Reinventing Maintenance Operations With cuOpt and Jetson Orin


Kawasaki Heavy Industries, Ltd. (Kawasaki) is a manufacturing company that's been building large machinery for more than a hundred years. With NVIDIA cuOpt™ and NVIDIA Jetson™ Orin, Kawasaki partnered with Slalom, Inc. to transform its track maintenance and inspection capabilities.


Kawasaki Heavy Industries


Slalom, Inc.

Use Case

Data Science


NVIDIA Jetson AGX Orin 64G

Current railroad maintenance and inspection solutions are obsolete and inefficient to manage at scale.

Historically, railroad track inspection remains heavily reliant on manual, visual processes that are labor-intensive and inefficient. Freight operations are halted so that specially dedicated cars can run the routes - an enormous lost opportunity cost. Well-trained inspectors must meticulously eyeball rails, tiIes, fasteners, and ballast to spot faults, follow stringent safety procedures, and record issues by hand. This manual method doesn't scale; it's limited by the number of qualified inspectors and their skills are better allocated in other areas of the business.

The North American Class I railroads consist of over 140,000 miles watched over by seven different competing corporations. While the federal government sets standards that they must all meet, it’s up to those companies to find, triage, and ultimately defend against entropy. The extensive system poses unique challenges in inspection and maintenance across millions of miles of diverse terrain. The status quo of track inspection contributes to over 1,000 train incidents annually. 

To truly transform railroad safety and efficiency, the industry needs a digital transformation.

“We are trying to change our business. The hardware business is a one-time business, whereas software and digital technology is a sustainable business.” — Ryoji Negi, Product Owner, Kawasaki

Railroad maintenance

Reimagining Growth in Profits and Safety

With the help of NVIDIA cuOpt to solve complex routing problems and incredible AI compute powered by NVIDIA Jetson AGX Orin 64G, Kawasaki is developing a digital platform, Kawasaki Track Maintenance Platform, that leverages artificial intelligence and machine learning at the edge and interfaces with track inspection devices such as digital cameras, lasers, and gyrometric sensors.

The platform allows users to leverage all existing data and analytics, and automated training intelligently improves ‌machine learning models at the edge. When combined with institutional knowledge, it allows maintenance teams to use the output to increase the safety and reliability of the operation. Users will aggregate and browse all rail data sources in one easy-to-understand view. See the problems on specific tracks in context and most importantly, take charge and send teams and materials to fix problems.

Rail companies that adopt the platform for automated track maintenance services can see an increase in safety and efficiency, and a decrease in time and costs for inspection and maintenance. Connecting sensor data with the on-the-ground expertise of railway staff ensures that valuable human capital isn’t left behind in the pursuit of automation.

With NVIDIA’s cuOpt and Jetson Orin, Kawasaki is achieving many benefits, including:

  • 26,000 hours saved per year = Estimated annual manager of track maintenance (MTM) hours saved per company = 5 hours per week per MTM * 52 weeks per year * 100 MTMs per company
  • $218.4M saved per year = Estimated annual cost savings from automating inspections = 150 inspectors per company * $208,000 * 7 companies

1 5 hr./week is the estimated time MTM’s spend manually creating repair schedules

2 100 is the estimated number of MTM’s employed per Class 1 rail company

3  150 is 50% of the estimated number of track inspectors (300) in a Class 1 rail company, where an MTM had 3 full time inspectors.

4  $208,000 is the calculated annual fully loaded labor rate


Kawasaki Track Maintenance Platform

Unlocking New Possibilities With cuOpt and Jetson

Rail companies that adopt Kawasaki Track Maintenance Platform for automated track maintenance services can see an increase in safety and efficiency and a decrease in costs for inspection and maintenance. Connecting sensor data with the on-the-ground expertise of railway staff ensures that valuable human capital isn’t left behind in the pursuit of automation. With the help of NVIDIA’s cuOpt and Jetson AGX Orin 64G, Kawasaki is able to build end-to-end solutions for the entire rail maintenance process, expanding to additional processes that are today, manual, or labor-intensive, and building data and AI-enabled platforms which can revolutionize the industry.

The platform represents a promising new revenue stream for Kawasaki. Even greater is the growth potential across Kawasaki when this digital strategy is implemented in other lines of business, such as robotics and hydrogen.

“Finding suitable solutions to NP-Hard problems, such as vehicle routing, is a highly complex endeavor and costly to any business since as the number of destinations, or nodes, increases, the time required to solve the problem increases exponentially.

With cuOpt, railroad maintenance managers can quickly generate schedules, work with other scheduling teams to ensure that rail lines will be safely shut down for maintenance operations, and repair crews to have an ideal schedule for their work, which in turn can make railroads safer and more efficient.”

Teddy Crane
Senior Engineer, Slalom

Ready to Get Started?

Learn more about NVIDIA cuOpt for achieving world-record accuracy and performance to solve complex optimization problems, save time, and reduce costs.