You're more likely to regret not doing research than you are to regret what you've done with it
Noah Vermeulen
Computer Science
Undergraduate Research in the Neighborhoods
The goal of this research was to evaluate the ability of different sensors on autonomous vehicles to prevent a collision with a human crossing the road perpendicular to the vehicle. Tests done by the International Institute for Highway Safety provided data for collision tests with autonomous vehicles moving at 12 mph and 25 mph with an adult test target and a child test target. The vehicles had different sensors installed, radio detection and ranging (RADAR), monochrome cameras, stereoscopic cameras, and a sensor fusion of cameras and RADAR. Binary values were given to whether or not the vehicle collided with the crash test dummy and all values were compared to RADAR statistics to determine the success rate of each sensor in a logistic regression model. Different values were calculated based on test speed, the size of the target, and the sensor type. The greatest success was seen at slower speeds with larger targets. The regression model results showed that autonomous vehicles are not advanced enough to safely and consistently prevent collisions without the assistance of a human driver.
Learn more about this and other research in the Neighborhood Engagement Centers.
.png)
