Transit agencies can leverage AI for predictive maintenance and other fleet management benefits 

To the passengers who rely on public transportation, transit agencies appear to have just one job: Make sure the trains, buses and other vehicles run on time. 

But the professionals who operate mass transit agencies know this one job is in fact thousands. Running on time entails perennial monitoring and maintenance, non-stop troubleshooting, complicated repair logistics, energy and emissions tracking and more. Hidden behind each punctual arrival and departure is an inordinate amount of data and fleet management. 

Managing rollingstock, city buses and other public transport assets can be grueling. Emerging technologies like artificial intelligence (AI) can significantly reduce this burden, but only when used strategically. Leveraging AI during just one portion of the asset lifecycle management (ALM) process may have moderate impact; threading it into most or all portions will have a maximum impact. Done right, traditional AI like machine learning and the newer generative AI technology can enable automation at scale and unlock valuable, data-driven insights that transform maintenance.   

ALM consists of five phases:

1. Plan
2. Deploy
3. Operate
4. Optimize
5. Dispose

Here are just a handful of ways transit agencies can leverage AI throughout this process.  

Plan

Before new rollingstock meets the rails, or a new bus hits the pavement, transit agencies must conduct complex calculus. Will this new asset pay for itself? In how long? And when will it become a liability?   

Machine learning-powered technology like digital twins can help take the guesswork out of the equation. Digital twins are virtual representations of a physical asset and its environment. They are informed by real-world data and can help simulate potential performance in the future. Digital twins are popular in the utility and manufacturing industries but can also be applied in urban planning and mass transit scenarios, helping determine if and when new fleets are necessary.       

Operate

Generative AI can accelerate productivity when it comes to asset management. Generative AI is commonly associated with consumer-facing chatbots, models that are trained on publicly available data, but generative AI models trained on specialty data, like a transit agency’s trove of maintenance information, can be equally impressive. These models can help automate time-consuming processes like creating and filling out work orders, drafting maintenance recommendations and more. This means less busy work for technicians.        

Optimize

A key part of running on time is optimization—ensuring fleets are performing at their very best, with the fewest possible defects and downtime. AI technology like computer vision is incredibly helpful at optimization. Drones, cameras and other devices equipped with computer vision provide speed and scale that humans cannot match, monitoring and analyzing rollingstock, buses and other assets in real time and collecting coveted performance data.  

This vast pool of data then enables predictive maintenance. Historically, agencies relied on reactive maintenance (waiting for a bus to break down) or proactive maintenance (inspecting rollingstock at a set time each month). Predictive maintenance anticipates future problems based on past data, helping head off defects and downtime well before they occur. For instance, IBM worked with Transport of London to manage the day-to-day maintenance efforts for over 10,000 internal technicians within the London Underground.   

AI has further applications throughout the ALM process, from measuring and managing emissions to prioritizing replacement-versus-repair decisions. As AI continues to evolve, new applications are on the horizon. Emerging AI agents promise to automate even more complex, multi-step processes.     

The work of transit agencies may never be as simple as passengers surmise, but by integrating AI solutions throughout the ALM process, agencies can significantly reduce the burden of keeping assets healthy and on time.