover well in high winds and its cargo shifted too much when the wing moved up and down, he explained.
Google began working on drones in 2011. It’s not clear when it began experimenting with the single-wing design, but Teller said half the team “knew it was the wrong answer” after eight months. After 18 months, about 80% of the team thought this, he added.
The team debated whether to scrap the design or test it quickly in public. Brin told the team they had five months to make deliveries to real people, which wasn’t enough time to come up with a different design, so the team went ahead with the Australia test with their existing prototype, Teller said.
“Even though Sergey’s five-month thing prolonged the problem, it created an end date for it,” Teller said. “It’s possible without that we would have extended the wrong thing even longer.”
Before the Australia tests, the team was already working on a new design that moves away from the single-wing-based approach, he noted
“They’re now working on that vehicle,” Teller said. He plans an update on the status of this new drone later this year.
Comments
Having autonomous cars is a 2D problem and its much more complicated than 3D space of the drones.
JD
Arrhhh the Germans! They know what they are doing! ;-)
Well guess what that video of the DHL delivery just proved what I said previously. First it was operated between depot to depot, not directly to the house. It operated over difficult terrain (water), it was a time critical delivery (medicine) of "high" value (someone's health was at stake). In my book that delivery was totally justified, and to be perfectly honest a platform I myself am working on to achieve, along with some other goodies.
But now give me a product that Amazon sells that I need urgently, so I can run the simulation on that!
-
Dez Socks
Hey buddy could you deliver those batteries I ordered from a lower altitude next time? Thx! ;-)
-
Bojan
I think the solution is the other way around, you make the Pizza (or tacos!) at home with your new oven that has a 3D printer robot chef built in. Faster, cheaper and safer per pizza and way more convenient. It's also connected to the fridge so it can pick out it's own ingredients, and order new ones when they're low on supply, which gets delivered normally by vehicle once a week. Plus I'd like to see the real economics of delivery of a $5 pizza with a $5000 drone. If the delivery company has to employ an actual pilot to monitor delivery then you end up with a $50 pizza (or a $10 a taco!). Who will buy that?
Operating a "swarm" of drones just makes the economics worse, simply because the equipment is being under utilized (at night or between lunch and dinner etc) which like the silly power grid we operate, increases cost of supply. What taxi will earn more money: five taxis that are constantly parked at the train station or the one driving fares around 16 hours a day? Economics one on one is $ per hour cost vs profit, so operating a bigger fleet might improve response time, but it decreases viability unless everything is operational all the time.
Another thing with drone deliveries is that you'd likely be increasing the amount of workers required for delivery of a single item not decreasing them. Lets say the local pizza shop deliver's with drones, for that they require a spot to land, so either a few dedicated parking spots in front of their shops ($), or they have to renovate their roof to incorporate a landing zone, put a staircase in etc ($). Then a technician, not pizza boy on min. wage, needs to service and operate the drones. Each drone will likely, in the first instance, have a pilot as well for monitoring purposes, who is underutilized as well as he waits for each single delivery trip, instead of taking a car and doing at least 3-4 deliveries at once. Then their's all the regulatory hoops, insurance, OHS and food safety...and the fact you'd have build in a heater of some kind to deliver a "warm" pizza as the airflow in winter will make it a delivery of Pizza Ice cream otherwise. ;-)
For drone deliveries to work it needs to deliver items of value, that are required urgently (at a premium) over difficult terrain where no other options are available, to make sense.
-
healthyfatboy
Automated cars are closer, much better than UAV's but still not quite good enough. :-)
Take an electric version of the schweeb I linked to above as an exmple. In city traffic most of the fuel consumption is stationary or near stationary cars with idling engines and rolling resistance (not so much aerodynamics at those speeds). Rolling resistance on a rail system like the schweeb are a order of magnitude lower than a car on a road. Plus it's less maintenance, weather immune if using a hanging rail, reduced congestion as they can butt up to eachother to form an aerodymamic train, which make them all go faster with the same energy, no monkeys who don't know how to drive or navigate or are preoccupied on their phones, no obstacle or pedestrian avoidance systems required as it's above ground, or vehicle crash testing etc etc. A rail system is also far simpler to automate, and construct. In fact the control system is essentially digital stop, go, left, right commands, with little dynamic input or control required.
Just as an energy comparison:
Essentially it's a individual point to point transportation system with the efficiencies (and greater) of rail.
Another thing worth mentioning is that with the schweeb its possible to increase traffic density per km of road/track by placing them end to end, in which case even a human pedaled (non-electric) schweeb has a much faster average speed than a car in the city, which means you get to your destination faster but at a fraction of the energy cost.
Coupled with a hyperloop you could then completely clear the skies of aircraft, apart from over uninhabited terrain like ocean and mountains where a rail system is more expensive to build. You'd also be able to transport goods and people with end to end convenience with little to no human interference. This means fresh from farmer to household, from distributed manufacturer to consumer (bypassing monopolies like Walmat and Amazon), from home to school or work using the same vehicle at different times and from household to recycling etc etc. Might sound funny coming from a UAV enthusiast, but the less things flying around the better, for the environment and our own well being. ;-)
What's more realistic is having fully autonomous cars going around town. Sure, you could say there will be more congestion but if cars are autonomous and we end up with robot taxis taking everyone everywhere instead of having a human drive, you could make everything much more efficient and not worry about congestion. Then you could also have delivery vehicles be autonomous and a ground robot could easily deliver a package to your doorstep. There was a pretty good article talking about how autonomous cars and taxi services could really change everything. You don't need a UAV for any of it. It's also harder to go and steal an autonomous car vs. taking a UAV down and pillaging it for parts. While the idea sounds cool, autonomous ground vehicles will change delivery, not UAVs.
Tacos...
Once they drop tacos, the world will change thier mind about drones.
Not pizza's tacos will change the world.
Is that picture like the quadcopter about to land?
JB> Only if we are unwise enough to let it. I'd really like to know of a product that is so important it can't be delivered "conventionally", or even better locally manufactured, that can achieve the same outcome for demand. Can you name a product that must be delivered to the door with such urgency?
There are many economic dimensions, urgency is only one of them. Energy costs, safety, convenience, size, price, availability of “local manufacture” (to use your expression), nature of geographical area, etc ... I don’t see it as black and white. In some cases it won’t make sense, (book delivery doesn’t seem very promising to me), in others it will.
Ultimately competitive market forces will decide ... at the economic margin, as usual. This may or may not be a good thing, but that’s another story.
Here’s an example (all too obvious) of drone package delivery that makes sense, imho ...
https://www.youtube.com/watch?v=AeWjnKfpXq4
PS:This page bookmarked! Will revisit on3/19/2025!
mP1
Or a schweeb: http://shweeb.com/
Less safety risk. Automated delivery of goods and people. No traffic congestion. No batteries. Can be used as transportation infrastructure for electricity/water/sewage/rubbish as well. Above ground so every area is for pedestrians only...pods can then go into trains or even better the hyperloop. Wouldn't be able to beat the speed that way and the same resource cover's nearly all the requirements, gets rid of cars and if you keep some pedals even keeps people fit. ;-)
http://www.wired.com/2015/02/construction-hyperloop-track-starts/
It would be far easier to build those vacuum tubes in Futurama that everybody seems to get sucked and transported about.
Martin and Joe
Check your emails for a previous post from Colugo from today, that video linked works no password required.
-
Gary
I'm not sure drone deliveries a viable economically...yet. Over the border "deliveries" are of course delivering high value products, but a drone might only deliver a $20-$50 item. A man in a van doing deliveries isn't that bad efficiency wise, and even per Wh per km, plus if they could be "uber-ed" with parcel sharing to gain more effective coverage it could be even better.
There's also two end points that need to be considered, one is manning the dispatch depot to load the drones and maintain/charge them and then there's the delivery side were safe landing and storage of the delivered product is required. Will we have drone compatible mail boxes to keep the deliveries secure until we get home? A delivery man loads his own van from the depot and even collects a signature on valuable items before delivery, as well as does pickups from other shops. In the depot I can imagine that drone staff might not be scheduled effectively either as they might have to wait around until the next drone returns, alternatively a large drone fleet is required that increases capital costs/overheads etc. to the point it could be more costly to setup than a van. Coverage in time is also an issue, with drones needing to return to base after every delivery whereas a man in a van doesn't. A lot of things will need to change to get it to work.
I still haven't a clue what products are worth delivery "drone style" in a rush. Any ideas?
-
Colugo
A 60% improvement is good. Is it possible to keep the wings permanently in a horizontal position and still VTOL like that?
-
1
-
2
-
3
-
4
of 4 Next