EDGE AI MOTORCYCLE ADAS SYSTEM

The Sixth Sense for Motorcyclists.

Eliminate blind spots and rear-end collisions with intelligent edge computer vision. Converts discreet tail cameras into an active, predictive safety layer with intuitive peripheral mirror alerts and real-time Time-to-Collision vectoring.

⚡ Test In-Browser Simulator 🛠️ DIY Build for Under ৳3,000

< 25ms

Inference Latency

170°

Dual Tail FOV

৳2,850

POC Hardware BOM

Stealth Tail Dual Pod & Digital Mirror

Seamless ThrottleIQ smartphone integration

ACTIVE ADAS

Engineered For Riders

The Indriyo Architecture

Bridging the white space between screenless radar and dumb display dashcams.

👁️

Active Blind Spot Detection (BSD)

Seamlessly monitors vehicles lingering in 120°–165° left and right rear quadrants. Triggers an immediate solid amber peripheral border so you know a vehicle is there without breaking head posture.

⚡

Predictive Time-to-Collision (TTC)

Uses Optical Expansion (Tau Theory) and relative velocity vectoring. Differentiates between a car pacing safely behind and a high-speed vehicle on a collision course (< 1.8s TTC).

🧭

Context-Adaptive Telemetry

Connects to your bike speed via OBD-II, GPS, or ThrottleIQ middleware. Scales sensitivity dynamically: stopped at a red light, approaching traffic triggers alerts from further away!

🚨

Intuitive Peripheral UX

The human peripheral eye is highly sensitive to color strobes. Solid Amber draws gentle attention; rapidly flashing Crimson Red instantly triggers evasive action before a rear collision occurs.

🏍️

Stealth Tail Pod Design

No bulky 360° dome cameras or awkward radar boxes. Dual waterproof miniature optical lenses tuck invisibly beneath the pillion seat or license plate bracket.

📱

ThrottleIQ Smartphone Bridge

Zero extra compute cost! Use your mounted phone as the AI brain via our drop-in Flutter module. Your phone ingests the ESP32 Wi-Fi feed, runs YOLO inference, and provides helmet audio warnings.

Market White Space

Indriyo vs. The Competition

See why Indriyo is the only system offering both active visual verification and edge predictive intelligence.

Feature / Metric	Indriyo (ইন্দ্রিয়)	Ride Vision 2	Garmin Varia Radar	Chigee AIO-5 Smart Dash
Architecture	Dual AI Vision + Screen	Screenless Mirror LEDs	Radar Box (LED Blips)	CarPlay Screen (Dumb Cam)
Visual Feed ("Trust but Verify")	YES (Live Digital Mirror)	NO (No Display)	NO (No Video)	YES (Video Only)
Object Classification	YES (Car, Bike, Bus, Truck)	YES (AI Vision)	NO (Mass Only)	NO / Basic
Time-to-Collision (TTC) Vectoring	YES (Optical Tau + Speed)	YES (Proprietary)	Basic Proximity Radar	NO (Proximity Only)
Tail Aesthetic Footprint	Stealth Mini Dual Lenses	Bulky Front/Rear Pods	Bulky Radar Brick	Standard Dashcam
Bangladesh DIY Cost	< ৳3,000 BDT	৳50,000+ (Imported)	৳30,000+ (Imported)	৳45,000+ (Imported)

Electrical Engineering

Motorcycle Safe Wiring Schematic

Protected from alternator voltage spikes, engine vibration, and battery parasitic drain.

[ Motorcycle Switched 12V Line (License Plate / Tail Light) ] │ ├───(+) Red Wire ──► [ 3A Inline Fuse ] ──► IN+ [ LM2596 Buck ] OUT+ ──┬──► (+) [1000µF Cap] ──► 5.1V [Left ESP32-CAM] │ │ │ └──► (+) [1000µF Cap] ──► 5.1V [Right ESP32-CAM] │ └───(-) Black Wire ───────────────────────► IN- [ Converter ] OUT- ──┬──► (-) [1000µF Cap] ──► GND [Left ESP32-CAM] │ └──► (-) [1000µF Cap] ──► GND [Right ESP32-CAM]

⚡

1. Calibrate Potentiometer First

Factory buck converters often ship set to 10V–12V. Turn the brass adjustment screw counter-clockwise until your multimeter reads exactly 5.1V before plugging in the ESP32s!

🛡️

2. Anti-Brownout Capacitors

Solder one 1000µF 16V electrolytic capacitor directly across the 5V and GND pins of each ESP32 board to absorb Wi-Fi transmission spikes and stop reboots.

🔒

3. Vibration Locking

Motorcycle engine RPM vibrations can drift adjustment screws. Add a drop of clear nail polish or silicone adhesive to lock the screw head permanently in place.

Open Engineering & CAD

3D Models, Quantized AI & Road Benchmarks

Free, open-source 3D printable meshes, lightweight INT8 models, and real road validation data.

📦

Dual-Camera Tail Pod

BINARY STL • 14.4 KB

Aerodynamic motorcycle tail pod with 15° outward splay for dual ESP32-CAMs, rain lens visors, cable relief canal, and M4 mounting brackets.

Download Pod STL ↓

☀️

Handlebar Display Hood

BINARY STL • 16.9 KB

Anti-glare sun visor cradle for 5" display/phone with 28mm glare hood, bottom USB cable pass-through, and dual 22mm/28mm bar clamps.

Download Hood STL ↓

📊

Real-Road Benchmark

HTML REPORT • 560+ FPS

Stress-tested against real motorcycle rear dashcam footage in daytime traffic and night rain. Sub-2ms edge latency with 100% frame throughput.

View Live Report ↗

🧠

Quantized INT8 Model

ONNX INT8 • 3.34 MB (72.7% Compressed)

Optimized YOLOv8n object detection engine quantized for mobile CPUs & edge NPUs. Slices latency to ~1.7ms per frame for instantaneous TTC warnings.

View Model on GitHub ↗