Option 1: Comparison of Geofencing Technologies

This summary compares LoRaWAN RSSI with higher-precision alternatives for a child-safety "electric fence."

Technology Typical Precision Reliability Best Use Case
LoRaWAN RSSI 10m – 30m Low Long-range "neighborhood" tracking where exact boundaries don't matter.
GPS + LoRaWAN 5m – 10m High (Outdoor) Large properties or hiking. GPS handles location; LoRa transmits the alert.
Ultra-Wideband (UWB) 0.1m – 0.5m Excellent Precise "invisible fences" for yards. High accuracy, very low latency.
Bluetooth (BLE 5.1+) 1m – 5m Medium Short-range (indoor/small garden) using "Direction Finding" (AoA).

Option 2: Technical Implementation & Hybrid Suggestions

If you are building a custom tracker (e.g., using an ESP32-S3 or similar), consider these three implementation paths for copy-paste planning:

1. The "Filtered RSSI" Path (LoRaWAN Only)

  • Method: Use a Kalman Filter to smooth the signal.
  • Logic: If (Average_RSSI < Threshold) AND (Duration > 5 seconds) -> Trigger Alarm.
  • Pros: Lowest power consumption; no extra hardware.

2. The "Reliable Outdoor" Path (GPS + LoRa)

  • Hardware: LoRa Module + Small GPS Module (e.g., u-blox NEO-6M).
  • Logic: The wearable calculates its own coordinates. If the GPS coordinate is outside a pre-defined polygon, it sends an "ALARM" packet via LoRa.
  • Pros: Not affected by signal bouncing or body blocking.

3. The "High Precision" Path (UWB)

  • Hardware: DW1000 or ESP32-UWB modules.
  • Logic: Uses "Time of Flight" instead of signal strength. It measures how long the radio wave takes to travel, providing centimeter-level accuracy.
  • Pros: True "Electric Fence" performance. If the kid crosses a specific line, you know instantly.

Technical Blueprint: 10KM LoRaWAN Geofence

1. The Scaling Problem (Distance vs. Signal)

In a 10 km radius, the signal follows the Inverse Square Law.

  • 0m to 500m: RSSI drops significantly (e.g., -40dBm to -90dBm). High precision.
  • 1km to 10km: RSSI drops very slowly (e.g., -110dBm to -120dBm). Low precision.
  • The "Dead Zone": Beyond 5km, environmental noise (weather, buildings) is often stronger than the distance signal.

2. Recommended System Architecture

Feature Specification Reason
Hardware ESP32-S3 + SX1262 LoRa High processing power + best-in-class LoRa sensitivity.
Logic GPS-Triggered LoRa Use GPS for the "Fence" logic; use LoRa to send the "Warning."
Spreading Factor SF10 or SF12 Required to maintain a stable link at 10km range.
Antenna 5.8dBi Fiberglass Increases gain to ensure the 10km boundary is reachable.

3. Implementation Logic (The "Smart" Fence)

To make a 10km fence reliable, do not use RSSI as the only trigger. Use this hybrid logic:

A. The "Keep-Alive" Pulse (RSSI)

The tracker sends a small packet every 60 seconds.

  • Logic: If the Gateway misses 3 consecutive packets, trigger an "Out of Range / Connection Lost" warning. This is your fail-safe if the device is destroyed or enters a dead zone.

B. The GPS Geofence (Precision)

The wearable (child's device) monitors its own GPS coordinates.

  • Logic:
    1. Define a Home_Center (Lat/Lon).
    2. Calculate Distance_to_Home on the device.
    3. If (Distance_to_Home > 10,000 meters) -> Send a High-Priority LoRa packet: "FENCE_BREACHED".

4. Advanced Filtering for RSSI

If you insist on using RSSI for the 10km boundary, you must use a Kalman Filter or Exponential Moving Average (EMA) to prevent false alarms.

Formula for EMA: $$RSSI_{filtered} = (\alpha \times RSSI_{new}) + ((1 - \alpha) \times RSSI_{old})$$ (Use $\alpha = 0.1$ for heavy smoothing at long distances).


5. Why RSSI alone fails at 10KM:

  • The "Jitter": At 10km, a child standing still will have an RSSI that jumps between -115 and -122.
  • The "False Trigger": An RSSI of -120 could mean "I am 10km away" OR it could mean "I am 2km away behind a large concrete building."
  • Conclusion: For a 10km goal, GPS over LoRa is the professional standard. RSSI is perfect for a 50-meter backyard fence, but dangerous for a 10,000-meter safety zone.

ref