SISURAKSHA – Intelligent Child Safety Platform

The Challenge

Why SISURAKSHA?

The Problem

Traditional school bus safety relies on manual observation. Drivers must focus on the road, causing dangerous delays in accident detection. Static routes waste fuel at empty stops, and manual attendance logs lack real-time visibility and data integrity.

No real-time hazard detection
Manual, unreliable attendance logs
Inefficient "ghost stop" routing
Privacy risks with cloud biometrics

Our Solution

A unified smart platform integrating IoT sensors (ESP32), Edge-AI , and Hybrid Blockchain providing proactive safety, verifiable attendance, and optimized routing entirely at the edge.

Real-time hazard & accident detection
Blockchain-secured RFID + Face ID
Attendance-Driven Dynamic Routing
PDPA-compliant Edge-AI processing

Project Objectives

01

Develop multi-sensor accident and emergency hazard detection

02

Implement ML-based window and footboard safety monitoring

03

Deploy Edge-AI biometric attendance system (PDPA compliant)

04

Design Attendance-Driven Dynamic Routing (ADDR) engine

Innovation

Research & Methodology

IoT Architecture

Edge-AI

Hybrid Blockchain

Cyber-Physical Systems (CPS)

Intelligent Transportation

System Architecture

L4

Presentation Layer

React.js / React Native · Parent · Driver · Admin Apps

L3

Backend Service Layer

Node.js/Express · PostgreSQL · MongoDB · REST API

L2

Edge Processing Layer

Raspbery PI · ESP32 · XAI Models · Local AI Inference

L1

Physical / IoT Layer

ESP32 · Optimus Air 780E LTE Module · MPU-6050 · NEO-6M GPS Module · PN532 NFC RFID Module · MQ-2 Gas Sensor · IR Proximity Sensor · Ultrasonic Sensor Module · FSR Sensor · Solenoid Door Lock 12VDC · 12V Relay Module · LCD Display with I2C Module · SD Card Module

Novel Contributions

SISURAKSHA integrates Edge-AI, IoT, Hybrid Blockchain, and spatial intelligence into a unified, privacy-preserving ecosystem. Its novelty lies in the synergy of these technologies to provide proactive, intelligent, and tamper-proof child safety.

Intelligent Footboard Monitoring

Movement-aware protection identifying unsafe usage during transit using computer vision and IR sensing. Triggers context-sensitive alerts based on vehicle motion.

Multi-state safety model

AI Window Safety Protection

Edge-powered YOLOv8 vision system detecting limbs or heads beyond safe window boundaries in real-time. Operates with ultra-low latency without cloud dependency.

YOLOv8 @ Edge

Hybrid Blockchain-Based Attendance Integrity System

Hybrid architecture combining Ethereum and MongoDB for efficient, audit-ready records. Features dual-factor (RFID + Face-ID) verification with SHA-256 hashing and automated tamper detection.

Hybrid Blockchain + Edge IoT

Context-Aware Egress (CA-AEES)

Automated door mechanism deploying only at zero velocity with dual-verification interlocks to prevent unsafe egress into active traffic.

< 300ms latency

Integrated Research Significance

Together, these contributions establish SISURAKSHA as a next-generation platform advancing Edge-AI, Cyber-physical response, Blockchain trust, and Spatial intelligence simultaneously.

Technologies Used

Hardware

ESP32Optimus Air 780E LTEMPU-6050 NEO-6M GPSPN532 NFC RFIDMQ-2 Gas Sensor IR ProximityUltrasonic ModuleFSR Sensor Solenoid Door Lock12V RelayLCD (I2C) SD Card

AI / ML

YOLOv8MobileFaceNetTensorFlow Lite RoboflowGoogle Colab

Software

Node.jsReact.jsReact Native Python/FlaskPostgreSQLMongoDBOSRM

Blockchain

Ethereum (Sepolia)SHA-256 HashingHybrid Architecture

Core Modules

System Features

Smart Hazard Guard

Multi-sensor fusion detecting rollover, fire, gas, and water hazards with a 60-second driver verification period to eliminate false alarms.

MPU9250MQ-260s Verify

Active Passenger Monitor

Real-time YOLOv8-powered audio alerts when students place limbs outside windows or stand on the footboard while the bus is in motion.

YOLOv8Audio AlertHall Sensor

Blockchain Attendance

RFID + Face-ID dual verification with SHA-256 hash chaining anchored to Ethereum for immutable, tamper-proof attendance records.

RFIDFace-IDEthereum

Dynamic Routing (ADDR)

Greedy nearest-neighbor algorithm with IRI road-roughness penalties and real-time graph pruning to eliminate ghost stops and reduce fuel waste.

OSRMGreedy NNIRI Penalty

Emergency Egress (CA-AEES)

Solenoid-actuated door with dual-verification: zero-velocity check + perimeter clearance. Deploys in under 300ms for critical situations.

SolenoidDual-Verify<300ms

Offline-First IoT

Resilient architecture that buffers all sensor data locally during connectivity loss, with 99.5% successful sync rate upon reconnection.

Local Buffer99.5% SyncCAN Bus

User Roles

Parent

Report child absences
Live boarding status
Nearest bus route map

Driver

Optimized pickup sequence
Real-time safety alerts
Emergency override controls

Administrator

Global fleet oversight
Vehicle approval & audit
Blockchain integrity checks

Academic Assessments

Milestones

Select an assessment to view its full details, dates, and marks allocation.

Prop Pres
6%

Prop Rep
6%

PP-1
15%

PP-2
18%

Final Rep
15%

4%

Final Pres
10%

Viva
10%

2%

Paper
10%

2%

Total Marks Distribution — 100%

Assessment 01

Proposal Presentation

Completed

Marks Allocated6% of Total

DeliverablePresentation

Phase Overview

This phase focuses on the fundamental research argument and initial design planning.

Assessment Criteria

Proven Gap / Creative Solution

35%

Clearly argue the knowledge gap using multiple credible sources and provide excellent justification for novelty.

Specialized Knowledge Application

30%

Clearly identify the research area and provide a critical evaluation proving the selection of the best technology.

Effective Communication

15%

Excellent structure, stage presence, and management of the Q&A session.

Commercialization Potential

15%

Sound evidence of business potential with many achievable user benefits.

Solution Implementation

5%

Brilliantly justified High-level System Architecture, comprehensive WBS, and realistic time estimates.

Total: 100 marks → Weighted as 6% of Final Grade

Assessment 02

Proposal Report

Completed

Marks Allocated6% of Total

DeliverableIndividual Document

Phase Overview

This phase evaluates the technical documentation and the academic rigor of the written proposal.

Assessment Criteria

Proven Gap / Creative Solution

35%

Extensive comparison of the research problem with similar products and services.

Specialized Knowledge Application

30%

Demonstrated full awareness of technologies and a critical evaluation of why they were selected.

Effective Communication

15%

Logical discussion of features, excellent formatting, and proper use of IEEE referencing.

Commercialization Potential

15%

Demonstrated sound evidence to prove business potential.

Solution Implementation

5%

Realistic workload distribution in the Work Breakdown Structure (WBS).

Total: 100 marks → Weighted as 6% of Final Grade

Assessment 03

Progress Presentation 1 (50%)

Completed

Marks Allocated15% of Total

DeliverablePresentation

Phase Overview

This phase shifts focus toward the actual implementation and the technical quality of the prototype.

Assessment Criteria

Solution Implementation

40%

Excellent design features, approximately 50% work completion, and application of appropriate standards/best practices.

Specialized Knowledge Application

25%

In-depth knowledge of technologies demonstrated through the current implementation.

Effective Communication

15%

Excellent body language, eye contact, voice projection, and commendable use of visual aids.

Proven Gap / Creative Solution

10%

Implementation clearly demonstrates the "Proof of Concept" for the proposed solution.

Commercialization Potential

10%

Highlight achievable user benefits that prove business potential.

Total: 100 marks → Weighted as 15% of Final Grade

Assessment 04

Progress Presentation 2 (90%)

Completed

Marks Allocated18% of Total

DeliverableLive Demo + Presentation

Phase Overview

This phase evaluates the near-final implementation, system integration, and advanced technical capabilities of the 90% completed prototype.

Assessment Criteria

Solution Implementation & Integration

40%

Near complete implementation (90%), seamless integration of modules, and rigorous application of testing standards.

Specialized Knowledge Application

25%

Advanced demonstration of machine learning, blockchain, and hardware algorithms working in tandem.

Evaluation & Results Analysis

15%

Clear presentation of testing metrics, accuracy, and real-world performance benchmarks.

Effective Communication

10%

Professional delivery, clear articulation of technical complexities, and strong defense during Q&A.

Proven Gap / Commercialization

Logbook Marks

100%

Total: Weighted as 2% of Final Grade

Assessment Summary

#	Assessment	Weightage	Status
01	Proposal Presentation	6%	Completed
02	Proposal Report	6%	Completed
03	Progress Presentation-1 (50%)	15%	Completed
04	Progress Presentation-2 (90%)	18%	Completed
05	Final Report	15%	In Progress
06	Final Report (group)	4%	Upcoming
07	Final presentation	10%	Upcoming
08	Viva	10%	Upcoming
09	Website	2%	Upcoming
10	Research paper	10%	Upcoming
11	Check Lists	2%	Upcoming
12	Logbook	2%	Upcoming
Total		100%

The People

Research Team

Group 25-26J-282 · Sri Lanka Institute of Information Technology

Group Leader M.H.S. Akther

IT22590312

Lead – Blockchain & Attendance

Hybrid blockchain RFID attendance and Segmented Polyline Analysis for routing.

EthereumOSRMMongoDB

View Progress

R.I.S.R. Pinto

IT22610102

Lead – Driver & Footboard Monitoring

Hall-effect movement sensing for footboard occupancy and AI driver fatigue analysis.

Hall SensorIR SensingState Machine

View Progress

A.M.D.C. Kumara

IT22600134

Lead – Egress & Biometrics

Automated CA-AEES emergency egress system and Edge-AI MobileFaceNet biometrics.

MobileFaceNetSolenoidPDPA

View Progress

T.A.P.K. Shameera

IT22606624

Lead – Accident & Window Safety

Multi-sensor accident/hazard detection & YOLOv8-based student behavior monitoring.

ESP32YOLOv8MPU9250

View Progress

Supervisors

Project Supervisor

Ms. Thisara Shyamalee
Lecturer
Department of First Year Division
Faculty of Computing, SLIIT

Project Co-Supervisor

Ms. Osuri Dunuwila
Assistant Lecturer
Department of First Year Division
Faculty of Computing, SLIIT

Individual Tracking

Member Progress

Group Leader M.H.S. Akther

IT22590312 Lead Researcher: Blockchain & Attendance

100%

Overall Progress

Completed Tasks

Hybrid blockchain framework development
OSRM-based matchmaking algorithm
Segmented polyline analysis implementation
SHA-256 hash chaining to Ethereum (Sepolia)

Achievements

100% tamper-detection rate in simulation

26% improvement in driver-student matching

18ms average polyline computation time

Module Progress

Blockchain Architecture100%

RFID Attendance100%

Polyline Analysis100%

Fleet Scaling98%

Next Milestones

Scale MongoDB schema for provincial fleet sizes
Real-world RFID scan batch testing (500+)

R.I.S.R. Pinto

IT22610102 Lead Researcher: Driver & Footboard Monitoring

100%

Overall Progress

Completed Tasks

Hall sensor pulse-based movement detection
IR step occupancy mapping
SAFE-WARNING-CRITICAL state machine
Transitioned GPS → Hall-effect sensing

Achievements

Reliable low-speed movement detection via Hall-effect

Footboard state machine with 3-tier alert levels

Challenges

False detections in driver monitoring due to lighting variations and bus vibration — ongoing calibration and filtering work.

Module Progress

Footboard Detection100%

State Machine100%

Driver Monitoring100%

Vibration Filtering100%

Next Milestones

Resolve driver monitoring stability in high-vibration
Integrate fatigue detection with safety state machine

T.A.P.K. Shameera

IT22606624 Lead Researcher: Accident & Window Safety

98%

Overall Progress

Completed Tasks

Sensor selection (MPU9250, MQ-2)
Accident detection logic implementation
YOLOv8 model training for window safety
60-second driver verification mechanism

Achievements

Integrated MPU9250 + MQ2 for comprehensive hazard mapping

60-second false-alert prevention system deployed

Current Work

Optimizing YOLOv8 confidence thresholds for varied lighting conditions

Module Progress

Sensor Integration100%

Accident Detection98%

Window Safety ML97%

Adaptive Thresholds96%

Next Milestones

Adaptive threshold learning with real road data
Extended field testing in varied conditions

A.M.D.C. Kumara

IT22600134 Lead Researcher: Egress & Biometrics

99%

Overall Progress

Completed Tasks

Solenoid-actuated door system prototyping
MobileFaceNet fine-tuning
CA-AEES dual-verification interlock
PDPA compliance with zero cloud transmission

Achievements

Sub-300ms emergency door deployment latency

97.32% biometric identification accuracy

100% PDPA compliance - zero biological data to cloud

Module Progress

Emergency Egress100%

Face Recognition98%

PDPA Compliance100%

Field Trials97%

Next Milestones

Extended field trials in varied weather
Demographic validation across student datasets

Repository

Document Archive

Presentations

Proposal Presentation Initial research proposal slides

Progress Presentation 1 Phase 1-2 results and findings

Progress Presentation 2 Phase 3-4 results and innovation

Final Presentation Pending – August 2026

Research Documents

Research Proposal SISURAKSHA initial research proposal

Project Charter Scope, objectives, and agreements

Literature Survey Edge vs. Cloud computing for biometrics

TAF Document Topic Assessment Form

Dissertations

Dissertation – Shameera (IT22606624) Accident & Window Safety Module

Dissertation – Pinto (IT22610102) Footboard Safety & Driver Monitoring

Dissertation – Kumara (IT22600134) Emergency Egress & Edge-AI Biometrics

Dissertation – Akther (IT22590312) Blockchain Attendance & Polyline Analysis

Technical Documentation

ESP32 Pin Mapping Hardware connection documentation

Hybrid Blockchain Schema MongoDB document schema design

OSRM Polyline Algorithm Routing algorithm technical spec

Arduino Source Code ESP32 firmware source files

Get In Touch

Contact Us

Institution

Sri Lanka Institute of Information Technology (SLIIT)

Department of Information Technology

Project Email

sisuraksha.info@gmail.com

Supervisors

Ms. Thisara Shyamalee · Lecturer

Ms. Osuri Dunuwila · Assistant Lecturer

Department of First Year Division · SLIIT

Research Group

Group 25-26J-282

Undergraduate Research Project · 2025/26