The New Road Standard—Predicting and Preventing Your Next Crash
Shirley Yu
Collingwood School
Floor Location : M 021 N

Nearly all drivers have turned left before and have run through at least one yellow light in their life. Therefore, anyone could be prone to accidents involving left turns.

Looking at statistics, around 1.35 million people are killed on the roads around the world each year according to the World Health Organization. 22.2% of all traffic crashes and approximately 61% of crashes at an intersection involve a left hand turn making it the leading “critical pre-crash event.”

Currently, drivers are taught to not rush through a yellow light and to always look for incoming cars to prevent left-hand turning crashes, but such guidance is insufficient in actually protecting drivers. First, many people still rush through yellow lights even if they are told not to. There is no strict punishment for doing so, and many people do it, not out of malice, but because they might be in a hurry. Second, when safety officials tell drivers to look around for oncoming cars, it is impossible to do so in certain situations where another car is blocking the driver’s vision.

This is where my project comes in. When the ToF sensor mounted on a traffic light sees a car that is not slowing down and may run the yellow light, it notifies the turning car’s driver through a display screen inside the car acting as a preventative measure and a second pair of eyes, allowing the driver to avoid a deadly crash.

This innovation involves a hardware and a software part. For the hardware, I designed the parts using a platform called Solidworks, 3D printed the parts, and built the simulation from wood planks. After the hardware aspect was completed, I went to program the software which has various algorithms for processing, filtering, manipulating, and sending data.
This project is a successful prototype and a proof of concept of a working ToF sensor that is run by a program that can accurately determine whether it is safe to turn left at an intersection. Based on the results of 50 test cases, it has 100% accuracy. Accuracy is determined by comparing whether the values outputted from my program are equal to the actual values they should be, verified by a graphing calculator.

The time tolerance and the turning time of a car are predetermined and held constant before running the program. Time tolerance is the extra time for a car to pass the intersection before the other car arrives or leaves to ensure safety. Turing time is the average time it takes a car to turn left. For the sake of testing, time tolerance was set to 0.75 seconds and turning time was set to 1 second. These values would differ in practice because real cars and roads are larger than the simulation.
This invention is built, programmed, and tested in a field that mimics real roads with a large remote-controlled car, simulating the innovation’s functions if put into action in real life.