How to Transform a Heavy Industry, One Sensor at a Time

Moving shipping containers is heavy work. Moving a traditional industry into the digital age is a different kind of heavy job. Software development agency GNAR took on the challenge and built an ops management platform for RMS Intermodal, one of the largest rail yard operators in the U.S. Their IoT solution gave RMS a data-driven view of their operations for the first time, resulting in a whole new definition of “efficiency” for the company.

Serendipity manifests a new idea

It all started at a wedding. GNAR Founder Brandon Stewart found himself chatting with Adam Gray, the son of RMS Intermodal’s President, and the conversation turned to RFID tracking. Brandon had done some work on an RFID solution for running events, and Adam was curious about the technology. Could it be applied to his father’s rail yards in order to harness the chaotic swirl of trucks and forklifts and cranes?

The two men put their heads together and spent the next few months researching the problem space. “We visited two local RMS sites in California and surveyed over 40 locations,” Brandon says. “We sat inside the trucks, and watched how people worked. We met with managers, crews, and executives to better understand their day-to-day challenges as well as the industry at large.”

The traditional way: radios, clipboards, and Excel to manage a lot of moving parts

Looking at rail yard operations, Brandon and Adam found that efficiency is measured by job completion speed and workforce costs. Fundamentally, it’s all about how fast the crew can get shipping containers on and off a train, with a number of specialized equipment involved in moving the heavy boxes. The goal is to have the right number of workers doing the right activities at the right moment.

The yard manager typically has a 24-hour window to prepare for an incoming train and can use a terminal operating system (TOS) — train industry software — to get the arrival time, container count, and track number. That time is precious, as once the train arrives, a flurry of activity begins.

The manager drives around the yard to corral the crew and delivers orders via two-way radio. Forklifts stack trailers alongside the train track, almost like “staging” the train. The crane then comes and starts pulling boxes off the train and dropping them onto the trailers. A hostler truck connects to the trailer and moves the container to a designated place in the yard, where it is unloaded and stored.

Throughout the process, the manager is trying to stay on top of all the moving parts, but, in a large yard that staffs up to 70 workers at a time, it can be tough. Managers are challenged to keep all workers actively engaged, but they often struggle to have direct visibility into basic operations, like which worker is operating which vehicle. This is primarily due to the long rectangular layout of the yards and rows of stacked containers (similar to a large outdoor warehouse). Meanwhile, the clock is ticking and the train needs to get unloaded as quickly as possible.

It doesn’t stop there. Once the job is complete, the manager fills out a clipboard full of paperwork, which then gets sent to the head office for record keeping and billing purposes, much of it stored in Excel spreadsheets. This adds extra time and overhead for managers as well as office staff.

The new way: an IoT platform for efficient rail yard management

Inspired by ride-hailing apps like Uber and Lyft, Adam, Brandon, and the GNAR team designed a solution that uses GPS coordinates streamed from Android tablets in each vehicle. Their new platform, Intrmodl, turns the vehicles themselves into connected IoT devices that could send real-time data to a central platform for processing and analytics. The driver app not only tracks its vehicle’s location in the yard across time, but also logs usage stats, like fuel level and engine hours, and travel paths and duration, as well as vehicle inspection details.

Managers now have a bird’s eye view of all their workers and vehicles in a dedicated manager’s app. During unloading activity, they can track precisely who’s doing what and take quick action to correct or fine-tune the activities. Managers can forgo the radio and communicate with one or the whole crew via the app, as well as stay on top of maintenance needs. The app also makes it easier for managers to perform audits and site inspections.

For upper management, the main platform provides business analytics that helps them track patterns in site activity and the overall performance of their operations across all their yards. Execs can see metrics the following day rather than waiting until the end of the month to see what happened yesterday. The data is useful to the business in a number of ways, such as informing profitability targets or bids to train companies. Brandon says, “Our data shows train companies how a yard adds even more value by demonstrating consistent levels of efficiency.”

Extracting meaning from a firehose of data

Every two seconds, live sensors in each vehicle stream massive amounts of data into the Intrmodl platform. One of the challenges for the development team was how to clean and distill all that data into a couple of bytes of really useful, actionable insights for the leadership to access quickly. “We maintain a three tier or three reservoir kind of setup,” says Brandon, “where the live data is constantly coming in and we're queueing it and making sure that we're digesting it and keeping everything in order.” They also had to figure out a way to flush data from the database once it’s no longer needed.

Another challenge was how to loop in signature moments in the workday, such as going on break. So, they set up sessions that define triggers for those signature moments to kick in and inform how data is captured during the break and ensures that work time data is more meaningful.

A third challenge was how to build a daily report that included a calculated overall metric for the day and allow it to be queryable. This would help managers quickly see if their productivity was on target — perhaps 10% ahead or 20% behind — compared to a rolling average.

The path to par: defining what top performance looks like

Previously, RMS had a hard time determining true benchmarks for performance metrics due to the lack of visibility across yard activities. “They never knew exactly what par was for them,” says Brandon. “They only had an idea of what par was. Now, they can identify more granular targets for specific vehicles and tasks that are based in real-time data.” Managers can set realistic performance expectations across the team, and reward star workers or coach underperformers. Metrics vary from yard to yard due to the variety of services that each yard offers to train companies.

The heavy-lifting is now easier for more yards, and more industries

RMS Intermodal’s digital strategy is paying off. With data insights readily available, the business can make better decisions and strengthen relationships with train companies and other partners. Today, Intrmodl has been fully deployed in six RMS yards, and about 50 others use some aspect of the platform.

Going forward, Brandon and the GNAR team are continuing to work on enhancements. They’re looking to take advantage of more tablet sensor capabilities to further map vehicle movements and workflows. They also plan to address improvements to billing and forecasting. Brandon also helped RMS hire their first IT team who will eventually take the platform in-house.

RMS is an innovator in a deeply traditional industry, but their ideas don’t end with rail yards. One exciting aspect of the Intrmodl technology is that it’s flexible enough to apply to other industry use cases, such as automobile transportation. RMS owns their own fleet of trucks and also operates yards for storing shipments of new cars.

The RMS story is only the beginning. Brandon can envision the IoT platform serving other traditional industries that orchestrate the movement of heavy things and vehicles. Airlines, shipping, construction, mining, warehousing — all can benefit from real-time data insights from a complex dance of moving parts. Wherever their clients want to put a sensor, the platform is ready to make sense of it.

Read the GNAR case study to learn more about Intrmodl on Heroku.

Listen to the Code[ish] podcast featuring Brandon Stewart and Yuri Oliveira: Monitoring Productivity Through IoT.