The Collision Detection Experiment in Analytics Lab

Submitted by Kim Thoman.

Collision Detection in the Analytics Lab

The collision detection experiment in the Geotab Analytics Lab reviews and confirms collisions that are automatically detected by the telematics GO device. Users can also use this tool to report collisions manually. Here is a link to “How to download and install Analytics Lab” in your Geotab database. After installing Analytics Lab, go to the Collision detection experiment and click try it.

Collision Detection – the Technical Stuff

Collisions are detected by the Geotab GO device upon any acceleration greater that or equal to 2.5 G, where G is 9.81 m/s2 (the acceleration due to the Earths gravity). This is classified as a collision-level event. Geotab GO device firmware will not use/Up/Down accelerometer data (also known as the Z axis) in this calculation. The calculation uses the magnitude of the hypotenuse between X and Y where X is Forward/Braking and Y is Side to Side.

When the GO device detects a 2.5 G event in any direction, the device is triggered to start recording at 100 Hz frequency. This detailed, high-resolution information will be reflected in the trip in MyGeotab, and includes acceleration, GPS speed, and brake data (if available).

About the Experiment

The Data Analytics Research team, using advanced analytics techniques, has developed a model capable of detecting collisions when they happen to present critical information on point of impact, magnitude to help users monitor, act, and take measures to mitigate future collisions.

How Does the Collision Detection Experiment Work?

Because it can provide an accurate scientific record of events, telematics data proves to be highly valuable for Fleet management. The model monitors the telematics data and applies what it has learned as signals in accelerometer data to detect and classify collision events. Due to the granularity of the data, the model can tell you many things about a collision such as a point of impact, trigger type, location, time, and much more. With this experiment, users could gain a broader perspective on the event by looking at the historical patterns for a specific driver or vehicle.

What is the Benefit to Fleets?

  • Fast and easy to use method to detect potential collisions.
  • Provides critical data to take Collision workflow decisions and actions.
  • Point of impact can provide a better understanding of the overall accident as it occurred.
  • Increasing overall fleet safety by unlocking patterns in-vehicle, locations, and conditions to reduce potential collision in the future.

For more detailed information on collision reconstruction, you can download this white paper from Geotab

Vehicle & Asset Telematics

Electronic Forms

Mobile Viewing

Routing

Vehicle & Asset Telematics

Electronic Forms

Mobile Viewing

Routing

Collision Reconstruction Limits Heartbreak Over Collision Damage

You can’t always prevent collision damage,

but you can prevent the heartbreak. Thanks in part to Geotab’s Collision Reconstruction add on, there is a happy ending to what could have been a very sad story.

I think we all have a vision of our dream car from our youth. For me it was always a little red rag-top. A couple of years ago I found her. She was a Chili Red Mini Cooper with a blue denim convertible top. Her name, chosen by her previous owner, was Rosie.

She had a lot of life left in her when I was recently hit from behind. The impact pushed me into the vehicle in front of me so I had collision damage both front and rear. As I dialed 911, I watched the other two very young drivers call their parents. Though the damage appeared superficial, my dream car was possibly going to be totaled, and ahead of me the negotiation with the insurance company loomed large.

Accessing the Collision Reconstruction Data

The first thing I did when I got home was fire up the the computer to pull the accident data. With the collision reconstruction add on, it took just a couple of minutes. The truth is often not exactly what we remember,and in this case, that was the case.

How could I have been so certain I was at a complete stop when I was hit from behind? The collision reconstruction data showed I was driving 6 mph and slowing to stop. I wondered in that moment if the other drivers had similar flaws in their memory of the event. Also, the data indicated an initial accident level impact at the rear of my vehicle propelling me forward at 6:07:06 PM. Two more minor spikes on the graph indicate backward motion at about half the force of the initial impact. I suspect one of those may have been the force of my roll bar deploying.

Processing the Claim

Processing of the claim for the collision damage was somewhat slow. The insurance company had to reach all three drivers involved for their statements before they could make a determination of fault. I had advised the insurance company that I had the data and could prove exactly what happened, but they had to go through their process. Apparently the crucial information they needed from the driver in front of me, was how many impacts he heard, relative to what he felt, to verify that the rear vehicle actually hit me before I hit him. I wondered if his memory was clear on that point, but was confident the data would back me up if it was not.

Had there been any question, the G force graph below and speed graphs above would tell the entire story. Had it ended up in litigation, the engineers at Geotab would have provided me with expert testimony in the form of a formal report explaining and validating the data from their collision reconstruction. Lucky for me, the insurance company just wrote a check to cover the collision damage.

Unfortunately,

The next day I received the dreaded call… after further review it was determined the frame was bent, and due to severe collision damage Rosie and I would not be seeing any more highway miles together. The insurance company settled, and I was off to find a new car. Knowing that I had solid facts acquired through collision reconstruction, allowed me to negotiate from a position of power, rather than accepting whatever the insurance company decided. I had a nice fat down payment in my pocket, and my former loan was paid off.

I promised a happy ending, so here it is. Not quite 2 weeks from the collision event, I was driving my new car. I could not find another Chili Red one, but British Racing Green can grow on a person. My new road pal is 2 years and 50K miles younger, has a turbo charger, and a far superior sound system. I am naming him Jack Hammer, after the salesman at the Mini dealer (I swear that is his real name).

Thanks Jack!

Thanks Geotab!