Published July 25, 2026
| Version v1
Proposal
Open
V2X - Collision Avoidance — Project proposal
Description
Project proposal. # Overview
The **V2X Collision Avoidance System** is an intelligent transportation platform that integrates **Vehicle-to-Everything (V2X)** communication, **embedded systems**, **artificial intelligence**, and **sensor fusion** to improve road safety through real-time collision prevention and advanced driver assistance.
The platform implements a complete **V2X ecosystem**, covering all four communication domains:
- 🚘 **Vehicle-to-Vehicle (V2V)**
- 🚦 **Vehicle-to-Infrastructure (V2I)**
- 🚶 **Vehicle-to-Pedestrian (V2P)**
- ☁️ **Vehicle-to-Network (V2N)**
---
# System Architecture
## 🚘 Vehicle-to-Vehicle (V2V)
The V2V subsystem is implemented on the **STM32F446RE** using **Bare-Metal C** and **FreeRTOS**.
It executes multiple cooperative ADAS algorithms by combining information received from neighboring vehicles with onboard sensor data to detect hazards and generate real-time safety alerts.
### Implemented ADAS Features
- Forward Collision Warning (**FCW**)
- Electronic Emergency Brake Lights (**EEBL**)
- Blind Spot Warning (**BSW**)
- Do Not Pass Warning (**DNPW**)
- Intersection Movement Assist (**IMA**)
- Safe Distance Warning (**SDW**)
---
## 🚦 Vehicle-to-Infrastructure (V2I)
The V2I subsystem runs on a **Raspberry Pi**, enabling communication with intelligent traffic infrastructure.
Computer vision is powered by **YOLOv8** to perform:
- Traffic signal detection
- Vehicle distance estimation
- License plate recognition
This enables vehicles to understand road conditions and traffic signal status in real time.
---
## 🚶 Vehicle-to-Pedestrian (V2P)
The V2P subsystem utilizes an **ONNX-based AI pedestrian detection model** running on the Raspberry Pi.
Unlike traditional V2P solutions, pedestrian detection is performed entirely through **computer vision**, eliminating the need for pedestrians to carry smartphones or dedicated communication devices.
---
## ☁️ Vehicle-to-Network (V2N)
Cloud connectivity is provided through **MQTT**, enabling:
- Real-time telemetry transmission
- Hazard notifications
- Remote monitoring
- Future Smart City integration
---
# Communication Layer
Wireless communication between vehicles and infrastructure is handled by **ESP32** modules using:
- ESP-NOW
- Wi-Fi
An **IPC Hub** running on the Raspberry Pi acts as the central message broker, routing data between:
- V2V
- V2I
- V2P
- V2N
- ADAS Dashboard
- Control Server
---
# ADAS Dashboard
A modern web-based **ADAS Dashboard** was developed using:
- HTML
- CSS
- JavaScript
with a **Python WebSocket Server** for real-time communication.
### Dashboard Features
- Live vehicle telemetry
- Radar visualization
- Object detection results
- Real-time safety alerts
- Vehicle status monitoring
Safety alerts are classified into multiple levels:
- 🟢 Safe
- 🟡 Warning
- 🔴 Critical
---
# Control Server
The platform includes a dedicated **Control Server** capable of issuing emergency commands.
When a critical collision risk is detected, the server can automatically trigger emergency actions such as:
- Vehicle stop command
- Emergency braking request
- Hazard notification broadcasting
---
# Sensor Fusion
To improve environmental perception and decision accuracy, the platform combines multiple sensing technologies:
- Ultrasonic sensors
- IMU measurements
- V2X communication
- AI-based computer vision
This unified sensor fusion architecture enables accurate, low-latency collision avoidance decisions in real time.
---
# Technologies Used
## Embedded Systems
- STM32F446RE
- Bare-Metal C
- FreeRTOS
- ESP32
- Raspberry Pi
## Artificial Intelligence
- YOLOv8
- ONNX Runtime
- Computer Vision
## Communication
- ESP-NOW
- Wi-Fi
- MQTT
- WebSockets
## Web Technologies
- HTML
- CSS
- JavaScript
- Python
---
# Project Highlights
- Complete implementation of all four V2X communication domains
- Multi-node embedded architecture
- Real-time cooperative ADAS algorithms
- AI-powered object and pedestr