Steve M Ashton

06/02/2022

There are openings in the Microsoft Excel for Crash Reconstruction class in August. Please sign up asap! Sign up at https://rich-llc.com/Training/Excel%20for%20TCR.html

06/02/2022

We had 20 in the I.P.T.M Motorcycle Crash investigation with students coming from as far away as Tennessee, Cleveland and Quebec, Canada. Excellent class! Mr. Sloan did a great job teaching and is very knowledgeable in the field of motorcycles!
Thanks to all who attended and special thanks to Cincinnati PD for providing us with 2 motor officers! Looking forward to Level II class!

04/19/2022

Next OTARA training is IPTM motorcycle crash Level 1 from May 23-27, 2022. Sign up at www.otara.org or www.iptm.org asap.

10/27/2021

NTSB and the "No Driver" Tesla crash, Spring FL.

On Saturday night, 17 April 2021, a 2019 Tesla zoomed around a 100-ish foot radius bend in a concrete cul-de-sac a couple hundred feet after leaving a residence. The road had a 30mph speed limit. The car failed to negotiate the turn, ran wide, hit a tree, and burned. And then burned some more. The attached diagram is from the NTSB report, showing the separation lines in the concrete roadway blocks drawn over an aerial photo. The location is here: https://goo.gl/maps/DcExxZ12gvuE2hz97

First responders found nobody in the front seat. Neither occupant survived. The tire marks reported by NTSB look to be following about a 320-foot radius. Recon-types reading this should already be thinking, "what speed can you negotiate a 100-foot radius at?" or perhaps "How fast can a car cut a 320-foot radius path?" The relationship between lateral traction used, path radius, and speed is the old Critical Speed equation: Speed = 3.87 * SQRT(radius * friction) I leave poking the buttons as an exercise for the reader at this point.

The NTSB inspection showed the steering wheel ring bent forward, as from an impact by an occupant.

The NTSB reported that they downloaded the car's burned EDR with the manufacturer's help. The data showed both front seats occupied, and the driver pressing the accelerator pedal up to 98% in the seconds before the crash, and speeds in moments before of up to 67mph.

Residential video shows the guys climbing in at the start of the trip, just before driving away.

So the occupants seem to have wound up in the back as a result of the jostling during the melee. Tesla appears at this point to be exonerated in this one, at least with regard to the "self driving", which was not involved.

Occam's razor still works: Rather than a multi-faceted failure of several safety systems AND a bad decision by the driver (to sit in the back and let the car do its thing), the truth is simpler: The driver made a bad decision to step on the throttle, and got to the corner faster than he expected at a speed much too high to negotiate the necessary path.

The 320-foot path, with lateral traction of 0.8g gets a predicted speed of about 62mph. Yep, physics works.

For all the public details, google "NTSB HWY21FH007"

Peace. -W

10/13/2021

09/21/2021

COLM and motorcycles.

Motorcycle crashes are subject to the laws of physics, and so momentum is conserved, BUT tire forces may be more important here than in most car-to-car crashes. More importantly, there is often a huge momentum disparity between a light bike and a heavy car. This means that some numbers you might try to calculate will be very sensitive to some of the other numbers you have.

One difficulty with COLM here is we need the weight of each vehicle, but if a rider separates from his motorcycle, engages the car somewhat, but goes off on his own path, we have to take that into account. We might be able to use a two-in-three-out sort of analysis if there's enough scene evidence to get his departure (vault) speed.

Failing that, we might consider ranging the effective weight that the rider/motorcycle brought to bear against the car. The low end of the range is the weight of the motorcycle alone (assuming the rider completely misses the car). The high end of the range is the total weight of rider and bike (assuming the rider hits the side of car and stops with the bike right there). Some authors over the years have suggested using half the rider's weight, but there is no basis for that, as far as I can tell, other than to say "It's within half of the rider's weight of being correct".

In some configurations, you might choose to run both of those bounding analyses, allowing one to say the speed was between V-low (when the rider's whole weight is involved) and V-high (when the rider just flies over the top and doesn't touch the car).

In my experience, the issues of "beyond a reasonable doubt" often make this sort of analysis barely useful in prosecution matters - it's good for helping understand what happened, but probably a little too fuzzy to use to convict someone, unless V-low is itself way over the speed limit. This technique might be good enough for a civil matter, though, where we're only working to the "more likely than not" or 51% standard.

Peace. -W

09/12/2021

Have we discussed the equal-spacing "requirement" for crush measurements lately? Though 6 equally-spaced measurements is the traditional data needed for the old CRASH3 software, it's not a mathematic necessity. The 4n6xprt team explored the whole matter in a presentation at the 2016WREX conference, and it's available on their website here:http://www.4n6xprt.com/WREX2016-4N6XPRT-poster.pdf
Peace. -W

08/31/2021

Today's Mathy Follow-Up to Yesterday's Quadratic Equation Discussion: Skinning the cat another way.

Iterative Solutions.

Sometimes we want a value that we can't just solve for in one equation. By that, I mean the equations available are generally meant to calculate from the thing we want to the thing we measured, not the other way around. Sometimes we can adjust those equations algebraically into something tractable (such as a quadratic form that we solved yesterday), but sometime we can't even do that.

Very complex systems of equations often have no "closed-form" solution, like when we use a dynamics modelling software package like MADYMO or one of the SMAC/CRASH3 variants.

Another option is to just propose a value that we think will be kinda close, solve the equations we have (or run the program) to find the thing we can measure, and see if our proposed solution (notice I didn't say guessed) would result in the right value. If not, adjust the proposed value and try again until we get close enough.

An example: Say a car turns left across a rider's path, and he flies cleanly over the hood, hitting the road and sliding some distance, coming to rest a total of 70 feet down range. We can use Searle, with the total distance (dtotal) or we can use a vault and slide, where we know dflight plus dslide = dtotal. We don't know the landing spot, but we know his departure speed from the vault is approximately equal to the speed from slide-to-stop (air resistance is negligible here). Propose a solution, and then refine the proposed value until we get a match, as shown in the example in the photo.

Don't let someone sucker you into calling it "A GUESS", because it's not, this is a time-tested, scientifically appropriate technique to find a solution to a complex system when no closed form exists which would allow a simple single-step solution.

Peace. -W

08/27/2021

Today's Etymology moment (not to be confused with entomology) - Photogrammetry.

From the link below, this may well be the official Greek roots that make up the word "photogrammetry":
PHOTOS = light
GRAMMA = something drawn or written
METR0N = to measure

I'm no specialist in modern photogrammetry, having done very little with it in years. Back then, we actually used a wet-film camera that allowed us to drop little clear sheet with a traced scene in the viewfinder to line up landmarks on a scene so we could measure a mark that never got measured originally, but did get documented on film. It all sounds antediluvian or at least antebellum, right? But it's true. Things have progressed.

As early as 1999, some practitioners were (apparently) worried that photogrammetry was dying - turning into an exercise in "computer vision". An article taken from a keynote speech on the topic is here:https://www.asprs.org/wp-content/uploads/pers/1999journal/jul/1999_jul_highlight.pdf

So, fb hive mind, is the article still relevant?

Peace. -W