At TheITService, one of my research focuses is on artificial intelligence. Recently, there’s been a lot of buzz about AI potentially taking over the world, much like in science fiction movies. As someone educated, trained, and experienced as an engineer and IT scientist, I’m somewhat skeptical of this idea. Therefore, I embarked on a project to determine if I could engineer the requirements (user inputs) and specifications (standard inputs) to train AI systems like Gemini and Bing. My goal was to develop a solution that could operate, document, and code for IoT and mobile applications using Leshan CoAP/LwM2M/OMA & IPSO. Following me at Leshan by training AI Gemini and Bing engine for details.
Hypothesis
- What solution do we get from the AI Gemini and Bing systems without training and with training the Requirements from the users and Specifications from Standards?
- What solution do we get from the AI Gemini and Bing systems with engineering the Requirements Specification to train the AI Gemini and Bing systems on following aspects:
- Operation including configuration to build and launch Bootstrap, Device Management and IoT Client)
- Documentation around the coding aspect of
- Coding
- The Specification comes from the literature researching into the OMA v1.0, OMA v1.1 and OMA v1.2 for CoAP, LwM2M and IPSO specifications to be used for the training of AI Gemini and Bing engine
- Acknowledgement to the generic user requirements implemented to the Leshan User Interface.
Why do we use CoAP/LwM2M/IPSO & OMA for the development
This is because the methodology provides
Architecture:
-
well-known architecture where many clients report to the Bootstrap and Device Management Systems. The Bootstrap System authenticates which device to be managed and the Device Management System can manage, operate, administrate and maintain many devices using the Light Weight Machine to Machine (LwM2M) protocol.
-
Two possible architectural Use Cases
-
IoT with South East Water Digital Meters: IoT Use Case
-
Mobile application with Mobile Phones: Mobile Use Csse
- Protocols and Security used in the development.
- CoAP/CoAPs: Protocol that operate in a constraint environment, i.e device can go sleep and wake up to save battery or device can operate in the unreliable and intermittent transmission media
- LwM2M: Knowledge and Wisdom accumulate over many years of how the Protocol between Machine to Machine works
- Register after the bootstrap operation
- Register Update
- Read/Write/Execute/Observable &Notification
- Firmware/Software download to the device
- Security with PSK, RSK, X500 Certificate between IoT device and Bootstrap/Device Management
Data Modelling
A consistent approach to model the data and communication using the LwM2M, thus resulting a consistent and concise APIs used in the development and maintenance of the application software.
- LwM2M IPSO (Server, Device, Firmware Upgrade, Location, Temperature, Test-Object)
- South East Water IPSO (App Data Container, Interval Data Delivery, Event Data Delivery, Delivery Schedule, Water Flow Readings, Daily Maximum Flow Rate Readings, Temperature Readings, Pressure Readings, Battery Level Readings, Communications Activity Time Readings, Low Battery Alarm, Daughter Board Failure Alarm
Methodology:
-
I develop the Digital Meter solution using Leshan CoAP/IPSO/OMA standards and technology with the followings:
- Leshan-1.5.0 support OMA v1.0
- Leshan-2.0.0-M16 support OMA v1.1 & OMA v1.2
-
Bootstrap sequence illustration
-
Device Management Default Data Model illustration
-
Device Management South East Water Model illustration
-
-
Reversed Engineering the PoC on the Digital Meter solution to engineer the specification to train the AI Gemini and Bing system to provide the solution
- Data Modelling (not in scope)
- Operation Leshan Operation training Build and Launch configuration for Bootstrap, Device Management and Client
- Documentation Leshan Documentation training for Bootstrap, Device Management and Client:
- Coding Leshan Coding training for Client
-
Validate AI Gemini and Bing results against development solution and the knowledge
Conclusion
- Can I engineer the specifications to train AI Gemini and Bing to be used for the development of CoAP/LwM2M/IPSO & OMA applications