Shuttle Tracker

What we want to achieve is a reliable, low-maintenance way to track shuttles without depending on the driver’s phone or vehicle GPS hardware. To do this, we shift our focus from tracking the shuttle itself to tracking the riders onboard. Each trip is defined by a unique route code, scheduled time, and a list of booked users. When a trip begins, users who booked it share their GPS location from their mobile devices. These location updates are tied to the route code and are sent every few seconds. The system continuously analyzes this data to detect when a group of users on the same trip are in close proximity within a predefined radius like 50–100 meters. If a certain percentage (e.g., 60–70%) of the booked users are close together, we consider them a cluster or a "bucket," and compute the average position of the group as the shuttle’s live location. This location is visualized on a dashboard where each cluster appears as a moving pin, showing details like the number of riders in it and how strong the cluster is (i.e., how many of the expected users are present). A live event log on the side updates as users join or leave a cluster.

To make the system even smarter, we monitor each trip's health: if the scheduled trip time arrives but no strong cluster forms within a set time window (say, 5–10 minutes), the system flags the trip as "no live signal"—suggesting the shuttle may be stalled or hasn't begun picking up riders. This alert is shown to admins and can even trigger a call or message to the driver. Over time, the system automatically transitions from "no tracking yet" to "cluster-based tracking" once riders start boarding, giving us a hands-off, highly scalable way to track shuttles and spot problems early—all without relying on drivers to comply or on hardware that might fail.