Post Detail

You’ve installed flashing lights, bells, and gates and can’t afford grade separation. Now what?

Post Detail | Features
You’ve installed flashing lights, bells, and gates and can’t afford grade separation. Now what?


When I meet with traffic engineers, here’s what they often tell me.

“Garreth, I’ve got a rail crossing that’s creating major traffic delays and safety risks. I’ve installed flashing lights, bells, and gates but drivers continue to go around the gates or get stuck on the tracks. I’m having trouble implementing signal preemption since the intersection is further than 200 feet away, not to mention the high cost. I can’t construct an underpass because I don’t have the budget. And even if I could afford it, there isn’t any room for an underpass. I don’t know what to do. Can you help me?”

Does this sound familiar to you?

You’ve probably wished for a low-cost option and wondered why one doesn’t exist. In the meantime, you follow the standard process for addressing issues at rail crossings as shown below. After installing flashing lights, bells, and gates for $400,000, you implement traffic signal preemption, pre-signals, or queue cutters (if possible) for $200,000 or more, and then you consider grade separation if feasible which often costs at least $50 million.

Figure 1: Typical process for implementing treatments at rail crossings

It doesn’t have to be this way anymore. Traffic information and management solutions exist as low-cost options to implement before preemption or grade separation. Now, you can integrate rail crossing information into your advanced traffic management system (ATMS), install a dynamic message sign to help drivers avoid blocked crossings, or adjust traffic signals in real-time when a train is approaching.

The figure below illustrates a new process to reduce traffic delays and collision risk at rail crossings. Start by implementing regulated systems such as flashing lights, bells, and gates then incrementally add new, low-cost traffic management treatments. Once you’ve exhausted options involving signs, signals, traffic management, and traffic signal operations, then move on to high-cost infrastructure-based options. Doesn’t this make more sense?

Figure 2: Low-cost options offer new process for implementing treatments at rail crossings

This isn’t a novel approach, though. Using traveler information and traffic management systems has been discussed extensively among transportation professionals. So why haven’t ATMS integration, dynamic message signs, and real-time traffic signal adjustments been available until now? The answer has to do with predicting traffic delays at rail crossings. Each of these options is severely limited or ineffective without these predictions.

Turns out predicting traffic delays at rail crossings is extremely difficult. Freight trains do not run on fixed and predictable schedules and railroad companies do not share their train data (for many valid reasons, I might add). If you don’t know when and how long a crossing will be blocked, you won’t be able to know when traffic will be delayed and for how long.

To solve this problem, we needed to collect accurate train data without installing equipment on rail property. Our first thought was to use radar or video cameras. After many tests, we were disappointed to find that these devices didn’t provide sufficient accuracy and reliability for this application. In fact, nothing on the market did. So, we did what we do best – innovate. TRAINFO developed special train detection sensors that can predict rail crossing blockages. Our sensors are installed off rail property and can detect trains with 100% accuracy. As difficult as this hurdle was to overcome, it only solved half the problem. We still needed to predict traffic delays.

This was a much harder problem to solve. We tried using advanced statistical and simulation models, but all fell short of the accuracy needed. We needed more analytical muscle. This was a problem for machine-learning and artificial intelligence. So, we created the first AI to predict traffic delays at rail crossings. Since it uses machine-learning, our predictions improve every single day as we collect new data and feed it to our algorithms.

At this point, you might be asking: “Can’t Google provide traffic delay information at rail crossings?” The short answer is no. For a longer explanation, click here.

The heavy lifting was now complete. We could collect train data and predict traffic delays at rail crossings up to 30 minutes before a train arrived. Next, we developed an Application Programming Interface (API) – a fancy term for software code that can communicate with other systems – which allowed us to integrate into ATMS to manage traffic on road networks impacted by rail crossings, inform motorists about traffic delays at rail crossings to help them re-route, and communicate with traffic signals to adjust their timing similar to preemption. Plus, we can integrate into computer-aided dispatch software for first responders.

And there you have it. An end-to-end system that accurately predicts when rail crossings will be blocked and the traffic delays caused by these blockages and that seamlessly provides this information to drivers and traffic management systems.

Back to the traffic engineers that I meet. When they ask if I can help them, I can confidently say “Yes!” TRAINFO’s products are low-cost, low-risk options that have reduced traffic delays and collision risk at rail crossings. They can discourage impatient drivers from going around gates by re-routing them away from blocked crossings and help manage traffic at a network-level to mitigate traffic delays. Since these products are not fail-safe, they cannot replace preemption. However, they can offer similar results for a fraction of the cost and may serve as an adequate alternative.

Garreth Rempel
Garreth Rempel

You’ve installed flashing lights, bells, and gates and can’t afford grade separation. Now what?