A portable tool to allow non-experts to identify key components and assets needing immediate or predictive maintenance and system that identifies known or unknown weaknesses in the supply chain.

Use case: A portable tool to allow non-experts to identify key components and assets needing immediate or predictive maintenance and automate and streamline the spare / replacement parts supply chain.

1.0. The primary focus of this project will be on the devices defined by the project partner. To this end, the AI-enabled portable maintenance identification and assistance tool will be adapted to:

1.1. Maintenance and Diagnostics: Enhance the maintenance procedures by quickly identifying faults or potential issues in the defined system. system components, thereby reducing downtime and increasing operational efficiency.

1.2. Training and Support: Provide real-time assistance and training to soldiers in the field, ensuring they can effectively maintain and troubleshoot the systems even in multiple scenarios and conditions,

1.3. Data-Driven Insights: Leverage AI to analyse usage and maintenance data, offering actionable insights to improve the reliability and performance of the system being maintained over time.

2.0. Data Collection and Computer Vision / AI Training Data Creation Process Objective: Develop a labelled image library encompassing parts, systems, and components, and establish both a training and test dataset.

2.1. Identify Components: Catalogue all relevant parts, systems, and components. Capture detailed images and specifications for each item using computer vision technologies.

2.2. Data Annotation: Utilise computer vision technologies to label images with precise metadata. Ensure labels include part names, serial numbers, and other distinguishing features.

2.3. Dataset Creation: Segment the data into training and test sets. Apply data augmentation methods to introduce variability in the images (e.g., different angles, lighting conditions).

2.4. Validation: Verify the dataset for accuracy and completeness. Modify labels and data distribution as needed.

3.0. Develop Process Map for Field => Supply Chain => Part Delivery for MendID Process Objective: Identify known or unknown weaknesses in the supply chain and recommend fixes and optimisations for the partner defined system.

Steps:

3.1. Data Collection: Use Siloed Retriever to gather comprehensive data on the supply chain for the system, including part identification, tracking, and delivery timelines.

3.2. Process Mapping: Use Siloed AI to analyse collected data and create a detailed process map. Identify key stages: field identification, supply chain logistics, inventory management, and part delivery.

3.3 Weakness Identification: Utilise Siloed AI’s advanced analytics to pinpoint weaknesses in the supply chain, such as bottlenecks, delays, and inefficiencies. Differentiate between known issues (historical data) and potential unknown issues (predictive analytics).

3.4. Optimisation Recommendations: Develop strategies for process optimisation, such as just-in-time delivery, automated inventory updates, and enhanced tracking mechanisms. Use predictive analytics to forecast potential supply chain disruptions and suggest pre-emptive measures.

3.5. Validation and Refinement: Validate the process map and optimisation recommendations with key stakeholders. Refine based on feedback and additional data analysis.

4.0. Define Prototype Development Plan for Adapting MendID for Partner Use Cases Objective: Define a detailed prototype development plan to adapt the MendID system for partner use cases.