Maritime Operational Intelligence: AR-IoT Synergies for Energy Efficiency and Emissions Control

Abstract

In response to mounting regulatory and environmental pressures, the maritime sector must urgently improve energy efficiency and reduce greenhouse gas emissions. However, conventional operational interfaces often fail to deliver real-time, actionable insights needed for informed decision-making onboard. This work presents an innovative Augmented Reality (AR) interface integrated with an established shipboard data collection system to enhance real-time monitoring and operational decision-making on commercial vessels. The baseline data acquisition infrastructure is currently installed on over 800 vessels across various ship types, providing a robust foundation for this development.

To validate the AR interface's feasibility and performance, a field trial was conducted on a representative dry bulk carrier. Through hands-free AR smart glasses, crew members access real-time overlays of key performance indicators — fuel consumption, engine status, emissions levels and energy load balancing — directly within their field of view. Field evaluations and scenario-based workshops demonstrate significant gains in energy efficiency (up to 28% faster decision-making), predictive maintenance accuracy and emissions awareness. The system addresses human–machine interaction challenges in high-pressure maritime settings, bridging the gap between complex sensor data and crew responsiveness. By contextualizing IoT data within the physical environment, the AR-IoT platform transforms traditional workflows into proactive, data-driven practices, accelerating compliance with International Maritime Organization (IMO) decarbonization targets.

1. Introduction

The maritime industry carries over 90% of world goods yet contributes approximately 3% of global CO2 emissions [1]. As decarbonization becomes imperative, stakeholders face IMO carbon-intensity targets [2,3]. Research spans carbon-aware vessel scheduling and berth allocation [4], onshore power supply cutting CO2 in Spanish ports such as Santander [5], Northern Sea Route assessments [6], blockchain logistics with policy frameworks [7] and sector-wide digital transformation [8] — alongside green-transformation success factors [9] and sixth-generation smart ports using automation, IoT and AI [10,11]. Yet translating complex real-time IoT data into actionable insights at the point of need, in dynamic high-stakes operations, remains persistently challenging.

Augmented Reality (AR) offers a promising, underexplored maritime solution [12,13]: embedding real-time data in the field of view enables hands-free, spatially contextualized interaction with complex systems, improving responsiveness, reducing cognitive load and supporting sustainability decisions such as emissions mitigation and fuel optimization [14]. AR–IoT integration has proven itself in manufacturing and automotive sectors, but maritime energy applications remain limited [15]; harsh onboard environments, intermittent connectivity and technology resistance slowed adoption [16]. Still, the potential to bridge machine intelligence and human decision-making — for predictive maintenance, energy efficiency and environmental compliance — is substantial.

This study introduces a large-scale AR-IoT platform deployed on over 800 vessels [17], empowering crews with immersive operational intelligence through real-time visual overlays, context-aware alerts and intuitive interaction for more proactive, sustainable operations. It addresses three research questions: how AR enhances crew interaction with real-time energy and emissions data; which technical and human-centered challenges arise during AR-IoT deployment; and how this convergence impacts sustainability performance in operational maritime contexts.

2. Contribution of the Work

This paper advances sustainable maritime operations with a novel AR–IoT integration: unlike monitoring systems isolating data from the point of action, it embeds real-time, context-aware information directly into the operational environment via AR smart glasses. Hands-free interaction with situationally relevant data — energy load, vibration anomalies, emissions thresholds — enables more responsive, efficient and environmentally conscious decisions. Innovation lies both in implementing AR interfaces under harsh maritime conditions and in large-scale deployment with empirical evaluation across 800 vessels.

Through scenario-based testing and field data collection, the study quantifies effectiveness in energy efficiency, predictive maintenance and emissions control — key levers for international decarbonization targets. The main contributions are: design and deployment of a wearable AR interface integrated with a vessel-wide IoT architecture for intuitive real-time interaction; empirical validation in real maritime settings with measurable response-time, accuracy and mitigation gains; insights into technical and human factors (adoption, interface design under variable lighting/noise, resilience in demanding marine environments); and scalability and sustainability potential, with strategic recommendations for fleet-wide AR-IoT scaling aligned with environmental compliance and digital transformation.

3. System Overview

The integrated AR-IoT system comprises three tightly integrated subsystems: (i) a wearable AR interface (smart glasses), (ii) an IoT-based data sensing and aggregation infrastructure, and (iii) a real-time data processing and visualization platform — engineered for maritime energy monitoring, maintenance and emissions control under vibration, lighting variability and safety-compliance constraints (Figure 1). The custom smart glasses, adapted from the SeeFar project architecture [18], provide spatially contextualized overlays through a transparent waveguide display: real-time data mapped onto physical components, visual/auditory threshold alerts, on-demand historical trends and diagnostics, hand-gesture and voice control, and crew localization via object detection (Table 1 details the Lingxi AW60 waveguide lenses, Intel RealSense D435 depth camera, ICM-20948 IMU, Snapdragon 845 SOM, BH1750/BME280 auxiliary sensors, wireless charging and Wi-Fi/Bluetooth connectivity).

Data flow spans mist, edge, fog and cloud layers for seamless flow under intermittent connectivity and heterogeneous sensors. Mist-layer smart signal collectors interface analog/digital sensors (4–20 mA, Modbus, CAN-bus) with conditioning, time-stamping and validation, forwarding via secure IEEE 802.15.4 mesh with maritime QoS layers to deck-level edge gateways performing high-frequency analytics (vibration pattern recognition, anomaly detection, embedded-ML predictive modeling [19]) and live AR distribution. The fog layer (onboard server with GPU/ML) stores historical data and triggers alerts; the cloud layer synchronizes via satellite for long-term storage, fleet benchmarking and optimization (Table 2). The cloud-native Decision Support Platform (DSP) consolidates all streams with time-series data, predictive outputs [20] and external inputs (weather, port scheduling) into KPI dashboards (fuel efficiency, vibration index, emissions per kWh), predictive maintenance alerts, geospatial 3D visualizations and automatic environmental compliance reports (Figures 3–4). Cybersecurity follows a multi-layered framework aligned with IMO Resolution MSC.428(98): TLS 1.3 ship–shore encryption, WPA3/Bluetooth 5.2 onboard, credential plus two-factor AR access, fog/cloud intrusion and anomaly monitoring, periodic scanning per ISO/IEC 27001 [21], and an API Gateway for secure vessel-headquarters exchange.

The UI is a spatially embedded information system built on minimalism, hands-free interaction and multi-modal feedback: only the most pertinent data displays per location, task and priority; control is via voice, hand gestures or side touchpad; color-coded visuals, audio and haptic vibration cues sustain situational awareness in low-visibility, high-noise spaces. Layered drill-down goes from summary indicators (asset status, emission scores, vibration indices) to time-series graphs and corrective actions on demand. Components (Table 3) — status badge overlays, metric cards, trend pop-outs, alert banners, task prompt windows, action recommendation bars — activate by context such as proximity, visual alignment, voice request or scheduled interventions (Figures 5–6).

4. Methodology

A rigorous mixed-methods design combined quantitative and qualitative techniques. A quasi-experimental design [22] with a pre-test/post-test control group design compared vessels without the AR system (control) against AR-equipped vessels (experimental) across three primary metrics — energy efficiency, predictive maintenance and user interaction. Prototype testing in a controlled environment with simulated vessel systems assessed latency, response times, communication stability and hardware/software integration against ISO/IEC 250002:2024 [23]; instead of synthetic data, two years of historical operational data from the same Panamax-class vessel were injected via a custom playback application, preserving authentic signal patterns, update rates and sensor noise. Field testing ran the AR interface on a single 225 m, ~75,000 DWT dry bulk carrier with 99.3% AR uptime and stable IoT acquisition within wireless tolerance limits.

Metrics (Table 4) covered fuel consumption and load balancing via difference-in-differences (DiD) analysis and regression models; downtime, predictive accuracy, MTBF and MTTR via maintenance-log comparison and confusion matrices; gaze tracking and task completion time via time-series and interaction-frequency analysis; and fuel savings, CO2/NOx reductions and operational efficiency via emission-factor analysis and vessel-specific KPIs. DiD fleet comparisons normalized pre-deployment fuel data for deadweight tonnage, average speed and ballast-to-laden ratio, excluding Beaufort > 6 voyages. Statistics used paired-sample t-tests (pre/post), one-way ANOVA across vessel types with Bonferroni correction, effect sizes (Cohen's d, partial η2), Shapiro–Wilk/Levene assumption checks, and qualitative thematic analysis of open-ended crew feedback triangulated with quantitative results.

Ethics approval included vessel-operator consent, GDPR and maritime data-protection compliance, crew briefing, minimal routine disruption, and anonymized, securely stored PII. Limitations: the six-month field window (complemented by the two-year historical prototype dataset) cannot fully capture long-term effects — a 12–18 month onboard study is planned; single-vessel AR trials limit environmental generalization despite the 800-vessel infrastructure footprint; and crew learning curves varied, potentially affecting early-phase effectiveness and motivating optimized training.

5. Use Cases and Results

Three representative use cases — energy optimization, predictive maintenance and emissions monitoring — were validated through scenario-based workshops with balanced cohorts (55% senior, 30% mid-career, 15% junior officers; engineering, deck and environmental-compliance roles). In energy optimization, AR overlays (fuel flow, generator load, efficiency indicators, contextual alerts, load-redistribution prompts; Table 5) replaced fixed-dashboard interpretation across three workshops with 18 engineers and deck officers: average response time to energy alerts fell 28% (4.6 to 3.3 min), corrective-action accuracy rose 13 points (78% to 91%), NASA-TLX workload dropped 25% (63.2 to 47.4) and SUS usability rose 10.5 points (71.0 to 81.5) (Table 6) — showing contextual on-location data bridges abstract information and physical action.

In predictive maintenance, AR overlaid equipment health, vibration/thermal readings, failure histories and inspection checklists at pumps, shafts, gearboxes and exhausts during walkthroughs (Table 7). Fourteen engineers diagnosing a simulated circulation-pump failure and overheating generator improved anomaly-detection time 40% (9.4 to 5.6 min), fault-localization accuracy 18 points (71% to 89%), decision confidence 1.2 points (3.2 to 4.4/5) and workload 24% (66.5 to 50.7) (Table 8) — transforming passive interpretation into information-rich inspection. In emissions monitoring, live CO2/NOx/SOx overlays on fuel and exhaust systems with geo-fenced ECA alerts, threshold-triggered audio-visual alarms, corrective recommendations (throttle, load, fuel switch), KPI trends and compliance snapshots (Table 9) let 12 fuel/emissions specialists act before exceeding limits: response time improved 33% (5.2 to 3.5 min), corrective accuracy 13 points (74% to 87%), compliance confidence 1.2 points (3.4 to 4.6/5) and workload 28% (62.8 to 45.1) (Table 10) — shifting emissions management from passive reporting to active mitigation toward IMO targets [27].

Challenges and lessons (Section 5.4): participant diversity — unfamiliar users showed reluctance while enthusiasts sometimes bypassed procedural fidelity (observer-amplified bias), and a restrictive subgroup resisted regardless — demands longer unsupervised deployments for true adoption assessment. Technically, emulated sensor streams and AR-rendering synchronization raised realism questions on latency and edge cases under variable connectivity, vibration and lighting, with legibility issues in dark/reflective engine rooms requiring adaptive contrast/brightness; and simultaneous multi-layer data risked cognitive overload, motivating context-aware content filtering. Still, spatially contextualized information was consistently valued — provided scale-up includes robust training, customizable interfaces and built-in safety measures.

6. Discussion

Operational intelligence: across all three use cases, AR-IoT enabled more intuitive, timely, context-rich crew interaction — a qualitative shift from reactive to proactive, data-driven decisions. Unlike removed dashboards, the platform embeds digital information in its physical context — situated operational awareness — minimizing cognitive fragmentation and placing intelligence at the point of need. Results extend prior maritime AR findings on navigation and situational awareness [12,13] into engineering and environmental domains with measurable outcomes, operationalizing the industrial-IoT operational-intelligence paradigm [28] through direct human–machine feedback.

Sustainability discourse: the study moves beyond theoretical IoT talk with a human-centered framework grounded in operational reality, casting AR as the enabling layer translating raw data into frontline-actionable information — with usability, workload and behavioral metrics keeping technical performance tied to safety-critical human experience. Predictive-maintenance gains parallel heavy-manufacturing AR results [29,30], but rarely validated under maritime vibration, interference and connectivity constraints — our empirical novelty. Proof-of-concept realism calls for the planned 12–18 month longitudinal deployment tracking reliability, trust and behavioral integration. Design, deployment and policy implications: adaptive interfaces scaling cognitive load by expertise, urgency and environment; longitudinal pilots with crew-centric onboarding; and per-vessel prerequisites — AR glasses, vessel-wide sensor networks (fuel flow, vibration, temperature, emissions analyzers), edge/fog compute, local Wi-Fi plus satellite backhaul, two-day role-tailored training plus one-month remote support.

Economics favor scaling: CAPEX USD 20,000–52,000 per vessel plus ~USD 4,000/year OPEX, against 3–6% fuel savings and 15–20% fewer unplanned maintenance events — USD 90,000–180,000/year for a Panamax carrier at ~20 tons/day and USD 500/ton bunker fuel, i.e., 18–24 month payback (faster in regulated zones/high-traffic trades). Alignment with green-port success factors [9], sixth-generation smart-port benchmarks [10], blockchain adoption [7], digital transformation [8] and route strategies [6] positions AR-IoT within wider Industry 4.0 maritime strategies. Human-factors analysis through Endsley's three-level situational-awareness model [31] (perception, comprehension, projection) and Wickens' multiple resource theory [32] explains faster, lighter decisions — with adaptive content filtering answering overload concerns, role-tailored onboarding cutting task times a further 12%, and emergency overlays (fire, failure, gas-leak hazards, evacuation routes, checklists, shore-specialist live streaming) extending augmented intelligence — technology supporting rather than supplanting human expertise — into time-critical response.

7. Conclusions

Integrating Augmented Reality and IoT presents a promising pathway for operational intelligence and environmental sustainability in maritime operations. Situating real-time data within the crew's field of view through wearable AR interfaces significantly improves energy management, predictive maintenance and emissions monitoring: faster decisions, reduced cognitive workload and increased task accuracy support both efficient operations and decarbonization objectives. Crucially, these results mark a shift in how operational information is accessed and acted upon — a tightly coupled feedback loop between environmental conditions, system performance and human response empowering sustainability-aligned decisions in real time rather than via post-event reporting or control-room assessments.

Initial deployment results are promising, but longer-term effects under varying maritime conditions need study; extended deployments across diverse vessel types and geographies will clarify scalability, robustness and behavioral integration, while adaptive interfaces responsive to expertise, task complexity and environment will drive broad fleet adoption. Advanced predictive analytics and machine learning could add emissions forecasting, fault prediction and voyage optimization; closer regulatory collaboration could embed such technologies in compliance mechanisms, digital certification schemes and incentive policies bridging regulatory ambition and practical execution. Digital innovation must play a central role in maritime sustainability: AR and IoT deployed at scale improve operational and environmental performance, laying groundwork for more efficient, intelligent, resilient systems aligned with the climate commitments shaping maritime transportation's future.