SMART INDIA HACKATHON 2025
EdgeDrive: Real-Time Safety & Diagnostics
Problem Statement ID - 25119
Problem Statement Title - Student Innovation
Theme - Smart Vehicles
PS Category - Hardware
Team ID -
Team Name (Registered on portal) - Byte&Code
Byte&Code EdgeDrive: Real-Time Safety & Diagnostics
Proposed Solution
🔧AI-Driven Predictive Maintenance
AI/ML models predict engine oil degradation, brake wear, battery health, and other critical failures before they happen.
🕹️ Gamified Driver Wellness
Innovation: Instead of just sending alerts, the system provides Remaining Useful Life (RUL) estimation for parts, reducing downtime and maintenance costs
A real-time “Driver Safety Score” is generated on the dashboard/mobile app.
Innovation: The platform uses gamification – drivers can unlock rewards, compete on leader boards, and receive AI-based coaching tips to improve safe
📡 Vehicle-to-Vehicle (V2V) Communication
driving habits..
Traffic Congestion Alerts – rerouting drivers before bottlenecks.
Emergency Vehicle Priority – ambulance/police broadcasts route to nearby cars, clearing the way.
Collision Prevention Alerts – especially useful during fog and smog , blind overtakes and lane changes.
🛡️ Geofencing & Security Owners define a virtual safety zone for their vehicle.
Innovation: Our system leverages edge-AI on OBU to filter and prioritize critical alerts, ensuring no delays in life-saving communication.
Instant alerts on mobile app & OBU dashboard if:
Vehicle exits predefined zone.
Vehicle is moved without authorization (theft detection).
🖥️On-Board Unit (OBU) – Edge Intelligence Device
Innovation: Uses AI-based anomaly detection (abnormal time/location movement patterns) for early theft detection, beyond static geofencing.
Edge-AI Processing → Ensures low latency, privacy, and reliability even without cloud.
Innovation: Unlike traditional telematics, our OBU supports AI model updates over-the-air (OTA), ensuring continuous improvement and adaptability.
@SIH Idea submission- Template 2
Byte&Code
TECHNICAL APPROACH
Technologies Used
⚙️Hardware:
Edge Device: Raspberry Pi 4, ESP32, CAN module(MCP2515)
Sensors: IMU Sensor, GPS,GSM, On Board Unit, LiDAR, CAM
module, Ultrasonic Sensor
Connectivity: BLE(V2V), 4G/5G (cloud)
💻Software:
Edge: MicroPython/C++ with anomaly detection.
Web Dashboard: React Native/Flutter for alerts, tracking,
leaderboard.
Cloud: AWS IoT Core, Flask/Django backend, AWS TimeStream,
TLS Encryption.
🤖AI/ML Framework:
Predictive Maintenance: Linear Regression, Gradient Boosting
(XGBoost, LightGBM), Random Forest.
Rash Driving Detection: Support Vector Machine (SVM),
Neural Networks.
Edge AI ( On-Device Interference): TensorFlow Lite, PyTorch
Mobile, ONNX Runtime.
Cloud AI ( Advanced Models): TensorFlow, PyTorch, Scikit-
learn.
Byte&Code
FEASIBILITY AND VIABILITY
✅ Technical Feasibility
AI/ML models can run on Jetson Nano / Raspberry Pi with real-time
performance.
Integration with IoT (Wi-Fi/Bluetooth) for notifications is practical.
✅ Economic Feasibility
Estimated cost: ₹9,500 – ₹15,000 (prototype level).
Affordable compared to high-end automotive safety systems.
✅ Operational Feasibility
User-friendly: Drivers get visual + buzzer alerts and radar-like display.
Works in day/night, light fog/smog, and detects blind-spot obstacles.
✅ Availability
Components available from local electronics markets
Open-source libraries available for sensor interfacing & ML training.
4
Byte&Code
IMPACT AND BENEFITS
🚗 Improved Road Safety
Alerts drivers about rash driving vehicles nearby,
reducing chances of accidents.
Detects blind-spot obstacles and warns in smog/low-
visibility.
👨👩👧 Lives Saved
Helps prevent collisions & fatalities, especially on
highways and in cities.
📡 Smart Assistance
Radar-like display improves driver awareness of
approaching vehicles.
Real-time alerts allow quick reaction in emergencies.
💰 Cost-Effective Solution
Affordable compared to advanced ADAS
(Automotive Driver Assistance Systems).
Can be adapted for low-cost vehicles in India.
🌍 Wider Impact
Supports Smart City & IoT transportation goals.
Promotes responsible driving & reduces road
accident cases.
Byte&Code
RESEARCH AND REFERENCES
Predictive Maintenance: AI-based RUL estimation proven in automotive industry to reduce
breakdowns.
V2V Communication: IEEE 802.11p / C-V2X standards enable low-latency vehicle-to-vehicle
alerts.
IEEE Standard 802.11p – Wireless Access in Vehicular Environments (WAVE).
SAE J1939 & OBD-II Protocols – Vehicle CAN data communication.
Research paper: “AI-based Predictive Maintenance for Automotive Systems” – IEEE, 2021.
Research paper: “Driver Behavior Monitoring using IMU Sensors” – Springer, 2020.
https://indianexpress.com/article/business/first-six-month-2025-over-29k-die-national-
highway-accidents-10176340/
