Project RAAC-AIS
Continuous automated validation to secure maritime data streams, utilizing the JDS Ship Monitor system.
Project description - short
The objective of the RAAC-AIS project is to convert crowdsourced data from the maritime sector into an industry-grade AIS data stream, using artificial intelligence and measurement technology for near-real-time validation.
RAAC-AIS - Real-time filtering of Crowdsourced Maritime Data
The project addresses a critical gap: no existing solution provides near-real-time, globally scalable anomaly detection for crowdsourced AIS streams. RAAC-AIS combines two complementary scientific approaches to solve this:
- Uncertainty Quantification (UQ): Each incoming AIS message receives a numerical trust score derived from timing deviations, spatial plausibility, MMSI consistency, and signal metadata.
- AI-Based Anomaly Detection: Scalable machine learning methods, including density-based clustering and physical ship-dynamics modeling, identify spoofed signals, falsified timestamps, and malicious data injection.
Developed by JDS Ship Monitor, the system runs on a fully cloud-native AWS infrastructure, processing up to 600 signals per second across approximately 100,000 tracked vessel objects in near real-time.
Why it matters
Safety & compliance
Protect navigation and critical port infrastructure.
Operational continuity
Early warnings reduce downtime, diversions, and claims.
Evidence you can trust
Time-stamped, forensically traceable records for audits and insurers.
Data Integrity Program
RAAC-AIS (Real-time Analysis, Assessment and Filtering of Crowdsourced vessel movement data from AISHub) is a funded research project by the Technologie-Beratungs-Institut GmbH (TBI) and incubated within the Regional Innovation Strategy for Smart Specialisation of the state of Mecklenburg-Vorpommern, targeting the cross-sectional technology areas of Big Data Analytics and Secure & Connected Systems.
Through this project, JDS Ship Monitor aims to process over 50 million AIS messages per day, drawn from the crowdsourced network, transforming raw, unverified signals into a high-fidelity information stream for the global research community.
What you can access
Research institutions and maritime innovation projects can partner with JDS Ship Monitor to access validated AIS data via three delivery formats:
- SFTP file export – Periodic batch files of filtered vessel states including position, speed, heading, and weather aggregation. Suitable for historical analysis and model training.
- Live data stream – Configurable real-time AIS stream with selectable field sets, vessel types, and geographic filters.
- API access – Virtual REST APIs delivering vessel state snapshots, provider statistics, and enriched ship metadata on demand.
All data is delivered separately from JDS's commercial production systems, ensuring full research reproducibility and no contamination from unverified sources.
Partner with us to enhance your maritime research with high-integrity data.
Regional Innovation Strategy for Smart Specialisation (RIS3)
RAAC-AIS is a research and development project by JDS Ship Monitor. The project is funded within the EFRE Fonds 2021 - 2027 of the State of Mecklenburg-Vorpommern from resources of the European Fund for Regional Development (ERDF) of the European Union.
The project is managed and supported by the Technologie-Beratungs-Institut GmbH (TBI).
- Project Management Agency: Technologie-Beratungs-Institut GmbH (TBI)
- Project Number: TBI-1-163-U-050
The project specializes in the cross-sectional technology areas of Big Data Analytics and Secure & Connected Systems as defined by the Regional Innovation Strategy for Smart Specialisation (RIS3).
About the RAAC-AIS Project
Project Content and Milestones
From data analysis to the implementation of the cloud-based architecture, the project follows a precise development plan. The subsequent milestones document the essential phases of this research work.
Project Objectives
The RAAC-AIS project focuses on real-time analysis, assessment, and filtering of crowdsourced vessel movement data. Using advanced measurement techniques and artificial intelligence, the platform automatically detects and clears data anomalies, time-stamp corruptions, and intentional spoofing events in quasi-real-time.
Evidence-based Research
Requirement Analysis: Developing stakeholder-driven engineering specifications for technical infrastructure, interfaces, and the data model.
Technical Translation: Formulating architectural plans for cloud processing modules and mapping cross-component infrastructure needs.
IT Structure Setup: Setting up isolated testing structures, centralized access tokens, secure credential vaults, and continuous integration pipelines.
Server Procurement: Calculating and pre-allocating necessary computing nodes, persistent volumes, and CPU times within optimized service brackets.
Project Management & Cloud Setup
Data Procurement: Acquiring deep-sea satellite feeds and global historical data sets to calibrate core behavioral models.
Raw Storage Management: Architecting block storage vaults and defining message header parameters to log precision server arrival times.
Vessel Database Building: Developing multidimensional searchable master record structures capable of tracking lifecycles and registration changes.
Weather Data Fusion: Integrating environmental models including wind vector paths and sea currents to analyze speed variations.
Station Management: Programming database logic to track provider behaviors and establish safety blocklists for malicious ingress channels.
System Stream Monitoring: Creating custom telemetry dashboards to monitor overall dataset health, cost frameworks, and operational thresholds.
AIS Data Decoding: Creating tailored ingestion layers to evaluate sentence structures, catch corrupt entries, and purge duplicate traffic.
High-Availability Partitioning: Deploying robust cloud tables indexed via high-performance geohashes and unique identifiers.
Security Auditing: Performing threat evaluations, assessing infrastructure vulnerabilities, and generating BSI compliance registers.
Data Management & Processing Infrastructure
Vesselstate Core Construction: Compiling microservices to stitch master static characteristics with streaming geographic coordinates.
SFTP Download Management: Engineering file-transfer gateways incorporating automated key rotations and specific output schedulers.
Multivariable Live Streams: Developing data piping connections across multiple protocol targets to pass structural data feeds securely.
Virtual API Gateways: Building elastic application layers protecting backend systems from high-volume automated requests.
Uncertainty Quantification: Executing numeric models to establish reliable confidence scores for unverified spatial entries.
Cloud Microservices Adaptation: Migrating abstract logic definitions into concurrent software processes across internal software services.
Stream Assessment Audits: Running joint code reviews to ensure live processing paths accurately categorize metadata.
Station Quality Rules: Evaluating historical coverage bounds and traffic flags to automatically isolate broken entry stations.
Field Correction Logic: Writing criteria protocols determining when individual attributes should be logically filled or rejected.
Algorithmic Logic & Verification
Demonstrator Proof of Concept: Formulating swift spatial indexing engines to support concurrent web visualizations.
Middleware Architecture: Establishing secure operational session tracking layers and implementing anti-bot defenses for web assets.
Frontend Engineering: Building interactive control layouts and map tiles to view live and cleaned maritime data layers.
Intellectual Property Review: Tracking work achievements to determine registerable engineering rights or protection paths.
Final Project Presentation: Disseminating accumulated research knowledge and presenting the unified project software platform.
Demonstrator & Results Exploitation
Provide maritime researchers with cost-effective access to globally filtered, archive-quality AIS data — previously only available through expensive commercial data streams.