Smart Crosswalks: Prevention, Visibility and Safety with IoT

In the modern urban context, pedestrian safety is a growing priority. To meet this need, we are working on an intelligent system for pedestrian crossings that integrates advanced sensors, dynamic LED panels and edge computing logic.

The system detects the approach of the pedestrian thanks to presence sensors and analyses incoming traffic in real time, measuring speed, distance and density of vehicles. Based on this data, it activates dynamic light signals to warn both the pedestrian and the driver, reducing the risk of accidents.

Thanks to local data processing (edge ​​AI), the system responds very quickly, even in the absence of an Internet connection, and can be configured for environments with different needs: from small intersections to high-traffic crossings. It is also possible to integrate an intelligent traffic light, activated only in conditions of real danger or excessive traffic.

Technical context information:

- Proximity sensors (IR, ToF or Doppler radar) to detect the pedestrian.

- Vehicle traffic analysis via radar sensors and cameras.

- Edge computing on MCU/MPU with AI accelerators for on-site calculations.

- Implementation of light signals via RGB LED panels with amplified visibility.

- Traffic light modules that can be activated on-demand.

- Power supply with integrated photovoltaic system for peripheral areas.

In the project we are developing, the heart of the intelligent crossing is a coordinated system in which sensors, LED panel and traffic light act synergistically to ensure pedestrian safety without penalizing traffic flow.

Detection occurs through a combination of presence sensors, such as Doppler radar or ToF sensors, which identify in real time the pedestrian approaching the crossing. At the same time, other sensors monitor the speed and distance of approaching vehicles, allowing the system to predict whether there is a potential risk.

If necessary, a dynamic LED panel is activated, positioned in a visible position for the pedestrian, which clearly signals the safe moment to cross. At the same time, a low-consumption traffic light dedicated to motor vehicles can be activated selectively: only when the data collected indicates a real risk or when the traffic flow is such that a forced slowdown is required.

All this processing takes place locally, thanks to an edge computing unit based on advanced microcontrollers and predictive analysis algorithms. This allows the system to function even in the absence of a cloud connection and to respond very quickly.

Furthermore, to improve the sustainability and energy independence of the system, we are evaluating solutions powered by photovoltaic panels with integrated batteries, also ideal for decentralized urban environments or areas with low infrastructure.