VisionSafe: Revolutionising Roadwork Safety with AI-Powered Traffic Control
Each year in Aotearoa New Zealand, lives are lost or permanently changed due to accidents at temporary roadwork sites. VisionSafe, an innovation by Revo Tech supported by NZTA’s Hoe ki angitū Innovation Fund, is changing that. By combining AI, computer vision, and edge computing, VisionSafe transforms the way temporary traffic lights operate—making them smarter, safer, and more efficient.
The Problem with Traditional Temporary Traffic Lights
At most roadworks, portable traffic lights rely on fixed timers or simple radar detection. They assume traffic volumes and speeds, often leading to unnecessary wait times, congestion, or even safety risks if vehicles are still clearing the zone when lights change.
Such systems don’t adapt to real-time conditions. They can’t detect stalled vehicles or red-light runners entering late. The result? Delays, frustrated drivers, and a dangerous environment for road workers.
Introducing VisionSafe
VisionSafe is a smart traffic control system designed for temporary roadwork zones. It uses strategically placed cameras and edge computing to:
Track vehicles entering and exiting work zones in real time.
Dynamically adjust green, red, and clearance phases based on actual traffic behaviour.
Prevent conflicting movements by ensuring zones are completely clear before switching phases.
In tests, VisionSafe cut vehicle wait times by up to 40% and cleared 10 vehicles through a zone nearly 50% faster than traditional systems—all without compromising safety.
Built from the Ground Up in Aotearoa
The VisionSafe system was locally conceptualised and developed by the Revo Tech team, who partnered with PB Traffic, Smart City Streets, and Media Design School to bring the project to life. Simulated environments were built in Unreal Engine 5 to train the system’s AI models, drastically reducing the time and cost of physical testing.
The project was led by Revo Tech’s founder Rawinder Singh and hardware AI lead Dr. Prakash Karn, with a small, agile team committed to solving big challenges on the road.
Field Testing Results
At a controlled test site in Silverdale, VisionSafe:
Successfully managed over 700 traffic cycles.
Maintained an average clearance time of just 19.65 seconds (compared to 36+ seconds for traditional systems).
Automatically responded to real-time conditions like red-light violations and stalled vehicles.
Remained stable and resilient even through 32 RF communication interruptions.
Won strong user confidence—77% of surveyed road users said it outperformed traditional systems, and 100% agreed it felt safer.
Smarter Traffic, Safer Work Zones
VisionSafe doesn’t just reduce wait times—it reduces tension. Shorter, more predictable delays result in less aggressive driver behaviour and greater compliance with road signage. For workers, this means fewer threats and lower risk. For drivers, it means smoother, safer journeys.
Its adaptive system also balances productivity and safety. Instead of maximizing vehicle throughput at the cost of risk, VisionSafe allows road operators to “tune the dial” based on site length, traffic density, or time of day.
What’s Next?
Revo Tech is working on:
Upgraded AI models and hardware for greater resilience and tracking accuracy.
Integration of emergency vehicle prioritization and per-vehicle clearance timing.
A blacklist database to flag frequent violators entering work zones.
Regulatory approval and broader deployment across New Zealand, starting with engagement from Auckland Transport.
With continued collaboration and support, VisionSafe is on track to become a new standard in roadwork safety—transforming not just traffic lights, but lives.