Hiking Guardian AI

• 

  Hiking Guardian AI – Full System Architecture

• 

  Medieval-style home screen with age-adaptive safety. “Start Hiking” enables real GPS + live map tracking with Gemini risk detection.

• 

  Live hiking tracker built in Flutter — AI check triggers Gemini safety analysis anytime.

• 

  Gemini detects sudden stop + chaotic HR fluctuations, triggering deep-reasoning safety analysis before SOS activation.

• 

  Multi-signal reasoning: HR volatility, elevation mismatch, and trauma-like patterns detected within seconds.

• 

  Gemini explains why it triggers emergency mode—extreme anomalies across HR, pace, elevation, and recovery curve.

• 

  Stable physiological state detected: zero HR volatility and consistent pace with Gemini deep-reasoning monitoring engaged.

• 

  Mild anomaly detected: rapid HR decline despite static elevation. Gemini recommends rest before continuing.

🌄 Inspiration

This project started from a very personal place. My father loves hiking. But as he gets older, I worry every time he goes out alone.

There are countless cases of older hikers collapsing from sudden heart issues, getting lost, or suffering falls where help arrives too late.

One day I thought:

“If I can build something that protects my father… maybe it can protect thousands of fathers, mothers, and older hikers too.”

That was the moment Hiking Guardian AI was born — not just an AI model, but a full mobile experience designed to keep the people we love safe.

⚙️ What it does

Hiking Guardian AI continuously analyzes:

Biometric signals (heart rate, volatility, sudden drops)

Environmental & weather conditions

Terrain strain (incline, elevation gain)

Movement patterns inconsistent with healthy activity

A personalized 7-day baseline for each user

Using multi-signal reasoning, Gemini detects:

Fall-like deceleration

Cardiac instability while stationary

Abnormal physiological stress

Subtle danger patterns that rule-based systems miss

Beyond the AI engine, the app also includes a Flutter-based real-time tracking system, featuring:

Live GPS movement

Clean grid-snapped path visualization

Golden glowing trail renderer made with CustomPainter

Auto pace / distance / elevation monitoring

One-tap emergency contact access

The system doesn’t just track your numbers — it protects the people you love.

🛠️ How I built it

I built this alone with one goal in mind: create something real that my father could use.

For this hackathon, I focused on:

Building the core AI safety engine using Gemini 3 Flash

Designing a reasoning pipeline that explains why Gemini triggers danger

Creating an anomaly-detection schema showing what simple rules can’t detect

Simulating aging-related stress cases and fall scenarios

Developing realistic emergency edge-case tests

In parallel, I started building the actual mobile app in Flutter, including:

A GPS tracking module with movement filtering

A golden glowing CustomPainter trail map

A 15-second Smart Monitoring Mode (auto Gemini checks for premium users)

A manual AI safety button for instant analysis

A simple Emergency Contact page for older users

A Hiking Record system with auto cleanup & saved routes

A full app structure (Main Menu → MapPage → Records → ResultPage)

The AI engine is fully functional today, and the Flutter integration is already partially completed.

🚧 Challenges we ran into

Teaching AI to understand multi-signal danger, not simple if-rules

Designing prompts where Gemini explains its reasoning transparently

Balancing zero false alarms with instant SOS detection

Creating UI flows simple enough for older adults

Filtering noisy GPS movement into a clean golden path

Integrating AI Studio logic with a mobile app that I am building alone

I’m not a full-time developer — I run a small restaurant as the owner, handling cooking, operations, and management every day.

Even while juggling the realities of running a business, I committed focused time to build something meaningful.

Between cooking, serving customers, and closing the shop, I designed and built this entire project alone — from concept, UI, and JSON schema to the AI safety engine and the Flutter mobile prototype.

Despite the constraints, I completed a working anomaly-reasoning system and a functional mobile app connected to Gemini — driven by one mission: protecting my father.

🏆 Accomplishments we're proud of

A working AI engine capable of real anomaly reasoning

A live mobile app prototype built with Flutter

Real-time GPS → Golden Trail Map renderer

15-second Smart Monitoring Mode for continuous protection

Emergency Contact UI designed with elderly users in mind

Realistic emergency simulations for older hikers

A meaningful, personal project built entirely by a single developer

📚 What we learned

AI can bridge the safety gap for older hikers

A single developer can build life-saving apps

Emotion-driven projects create the most authentic solutions

Flutter makes building real prototypes accessible even under time pressure

True safety isn’t about tech — it’s about people

🚀 What's next for Hiking Guardian AI

Complete full Flutter integration

Voice assistant mode for older users

Wearable device connections (HR strap, smartwatch)

A simple built-in radio for hiking

A Family Guardian Dashboard

More accurate fall + cardiac anomaly detection

Offline-safe mode (AI fallback + cached baseline)

My mission started with my father… but now, I hope it helps every family who worries about someone they love.

Built With

• dart
• flutter
• gemini-1.5-flash
• gemini-3-flash
• google-ai-studio
• gps
• open-meteo