Connected IoT sensor network prevents road hazards on Japanese highways in real time

PROJECT OVERVIEW

ABOUT THE CLIENT

This case study is based on a real VNEXT project. The client's identity and certain project details have been anonymized or generalized for confidentiality purposes.

Our client designs the quieter side of highway safety: tools that guide drivers without them ever needing to think about it, and sensor networks that keep watch over a highway's own infrastructure long after installation. Expressway operators depend on both halves of that work, the visible guidance system drivers interact with directly, and the invisible one monitoring equipment scattered along miles of road. For a business built on things working reliably out of sight, the software behind each system has to be every bit as dependable as the hardware it serves. Exact financials are withheld for confidentiality.

THE CHALLENGE

A highway exit is an unforgiving place to get software wrong. Drivers moving at speed need guidance that responds correctly every time, and the camera based confirmation system built to support them had to hold up against business logic layered several steps deep, the kind that rarely behaves the same way twice once something goes slightly off script.

That difficulty showed up most clearly in testing. Some of the trickiest reported issues simply refused to reproduce inside the team's own test environment, behaving one way on a desk and another way entirely once the system was running in the field. Chasing a bug that will not repeat itself on demand is one of the more demanding problems in software testing, and here it sat at the center of the project rather than at its edges.

Running alongside that work was a second, quieter system: a network of sensors installed along the highway itself, each one watching for signs that a guardrail, a barrier, or a piece of roadside equipment had fallen or shifted out of place. Keeping that network visible and trustworthy meant building software that operators could rely on exactly as much as the physical sensors it was reporting on.

PROJECT OBJECTIVE

The client wanted two quieter parts of highway safety looked after as one: a camera guided tool to help vehicles find their way at the exit, and a watchful sensor network tracking the health and whereabouts of fall detection devices along the road itself, both held to the same standard of software an operator could trust without needing to double check it.

APPROACH AND METHODOLOGY

Turning vague reports into confirmed, testable cases

A reported issue rarely arrived as a clean, reproducible bug description. More often it arrived as a account of behavior that still needed to be traced back to its cause. Rather than guessing at what a report meant, the team confirmed specific requirements directly with the client first, then built a detailed test case for each piece of business logic involved, so that by the time anyone touched the code, there was a shared, agreed definition of what correct behavior actually looked like.

Reproducing field conditions the test environment could not

When a ticket refused to reproduce through normal testing, the team changed its approach rather than its effort. It went directly into the source code and requested real data from the production system, using both together to reconstruct the exact conditions under which the issue actually occurred. That shift, from testing against assumptions to testing against real system behavior, is what eventually turned stubborn, unreproducible tickets into fixed, verified ones.

TEAM SCALE

VNEXT staffed this engagement with a compact team of four, small enough to stay close to the detail of two technically demanding systems at once, with each layer of the stack, backend, frontend, and testing, given a dedicated owner rather than a shared responsibility.

• Project Manager: kept requirement confirmation and delivery moving across both systems at once
• Backend developer: built the API layer and the sensor data handling logic that the fall detection dashboard depends on
• Frontend developer: built the web interfaces for both the exit guidance tool and the sensor monitoring dashboard
• Tester: worked from the detailed, confirmed test cases to verify business logic that had already proven difficult to reproduce once

THE SOLUTION AND TECHNOLOGY STACK

What came together in the end covers both sides of highway safety, guidance for the vehicles passing through, and a steady watch over the sensors keeping the highway itself in check. Each system solves a different problem, but both had to meet the same bar: software an operator could trust without needing to verify it manually every time.

What was built

• Highway exit guidance system: gives camera based confirmation and guidance support for vehicles at highway exits, built so an operator can trust what it reports without checking it by hand
• Fall detection sensor dashboard: tracks the status, angle, temperature, battery level, and GPS location of fall detection devices installed along the highway, turning scattered hardware readings into one legible picture of equipment health
• Map based device visualization: plots every monitored device on a map by its installed location, so an operator can see at a glance where attention is needed rather than scanning a table row by row
• CSV data export: lets operators download sensor readings directly for further analysis, keeping the data portable rather than locked inside the dashboard

Technology Stack

THE RESULTS

Each system carried its own quiet complexity, yet both found their way to delivery as one coordinated effort.

• Delivered on schedule: both systems shipped within the committed timeline despite complex, hard to reproduce logic
• Issues resolved against real behavior, not guesswork: tickets that could not be reproduced in testing were traced and fixed using source code review and real production data
• Two systems, one coordinated team: exit guidance and sensor monitoring were delivered together under a single small team rather than as separate, disconnected efforts
• Full device visibility for highway operators: every fall detection sensor's status, health, and location is visible from one dashboard instead of being tracked manually

For highway infrastructure where a missed signal can mean a missed hazard, a system that tells operators exactly which device needs attention, and where, turns raw sensor data into something people can actually act on. With both systems now delivered as one coordinated effort, the client has a foundation built to extend further, more sensors, more routes, more of the highway network brought under the same steady watch, without starting the next piece of work from zero.