Wednesday January 23, 2019

Researchers at Iowa State University Use TensorFlow to Analyze Traffic Patterns

The Iowa DOT and Iowa State University researchers have teamed up to use artificial intelligence (AI) and machine learning (ML) to analyze traffic pattern data such as speed, accidents, weather conditions, congestion, and more. This is accomplished by using the TensorFlow framework from Google to analyze visual data from stationary cameras and cameras mounted on snow plows. The AI can also analyze data from radar detectors. Machine Learning takes the data and makes conclusions about road conditions and in turn helps get first responders to the scene of accidents faster.

Officials in Iowa say machine learning could also be used to predict crash risks and travel speeds, and better understand drivers' reactions or failures behind the wheel. In California, snow may not be an issue, but traffic certainly is, and college students there used TensorFlow to identify pot holes and dangerous road cracks in Los Angeles.