Train Collision

Accident Reconstruction Case Study

A mechanical pick-up truck driver is struck by a train as it crosses the railroad tracks in a rural part of the Midwestern United States. A number of factors were seen as possible causes, including vehicle speeds, road conditions and visibility. Kineticorp team members conducted on-site testing and analysis, and produced several scenarios showing how this crash occurred and how it may have been avoided.

Video Transcription

Neal Carter: As a train approached in this case, a pickup truck was crossing the tracks. This was an area where there wasn’t a gate, or there weren’t flashing lights to indicate when a train was coming, kind of a rural area.

William Neale: It’s just a packed earth roadway, with a slight grade to the hill, and then the train tracks, of course, sit higher and go across. In this crash, there is video from the front of the train, so you can actually see the utility truck come into view, so you can analyze the video to at least determine what a relatively unarguable fact about the case. What was the visibility of the train? What was the train’s visibility of the truck? What actions, if any, did the truck driver take in avoiding the crash? The video was able to help answer some of those questions. One of my colleagues, Seth Miller, and I went to the scene. We were there to obtain a couple of things. One, we needed to obtain scan data of the scene. The scan data allows us to build a computer environment of the entire area, and we use that computer environment to analyze the video.

Seth Miller: We thoroughly document the area. We 3D scan up and down the train tracks, took a ton of photos, but there was a particular tree stump in question that people were saying, “This tree was there at the time of the accident.” When we go there, it’s just a stump, and that was kind of the question. Did this tree block this guy’s view? After further analysis, we see that this stump has been a stump for years. It’s all decayed on the inside, so that wasn’t the one in question.

William Neale: We went back to photographs taken by the police, and you can see the same stump that we found when we were out there. The police had photographed. In fact, the other experts that were saying that that stump had been cut down had it in their photographs too, so clearly, this stump was there prior to the crash. One of the reasons we wanted to do deceleration testing is to figure out what a typical vehicle, or in this case, the mechanical truck, how quickly could it have stopped had it applied full brakes. We did deceleration testing at the scene, where we set up both cameras and a data acquisition device, called Harry’s LapTimer. It measures your position over time, so you can calculate speed, and then of course you can calculate change of speed over change in time, which is going to be deceleration. What we found from the testing is this is not a concrete or an asphalt paved surface, that has a bit more of a grip to it than the packed earth, so it takes longer to stop on this surface. But even though it takes longer, there’s still some available friction, and given the speed that we know this truck was coming in in the video, and the available friction that the driver had to stop, had he applied full brakes, he definitely would have avoided the crash.

Seth Miller: We just wanted to see, if the guy was going the speed limit and slammed on his brakes, how long would it take him to stop. We did multiple runs to kind of get an average of that. Once we tracked the video, and we obtained train speeds, and vehicle speeds, and we established line of sight when the driver of the truck should have been able to see the train, with that information, we were then able to test out different scenarios.

Neal Carter: I took the motion of the pickup, and I simulated that motion, and that played a key role in this, because we knew the location of the truck through time from the video, but we wanted to see, okay, is that consistent with a driver braking as he approached the tracks? What we found was it was consistent with very light braking or no braking at all, even just kind of coasting. It was a pretty heavy truck, so just coasting, that would decelerate quite a bit, which is an indication that the driver didn’t even look to see the train coming, unfortunately. In one scenario, we figured out when the driver of the truck could have seen the train, and gave a one-and-a-half second perception response time, and braked the truck fully, and we found out that the truck stopped about 18 feet short of the track. In another scenario, we said, okay, maybe one-and-a-half seconds isn’t enough to allow him to react, so we actually gave him a longer reaction time and full braking, and found that he could still stop short of the train tracks. In the last scenario, we gave the truck driver one-and-a-half seconds to respond to the train, and then found that he could actually still avoid the train even by lightly braking. In other words, had the driver seen the train, reacted to the train, and even lightly braked, he wouldn’t have crossed the train’s path, and ultimately, the accident wouldn’t have happened.

William Neale: The conclusion was we’re not even asking this driver to maximize the braking to avoid the crash. They just have to be reasonably attentive and apply reasonable braking, and they would still avoid the crash.

The Process:

Kineticorp’s William Neale and Seth Miller conducted a scene inspection at the location of the crash to gather additional evidence regarding the circumstances of this crash. Based on video evidence provided by the train company, they were able to determine both the pick-up truck’s speed and the train speed using two-dimensional pixel-tracking and three-dimensional object tracking. They conducted deceleration testing of an exemplar truck to determine how much braking would be necessary for the truck to stop before reaching the tracks, given the vehicle speeds and visibility through the nearby foliage.

The Result:

The testing and video analysis were used to show that given the physical evidence, aligned with simulations and visibility, that the driver of the pick-up truck likely never looked to see the train approaching as it crossed the tracks.

The Team:

William Neale, Seth Miller, Neal Carter

Related Case Studies, Content & Research:

Pixel Tracking Research
SAE 2016-01-1478 – Determining Position and Speed through Pixel Tracking and 2D Coordinate Transformation in a 3D Environment
SAE 2014-01-0464 – Vehicle Acceleration Modeling in PC-Crash
SAE 2013-01-0788 – Video Projection Mapping Photogrammetry through Video Tracking

Back to Top