backtop


Print E-mail del.icio.us 20 comment(s) - last by Diesel Donkey.. on Dec 16 at 5:33 PM

Virtual dummy will be the most complex virtual human model ever created

Crash testing vehicles to ensure the safety of occupant restraint systems and the overall design of the vehicle is nothing new. Automakers and the U.S. Government regularly crash test vehicles to study how protected the occupants will be in an accident.

One thing that many of the crash tests have in common is the use of crash test dummies. These dummies are used to get an approximation of how a human would be affected in a crash inside a vehicle. The test results give automakers an idea of how occupant restraints work and how likely drivers are to survive a crash in the vehicles.

One of the downsides to crash testing with dummies is the cost and the fact that dummies are much less complex than humans are. Each crash test dummy costs somewhere in the range of $5,000 to $100,000 and only come in three sizes that approximate people. The average life span for a dummy is only ten years and the dummies must be repaired after each crash test.

A group of automakers and auto suppliers have banded together to form the Global Human Models Consortium (GHMC) with the goal of funding the development of a very complex and very realistic computer model of a human to use in virtual crash testing.

Two teams of engineers from the University of Virginia Center for Biometrics will play major roles in the development f the virtual dummy. One of the engineers, Richard Kent said in a statement, "Already, cars and their safety systems are designed on computers. It's logical that we would create a virtual crash test dummy that would allow us to test these safety systems before they are ever physically built."

Kent is heading a six-member team that will be responsible for creating the highly detailed and realistic model of the human thorax and upper extremities including the ribcage, ligaments, heart, and lungs. Another engineer from the University, Jeff Crandall, will be leading another team that is focused on developing the model for the virtual pelvis and lower extremities.

Each team was awarded $3 million to complete the models by the GHMC. Researchers at six additional universities and institutes are working on models of other body parts including the head, neck, and abdomen.

One researcher on the project at the University of Virginia, Damien Subit said, "Eventually all of these models will be joined together to create the most sophisticated and lifelike simulation of the entire human body ever assembled for safety testing."

Not only will the virtual dummy be much cheaper, basically free to use, after the development is complete; it will be much more accurate and the researchers say it will allow much more detailed crash tests to help make automobiles safer. Kent said, "This will be an adaptable, cost-saving system that will provide amazing insight to body injuries for improving auto safety."

The researchers also see uses for the virtual dummy models outside of automotive crash testing. Kent believes that the detailed dummy could be used in the design of safer sporting goods and in medical schools for students to use in studying trauma.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

We should just
By Cuddlez on 12/12/2008 4:14:53 PM , Rating: 2
use people on death row in crash tests. Way more reliable than a crash test dummy. And cheaper too because they don't need to be repaired...

But seriously, so this would completely replace physical crash tests? Or is it just to be used to fine tune the safety system so they don't have to do as much physical testing?




RE: We should just
By Murloc on 12/12/2008 5:17:59 PM , Rating: 2
I think not, but with simulated crash tests you can collect much much more data about the impact and the structural consequencies on the car and the dummy, and you can test how would the car look like after falling from a 100 meters height, because it's fun.
Real tests for the verifications will be always there I think.


RE: We should just
By HostileEffect on 12/12/2008 6:48:19 PM , Rating: 2
I hope one day we can just load up a computer test and have it tell us everything we would ever want to know. Do you really need to shoot a rocket at a superb tank just to figure out that its going to blow up?


RE: We should just
By foolsgambit11 on 12/12/2008 7:07:27 PM , Rating: 3
I'm sure NVidia(tm) PhysX(c) will make it possible. Or, at least, you'll get to see persistent shrapnel and billowing smoke, allowing a more immersed virtual experience.


RE: We should just
By LRonaldHubbs on 12/14/2008 8:03:17 PM , Rating: 2
This isn't a replacement for real-world testing, it's just a supplement for early on in the design process. By simulating crashes they can test their overall design and fix any glaring problems before building it. Then after the vehicle is built they still have to do real-world tests to catch problems they didn't consider in the models. It doesn't matter how sophisticated the software is, it can't account for things that are unexpected or have yet to be witnessed in previous tests -- things poor welds or other assembly errors. Real-world testing is still needed to find those problems and won't be replaced any time soon, if ever.

As for blowing up a tank to test a new rocket, they absolutely will continue to do so, even after the automotive industry hypothetically phases out realworld testing. The military is far more stringent about QA than any civilian/commercial organization.


RE: We should just
By Diesel Donkey on 12/12/2008 8:21:07 PM , Rating: 1
I think a parallel could be drawn to jet engine testing.
Based on what a higher up at Boeing who visited my physics department a few years ago said, if this becomes anything like jet engine testing, then you run a whole bunch of simulations before doing the actual physical testing. This must be done in the jet engine industry because a jet engine is a wee bit expensive to destroy. So, the folks at Boeing do enough (months and months of) finite element analysis testing with physics simulations that they pretty much know exactly what will happen before they launch wild turkeys or whatever into the turbines for real. Granted, a $10k dummy is a lot cheaper than a ~$10m jet engine, but the same idea applies.

Actually, I'm more excited about the possible uses outside of the auto safety testing industry for such realistic human models. I'm imagining much more comfortable chairs and mattresses.


RE: We should just
By wordsworm on 12/14/2008 8:29:28 PM , Rating: 2
Well, considering that a single jet can have up to 800 dummies in it, that's $8,000,000-80,000,000 they could save on airline crash tests.

Imagine if the military gets their hands on these, they'll be able to test the effectiveness of their weapons in simulations.


RE: We should just
By Diesel Donkey on 12/15/2008 3:06:06 PM , Rating: 2
quote:
Well, considering that a single jet can have up to 800 dummies in it, that's $8,000,000-80,000,000 they could save on airline crash tests.


I'm not sure if you're arguing against my point or simply making another, but I was actually referring to the destruction of a single engine as opposed to the destruction of an entire plane.


RE: We should just
By wordsworm on 12/16/2008 7:39:36 AM , Rating: 2
quote:
I'm not sure if you're arguing against my point or simply making another, but I was actually referring to the destruction of a single engine as opposed to the destruction of an entire plane.


You seemed to be undermining the potential expenses saved by impressing us with a more expensive device. However, since the number of dummies inside an aircraft numbers in the hundreds, I'm arguing that by using the virtual dummy, they're saving an equivalent amount of money. It was a bit of tongue in cheek really. But here's a link to a vid: http://www.boingboing.net/2004/10/17/plane-full-of...


RE: We should just
By Diesel Donkey on 12/16/2008 5:33:59 PM , Rating: 2
quote:
You seemed to be undermining the potential expenses saved by impressing us with a more expensive device.


Quite the opposite, really. I was just trying to draw a parallel to another industry in which money is saved by using computer simulations a lot before actually breaking stuff.

quote:
I'm arguing that by using the virtual dummy, they're saving an equivalent amount of money.


Ah, gotcha. That's a pretty amazing video, by the way. If you could do that test in a computer simulation with 3 jet engines AND all those dummies at once, then you'd really be saving some serious money!


RE: We should just
By Bruneauinfo on 12/13/2008 1:02:19 AM , Rating: 1
you can crash a new part and find out it's flawed before you waste money tooling and building it.


virtual dummies
By Dreifort on 12/12/2008 3:32:21 PM , Rating: 5
"Sir, I'd like to show you our newest model, the J10... it has top speed of 210 mph, 0-60 in 4.9 secs and has the nations top virtual safety ratings by the virtual safety industry. So it virtually saves you from almost any wreck."




RE: virtual dummies
By Motoman on 12/12/2008 3:42:23 PM , Rating: 4
Sweet. I virtually want one.


RE: virtual dummies
By japlha on 12/12/2008 4:48:09 PM , Rating: 2
Why not just use convicts sentenced on death row?


RE: virtual dummies
By Motoman on 12/13/2008 1:26:46 AM , Rating: 2
Thank you, Dr. Mengele.


Realistic gore
By johnadams on 12/12/2008 7:23:55 PM , Rating: 2
I can imagine computer games of the future (GTA VI?) borrow this technology and incorporate it into their human models. More realistic gore!




RE: Realistic gore
By icanhascpu on 12/13/2008 5:50:14 AM , Rating: 3
More realistic whore!


A few corrections...
By puckalicious on 12/13/2008 10:29:15 AM , Rating: 2
Virtual crash test dummies have been in use for many years already, but the important difference here is this new model will more closely model the human body. The existing models closely model the physical crash test dummy.

This is new and exciting research, but that's all it is for now since legal requirements use the crash test dummy. The hope is that this research leads to more detailed physical dummies, much like human cadaver research has lead to the current understanding of human injury.

This will not change how cars are engineered for production probably for quite some time. But I think it will lead to a more detailed understanding of how certain injuries happen in the field. UVA is known for very good research in human injury relating to car safety, keep up the excellent work guys.




Imports and Startups
By abscoder on 12/13/2008 7:00:01 PM , Rating: 2
I hope this ends up completely replacing real crash tests.

There are a few vehicles I'd like to import from Europe, but it's too much of a hassle (not to mention expensive) because the manufacture hasn't submitted their models to US crash tests. I can see this expediting the entire road-legal certification process.

Also, it could help start-up automakers by removing the expense of destroying real cars of limited production.




could go either way
By jlips6 on 12/13/2008 11:04:10 PM , Rating: 2
this could be used to benefit mankind by allowing people access to newer, more specific data, or it will be used for pornography. It's 50-50




"I f***ing cannot play Halo 2 multiplayer. I cannot do it." -- Bungie Technical Lead Chris Butcher











botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki