Topics Related to Analysis of Motor Vehicle Collisions
- Site Admin
- Posts: 1598
- Joined: Thu Jun 18, 2009 12:37 pm
Recently posted on NAPARS Facebook page
- Today's History Moment: RICSAC Tests: In the 70's, it became clear that computer models of car crashes might be viable, but everyone needed good input data to work with. So a series of staged crashes were run called the "Research Input for Computer Simulation of Automobile Collisions", or "RICSAC tests" for short. They are a little dated now. Many of the people doing this work weren't born when they crashed these 1974 cars into each other, for instance, but physics is physics. If you have the data on the cars (dimensions, weights, inertias, tire parameters, etc), any new model should still fit their results. Sadly, it's not that simple, because some of the accelerometers weren't at the CG, and other issues of actual testing that still leave a little wiggle room when analyzing stuff. Modern software companies still use these tests to demonstrate their ability to get accurate results. So when you see "RICSAC" know you're headed for crash test comparisons.
- Having delved deeply into RICSAC in our 1997 SAE paper to adjust the impact speed changes by accelerometer location i still believe RICSAC serves as a good test course for software BUT one must publish ALL inputs and how they determined each (since knowing the results makes "validation" easy. SMAC has been validated many times and we validated SMAC blind in 2003 (with +/110% results).
Please be careful calling scientific data 'old' (and some of us involved in RICSAC are still around like Ray! and i as an interested college kid))
Realize Sir Isaac's Laws is a wee bit dated too but still holds true (science is science!)
The Recon community has such a growing body of full scale tests with varying amounts of documentation available. It would be nice if there were standards made to record ALL information for each test like as done for the RICSAC tests (and technology so much better now!) like recodrding any post impact steer and braking to keep in test area? record it! likewise make public all the information in some form of public 'validation database'.