Driving predictive analytics for the IoT and connected vehicles

Navistar—one of the leading manufacturers of commercial trucks, buses, defense vehicles, and engines—is widely known for its flagship International Truck and IC Bus brands. Unscheduled maintenance and vehicle breakdowns account for a large share of total costs for vehicle owners. For a company with the motto “Uptime is our mission,” the status quo in terms of trying to avoid breakdowns and addressing them as they occur wasn’t really the way to go. Typically, vehicle manufacturers schedule vehicle maintenance based on miles traveled or time since last appointment. But these were very rudimentary and only two of thousands of data points that can signal the need for a specific type of maintenance.

In response to the growing number of vehicles in their fleet and the mounting cost to maintain them, Navistar built OnCommand Connection, an IoT-enabled remote diagnostics platform, on Cloudera Enterprise. The platform brings together over 70 telematics and sensor data feeds from more than 375,000 connected vehicles in real time, including engine performance, truck speed, acceleration, coolant temperature, and brake wear. This data is then correlated with other Navistar and third-party data sources, including meteorological, geolocation, vehicle usage, traffic, historical warranty, and parts inventory information. The platform currently stores over 110 TB of data and uses machine learning and advanced analytics to automatically detect engine problems early and predict maintenance requirements. Fleet and vehicle owners can now monitor truck health and performance from smartphones or tablets, prioritize needed repairs, and quickly identify the nearest dealer service locations that have the relevant parts in stock, available technicians, and available service bays.

Navistar is collecting data every 5–10 seconds from hundreds of thousands of vehicles in real time and needed a next-gen data management platform that could not only scale effectively to the volumes of data but also handle both real-time and streaming IoT data sources. The company uses Cloudera Enterprise to handle close to 20 million records a day, measuring everything from vehicle speed, acceleration, fuel economy, geolocation, idle times, and potential failures, and recommend corrective measures. Navistar is also able to drive predictive analytics and machine learning on all of its data.

Steve Otto walks you through Navistar’s journey to build the OnCommand platform. He also delves into how the company is driving predictive analytics and modeling on terabytes of data from connected vehicles. Some of the quantifiable key business outcomes that they have experienced include:

• Lowered maintenance costs: By collecting and analyzing vehicle telematics data using Cloudera, Navistar has been able to reduce maintenance costs and vehicle downtime by almost 40%. One Navistar customer reduced the maintenance cost-per-mile for its vehicles from 12 to 15 cents to less than 3 cents.
• Effectively scale operations: Navistar has been able to effectively scale its coverage of connected vehicles from a few thousand vehicles to more than 375,000 connected vehicles today and is set to grow exponentially in the future.
• Launch new services: By utilizing the same platform, Navistar also provides predictive maintenance services to thousands of school buses to identify fault codes and determine exactly what maintenance is needed, thereby helping keep kids safe.