AIDI · Applied outputs
Software systems, platforms, physical prototypes and decision components developed through applied R&D.
The catalogue brings together 30 technical results transferred from the NEXUS and ILLIANCE work: 16 from NEXUS and 14 from ILLIANCE. It includes IoT platforms, Control Towers, digital twins, optimisation, forecasting, computer vision, edge computing, diagnosis and analytics systems.
Descriptions were normalised for the new AIDI website and contain no links to the legacy websites. Detailed technical information, versions and scope can be updated as the results evolve.
Core IoT platform for collecting, processing, visualising and managing physical and virtual device data. It supports common IoT protocols, dashboards, device and asset management, alarms, rule-based automation, integrations and APIs for scalable solutions.
Module that replays historical data as current streams for realistic platform testing and demonstrations, with calendar alignment options and configurable accelerated injection speed.
Unified operational-intelligence environment that brings maritime analytics, optimisation, simulation and monitoring components together in a real-time Control Tower.
Interactive benchmarking tool combining Multi-directional Efficiency Analysis, visual analytics and a conversational AI assistant to identify equipment inefficiency drivers and support operational decisions.
Python-based digital twin using discrete-event simulation, battery charge/discharge models and optimisation to compare terminal scenarios and stress-test resilience across quay, yard and landside operations.
Web application combining machine learning and explainable AI for prediction, root-cause analysis and actionable recommendations, including SHAP, PDP and ICE visualisations and external weather data.
Decision-support system for berth allocation and quay-crane assignment, integrating real-time vessel data, operational constraints and optimisation to reduce waiting time and improve resource use.
Real-time terminal monitoring dashboard for vessel arrivals and queues, berth occupancy, quay-crane availability and service status, plus historical-period exploration.
Analytics platform that ingests real-time container events, classifies import/export/transshipment flows, predicts dwell time and uses an algebraic-topology-based anomaly index to flag disruptions and emerging risks.
AI-powered intelligence platform that monitors global news and events, then discovers, filters, summarises and categorises developments relevant to maritime and port operations using configurable topics and language models.
Continuous-improvement engine that generates periodic reports, identifies good and poor operating practices and computes indicators reflecting the relevance and urgency of potential improvements.
Dashboard for exploring selected Port of Sines KPIs, developed under a systematic scientific study of user-interface and user-experience usability.
Control Tower component providing real-time visibility of truck movements and transport capacities using MQTT event processing and AnyLogic-based simulation to track fleet location, load state and utilisation.
Forecasting platform for sparse, zero-inflated transport-capacity time series, using meta-ranking algorithms to select and combine models and exposing prediction capabilities through an API.
Interactive web and Control Tower component for configuring, predicting and optimising fleet arrival windows, routes and speeds, including Monte Carlo stress testing under traffic and weather disruptions.
Multimodal freight quotation engine that evaluates direct and multi-leg alternatives using cost, CO₂ emissions and transit time, with interoperable APIs and provider plugins.
Core IoT platform for collecting, processing, visualising and managing industrial device data, with standard protocols, dashboards, alarms, rule-based automation, integrations and APIs.
Unified real-time environment that integrates manufacturing analytics, optimisation, simulation and monitoring applications into a shared operational view.
Automated Exploratory Data Analysis system for heat-pump manufacturing data, combining AutoML, local LLM-based metadata enrichment, an ML-readiness metric and FAIR data-maturity reporting, tested with real Nexeed MES data.
Web application for filtering, selecting and analysing Nexeed MES data to model manufacturing-process factors such as sensor variation, assembly errors and rework, validated with real heat-pump production batches.
Edge-computing computer-vision system for operator ergonomics and safety, including danger-zone detection, task monitoring, picking navigation and hand/glove tracking, with laboratory validation.
Edge-computing vision module for task-quality control, using the industrial validation pipeline for task monitoring, picking navigation and hand/glove tracking while keeping processing close to the production environment.
Time-series analysis application for modelling environmental variables that influence heat-pump production tests, including pressure, temperature and humidity, integrated with the diagnosis framework and Control Tower.
Microservices-based supervisor aggregating diagnosis modules for failure diagnosis and interpretable Root Cause Analysis, using real industrial data and Belief Propagation components on AIDI computing infrastructure.
Operational reduced-scale production-line prototype at AIDI, combining object detection, human and hand movement tracking, edge sensing and retroprojection with visual feedback.
High-volume trial-data storage and processing platform using NoSQL storage, asynchronous batch processing, a Python API, web monitoring and AnyLogic-based real-time production-line simulation.
Docker microservices platform for spare-parts consumption forecasting, with forecast storage, REST API, periodic hyperparameter retraining and integration with SAP/Bosch data feeds.
API/application supporting product and maintenance decision workflows with dashboards, forecasts and model rankings, including local language-model extraction and sentiment signals for time-series enrichment.
Application for reconstructing and analysing workflows to identify process patterns and bottlenecks, designed for integration with the wider AIDI analytics and operational-data stack.
AI-powered intelligence platform that monitors global developments relevant to heat-pump operations, using configurable topics and language models to discover, summarise and classify supply-chain, geopolitical, labour, regulatory and market events.