With this year’s Crash Conference 3 fast approaching, we wanted to share with you an article written by Sam Kodsi, B.Eng, P.Eng., and Dwayne Toscano, B.Eng., about a crash test we did last year at OIAA’s Provincial Claims Conference. The article was published in the September 2015 issue of Without Prejudice and is reproduced below with permission.




When the OIAA’s Provincial Claims Conference organizing committee invited us to conduct a live crash test at their event in Huntsville, Ontario, we were thrilled about the opportunity to expand the crash test knowledge base and educate attendees. We would like to share a behind-the-scenes look at how we planned for the crash test, and share the data we gathered from it.

Why are we, as forensic experts, so interested in crashing cars? Reconstructing collisions as part of real world investigations requires the use of certain educated assumptions, and relevant crash test data can improve the certainty of experts’ conclusions.

Assumptions made during a reconstruction include the duration of the crash pulse (assumed to be an average of 100 milliseconds) and the restitution (elasticity) of the collision (assumed to be about 0.1 to 0.3). The purpose of this crash test was to compare the typical values of the crash pulse duration and restitution to assess their applicability to a common crash scenario. With the knowledge gained from this particular crash test, our certainty in evaluating and analyzing similar events can be refined.

Planning & Preparation

More than a quarter of real world collisions involve a turning movement, but very few crash tests have been conducted where the vehicles are offset such that they collide front bumper to rear wheel. We knew this was a gap we could fill while also conducting human subject research. We decided on a crash test design that would emulate an intersection collision between a through and a left turning vehicle, where the front right of the bullet vehicle impacts the right rear wheel of the target vehicle (see Photograph 1).

Photograph 1 and 2

Our plan, based on hundreds of hours of designing, analyzing, peer-reviewing, and simulating variations of the crash test, was for the test subject (Sam Kodsi) to accelerate the bullet vehicle (the silver Chevrolet) up to an impact speed between 45 and 50 km/h. Simulations predicted that the Chevrolet and the test subject would experience a rearward speed change in the range of 10 to 14 km/h. Based on available research, the likelihood of the test subject sustaining anything beyond a minor injury in the proposed crash test was ascertained to be less than five percent. The test subject was informed of the probability of injury and provided his informed consent to proceed with the crash test. As part of the guidelines on human subject testing, the crash test driver could withdraw consent at any time prior to the crash test.

Test Day Preparation

The first order of business was to run acceleration and braking tests to confirm that the Chevrolet could be accelerated up to the desired impact speed of 45 to 50 km/h. After several runs, Sam was able to consistently achieve a speed of just under 50 km/h. During one of these tests the Chevrolet’s black box recorded an event (due to extreme braking), which indicated a speed of 48 km/h and was in general agreement with the speed measurements from other instruments (high speed GPS loggers).

Crash Test Instrumentation

The Chevrolet and the test subject were instrumented with high speed accelerometers, GPS, and other data loggers to thoroughly capture the vehicle and occupant motion during the entire event. Outside of the vehicle, video cameras, including a high speed camera, were used to capture the crash test dynamics. The driver’s door of the Chevrolet was removed to give the video cameras an unobstructed view of the test subject’s motion during the crash. The passenger seat was removed to allow more convenient access to the black box and other instruments.

The Crash

GPS data and video analysis of the crash indicated that the impact speed of the Chevrolet was approximately 48 km/h. The black box of the Chevrolet indicated that the vehicle was travelling at 30 mph (48.3 km/h) at 1 second prior to airbag deployment. Figure 1­ illustrates the entirety of the vehicle’s speed before, during, and after the crash test. Figure 2 illustrates the crash pulse recorded by the accelerometers during the crash.


Due to the higher than normal accelerations experienced by the Chevrolet during the crash, the airbag control module made the decision to deploy the airbags only 30 milliseconds after the Chevrolet struck the Toyota. At the time the airbags were commanded to deploy, the crash had not yet reached its maximum severity. Photograph 2 (above) shows the airbag deployment during the crash test. The speed change experienced by the Chevrolet during the collision was approximately 12.1 km/h (as measured at the Chevrolet’s centre of gravity).

A summary of the data we obtained is illustrated here:



What We’ve Learned So Far

The following are some differences between our crash test and a typical vehicle-to-vehicle crash:

  • The maximum deceleration (braking) rate of the bullet vehicle on the interlock surface was about 0.7 g, which is less than the maximum deceleration of a passenger vehicle on a dry asphalt surface by about 10%.
  • The measured crash pulse duration of 76 milliseconds was about 25% shorter than the typically assumed average value of 100 milliseconds. The shorter than average crash pulse was due to contact between relatively stiff components on the incident vehicles (i.e. the bumper of the Chevrolet and the wheel of the Toyota).
  • Peak and average accelerations can be calculated using the crash pulse duration. The below-average crash pulse duration resulted in higher acceleration values than would have been calculated if the average 100 millisecond crash duration was assumed [Note: Research shows that peak and average accelerations during a crash are relevant to the crash severity, and subsequently, the potential for injury].
  • The elasticity of the collision (almost 0.5) was significantly higher than the typically assumed range (0.1 to 0.3) given the impact speed. In other words, the collision was “bouncier” than typically assumed due to the stiffness of the components coming into contact.

Every properly conducted crash test becomes another data set in the knowledge base of accident reconstructionists. The more data points researchers add, the more clearly and confidently we can delineate trends and values and apply them to real world investigations. The information from this test will be added to the pool of crash test databases that we rely on in our forensic work, allowing our experts to further refine their opinions.

Kodsi Forensic Engineering would like to thank the OIAA Provincial Claims Conference ‘s organizing committee for inviting us to conduct a crash test. We would also like to thank all who came to our event to learn about crashing for research. We invite anyone interested in our upcoming Crash Conference, where we will be conducting several live crash tests at various impact speeds and orientations, to go to our Crash Conference 3 registration page to find out more. We would love to see you on May 12th!


Share This
Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now