— EPO application EP25205451.5.

Helm Order Monitor (HOM) uses a beam-forming microphone array, noise-robust speech recognition and live rudder-angle data to instantly verify every spoken helm command during manual steering. It triggers “WRONG HELM” or “CHECK THE RUDDER” alarms if orders are mis-executed or delayed, and incorporates Speech Emotion Recognition to generate a Unified Stress Score for early warning of high-stress conditions—without ever overriding human control.

Aware Mate uses privacy-by-design computer vision (YOLOv8 + RetinaFace/ShuffleNet) to monitor eyelid closure (PERCLOS), gaze, head pose and posture locally on the vessel, detecting fatigue and distraction before they cause errors. Both modules run entirely offline, integrate with existing Bridge Navigational Watch Alarm Systems (BNWAS), and are engineered to comply with GDPR and the EU AI Act’s “high-risk” AI requirements.

Founded in 2021 by captain Hrvoje Mihovilović after three decades at sea, ELNAV.AI has secured €500K in non-dilutive grants, five innovation awards, and strategic partnerships with Fraunhofer IIS, Audeering GmbH, Axis Communications, and the Faculty of Maritime Studies in Split. The company is preparing for 2025 sea trials and Det Norske Veritas type-approval, targeting a global market of over 60,000 SOLAS-class vessels. Its business model combines one-off hardware sales with recurring software licences, and an OEM integration roadmap for scale.

Beyond bridge safety, ELNAV.AI is developing environmental protection technologies, including Project Sentinel—an AI-enabled underwater glider system for real-time detection of illegal marine discharges—and Project Sono Aegis, a distributed-acoustic-sensing (DAS) early-warning “guardian” cable for protecting submarine infrastructure.

ELNAV.AI’s vision is to make real-time, privacy-respecting human-factor sensing and autonomous environmental monitoring a safety standard on every vessel, reducing accidents, lowering insurance costs, and safeguarding both crews and seas.