Home Skip to main content

Research Shows Machine Learning and Data Can Optimize Snowplow Life

Posted: Apr 5, 2021

Researchers at the University of Utah implemented machine learning techniques to optimize snow plow truck life. Snowplow trucks serve a crucial role in winter highway maintenance activities of deicing and snow removal, activities that are essential to public mobility and safety. The researchers showed that random forest modeling, a machine learning data analysis technique, can help agencies predict truck performance and identify factors that lead to performance depreciation. They found that a number of trucks may still perform well at the "optimal" replacement year determined by the model. As a result, replacing all trucks at that age could be a significant waste of resources. If truck performance can be monitored and predicted with high resolution and high accuracy, trucks can be replaced in a timely manner to optimize useful life while minimizing maintenance expense. A key is balancing the repair and maintenance costs, which tend to increase over the life of a truck with the price of purchasing a new truck exceeding $150,000.

Xiaoyue Cathy Liu, Ph.D.
University of Utah

Lifecycle Assessment Using Snowplow Trucks' Automatic Vehicle Location Data

NDSU Dept 2880P.O. Box 6050Fargo, ND 58108-6050