Multifunction devices operating in offices may sometimes produce unpleasant noise during operation, requiring a customer engineer (CE) visit for repair and service recovery. However, there have been cases where the abnormal noise could not be reproduced during the CE’s maintenance visit or insufficient information prevented the identification of the faulty component, resulting in prolonged resolution times and causing inconvenience to customers.
To resolve this issue, we have equipped the multifunction devices themselves with new noise detection and identification capabilities and built a cloud-based system that collects and analyzes the characteristics of an abnormal noise from a device. Then, we have launched an operational system that enables CEs to quickly replace noise-causing components using diagnostic results generated by the cloud-based system (Fig. 1).
Two sound sensors installed inside a multifunction device collect operating sound only while the device is in operation. The sound data is analyzed using an AI engine for abnormal noise detection embedded in a newly developed ASIC within the device to determine whether an abnormal noise is present. Furthermore, by optimizing the timing of noise collection and the sensor placement, we have improved the abnormal noise detection rate (Fig. 2).
If an abnormal noise is detected, the operating sound data is converted into binary data that cannot be reconstructed into the original sound data. This binary data, along with the operational log data for each component, is transmitted to the cloud-based Remote Diagnostics System.
Diagnostic data transmitted from the multifunction device is aggregated in the Remote Diagnostics System along with the operational log data. The system analyzes this data to identify potentially noise-causing components.
CE technicians can identify suspected parts suggested by the system before inspecting the actual device, allowing them to prepare the appropriate replacement parts in advance before visiting the customer’s site. This enables rapid resolution of abnormal noise issues during the first visit and helps reduce environmental impact by minimizing the number of service visits.
The remote abnormal noise diagnosis technology leverages the Orchestration agent, which is one of the core AI technologies within FUJIFILM Business Innovation's AI technology brand, REiLI Business.
REiLI Business represents FUJIFILM Business Innovation’s approach to accelerating value creation and innovation through AI-enabled products and solutions.
It brings together AI capabilities with the knowledge, expertise and know-how accumulated within organizations to help customers address real-world business challenges, enhance the way they work and create new value.
Built on the Fujifilm Group’s long-standing core technologies, including image processing and natural language processing, REiLI Business offers a suite of AI agents. These AI agents structure and leverage information, knowledge, and know-how within organizations to support business growth.