Smart Inspection
An energy company, due to the complex maintenance and operation of many critical devices, requires operators to use mixed-reality(MR) headset with AI assistance to ensure that each operation step is compliant and reliable.
Edgenesis build an IoT platform and helped them:
- Reduced API development effort for MR headset integration by 80%.
- Reduce Edge AI platform integration by 50%.
- Provided vendor-agnostic support for MR headset adoption.
- Empowered the creation of a scalable Edge AI model to enhance the functionality of MR headsets.
Challenges
MR devices and AI models are coupled with App. App development does not scale, because any changes are deeply coupled with device and AI models.
- Slow integration of MR headsets: It takes months for engineers to familiarize themselves with a single MR headset. Different device vendors make this process even slower, for example, Microsoft's HoloLens, Apple's Vision Pro, and Meta's Quest Pro.
- Coupled AI model: The MR AI powered App is deeply coupled with AI models. Any changes from AI model require changes to the App.
- Not reusable: The MR headset integration for one app is not reusable for other apps since business logic is constantly changing.
- Not scalable: The trained AI models are combined with the app itself. It is not possible to scale them for usage in other instances.
Edgenesis Solution
Simple MR Headset Integration
Integration with any MR headsets like Hololens2 and Vision Pro into one IoT Platform, any MR AI powered App development are based on standard IoT platform headset API. App developers no longer need to get into specific MR headset integration. They can focus on business development when MR headset operation becomes some single web API calls.
Flexible AI Integration
The MR event related AI inferences can hide into the IoT platform which linked to IoT device telemetry service. The App can directly access the inference result via IoT platform's Web API. This makes AI integration as flexible as a Web API call.
Reusable
The MR headsets function as an API service, always available for app development. Developers can integrate them without having to build them into the app itself. So do AI integration.
Scalable
AI models are provided as an API service, capable of scaling to multiple instances as traffic increases.
Result
With the support of Edgenesis, they can seamlessly connect with flexible MR devices and rapidly evolve AI models, achieving digital and intelligent safeguards in power station operations to ensure 100% safe production across major power stations.
