Sharper Insights for Better Decisions

The Future of Bus Performance: AI-Driven KPI Dashboards for Smarter Network Management 

The quality of the service delivered to the passengers directly impacts the revenue generated for the bus operator. Provide the service within the quality threshold defined in the contract and the commensurate payment will be made. However, things are not always that easy, situations will arise where the bus operator needs to take action to guarantee the level of service. As congestion builds during the day, as is always going to happen in a busy metropolis, the buses start to become delayed and the operator needs to take this into account and manage the situation. This is in part anticipated in the time table, but it also needs to be managed in real time to adjust as the situation develops. 

An accident that closes the road is one thing and is generally considered to be outside the control of the operator, but normal congestion will always happen and needs to be managed.

Introducing Our AI-Powered Prediction System 

The new Trapeze fleet management system incorporates an AI based prediction system. This will learn throughout operation and start to build correlations between the situation now and what it has been in the future. If every day this particular link builds up to be a 5min delay, then the AI system will recognise this trend and use it to enhance the prediction accuracy.

The system also recognises the correlations between different parts of the network, a build-up in this area, generally results in a build-up, 10 minutes later, in this area, thus the predictions will already start to be adjusted for the later buses.

The prediction system also recognises the dispatcher actions taken to manage the network. Therefore, if a diversion is entered, the prediction system will recalculate the arrival time for the revised route and publish this to the real time prediction consumers ultimately the public.

Turning Predictions into Real-Time Performance

BFMS uses this prediction information to take the monitoring of KPIs one step forward in real time. The system continually calculates the real time KPI information for each bus and route, using the output from the AI prediction algorithm and provides this in a real time dashboard to the service controller.

Because BFMS is making predictions as to when the buses will arrive at the stops, and these predictions consider the service control actions that are in place, the system makes a prediction of the likely KPI for the next 60 minutes using this prediction information to formulate a likely KPI value. When the service controller sets up a path dispatch, this is incorporated into the real time arrival predictions and therefore also updates the real time KPI dashboard, providing an approximation of the performance result of the path dispatch.

Service controllers introduce changes to the network in real time, adjusting the path the buses are taking, if required, as well as curtailing trips, if necessary. These changes are included in the updated prediction calculations, provided to the public as well as in the real time KPIs.

This dashboard provides information on high and low frequency routes, using the appropriate calculation for headway and scheduled services. The dashboard monitors several key indicators including mileage and trips in the past hour and the current hour, highlighting worst performing routes for example.

The real time dashboard provides a central point for monitoring the performance of service delivery with historic information from the current day available as well as the predictions. The actual operated mileage, number of trips as well as punctuality or headway performance are clearly visible and accessible.

The view also provides information on the worst performing routes to enable more attention to be focused where needed to enable the service quality goals to be achieved.

From Performance to Passenger Experience

A good service delivers happy customers and the necessary revenue under the franchising system to make the activity profitable for all concerned! Giving the service controllers visibility of actual performance as well as an indication of how the service is looking into the future will enable operators to take actions more intuitively to manage the quality of bus services delivered to the public. Ultimately, implementing these advances will support LTA in achieving your goal of delivering more efficient service and creating a richer, more enjoyable passenger experience.