SquadSync

Inspiration

During the summer, I went on a trek at Mt. Whitney with a group, and at one point, I got separated from my friends. That's when I realised that in remote locations like trails and national parks, there’s often no signal and no way to contact anyone. For a while, I was completely cut off, with no connectivity, and the group had no way of knowing where I was. Thankfully, I eventually managed to find my way back, but the experience made something very clear: getting lost in national parks is common, and people often go missing simply because there’s no way to stay connected.

Connectivity tends to fail at the exact moments when people need coordination the most, such as during emergencies, large events, and especially in remote outdoor environments. That experience led us to ask: What if groups could stay aware of each other’s locations even without internet or cell service?

This inspired SquadSync, a peer-to-peer, completely offline positioning and coordination system designed to keep small groups connected, informed, and safe, no internet, no WiFi, and no servers required.

What it does

SquadSync forms an offline mesh network between nearby devices and estimates real-time positions using:

• Bluetooth RSSI distance estimation
  
• GPS readings
  
• Step-count

The system allows a group to see each other's approximate positions, movement paths, and distances using only device-to-device communication.

All computation and communication occur locally, entirely offline.

Features

• Mesh Network Simulation: Devices connected in a mesh topology where devices can communicate through intermediate nodes
• Network Awareness: All devices are aware of all connected devices in the network
• Connection Monitoring: Real-time tracking of connection strength between devices
• Weak Connection Detection: Automatic warnings when device connections become weak (<50%)
• Disconnection Alerts: Critical warnings when devices disconnect from the network
• GPS Tracking: Location tracking for all devices
• Step Count Integration: Uses step count data to estimate where disconnected devices might have moved

Challenges we ran into

• Extremely noisy BLE RSSI signals
  
• BLE RSSI signals with phone are extremely short ranged
• Simulating realistic GPS drift patterns
  
• Designing a peer-to-peer system without centralized timestamps
  
• Balancing the weight of BLE vs GPS vs step-count in the fusion pipeline
  
• Ensuring the system remains fully offline and scalable

Accomplishments that we're proud of

• Created a working offline positioning system on laptops
  
• Achieved ±1–3 meter relative accuracy in simulation
  
• Built a reusable multi-sensor simulation environment
  
• Implemented BLE + GPS + PDR fusion within hackathon time
  
• Designed an extensible architecture for real phones
  
• Delivered a functional MVP without any hardware dependencies

What we learned

• BLE RSSI requires heavy correction to be usable
  
• Step-count odometry improves indoor accuracy significantly
  
• Multi-sensor fusion outperforms any single sensor
  
• Movement simulation helps test algorithms faster than real hardware

What's next for SquadSync

• Port the simulation to real Android and iOS devices
  
• Build an offline map generation layer
  
• Add safety alerts using local ML models
  
• Expand the system for search-and-rescue teams, campus groups, and events
  
• Turn SquadSync into a full offline mesh navigation framework

Built With

• folium
• python
• streamlit