Non-Invasive HEC Early Warning System The Problem: A Crisis of Fragmentation The Western Ghats of Karnataka are experiencing a severe ecological shift. In districts like Uttara Kannada, interior forest cover has declined from 64.4% to 25.6% over the last four decades. This fragmentation, combined with over 34,000 reported conflict cases in five years, has made the "0-minute warning" a lethal reality for local communities.
The Solution: Digital Sentinel Logic Our project moves beyond physical barriers to provide a proactive, non-invasive safety window.
- Seismic Signal Analysis (Simulated) The Seismic_Elephant_vs_Truck_Simulator.html demonstrates the core AI/ML logic:
Edge Processing: Utilizes a Butterworth-style IIR low-pass filter (fc = 12Hz) to strip away high-frequency noise from vehicles.
Pattern Recognition: Implements an FFT-based analysis to identify the specific 1–10 Hz signature of elephant footfalls.
Accuracy: Distinguishes rhythmic megafauna movement from background seismic noise.
- The Digital Twin & Command Center The interactive2.html provides a full-system simulation:
Advanced Warning: Calculates an 8.8-minute warning window—a 540% improvement in readiness.
LoRa Connectivity: Simulates the transmission of alert packets across 5–10 km without cellular dependence.
Impact Visualization: Shows the localized siren and SMS alert chain that allows farmers to reach safety before an encounter.
Technical Specifications Node Cost: ₹2,650 (~$32 USD) per unit, compared to ₹1.5 Lakh+ for traditional fencing.
Power: 24/7 off-grid operation via 6V solar panels and Li-ion batteries.
Detection Range: 150 m radius per geophone node.