5 ms·

There is no demonstration on youtube of a drone carrying a few pounds at 100mph. >> zig zagging through buildings and alleyways - fully autonomously Also neve

There is no demonstration on youtube of a drone carrying a few pounds at 100mph. >> zig zagging through buildings and alleyways - fully autonomously Also never been done

lazide 4y ago
Zigzagging through a forest fully autonomously (with no GPS) - there are videos on YouTube of a Swiss company doing similar. [https://youtu.be/m89bNn6RFoQ https://youtu.be/m89bNn6RFoQ] The alleyways are a simpler version [https://amp.france24.com/en/live-news/20220504-drone-swarms-can-now-fly-autonomously-through-thick-forest https://amp.france24.com/en/live-news/20220504-drone-swarms-...] Here is a YouTube video of a demonstration with a hexcopter capable of carrying 10+ lbs. [https://youtu.be/ODQT_hQsQgc https://youtu.be/ODQT_hQsQgc] And here is a dude flying a DJI Inspire (not the same as above) carrying 5 lbs. [https://youtu.be/DKr-yj68VJs https://youtu.be/DKr-yj68VJs] Folks have been using those and similar (cheaper) drones to drop Grenades on tanks in Ukraine. As the autonomous flight is harder with smaller drones (less computing payload available, worse sensors), the only reason you’re not seeing both of these parts together in urban environment is the implication (for national security). It is there.
- CraigJPerry 4y ago
  >> Zigzagging through a forest At speeds of up to 40kmh, the claim was 100mph+ >> The alleyways are a simpler version Here's the video of that https://www.youtube.com/watch?v=Lr7L2t-svJQ&t=16s https://www.youtube.com/watch?v=Lr7L2t-svJQ&t=16s - that's below 5mph and carrying an external payload of 0 grams, and the reason it's in a forest is because you can't fit them all in an alleyway and you need multiple in order for them to work. >> Here is a YouTube video of a demonstration with a hexcopter At 22mph, the claim was 100mph+ >> And here is a dude flying a DJI Inspire At around 20 mph max but mostly hovering and for 8 minutes at that before running out of juice >> Folks have been using those No. People like yourself can easily dream this concept. It stands to reason, why wouldn't it be possible? The truth is that battery technology is no where near good enough yet. It's the first idea tried in Syria, it's been tried thousands of times at this point and yet we have more documented cases of the aggressors killing or injuring themselves than of the 3 documented cases where this actually worked in practice - all 3 of which i think everyone can agree from the footage are pure fluke. Zero aiming was possible, it was just release and hope for the best in all 3 cases. What people are actually using them for is for recon and aiming of artillery. They're effective in these roles and that's why they're used for this rather than strapping heavy explosives to them. >> autonomous flight is harder with smaller drones (less computing payload available That's not true. Compute isn't a restriction in small drone autonomy. A basic F4 is wayyyyy overkill for flight stability and control. Waypoint navigation is even lower power requirements since it amounts to a simple loop not much more complex than: 1. Get current GPS coordinates 2. Get current compass heading 3. Send directional input via serial peripheral to F3 or F4 flight controller 4. Repeat The flight controller is doing far more processing than this - gyro input & smoothing for each of 3 axes, accelerometer (same), GPS location at approx 10Hz, optionally barometric input (rare and usually not necessary). These inputs are often processed through kalmann filter before finally calculating the torque command to issue to each ESC's microcontroller. This only happens at up to a few hundred Hz so you can see compute power is not a restriction. In larger aircraft it's harder because you can't beat physics, in a small drone i can change the amount of torque applied to a propeller 4+ times per rotation in toy off the shelf drone tech today. With larger craft that have non-electric motors this is simply not possible. You can't instruct a mechanical engine to regulate its power output with the same fidelity of an electric motor. >> worse sensors The only required input sensor is a gyro - in practice we use accelerometers and GPS in autonomous drones but there are ways of avoiding the need for these (e.g. visual inputs / optical flow as a correction for drift). MEMS gyros are cheap and effective, what better sensors do you need? Your reasoning is uninformed / dreamed up based on watching too many movies and not being familiar with the actual discipline. Better algorithms are a restriction. The science / theory part of implementation was achieved by all projects fairly quickly (ardupilot, clean/betaflight, etc. etc.) - but the art part, the how it feels and performs in practice, is an unsolved problem under continual development. >> drop Grenades Too heavy given current battery technology and lack of aiming. These are not effective weapons which is why they have limited success. More people have been killed in war zones in the past decade by garden spades than toy / photo drones with explosives. TL;DR until batteries technology moves forward by a few generations, the concept of a small drone as an offensive weapon will not be better than that offered by something like the switchblade drone. Which when you cut through the excitment, is less valuable than conventional weaponry in most fight scenarios.
  - lazide 4y ago
    FYI, your argument refutes itself. Compute is not a limiting factor in a small drone for basic flight operations, it is a limiting factor for real-time optical guidance (aka in GPS denied areas and real-time object/target recognition). Which is exactly what I was referring to. Those examples yes are not showing 100mph. I’m afraid the video I’m referring to has been pulled, and the in-person demonstrations I’ve been asked to not talk about or publicize (now that I follow up with the parties involved). The battery technology is more than sufficient for what I’m describing. My M300RTK has no issues carrying 5+ lbs for 15 minutes or 1-2 lbs for 40 minutes, which is more than sufficient, and scale is more than adequate here. You’re a couple years behind on this assessment. I literally have a 10 lb ‘hunter killer’ autonomous anti-drone drone from a company I’m not going to name sitting a few desks away from me right now.
    - CraigJPerry 4y ago
      >> I’m afraid the video I’m referring to has been pulled, and the in-person demonstrations I’ve been asked to not talk about or publicize Interesting. Let's take a second to take this from back of a fag packet physics. To produce a 100mph drone is easy - lets take one of mine (there's nothing special about it) as a starting point and then lets see what we can change to achieve load carrying while retaining this speed. Each motor & prop combo produces 1.5kg of thrust. They do this while drawing 40amps each from a 6s - which they can sustain for up to around 10 seconds before burning out MOSFETs on the ESCs (not to mention deforming the battery plug due to 160amps passing through a 60amp connector). This craft has a weight of just under 350 grams and when coupled with an 1800mah 6s, has an all up weight just under 650 grams. We want to carry "several pounds" - lets make it easy, lets say 2lbs - and retain that 100mph speed. Adding 10 grams prevents us hitting 100mph, we need to scale up our thrust. First problem - there's no other prop pitch or blade count for this motor size which will increase thrust while retaining prop speed, so we need to move up a size in motor. Since we need to carry 2lbs, you might think we only need each motor / prop combo to produce at an extra 0.5lbs but in practice we're going to need more than this because we need to sustain the aircraft vertically, whilst also providing enough thrust horizontally to reach the target speed, so in practice we will need more than an extra 8lbs of thrust (remember how 6kgs of thrust is needed to carry a 650g craft to 100mph - you don't just get to add 2lbs of thrust and call it done). Problem - we need a motor / prop combo that produces above 2.5kg thrust each - so we're into 14 inch propellers and we need a motor heavy enough to take around 60amps at 6s (i'm being conservative to try and make the maths work out in your favour but i hope we can both agree that only 60amps is pure fantasy here). That's a new problem, our weight went up and so we need even more thrust to compensate. That's another new problem, we have more thrust but now we need a power supply capable of a higher discharge rate - that kind of battery entails even more weight. But now there's a new problem, our larger props mean we're well away from the happy 300-400mph tip speed and now we're approaching the speed of sound - which of course means we can't fly. But now there's a new problem, the battery that can supply that discharge does exist, the only problem is it weighs almost 2kg by itself. And so we're back to you can't fly 2lbs payload at 100mph on today's battery tech. Camera operators have been trying hard for years at this point to put a RED or some other expensive heavy camera / lens combo right next to race cars on track. There are two solutions for that today: 1 - fly the heavy lens and get that sweet image but do it at 20mph or 2 - keep up with the 100mph+ action but film it on a gopro or similar.

There is no demonstration on youtube of a drone carrying a few pounds at 100mph. >> zig zagging through buildings and alleyways - fully autonomously Also never been done

lazide 4y ago

Zigzagging through a forest fully autonomously (with no GPS) - there are videos on YouTube of a Swiss company doing similar. [https://youtu.be/m89bNn6RFoQ https://youtu.be/m89bNn6RFoQ] The alleyways are a simpler version [https://amp.france24.com/en/live-news/20220504-drone-swarms-can-now-fly-autonomously-through-thick-forest https://amp.france24.com/en/live-news/20220504-drone-swarms-...] Here is a YouTube video of a demonstration with a hexcopter capable of carrying 10+ lbs. [https://youtu.be/ODQT_hQsQgc https://youtu.be/ODQT_hQsQgc] And here is a dude flying a DJI Inspire (not the same as above) carrying 5 lbs. [https://youtu.be/DKr-yj68VJs https://youtu.be/DKr-yj68VJs] Folks have been using those and similar (cheaper) drones to drop Grenades on tanks in Ukraine. As the autonomous flight is harder with smaller drones (less computing payload available, worse sensors), the only reason you’re not seeing both of these parts together in urban environment is the implication (for national security). It is there.

CraigJPerry 4y ago
>> Zigzagging through a forest At speeds of up to 40kmh, the claim was 100mph+ >> The alleyways are a simpler version Here's the video of that https://www.youtube.com/watch?v=Lr7L2t-svJQ&t=16s https://www.youtube.com/watch?v=Lr7L2t-svJQ&t=16s - that's below 5mph and carrying an external payload of 0 grams, and the reason it's in a forest is because you can't fit them all in an alleyway and you need multiple in order for them to work. >> Here is a YouTube video of a demonstration with a hexcopter At 22mph, the claim was 100mph+ >> And here is a dude flying a DJI Inspire At around 20 mph max but mostly hovering and for 8 minutes at that before running out of juice >> Folks have been using those No. People like yourself can easily dream this concept. It stands to reason, why wouldn't it be possible? The truth is that battery technology is no where near good enough yet. It's the first idea tried in Syria, it's been tried thousands of times at this point and yet we have more documented cases of the aggressors killing or injuring themselves than of the 3 documented cases where this actually worked in practice - all 3 of which i think everyone can agree from the footage are pure fluke. Zero aiming was possible, it was just release and hope for the best in all 3 cases. What people are actually using them for is for recon and aiming of artillery. They're effective in these roles and that's why they're used for this rather than strapping heavy explosives to them. >> autonomous flight is harder with smaller drones (less computing payload available That's not true. Compute isn't a restriction in small drone autonomy. A basic F4 is wayyyyy overkill for flight stability and control. Waypoint navigation is even lower power requirements since it amounts to a simple loop not much more complex than: 1. Get current GPS coordinates 2. Get current compass heading 3. Send directional input via serial peripheral to F3 or F4 flight controller 4. Repeat The flight controller is doing far more processing than this - gyro input & smoothing for each of 3 axes, accelerometer (same), GPS location at approx 10Hz, optionally barometric input (rare and usually not necessary). These inputs are often processed through kalmann filter before finally calculating the torque command to issue to each ESC's microcontroller. This only happens at up to a few hundred Hz so you can see compute power is not a restriction. In larger aircraft it's harder because you can't beat physics, in a small drone i can change the amount of torque applied to a propeller 4+ times per rotation in toy off the shelf drone tech today. With larger craft that have non-electric motors this is simply not possible. You can't instruct a mechanical engine to regulate its power output with the same fidelity of an electric motor. >> worse sensors The only required input sensor is a gyro - in practice we use accelerometers and GPS in autonomous drones but there are ways of avoiding the need for these (e.g. visual inputs / optical flow as a correction for drift). MEMS gyros are cheap and effective, what better sensors do you need? Your reasoning is uninformed / dreamed up based on watching too many movies and not being familiar with the actual discipline. Better algorithms are a restriction. The science / theory part of implementation was achieved by all projects fairly quickly (ardupilot, clean/betaflight, etc. etc.) - but the art part, the how it feels and performs in practice, is an unsolved problem under continual development. >> drop Grenades Too heavy given current battery technology and lack of aiming. These are not effective weapons which is why they have limited success. More people have been killed in war zones in the past decade by garden spades than toy / photo drones with explosives. TL;DR until batteries technology moves forward by a few generations, the concept of a small drone as an offensive weapon will not be better than that offered by something like the switchblade drone. Which when you cut through the excitment, is less valuable than conventional weaponry in most fight scenarios.
- lazide 4y ago
  FYI, your argument refutes itself. Compute is not a limiting factor in a small drone for basic flight operations, it is a limiting factor for real-time optical guidance (aka in GPS denied areas and real-time object/target recognition). Which is exactly what I was referring to. Those examples yes are not showing 100mph. I’m afraid the video I’m referring to has been pulled, and the in-person demonstrations I’ve been asked to not talk about or publicize (now that I follow up with the parties involved). The battery technology is more than sufficient for what I’m describing. My M300RTK has no issues carrying 5+ lbs for 15 minutes or 1-2 lbs for 40 minutes, which is more than sufficient, and scale is more than adequate here. You’re a couple years behind on this assessment. I literally have a 10 lb ‘hunter killer’ autonomous anti-drone drone from a company I’m not going to name sitting a few desks away from me right now.
  - CraigJPerry 4y ago
    >> I’m afraid the video I’m referring to has been pulled, and the in-person demonstrations I’ve been asked to not talk about or publicize Interesting. Let's take a second to take this from back of a fag packet physics. To produce a 100mph drone is easy - lets take one of mine (there's nothing special about it) as a starting point and then lets see what we can change to achieve load carrying while retaining this speed. Each motor & prop combo produces 1.5kg of thrust. They do this while drawing 40amps each from a 6s - which they can sustain for up to around 10 seconds before burning out MOSFETs on the ESCs (not to mention deforming the battery plug due to 160amps passing through a 60amp connector). This craft has a weight of just under 350 grams and when coupled with an 1800mah 6s, has an all up weight just under 650 grams. We want to carry "several pounds" - lets make it easy, lets say 2lbs - and retain that 100mph speed. Adding 10 grams prevents us hitting 100mph, we need to scale up our thrust. First problem - there's no other prop pitch or blade count for this motor size which will increase thrust while retaining prop speed, so we need to move up a size in motor. Since we need to carry 2lbs, you might think we only need each motor / prop combo to produce at an extra 0.5lbs but in practice we're going to need more than this because we need to sustain the aircraft vertically, whilst also providing enough thrust horizontally to reach the target speed, so in practice we will need more than an extra 8lbs of thrust (remember how 6kgs of thrust is needed to carry a 650g craft to 100mph - you don't just get to add 2lbs of thrust and call it done). Problem - we need a motor / prop combo that produces above 2.5kg thrust each - so we're into 14 inch propellers and we need a motor heavy enough to take around 60amps at 6s (i'm being conservative to try and make the maths work out in your favour but i hope we can both agree that only 60amps is pure fantasy here). That's a new problem, our weight went up and so we need even more thrust to compensate. That's another new problem, we have more thrust but now we need a power supply capable of a higher discharge rate - that kind of battery entails even more weight. But now there's a new problem, our larger props mean we're well away from the happy 300-400mph tip speed and now we're approaching the speed of sound - which of course means we can't fly. But now there's a new problem, the battery that can supply that discharge does exist, the only problem is it weighs almost 2kg by itself. And so we're back to you can't fly 2lbs payload at 100mph on today's battery tech. Camera operators have been trying hard for years at this point to put a RED or some other expensive heavy camera / lens combo right next to race cars on track. There are two solutions for that today: 1 - fly the heavy lens and get that sweet image but do it at 20mph or 2 - keep up with the 100mph+ action but film it on a gopro or similar.