Our Solution

Our Solutions

R.O.M.A.'s solutions are easy and cost-effective to add to existing infrastructure. They take over when standard traffic lights fail, especially during emergencies such as wildfires or hurricanes, to efficiently direct a large flow of traffic and save countless lives. The hardware and software platforms are highly integrated to provide critical support for rescue and evacuations. To learn more about our offering, please see our Innovation Video.

The Control Unit

The brain of our system is a Raspberry Pi 5 microcontroller, which functions as the decision-making hub. The Pi runs a lightweight machine learning AI environment that manages sensor inputs, decision making, and communicates with nearby modules. Using a compact computer-vision model optimized for smaller devices, the Pi continuously estimates congestion levels at the intersection, and when traditional systems shut down, this AI model steps in to govern the traffic-signal cycle, selecting the optimal green time, predicting surges, and preventing backups before they form, all to ensure proper evacuation during emergencies.

The Power Unit

To solve the issue whereby traffic lights can not function during power outages, we utilize solar panels and Lithium-Ion batteries to power the light signals. Our systems provide up to 48 hours of power, which is significantly longer compared to about 6 hours from existing conventional systems. Our inbuilt AI software allows Traffic Engineers to remotely manage batteries and take corrective actions as necessary.

AI-Based Decision Making

TinyML is a software approach of running machine learning on small, low-power embedded devices, such as microcontrollers, so they can make intelligent decisions locally without accessing powerful computers or the cloud. Unlike traditional ML, which runs on servers or laptops, TinyML compresses trained models and trained data to run efficiently on small microcontrollers and devices.

Sensory Input

R.O.M.A. relies on multiple sensors to notify our systems the change of the environment. For example, the "eye" of our systems comes from a high definition camera, which feeds the AI module to detect vehicles and traffic conditions. During emergencies, first responder vehicles are given priorities to clear the intersections based on image signals to ensure that they can reach the rescue sites faster.

Cross-Device Communication

R.O.M.A utilizes LoRa (Long Range) for communications with the Traffic Control Centers and between intersections. It allows intersections to act as relay nodes to share traffic/emergency data across the city. For example, when a firetruck approaches an intersection, Node A detects a siren pattern and triggers “green corridor” locally. Using LoRa, Node A sends an “Emergency Approaching” packet to Nodes B & C within 1-2 miles. Those nodes pre-emptively shift to green, forming a coordinated rescue route without the need for a central server.