Full Stack · Electronic · AI

← Back IoT / Competition

S.D.A.S

Smart Driving Awareness System

Embedded Systems Engineer 3 months Shell Selamat Sampai Varsity Challenge 2024
SDAS hardware mounted in car dashboard
At a Glance
3 Safety Gates alcohol + identity + speed
<1s Response Time gate to decision
~0% False Positives in testing
National Competition Shell 2024
Overview

What is it?

An in-car IoT safety system that prevents drunk driving, monitors speed, and identifies the driver. Built for a national competition in 3 months.

SDAS hardware product shot
The Problem

What needed solving?

Drink-driving remains a leading cause of road fatalities in Malaysia. Existing solutions are expensive, cloud-dependent, or easy to bypass. None address sobriety, identity verification, and speed monitoring in a single unified system.

No single device addressed drunk driving, identity, and speed at once.
Process

How it was built

01
Safety Requirements Define safety gates and Malaysian road safety standards
02
Sensor Calibration Calibrate MQ-3 alcohol thresholds and GPS speed limits
03
Firmware Development Write Pico firmware for sensor fusion and gate logic
04
Dashboard UI Build driver-facing status display and alerts
05
Integration Testing Test all 3 safety gates under real conditions
06
Competition Demo Present and demonstrate live at Shell Varsity Challenge
Solution

How we solved it

Designed a pre-ignition gate: the car won't start until the driver passes a MQ-3 breath alcohol test. HuskyLens handles facial identification so only registered drivers can operate the vehicle. A GPS module triggers configurable audio speed alerts in real time. The entire system runs on a Raspberry Pi Pico.

SDAS electronics internals
Gallery

Build & demo

Technology

Why these technologies?

TechnologyWhy we chose itRole in system
Raspberry Pi PicoDual-core ARM microcontroller with ultra-low power draw, suitable for always-on installation in a car without draining the battery.Sobriety gate + sensor fusion
HuskyLensDedicated AI camera module for face recognition that runs entirely on-device with no server, no API, no internet required.Facial ID gate
MQ-3 SensorIndustry-standard electrochemical alcohol sensor with calibrated sensitivity in the breath-test range (0.05–0.5 mg/L).Alcohol detection
GPS ModuleStandalone NMEA-output GPS provides speed data without cellular dependency and works in any location.Speed monitoring
Performance

Key metrics

Safety Gates Activealcohol + identity + speed
3
Gate Response Timegate to decision
<1s
False Positivesin testing
~0%
Results

What we achieved

Pre-ignition alcohol gate blocks drunk drivers before the car starts

HuskyLens facial ID ensures only authorized drivers can operate the vehicle

Real-time GPS speed alerts with configurable thresholds

Competed nationally in Shell Selamat Sampai Varsity Challenge 2024

Full system assembled and demonstrated within a 3-month timeline

Stack

Technologies used

Raspberry Pi Pico HuskyLens AI GPS Module MQ-3 Sensor JavaScript HTML/CSS
Up next

Electrofuel.co

F&B Startup Business Architecture

View case study →

Let's build
something together.