IoT-Based Automatic Vehicle Counting and Traffic Forecasting

Clients

Techstack

Purpose

Planning roads and transport well depends on knowing how many vehicles actually pass a given point — but manual traffic counts are tedious, costly, and quickly out of date. A research team set out to prove that a small roadside device could do the counting automatically, around the clock, and turn that stream of data into a forecast of what traffic will do next.

Description

We built a roadside unit that watches passing traffic through a camera, recognizes each vehicle with a vision model, and tallies them as they go by — logging the counts to the cloud so anyone can monitor them live from anywhere. Those accumulated counts then feed a forecasting model that projects future traffic flow to support smarter transport planning. It is a self-contained blend of on-device AI and cloud monitoring — the kind of applied research we build end to end.