4 ms·

Ask HN: What are the major problems in drone technology?

7 points by zxcvvcxz 12y ago

Is it... Height? Speed? Range? Battery life? Lack of autonomy? Size? Reliability? Other? In particular, how are these limitations hurting potential applications?

joshdance 12y ago
Reliability. You don't want 20 lb blenders falling from the sky.
- informatimago 12y ago
  You don't want them to fall on people, or brittle devices. But otherwise, they could fall onto a lot of places and things without further consequences. If you consider a city like Los Angeles, with 3792621 inhabitants, and a surface of 1215e6 m², assuming each person takes 1 m² (ie. the person-drone cross section), that gives for fall points equidistributed over the surface a probability of hitting a person of 3792621/1215e6 = 0.0031; only 0.31% per fall! To hit a person with 1/2 probability, you'd need a rainfall of 160 drones! Will your drone fly over crowded areas? Most people are inside houses most of the time anyways, and there's no reason to fly over the sideways. Now, you must also consider other areas to avoid landing (or falling) on. Cars, (streets), random machinery (eg. air conditioners), power lines, etc. But all in all, that still leaves a lot of places where to crash with no consequence, if you need to crash. IMO, there's a lot of unreasonned fears about the subject (as there is about anything else; and people wonder in other threads, why the belief that humans use only 10% of their brains persists...).
  - joshdance 12y ago
    True. However, current drone usage is to fly over interesting areas. These areas are interesting because of natural beauty (ocean, lake, mountain etc) or because of people. Crowds, Times Square, sporting events, etc. For example me using a drone at Park City filming my friends snowboarding has a much larger risk of hitting someone that on the empty desolate mountain 1 mile over. But no one flies drones over on that empty mountain because it is boring. I don't know what the percentage of hitting someone might be but certainly higher for current drone usage than flying a drone over a random city with equally spaced inhabitants. :) And kudos on your calculations, I had never seen this done before. This is the type of math that "news" stories need to have.
- matheweis 12y ago
  This, although let me add a slightly different reason... You also don't want your $3,000 piece of equipment to irrevocably fly off.
schmidtc 12y ago
I started a UAV company several years ago. The biggest hurdle we faced was governed regulation. Not only was it nearly impossible to operate the things I the US, it was also exceedingly difficult to get an export license. I haven't been paying attention since I left the comply a few years ago, but it sounds like the regulations are getting better. On the technical side we faced a few chanllanges. The first was reliability. Our planes were small <10 lbs, but would enevitly go down and be very difficult to recover. After the first couple crashes we designed a protective shell around the expensive bits. Causes of the crashes we always unique, from a loose connection, to interference. The next biggest problem was battery life, ours was an electric system and getting more flight time in the tiny platform was difficult. This of course limited the range which limited it usefullness. Otherwise, we solved the autonomy with off-the-shelve components (albeit military grade). Otherwise most other issues were resolved with software (flight planning, data recovery, etc). Our application was aerial imaging, I imagine other applications would have another set of problems.
- schmidtc 12y ago
  Forgot to mention, price was another difficult problem. We were trying to build a low cost system for academic use. The single most expensive component was the flight computer, there were some low cost alternatives, but at the time they didn't perform we'll enough. Perhaps this has changed.
notwedtm 12y ago
I have been building drones for a few years now as a hobby. From a consumer standpoint, the pricing has come down substantially. You can build a fully autonomous quadcopter for ~$500 these days (minus the transmitter). The problems I see are regulation, and battery performance. Regulations are really up in the air right now. There have been two rulings by judges that the FAA's current policies do not cover restriction of commercial usage for drones. The FAA is actively fighting this. I feel that over the next few years, we'll see some solid rules set down, and more than likely they'll be extremely strict to start. Statistically speaking, the FAA has an amazing track record, and letting anyone into the skies has a real chance of marring that record substantially. Cell power density is getting better and better, but it still takes a lot of work to get over an hour flight time on a multirotor. Thus they tend to be used for on-spot and precision flying shots for video. Fixed wing drones on the other hand have amazing range (I've seen 70km+ flights), but are less capable in the precision flying that a multirotor provides. On this topic, I am a huge fan of what Chris Anderson is doing over at 3DRobotics. Their open-source APM is the flight controller I choose for most of my projects, and it is extremely configurable. Hope this helps :)
bliti 12y ago
For what application? The word drone is usually used to define a flying autonomous robot, but it applies to ground, and water based ones too. [0] For flying types: - Regulation. Which is understandable, due to the safety risks of autonomous flying machines. - Energy storage/consumption. Electric drones are the most common now, but there could be a move towards Petrol based ones once technology catches on (lighter/more efficient engines). For land type: - Mobility. Right now there is a lot of focus on leg-based and weg-based[0] designs. - Vision. Object recognition is slowly coming along, but the POV of a land based robot makes it a bit more difficult. - Costs. For water type: - Costs. - Regulation. [0]From my experience. [1]Wheel-leg hybrid design.
- BWStearns 12y ago
  regarding energy storage/consumption and the potential movement to fossil fueled drones: http://en.wikipedia.org/wiki/Wave_disk_engine http://en.wikipedia.org/wiki/Wave_disk_engine I can't find the original article I saw this on but it seems like one of those things where even though the big market is probably smaller hybrid cars, might have application in commercial drones since it can potantially save a lot of weight vs more conventional engines.
  - bliti 12y ago
    There are radio controlled model airplanes that use a rotary engine[0] (which is in the same spirit of the Wave disk engine). The problem with this rotary engine is that it uses an alcohol-based fuel[1]. I could see someone adventurous enough trying out this mini-rotary in a drone. It would be highly expensive, and difficult to tune, but ultimately fun. [0]http://www.osengines.com/engines-airplane/osmg1401/ http://www.osengines.com/engines-airplane/osmg1401/ [1]http://rcvehicles.about.com/od/nitro/f/nitrofuel.htm http://rcvehicles.about.com/od/nitro/f/nitrofuel.htm
thejteam 12y ago
The DOD has a fair number of SBIR topics each release on various types of autonomous vehicles, including air, land, and sea based drones. Look at http://www.dodsbir.net http://www.dodsbir.net and look through the recent archives, say the last two years, and you will see what the DOD thinks some of the major challenges are.

7 points by zxcvvcxz 12y ago

Is it... Height? Speed? Range? Battery life? Lack of autonomy? Size? Reliability? Other? In particular, how are these limitations hurting potential applications?

joshdance 12y ago

Reliability. You don't want 20 lb blenders falling from the sky.

informatimago 12y ago
You don't want them to fall on people, or brittle devices. But otherwise, they could fall onto a lot of places and things without further consequences. If you consider a city like Los Angeles, with 3792621 inhabitants, and a surface of 1215e6 m², assuming each person takes 1 m² (ie. the person-drone cross section), that gives for fall points equidistributed over the surface a probability of hitting a person of 3792621/1215e6 = 0.0031; only 0.31% per fall! To hit a person with 1/2 probability, you'd need a rainfall of 160 drones! Will your drone fly over crowded areas? Most people are inside houses most of the time anyways, and there's no reason to fly over the sideways. Now, you must also consider other areas to avoid landing (or falling) on. Cars, (streets), random machinery (eg. air conditioners), power lines, etc. But all in all, that still leaves a lot of places where to crash with no consequence, if you need to crash. IMO, there's a lot of unreasonned fears about the subject (as there is about anything else; and people wonder in other threads, why the belief that humans use only 10% of their brains persists...).
- joshdance 12y ago
  True. However, current drone usage is to fly over interesting areas. These areas are interesting because of natural beauty (ocean, lake, mountain etc) or because of people. Crowds, Times Square, sporting events, etc. For example me using a drone at Park City filming my friends snowboarding has a much larger risk of hitting someone that on the empty desolate mountain 1 mile over. But no one flies drones over on that empty mountain because it is boring. I don't know what the percentage of hitting someone might be but certainly higher for current drone usage than flying a drone over a random city with equally spaced inhabitants. :) And kudos on your calculations, I had never seen this done before. This is the type of math that "news" stories need to have.
matheweis 12y ago
This, although let me add a slightly different reason... You also don't want your $3,000 piece of equipment to irrevocably fly off.

schmidtc 12y ago

I started a UAV company several years ago. The biggest hurdle we faced was governed regulation. Not only was it nearly impossible to operate the things I the US, it was also exceedingly difficult to get an export license. I haven't been paying attention since I left the comply a few years ago, but it sounds like the regulations are getting better. On the technical side we faced a few chanllanges. The first was reliability. Our planes were small <10 lbs, but would enevitly go down and be very difficult to recover. After the first couple crashes we designed a protective shell around the expensive bits. Causes of the crashes we always unique, from a loose connection, to interference. The next biggest problem was battery life, ours was an electric system and getting more flight time in the tiny platform was difficult. This of course limited the range which limited it usefullness. Otherwise, we solved the autonomy with off-the-shelve components (albeit military grade). Otherwise most other issues were resolved with software (flight planning, data recovery, etc). Our application was aerial imaging, I imagine other applications would have another set of problems.

schmidtc 12y ago
Forgot to mention, price was another difficult problem. We were trying to build a low cost system for academic use. The single most expensive component was the flight computer, there were some low cost alternatives, but at the time they didn't perform we'll enough. Perhaps this has changed.

notwedtm 12y ago

I have been building drones for a few years now as a hobby. From a consumer standpoint, the pricing has come down substantially. You can build a fully autonomous quadcopter for ~$500 these days (minus the transmitter). The problems I see are regulation, and battery performance. Regulations are really up in the air right now. There have been two rulings by judges that the FAA's current policies do not cover restriction of commercial usage for drones. The FAA is actively fighting this. I feel that over the next few years, we'll see some solid rules set down, and more than likely they'll be extremely strict to start. Statistically speaking, the FAA has an amazing track record, and letting anyone into the skies has a real chance of marring that record substantially. Cell power density is getting better and better, but it still takes a lot of work to get over an hour flight time on a multirotor. Thus they tend to be used for on-spot and precision flying shots for video. Fixed wing drones on the other hand have amazing range (I've seen 70km+ flights), but are less capable in the precision flying that a multirotor provides. On this topic, I am a huge fan of what Chris Anderson is doing over at 3DRobotics. Their open-source APM is the flight controller I choose for most of my projects, and it is extremely configurable. Hope this helps :)

bliti 12y ago

For what application? The word drone is usually used to define a flying autonomous robot, but it applies to ground, and water based ones too. [0] For flying types: - Regulation. Which is understandable, due to the safety risks of autonomous flying machines. - Energy storage/consumption. Electric drones are the most common now, but there could be a move towards Petrol based ones once technology catches on (lighter/more efficient engines). For land type: - Mobility. Right now there is a lot of focus on leg-based and weg-based[0] designs. - Vision. Object recognition is slowly coming along, but the POV of a land based robot makes it a bit more difficult. - Costs. For water type: - Costs. - Regulation. [0]From my experience. [1]Wheel-leg hybrid design.

BWStearns 12y ago
regarding energy storage/consumption and the potential movement to fossil fueled drones: http://en.wikipedia.org/wiki/Wave_disk_engine http://en.wikipedia.org/wiki/Wave_disk_engine I can't find the original article I saw this on but it seems like one of those things where even though the big market is probably smaller hybrid cars, might have application in commercial drones since it can potantially save a lot of weight vs more conventional engines.
- bliti 12y ago
  There are radio controlled model airplanes that use a rotary engine[0] (which is in the same spirit of the Wave disk engine). The problem with this rotary engine is that it uses an alcohol-based fuel[1]. I could see someone adventurous enough trying out this mini-rotary in a drone. It would be highly expensive, and difficult to tune, but ultimately fun. [0]http://www.osengines.com/engines-airplane/osmg1401/ http://www.osengines.com/engines-airplane/osmg1401/ [1]http://rcvehicles.about.com/od/nitro/f/nitrofuel.htm http://rcvehicles.about.com/od/nitro/f/nitrofuel.htm

thejteam 12y ago

The DOD has a fair number of SBIR topics each release on various types of autonomous vehicles, including air, land, and sea based drones. Look at http://www.dodsbir.net http://www.dodsbir.net and look through the recent archives, say the last two years, and you will see what the DOD thinks some of the major challenges are.