1x39: Ambitious but Rubbish

That’s what I meant in saying Top Gear is dead. The BBC can continue to make a program called “Top Gear” but the format will need to change significantly and it will have a different audience.

The existing show works because of the interplay between the three, though admittedly likeable, idiots.

The new Top Gear, if it succeeds, will be a very different show: perhaps it will be about cars.

I honestly don’t see why. Driverless cars are expected to go out onto the road and avoid crashing into pedestrians, bridges, and other cars. Assuming that that’s possible, how can it be more difficult to avoid crashing when you don’t have to stick to a road but can instead move anywhere in three dimensions?

On drone deliveries and energy costs
Drones are rather ‘fuel’ inefficient because flying is hard and Gravity sucks, as a result I don’t think we will see a future where drone delivery of low value Items is the norm as it will not be cost effective. Clearly they have a role for high value, time critical items.

Regretfully Jono will have to endure Neat Gin on occasion.

However delivery by road transport is clearly viable, Autonomous (electric) vehicles and better route optimisation will have huge impact here. The future for White van man does not look so good.

By the look of this Amazon Prime Now thing, ordering an item and getting it an hour later costs six dollars or something. Surely six bucks is enough to pay for a drone flying to my house and back, and thus it’s economically viable? I agree that many people will not want to use the service because they don’t want to spend six dollars delivery on a three dollar bottle of tonic, but that doesn’t make it non-cost-effective, it just means that only desperate people will use it :slight_smile:

I don’t think so. I would presume that computer vision, GPS, and maybe some form of radar will take care of this.

I am not a monster, @bminish. I would put an ice cube in there.

2 Likes

After reading the post from @dotwaffle, the size of what would be able to be delivered would be determined by the size of the drone. As you mentioned in a show before about cameras, you can’t get past the laws of physics. Not that I’m thinking that a wall safe would be able to be delivered in such a way, but maybe even that bottle of tonic @jonobacon was needing might require a substantial vehicle, in this context of a delivery service.

Now, what I think would be so cool, albeit highly impractical, is the old tube systems that department stores had when I was young. The sales person would get the order with the money, put it in this round carrier, stick it in this tube that had a vacuum on it, and off it would go to the cashier. I guess it would require a rethink of pizza, having to roll it up to stick in a tube carrier!

I would like to hear the thoughts of this drone delivery idea from the perspective of an air traffic controller!

I rather doubt that the costs per delivery via drone of a few Kg of stuff can be gotten down to 6 Bucks a time

Delivery drones will need to be 'copters ’ of some kind since unlike Military UAV’s delivering the payload at high velocity towards the target is suboptimal.

Copters are a pretty inefficient way of flying, a quick websearch suggests that a Helicopter uses about 1.4 Litres of Fuel Per Km. My Diesel VW Golf managed 0.05L per Km on Yesterday’s Drive from the South of the UK to the west of Ireland.

I suspect both modes of transport will have proportionate efficiency gains by going electric. In both cases there is a lot of weight savings to be made by not having a driver along with all the safety systems designed to reduce risk to the driver when fail happens.

Our Family car is a Nissan Leaf, We get about 160W per Km out of our Nissan Leaf and that’s not just carrying me, the Mrs, the kids and our stuff, it’s also got a lot of weight tied up in stuff like seats, airbags, soundproofing , ‘user interface’, Climate control etc.
An autonomous delivery vehicle will do much better than this, be cheaper to manufacture and not have a driver to pay. The costs are also potentially split across multiple deliveries .
In terms of safety it merely has to manage to come to a controlled stop if systems fail.

A Delivery drone will also gain very significantly in efficiency over a Helicopter but when considering individual deliveries each journey has to pay it’s cost in full.
In addition to this Drones will have much higher costs per hour tied up in keeping them in safe working order. Occasionally Dropping Jono’s Tonic delivery with attached drone on members of the public from 500M up will prove unpopular.

Drones:
I think the development of drones is going to be similar to that of planes. First there are just a bunch of random people flying planes from a field. Later come the first ones that can carry a few people (or objects in the case of drones), but a very expensive, short ranged, kinda dangerous and not very practical in any sense expect maybe time. With the beginning of commercial flights there arises a need for infrastructure and regulations. We get airfield (or landing pads) and people decide on fixed routes, air highways so to say. The next step is advanced air control from the ground towers by radio or what ever fancy technology you want to use. It slowly evolves to be a very useful way of transportation and becomes available for common people. Then lots of accidents are going to happen, people get scared of possible terrorist threats or something like that and there will be a massive upgrade in security.
That would sum it up for me. In densly populated areas I would see a system where roof tops become lading spots and whenever you get a drone package you will be called in advance with something like this: Requesting delivery of package to your roof, if request accepted press 1, if request denied press 2 etc. I think it will take a while before fully implemented, but I expect it here within 3 decades.
Ow, I almost forgot to mention that I just ignored the military involvement in commercial flight, because yeah… The science behind bombing drones is completely capable of doing the things amazon wants to do… Droping packages… Sadly not the ones I’d like to be dropped though…

Clarkson:
I find it troubling that they waited so long long before they fired him. I have the feeling they wouldn’t have fired him if he had not been slightly vocal about the higher ups in the BBC in his column and that is one of the things why I think the BBC is the bad guy at the moment. Of course Clarkson is not above the law, but it is understandable (!= acceptable) that he is under a lot of pressure and he sometimes expresses it to someone who has no further involvement in that pressure. He is human too. I think there should have been some form of punnishment, but the BBC failed in doing that the right way, presumably because their ego was in the way.

I did enjoy the show. The level of seriousness was higher than normal, I would have nooo clue at all why though…

1 Like

Think about this: Roads are essentially one dimensional – all you have to worry about is that which is directly in front of you. If you want to overtake, you signal a bright light and move out to overtake, then resume the path. Nothing can collide into you in the other direction because they are on the other side of the divider.

With a flying vehicle, you’re now dealing with three dimensions. You may have a collision from any direction, and collisions aren’t just “hitting physically”, you can easily be affected by wake etc. The aviation world solves this by creating airways – essentially, everyone has to go down a set path and depending on your direction of travel you travel at a different altitude.

Only problem with this is that everything travels at different speeds so you need to have adequate spacing between the aircraft and in high traffic areas have “controlled” areas where everything has to be registered with a central authority who is responsible for collision avoidance. The main problem with doing this electronically is that pilot reports are incredibly inaccurate. Your GPS may be able to give you a reading to a few square feet when you’re stationary, but when travelling at speed it’s hard to relay that information in a timely fashions so that it’s current and accurate. Certainly altitude information is really difficult to convey as GPS is not that accurate (even with WAAS) and altimeters are generally only accurate to a couple of hundred feet unless you pay a LOT of money).

Basically, drones are expensive and entirely not suited to co-ordinated outside flying in a big mesh, especially when the software hasn’t been ratified by the FAA etc. A driverless vehicle, however, is relatively easy to do and is practically here today.

In terms of navigating in 3d space and having AI good enough to do this I think this will become a solvable problem although we are by no means there yet.
Radar altimeters are no longer a hard problem, GPS gets better once you combine Glonass and Galileo along with smart beacons, internet derived Assisted GPS etc.

However there will be lots of fail along the way. Fail may lead to regulatory interventions that limit what is permitted in inhabited areas.

I think the energy efficiency problem is not going to be solved, it will improve but the Physics involved in Flight are ‘inconvenient’ in terms of energy and this will restrict the use of delivery drones to High value, time sensitive Items such as Medicines and High explosives.

OK. I grant all of your points as true, certainly. However, it still seems to me that the problem may not be quite as hard as you make out. Roads may be one-dimensional, but collision avoidance on roads is too – if a pedestrian steps out in front of you you can only really evade by swerving to the right (or to the left in non-UK places). If a bird tries to fly into your drone you have six degrees of freedom to play with – even if we assume you can’t instantly reverse direction, you can still evade up, down, left, and right. Sure, other stuff can also crash into you from any direction, but these two things ought to balance one another out at least a bit, I’d have thought? Similarly, aeroplane airways are set up that way at least partially because two aeroplanes crashing is a hysterically bad thing to happen: aeroplanes normally have passengers on them; being on an aeroplane which crashes in mid-air is a pretty much guaranteed flaming death for everyone on board. Drones are less of an issue here; yes, we don’t want them to crash into one another, and yes it’d probably be fairly unpleasant for you were you standing in a field when two drones collided 200 metres above your head, but a drone failing in midair means that Jono doesn’t get a bottle of tonic, rather than guaranteed death for three hundred passengers and the complete destruction of anything at the crash site. So the airways could be considerably smaller and closer together, if airways are needed at all – the whole point here is that drones are aiming to be autonomous and react to things around them be avoiding, rather than relying on air traffic control to make sure there’s nothing that needs avoiding, right?

[quote=“sil, post:33, topic:10405”] yes it’d probably be fairly unpleasant for you were you standing in a field when two drones collided 200 metres above your head, but a drone failing in midair means that Jono doesn’t get a bottle of tonic, rather than guaranteed death for three hundred passengers and the complete destruction of anything at the crash site. So the airways could be considerably smaller and closer together, if airways are needed at all
[/quote]

Collision avoidance is not such a hard problem since if drones operate a transponder system (like modern aricraft currently do) then avoiding collisions becomes a computational problem. A lot of current procedure around flight paths pre-dates the real time situational awareness that transponders can provide and was designed around what was practical for air traffic controllers to keep track of reliably in a pre-computer world.

Any drone failure results in a ground impact, it does not take a mid air collision. Gravity sucks hard.

I think the issue here is that most of the people located in areas where drones have a hope of becoming a mass market delivery platform, also have a lot people (& property) in them.
Jono’s Tonic (with attached delivery drone) crashing through my conservatory is likely to enrage the ‘something must be done’ brigade. The Kinetic Energy of a free-falling payload from a few 100M Up is substantial and it will not take many such failures for the pitchforks to come out.

On the other hand there’s probably a real role to be played in ferrying Blood bank products or Donor Organs around, Here the risks of fail can be better offset against the risk to the public in an urban area.

Great show! One note… You guys mentioned that you hadn’t heard of any technologies that took an AGPL piece of software, forked it, and then re-sold services around it on the back of the original. I think a good example of this is TokuMX. This is a fork of MongoDB that has its own storage engine that adds things like compression, a fractal tree index, etc.

It makes for an interesting relationship between MongoDB and TokuTek. On the one hand, TokuTek expands the overall MongoDB community. On the other hand, companies will often have a bake-off between the two pieces of software. Anyhow, do a little searching about the two technologies - interesting stuff there.

I hadn’t heard of them at all; nice recommendation, thank you, and I’ll look into it!

Also: somehow did not know mongo was AGPL rather than just GPL.

Slight clarification - They didn’t even fire him. They merely did not renew his contract[1] (A contract which was apparently coming to the end anyway). This was in my opinion - literally the most face saving farcical decision that the BBC could have made. If I had finished a days work, walked over - ranted for 20 minutes at a colleague and then punched them so hard in the face they had to go to A&E - I reckon my employer would have done a bit more than “regrettably” chose not to renew my contract. Clarkson deserved a proper sacking, and in my opinion, an assault charge.

Given that Clarkson will be picked up within months, probably for similar money, by a competitor anyway, with next to no ramifications - all this sordid situation has proved is that you really can have enough money to be above the law!

[1] BBC Director-General's statement regarding Jeremy Clarkson - Media Centre

2 Likes

Hi there.

First, we are not waiting for drones because they are here, just two examples:

  • military applications are letally reliable
  • modern manufacturing operations have unmanned production steps

Second, Jeremy Clarkson’s story: I’m a long time Top Gear fan, however such a rude behaviour cannot be ignored or simply forgiven by an employer in general, and by a relevant one as BBC is in particular.

Third, few words more on openSUSE, it’s not fair to discard it after a short test; IMO if you evaluate a distro you never ran for just few days you will dislike it the more it differs from the distro you are used to, because it asks you for changing your habits.

Amusing show guys!

gabriel

Disagree.

Fun facts about drones

  • Don’t work well in bad weather
  • Not very manueverable
  • No air to air capability
  • Very, very expensive.
  • much larger than you think.
  • crash more often than you think.
  • Effective industrial robotics is only available to very large companies. It is quite expensive.

I don’t think we are anywhere near having drones in cities, because first we would have to live in a world where it was ok for drones to crash in a city, which is a world we do not live in.

quadcopter drone != drone capable of deliveries.

Delivery drones must be much, much larger to form any kind of fuel efficiency with a load of Amazon type stuff, and therefore would carry lots of fuel. That burny explody stuff that melts building girders. That stuff. Or batterries, the lithiumy explody kind. Think of UPS sized thing crashing into the side of a building, filled with jet fuel. Not a pretty picture.

Also, the things would most certainly not be autonomous, which is something not even the military would trust with their drones. Also computer vision algorithms cannot pilot a drone of any size, because it a 2 ton fire bomb. You do not let machine learning - which is a statistical practice govern a 2 ton firebomb. One does not leave the fate of a 2 ton firebomb to statistics. Besides, standard aeronautics - with GPS and human pilot - succeed at slamming much safer aircraft into mountains all the time.

This doesn’t even scratch the privacy, crowding, and regulatory issues.

To be honest, delivery by ballistics would be safer and more practical, and honestly easier to legalize. Amazon cannon anyone?

1 Like

THIS IS A THING THAT I WANT TO BE HAVING