When a classical legacy system such as a manufacturing production line, gets upgraded with a deterministic module replaced by an AI module, what are the criteria to consider, and how to adapt methodology from existing standards for software engineering (such as ISO-IEC-JTC1-SC6 12206)?
Without loss generality, we can consider quantity, quality and cost of products, and cost of upgrade. Our destination is raise profit via upgrade retrofit, and reduce cost of upgrade retrofit, therefor that renewed system has flexible adaptation for similar products as many as possible is necessary.
Thanks for your input. The question is about AI specific aspects.
Dear Renaud Di Francesco ,
I would use the scenario from Automotive Industry,
Many organizations are still continuing with their legacy codes and vehicle infrastructure. In parallel, adapting the new platforms module-wise which are required to manufacture next generation vehicles. In this case, they are adapting new platforms module-wise to address the economic and technical feasibility. E.g. Integration of AUTOSAR Classic and Adaptive platforms, CoherenSE and EUCD platforms.
Specific to your question on AI, Since many AI features for infotainment and ECU functionalities are now enabled through ADAS applications on legacy platform which is now converted to new hybrid platform which can be interfaced through wrappers.
In many scenarios, Legacy codes and platform has sufficient potential to go with L3 or L4 AI autonomy but for sustainability, it has to adopt the new powertrain and respective platforms/framework to address the challenges countered in period of time.
In short, Some of primary criteria such as hardware and software platform's technical feasibility and legacy design constraints need to be considered for sustainability. And to adopt methodology from existing standard, you can refer documents of AUTOSAR Adaptive platform( https://www.autosar.org/standards/adaptive-platform/) which proposes methodology to interface with Classic platform as well as non-standardized platforms. This is just a reference, hope it helps.
Regards,
Jai Deo Tiwari
Dear Renaud Di Francesco ,
Look the link, maybe helpful.
https://diginomica.com/the-retrofitting-outlook-for-ai-in-the-enterprise-2019-2020/
https://appdevelopermagazine.com/implementing-ai-technology-while-overcoming-legacy-systems/
Regards,
Shafagat
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence