Advanced & Sustainable Manufacturing

Integrating A*STAR I2R’s capabilities in Artificial Intelligence, Computer Vision, Data Analytics, Robotics, Connectivity, Sensory, Cybersecurity and Sustainability R&D, we are providing Industry 4.0 and Industry 5.0 capabilities to the Manufacturing, Engineering and Aerospace MRO industries.

We co-develop cutting-edge technological solutions for digitalisation with our industry collaborators, to enhance productivity and improve operational efficiency. These include vision-based anomaly detection, part-identification, real-time process optimisation & control, next-generation logistics and supply chain, large-scale automated inspection, and predictive maintenance.

We focus on the research and development of 3D computer vision technologies to transform manufacturing by enhancing precision, automation, and safety. Manufacturers can use these technologies to improve or automate production processes.

We specialise in the following areas:

3D object identification

Incremental learning and continual learning were used to identify industry products or assets for product registration and verification, as well as tracking the assets to monitor repair status in MRO processes.

3D segmentation

A 3D deep learning technology that requires less training data has been developed. It enables more robust 3D defect detection and quantification without the need of a lot of sample data, which can often be defective. This results in more reliable and quantifiable defects detection in manufactured items.

3D reconstruction

Specialised techniques such as structured light, stereo vision, Time-of-Flight (ToF), and laser triangulation, enables the detection and measurement of dimensional defects in products by accurately reconstructing their 3D models. It also enables the reconstruction of 3D models of factory assets to enable the development of digital twins.

3D mapping

3D mapping technology enables the understanding of our world from a spatial perspective for urban planning, navigation, and scientific research.

AWTO Platform under deployment

3D reconstruction and digital twin in manufacturing

Using generative AI to assist technicians in the MRO processes, proprietary, slim, and affordable AI expert systems with domain-specific knowledge can be customised for various industry sectors.

Human-AI collaboration combines human expertise with AI's analytical power to enhance problem-solving and decision-making. Automated visual inspection uses AI to analyse visual data, to detect anomalies, defects, or patterns in real-time, significantly improving quality control in manufacturing and other industries.

Automated MRO (Maintenance, Repair, and Overhaul) task tracking employs AI-driven systems to streamline and optimize the monitoring and management of maintenance activities, facilitating operational efficiency and minimizing downtime.

One key trend is integration of Visual Intelligence, Computer Vision, and AI technologies to digitalise and automate manual inspection processes, facilitating the seamless transition from labour-intensive manual inspections to more efficient and precise automated systems.

Research in AI visual inspection spans across various key areas, focusing on enhancing inspection capabilities for large surfaces like aircraft and car exteriors, as well as intricate component assemblies such as aircraft engine oil-bearing supply tubes. It also includes AI-enabled MRO task tracking and monitoring, which involves developing models deployable on wearable or mobile devices. These devices offer real-time augmented reality instructions, issuing reminders upon error detection, and monitoring task progress seamlessly.

Defect Detection

Activity Recognition and Task Monitoring

We focus on digital and sustainable manufacturing solutions such as digital twin systems, AI-powered predictive maintenance, supply chain optimisation, and sustainability, across sectors like pharmaceuticals, aerospace, food and beverage, machinery, semiconductors, and supply chain management. Our digital twin systems integrate real-time data for process visualisation, while AI algorithms enhance predictive maintenance for equipment reliability. We optimize supply chain operations and use time series forecasting for efficient resource allocation. Prioritizing sustainability, we implement energy optimization and waste reduction strategies. Our efforts support Singapore's mission to lead in smart manufacturing and align with global Industry 5.0 trends.

We focus on digital and sustainable manufacturing solutions across sectors like pharmaceuticals, aerospace, food and beverage, machinery, semiconductors, and supply chain management. We focus on four key areas: digital twin systems, AI-powered predictive maintenance, supply chain optimization and time series forecasting, and sustainability in manufacturing.

Our digital twin systems integrate real-time data for process visualisation, while leveraging AI algorithms to enhance predictive maintenance for equipment reliability. We also optimise supply chain operations and conduct time series forecasting for efficient resource allocation. Lastly, prioritising sustainability, we implement energy consumption optimisation and waste reduction strategies throughout the product lifecycle.

The capability of digital and sustainable manufacturing drive us towards the trend of Industry 5.0 and the broader ecosystem of smart manufacturing.

There are 4 areas that Digital and Sustainable Manufacturing Unit focuses on:

Digital Twin Systems: Focuses on developing digital twin platforms that integrate real-time data and provide visualisation for manufacturing processes. Additionally, it includes deploying AI models on cloud or on-premises infrastructure and optimising digital twin simulation capabilities.
AI-Powered Predictive Maintenance: Specialises in analysing sensory time series data and overcoming data challenges such as scarcity, imbalance, and low-quality data using AI techniques. This area also involves implementing deep learning for anomaly detection, predicting remaining useful life, and developing explainable AI models and smart sensing solutions.
Supply Chain Optimisation and Forecasting: Concentrates on real-time monitoring and control of supply chain operations to minimise waste and maximize efficiency. It also includes resource scheduling and process optimisation for manufacturing, as well as utilising generative AI for time series forecasting such as demand, price, and load forecasting.
Sustainability in Manufacturing: Focuses on optimising energy consumption, reducing carbon emissions, and minimising waste in manufacturing processes. This area also involves developing tools for tracking the product lifecycle and implementing blockchain technology for transparent supply chains.