Vessel Guard-AI Based Human Factor Risk Detection

Created by team MH4seafarers on February 07, 2026
Autonomous Robotics Control in Simulation

Vessel Guard is an AI-assisted maritime safety system designed to prevent human-error-related accidents caused by fatigue, stress, and reduced operational readiness onboard vessels. Life at sea exposes crew members to shift work, sleep fragmentation, vibration, isolation, and high-risk tasks in deck, engine, and bridge operations. While many shipping companies provide access to teletherapy with reactive tools and disconnected from daily operational safety decisions. They do not integrate fatigue thresholds, maritime regulations, or structured escalation to command. Vessel Guard addresses this by combining short daily crew check-ins with a deterministic rules engine based on MLC fatigue standards and wellbeing scales. The system translates human factors into operational risk levels ) and escalates only high-risk cases to the Captain or designated officer. The AI layer does not diagnose or classify risk. It extracts structured data and generates safety-focused summaries, while final decisions remain human-controlled. This human-in-the-loop design ensures regulatory alignment, privacy protection, and institutional credibility. Vessel Guard is designed for maritime sector seeking proactive duty-of-care compliance, improved safety culture, and measurable reduction of fatigue-related incidents—without adding administrative burden to crew.

Category tags:Additional links:

"Description: Maritime safety system for crew fatigue. Daily check-ins, deterministic rules engine based on MLC fatigue standards, AI extracts data but humans decide. For shipping companies. Features: Crew check-ins MLC fatigue standards Risk level translation Escalation to Captain Human-in-the-loop Maritime regulatory compliance Demo: http://136.244.86.246/login Pros: ✅ Unique domain (maritime) ✅ Real problem (fatigue at sea) ✅ Working demo ✅ Human-in-loop design"

avatar

Sanem Avcil

"Overview: This is a solid idea. I see its potential and how it can help, but it relies on manual data entry. My first question is, how do you define stress level? How can a person say he is at level 1 or 10? Is there any validation or verification of user input? The employee can intentionally put in higher values to relax more, or input small amounts to get extra pay (even though the person can have higher fatigue). So, how can your application prevent such cases? Pros: Solid foundation and understanding of the problem. Cons: I am not sure if it is using AI. Based on the demo, it is just simple if-else checking on the backend side, which then gives you a green or red result. "

avatar

Anton Kiselev

Lead Backend Developer