Trade, economy, and fleets are greatly impacted by asset status and management. Transportation of food, medicines, consumer goods, and industrial products all depend on the effective operation of fleets, which range from automotive vehicles to airplanes and container ships.
Over years of interactions with various customers and associated stakeholders, Hitachi has learned that a critical challenge in the transportation industry is to maintain the fleet assets in good condition to increase safety and avoid downtime. As such, it is crucial to proactively identify possible defects and reduce the impact of degradation in assets by an effective maintenance process. Inspection is a key practice to effectively find defects and streamline the maintenance process. Nevertheless, the inspection process requires appropriate data collection and the associated quality of assessment greatly impacts the efficiency of maintenance plans and of repair time and cost. In the current transportation industry, inspection is mostly a manual process where human operators visually inspect and identify defects in assets. Therefore, the current process can be improved by considering the shortage of skilled labor, subjectivity, and inconsistent procedures.
Consequently, Hitachi’s Global Center for Social Innovation – North America is developing a visual inspection system in the transportation domain to address these challenges using AI, machine learning, and computer vision. Some of the key technologies powering the system are:
- Inspection planning service: Automatic inspection planning to guide operators based on the type of inspection, type of asset, or detected defect.
- Inspection library: A library of plans to enable consistent, automatic, and scalable inspection execution based on the type of inspection or type of asset.
- Automated inspection: Enabling robots and drones to perform an inspection process based on the plan generated by the planning service.
- Defect analysis: An AI-based analysis system for multi-level defect detection that enhances accuracy and reduces false positives.
- Vehicle defect library: A library of images related to assets of interest and their defects to enable identification and localization of defects based on AI techniques.
Hitachi is integrating these technologies into an end-to-end visual inspection system for defect detection with high accuracy along with a consistent and repeatable process. The system incorporates new technologies for systematic data collection via robotics, human operators, and fixed instrumented cameras as well as drones; uses AI to analyze and identify defects; and also integrates design thinking to convey the results of the system to decision makers.
Hitachi is working on integrating this system with its other fleet maintenance and repair solutions like an AI-based repair recommendation engine, part inventory, and an enterprise resource planning (ERP) system to develop an end-to-end inspection and repair service system.
(Hitachi America, Ltd.)