NETRA: A Decentralized LoRa-MANET Ecosystem for AI-Driven Collaborative Reconnaissance in GPSDenied Environments

Modern military operations frequently encounter communication blackouts and GNSS-denied environments caused by electronic warfare and infrastructure destruction. This paper proposes NETRA (Networked Emergency Tactical Radio Awareness), an integrated helmet-mounted ecosystem designed for resilient indigenous soldier-level situational awareness. NETRA acts as a fallback communication system. The system fuses three core capabilities. First, a Hybrid Positioning Subsystem combining GPS for macro-localization and Ultra-Wideband (UWB) for less than 30 cm intra-squad relative tracking in jammed environments. Second, an Edge-AI Reconnaissance engine utilizing quantized YOLO-based models on ESP32 hardware to detect human, weapon, and tank threats with less than 1 second latency. Third, an Encrypted LoRa-based Mobile Ad Hoc Network (MANET) for multi-hop tactical data relay. By transmitting 16-byte metadata packets instead of raw video, NETRA bypasses LoRa’s bandwidth constraints while maintaining end-to-end AES-128 security. Experimental prototyping demonstrates a 1 km Non-Line-of-Sight (NLOS) range, 8.7 km Line-of-Sight (LOS) range, and an energy-aware operational endurance of 13.3 hours. This architecture establishes a persistent shared knowledge bank, delivering critical intelligence via an eyes-free, in-ear audio interface to enhance decision-making in infrastructure-denied tactical theaters.