EV Collision Warning System

An embedded ADAS-style safety project for real-time obstacle awareness and collision-risk warning using radar + ToF sensing, edge preprocessing, signal processing, and machine-learning classification.

Project Snapshot

• Domain: Embedded systems, automotive safety, ADAS, robotics safety
• Edge sensing stack: HB100 Doppler Radar + LM358 signal conditioning + ESP32 ADC node
• Distance sensing: ToF sensor path (project materials reference VL53L0X/TFLC02-class ToF usage)
• Main compute: Raspberry Pi 5
• Core analytics: FFT-based relative speed estimation, sensor fusion, TTC calculation, Random Forest collision classification
• Warning output: Haptic alert using motor driver + vibration motor

System Architecture

Processing Workflow

Engineering Flow

HB100 Radar -> LM358 Amplifier -> ESP32 ADC + Filtering -> USB Serial -> Raspberry Pi 5
								   ^
								   |
							   ToF Distance Sensor

Raspberry Pi 5 -> FFT Speed Estimation -> TTC Calculation -> Random Forest Classification
		  -> Collision Risk Assessment -> Vibration Motor Alert

Implemented Components

• ESP32 firmware node for ADC acquisition, filtering, and serial telemetry output.
• Raspberry Pi 5 pipeline for feature processing, fusion, and risk logic.
• FFT-based speed estimation as a practical Doppler analysis component.
• Random Forest risk classification integrated in project decision flow.
• Haptic alert path for non-visual warning feedback.

Repository Documentation

Core Architecture

• Architecture Diagram
• Hardware Architecture Documentation
• Software Architecture Documentation

Algorithms and Risk Logic

• Sensor Fusion Documentation
• Collision Risk Assessment Documentation
• Raspberry Pi 5 Processing Workflow
• ESP32 Firmware Documentation
• FFT-Based Relative Speed Estimation Documentation
• Random Forest Collision Classification Documentation

Project Outcomes and Planning

• Results Documentation
• Future Scope Documentation
• Project Structure Documentation

Source Artifacts Included

• firmware_esp32.py.py (ESP32 firmware source)
• EVT Project report .docx (1).pdf (project report)
• Collision-Warning-System.pptx (project presentation)

Recruiter-Facing Positioning

This repository demonstrates practical design choices across the embedded-to-intelligence stack:

• Analog sensor conditioning and microcontroller acquisition
• Real-time serial telemetry integration
• Applied signal processing for radar interpretation
• Feature fusion and ML-based risk classification
• Safety-oriented actuator feedback for human-machine interaction

It is suitable for portfolios targeting EV startups, ADAS teams, robotics companies, and embedded systems roles.