Fleet Management News & Business Info | Commercial Carrier Journal
Issue link: http://read.dmtmag.com/i/716013
84 commercial carrier journal | august 2016 TECHNOLOGY: MACHINE LEARNING takes a company vehicle home and fuels soon afterward with- in a few miles of home. Bring- ing in telematics data might show a fuel transaction occurred while the company vehicle was at the driver's house and its fuel gauge did not change. "You start to come up with patterns that make sense, but patterns can change over time," Thearling says. Wex ClearView also uses real-time analysis to show if drivers are making good purchase decisions. Fleets can view reports that show where drivers purchased fuel with a cost comparison to nearby locations. The analytics team also is developing an automat- ed messaging system for fleets to send text reminders to drivers when company policies are violated, such as buying premium versus regular gasoline or stopping at higher-priced locations. Thearling plans to include a feature in the messaging sys- tem that can send escalated alerts. On the first infraction, a text would politely remind the driver of the error. On the second infraction, a message would use all-capital letters in the heading, and the third time, the driver would get an all-caps text. To create the machine-learning algorithms, Wex Clear- View uses business intelligence platforms such as Tableau to identify data patterns. "Modern data analysis tools for machine learning create an ability to look at data in a more intuitive, interactive way through visualization," Thearling says. "You can start to distill the data to see what patterns are most interesting." For more complex analysis, Wex ClearView uses statis- tical programming packages such as R and Python. For larger projects that involve huge volumes of data, the team moves the processing to cloud servers from Amazon Web Services. Taking command In 2015, Navistar was considering new lighter-duty Class 8 truck designs for its International Truck brand. Specifically, the company saw an opportunity in regional applications. On average, Class 8 trucks in this market have less long- term usage – in terms of miles and hours driven – than long-haul applications. In concept, a truck with light- er-weight components would cost less and deliver greater fuel efficiency for regional freight haulers without tradeoffs in durability. In years past, Navistar likely would have conducted an expensive market study to test the concept. In this instance, an analytics team used real-world telematics data to reach a conclusion. The compa- ny's in-house analytics team was assembled in 2014 after the company introduced its remote diagnostics system, OnCommand Connection. The system provides access to vehicle and engine data through an interface with the telematics systems used by its fleet customers, such as PeopleNet and Omnitracs, that can collect data on all truck makes and models. An interesting pattern emerged by studying data from the target market. Regional fleets indeed were using trucks less frequently compared to those in the long-haul market, but trucks were being used for severe purposes, such as pulling trailers through cornfields during harvest season. "Those are things you don't think of," says Andy Mint- eer, director of Internet of Things analytics and machine learning for International Truck. These and other market insights helped the company avoid the further costs and potential future problems of lightweight truck designs. Besides supporting internal customers in truck de- sign and engineering, the analytics group uses advanced statistics and machine-learning techniques to benefit its external customers. The analytics team created a predictive model that iden- tifies potential maintenance issues for customers, often be- fore their "check engine" lights appear. The model predicts failures for more than 40,000 combinations of diagnostic trouble codes by vehicle make, model and year. The model is run weekly, and customers receive an alert if any equipment has a DTC, or combination of DTCs, above the danger area for imminent failure. When alerts are found for International trucks, the company's customer service group can address the problem directly with fleet The Exception Module in the Wex Online platform shows patterns in fuel transactions that indicate waste and fraud. Besides supporting internal customers in truck design and engineering, Navistar's analytics group uses advanced statistics and machine-learning techniques to benefit its external customers.