Jump to content
Sign in to follow this  
gonk

Swarm Technology

Recommended Posts

This looks interesting...

Now all they need is to be self replicating..:butbut:

Share this post


Link to post
Share on other sites
I want one! Looks like something from Half life 2

Only one ? I want 50....lol I can use them to scare the neighbours cats...

Share this post


Link to post
Share on other sites

You would scare the whole neighbourhood, not just the cats

Share this post


Link to post
Share on other sites

Hi all

I presume that everyone went on to see all the other videos by the same source. If not then you should.

http://www.youtube.com/user/TheDmel

I pointed out in a previous thread that these will replace soldiers and rifles in war.

The singularity is near now.

Kind regards walker

Edited by walker

Share this post


Link to post
Share on other sites

Very nice videos

of course from gamers perspective it is funny to see the old stretching of truth , anouncer seems to imply realtime pathing where i think we can actually see its what we know as scripted AI .

I am sure they will get there , but maybe we can equate to FPS world how somethings can be done in a small area but when we get to world size there are a lot of penalties that must be overcome. .

i will keep following this its very interesting especially as i have only recently began to mess ith unit record and play functions in RV engine i can see so many similarity :)

Share this post


Link to post
Share on other sites

Hi all

In reply to Thromp. Sorry but I must respectfully inform you that you are wrong. The proof that pathing is real time is in the dynamic reaction to flying through the thrown hoop in one of the other videos. The figure of eight also requires dynamic pathing as it is subject to chaos.

Kind regards walker

Share this post


Link to post
Share on other sites

I think you have more confidence than me that the fly through hoop is not pre programmed ! I will wager the loop is added after the flight path is programmed at this early stage too , but even so I think with the cameras and tech scanning the room , I believe they can achieve it , however the demo is definitely a pre determined flight path , like I said people who code games have been pathing such manoeuvres in x,y,z for years .

As for figure 8 well it's the best case of pre determined pathing I saw :) .

Edited by Thromp

Share this post


Link to post
Share on other sites

now, make these flying things to operate hours or days, slap on them some weaponry and make use of theirs rapid maneuvering abilities to evade incoming projectiles ...

await ARMY order major number ;)

Share this post


Link to post
Share on other sites
Hi all

...

I pointed out in a previous thread that these will replace soldiers and rifles in war.

The singularity is near now.

Kind regards walker

I'll need to see them working in a real world environment with wind, rain, temperatures, mud, dust, em scatter and shockwaves bouncing them around before I'll accept that sort of prophecy.

Ive seen a mild wind take my micro helicopter and blow it clear across the garden in a few seconds and that weighed about as much as the entire "swarm".

Share this post


Link to post
Share on other sites

there are other videos from same developer

showcasing how these quad-micro-copters can recover from rapid deployment

(throw them into air upside down or with rotation with and without enabled engines)

other videos show rapid and adaptive maneuvering even via and around moving objects

ofcourse that's all inside 'room and hall sized area'

but if there are good sensors onboard (active/passive)

then i assume it may work outside just fine

Share this post


Link to post
Share on other sites

They should make them bigger and attach them a machine gun or something,securing your house would have a new meaning.

Of course in that case pray that they have really good sensors and they won't pew-pew you down by mistake or start stalking you around the house.

Share this post


Link to post
Share on other sites
there are other videos from same developer

showcasing how these quad-micro-copters can recover from rapid deployment

(throw them into air upside down or with rotation with and without enabled engines)

other videos show rapid and adaptive maneuvering even via and around moving objects

ofcourse that's all inside 'room and hall sized area'

but if there are good sensors onboard (active/passive)

then i assume it may work outside just fine

I have seen the same videos. I still don't think it will work. I'm not a programmer, just a Production Design Engineer but after 10 minutes thinking I can see a lot of issues to overcome before it could come even close to being practical with today's technology.

Unless of course they come up with some ultra light weight/long endurance batteries and equally (and unbelievably) efficient rotors that can counteract the force of gusts of winds and shock waves etc. And let's not forget the obvious need for either: a robust high-speed secure data link or an onboard - and artificially intelligent - processor capable of path finding and deciding what is important vs what's just and obstacle. Then we have the need for a storage device with such a low power draw as to allow sustained flight and real time/stored programs and video.

Revisiting the control issue for a moment; the real time vs. programmed path question.

Real Time then we go back to the power issues around signal integrity for data links. Most high speed sensor links require quite a lot of power to send high bandwidth signals over any meaningful distance. Real time links require constant transmission = lots of power draw we end up back at the ultra light weight high-power cell issue.

The power of the transmission can of course be compensated by having more powerful relays. E.g. an orbiting UAV (ala Bin Laden raid - RQ-170 anyone?) to receive and send on the signal via a secure high power data link.

Pre Programmed Control This is even more challenging. The hardware is less power hungry but its far more complex due to the control method. Essentially the flaw in the plan is the very complex computing issue and again power. To do this with any sort of chance of it working you'd almost need an AI contained in each nanobot or split across the swarm. Path finding, obstacle detection, sensor management, swarm path and positioning (as much for general movement as for composite sensors or imaging). Something I'm sure would be possible on a very small chip. But again it would need a lot of power and additional sensors to run.

Processing Power - I did read some time ago about distributed or at least composite imaging. ie taking lots of low resolution images, taken by a "swarm" of sensors, optical and EM etc being pieced together to produce a usual high res image. But it took tremendous processing power to piece all the images together in a 3d matrix. And that was for static sensors. To do the same for a swarm of nanobots all charging around the sky? Well that's going to be fun. Especially if you want it in real time imagery from composite optics.

The Power draw issue - Lets just use the average modern cheap lightweight cell phone as a standard - purely for the purpose of discussion. I know that military spec systems are better but for the moment lets stick with a common item we can all relate to. Let say it has a range of about 5km at best in ideal conditions on a low bandwidth voice only call (I'm guessing here because I'm sat in a train station and dont have a good wifi or Mobile net connection). We need something a bit better if we're sending real time control and video back and forth. Lets say 2G/2.5G type bandwidth for a very low res image and basic control. For a reliable robust connection, let's say 3km max range. For a lot of data traffic you're constantly transmitting, drawing a lot of power. The average phone sending at 2.5G speeds and faster runs out of juice pretty quick (My old blackberry 8310 could manage about 35mins when used as a modem or as a GPS device. My 9800 Torch with a new battery lasts about 70 mins on the same load - but they are large-ish and weighty - in nano terms - items). Add to the demands of the motor, computing power to co-ordinate the control inputs, geographic location GPS/INS finding/positioning. Position keeping within the swarm etc. Any battery isn't going to last long.

Ask any RC pilot how long a high performance Li-Ion battery lasts in your average electric RC helicopter? Mine lasts about 8-10 mins tops. Even with a high performance military spec power cell you aren't getting much more than that.

Now, Environmental conditions. - This is my main concern. I've raced/flown RC cars and aircraft around for about 28 years in one form or another, in all conditions.

I even have a petrol helicopter equipped with a wifi camera. (An old Nexus 30 if anyone it interested) that last max 12 mins flying time.

The one constant in all that time is that the environment/weather can seriously mess with your kit. Dust/dirt gets in, things get broken humping all the kit around just to set up. Unexpected high winds take your aircraft outside of your control range in seconds. Water ingestion kills any and all kinds of motors dead no matter how much you weather proof it. And as any squaddie will tell you the more complex the "toy" the harder it is to use in any practical way

I doubt it will happen for a long time to come if at all. Even then I doubt they will be nano-anything or large particularly large swarms.

Edited by RKSL-Rock
typos and a small addition/clarification

Share this post


Link to post
Share on other sites

there are already quadcopters capable of flight time between 30 to 300 minutes (i remember some doc about it being tested for police and traffic issues overwatch in germany)

that includes flight in wind up to 100km/h with stabilized camera and ability keep within 50x50x50 meters cube

take in mind what you see are micro concepts so they more use in urban environment and considering how nanotechnology progress

(already some experiments with batteries on 10x capacity of best on market)

i can see utilization of these drones to e.g. find the enemy, usage of 'drug' incapacitation darts or just info gathering ...

you deploy swarm, lose some yet rest will still obtain the info and may even disrupt enemy perimeter by taking out some guards etc

Share this post


Link to post
Share on other sites

Bad thing about robots replacing humans in war: robots have no emotion.

Share this post


Link to post
Share on other sites
Bad thing about robots replacing humans in war: robots have no emotion.

LoL, I think a robot would be more compassionate with you in war then a human :biggrin:

Share this post


Link to post
Share on other sites
there are already quadcopters capable of flight time between 30 to 300 minutes (i remember some doc about it being tested for police and traffic issues overwatch in germany)

that includes flight in wind up to 100km/h with stabilized camera and ability keep within 50x50x50 meters cube

take in mind what you see are micro concepts so they more use in urban environment and considering how nanotechnology progress

(already some experiments with batteries on 10x capacity of best on market)

i can see utilization of these drones to e.g. find the enemy, usage of 'drug' incapacitation darts or just info gathering ...

you deploy swarm, lose some yet rest will still obtain the info and may even disrupt enemy perimeter by taking out some guards etc

Yes Ive seen the same. But they all weigh between 300g and 2000g. Considerably more than the cute little things we're seeing in those videos.

Speed = higher power draw = less endurance

Liverpool's Police have a nice little quadrotor too.

air-surveillance-drone.jpg

Its nearly 1m across. Again not nano scale.

I've seen it demonstrated at a model meet in the GMEX in Manchester. The station keeping feature is impressive but the power drain is high. The demo only lasted for less 10 mins and they had to take it down the wind (in an urban environment) was too much. Much over 7kts wind and it becomes too dangerous to use over the public. The video resolution was also quite poor by modern standards. 640x480 at 15fps with 10x digital zoom. About the same as my £380 unit.

They use (very expensive) milspec Li-polymer batteries which weigh about 400g each. Also again not exactly nano sized. My use of a common cell phone battery was an ill explained attempt to compare/estimate the power a nano size mil-spec battery to the larger types you refer to. (Hey I was on a crowded moving train when I wrote it I missed some bits give me a break :P)

What do you think the endurance of a cell that could fit in 5-7cm a nano-quadrotor is?

They are only relatively short range about 750m max (from the transmitter) and use a similar wifi system for video sending to what you can buy in most electronics stores. There is a system in the UK that uses the mobile phone network instead though.

On the subject of micro size... take another careful look at that video. Especially the walls, the white material, the IR cameras? Take a close look at the swarm, notice the LEDs? It's a motion tracking system (similar principle to TrackIR) that controls the swarm. None of the computing or sensor control is actually on board the 'bots' in the swarm. Outside of this controlled environment they aren't going to work as they are now. They need onboard control systems, positioning sensors, transmitters etc before they can be even semi autonomous.

Now if you scale it all up to fit all that kit on it with today's tech it stops being nano. Which is what i said before.

I doubt it will happen for a long time to come if at all. Even then I doubt they will be nano-anything or large particularly large swarms.

EDIT - http://www.microdrones.com/produkt-md4-1000-behoerden-en.php

Thats the bunny.

[h=1]Technical Specifiactions UAV md4-1000[/h]

[TABLE=class: full]

[TR]

[TD]Climb rate[/TD]

[TD]7.5m/s[/TD]

[/TR]

[TR]

[TD]Cruising speed[/TD]

[TD]15.0m/s[/TD]

[/TR]

[TR]

[TD]Peak thrust[/TD]

[TD]118N[/TD]

[/TR]

[TR]

[TD]Vehicle mass[/TD]

[TD]approx. 2650g (dep. on configuration)[/TD]

[/TR]

[TR]

[TD]Recommended payload mass[/TD]

[TD]800g[/TD]

[/TR]

[TR]

[TD]Maximum payload mass[/TD]

[TD]1200g[/TD]

[/TR]

[TR]

[TD]Maximum take-off weight[/TD]

[TD]5550g[/TD]

[/TR]

[TR]

[TD]Dimensions[/TD]

[TD]1030 mm from rotor shaft to rotor shaft[/TD]

[/TR]

[TR]

[TD]Flight time[/TD]

[TD]up to 70 minutes (dep. on load/wind/battery) <------ The dependancies are very telling.[/TD]

[/TR]

[TR]

[TD]Battery[/TD]

[TD]22.2V, 6S2P 12.2Ah or 6S3P 18.3Ah LiPo[/TD]

[/TR]

[/TABLE]

Edited by RKSL-Rock

Share this post


Link to post
Share on other sites

ok. Is a more "serious" version confirmed for ArmA 3 yet? :p

Share this post


Link to post
Share on other sites
Bad thing about robots replacing humans in war: robots have no emotion.

Good thing about robots replacing humans in war: robots have no emotion :)

Share this post


Link to post
Share on other sites

Great piece of technology, reminds me of the '"Fix-it" Robots' from Batteries not included. :p

Share this post


Link to post
Share on other sites

This is cool. Finally came to life. I remember hearing about it years and years ago, and something about making literally a robot fly for the purpose of getting in almost any opening. Nice find.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×