A Scalable Web Platform for AI-AugmentedSoftware Deployment in Automotive Edge Devices via Cloud Services

Abstract

Integrating artificial intelligence into automotive
applications demands a scalable web platform for AI-augmented
software deployment in automotive edge devices. Such a platform
not only facilitates a range of AI applications but also leverages
the computational and storage capacities of cloud services to
enhance platform scalability. Ensuring seamless operation during
driving scenarios requires careful consideration of the diverse
features of automotive edge devices, particularly their often
infrequent and time-limited internet connection. AI-augmented
algorithms deployed in the vehicle must consequently be cognizant of both the capabilities and connectivity constraints of the
target device. The implementation of the architecture is based
on a modular web platform concept comprising eight modules.
These modules define interfaces and facilitate their interconnection, either within the platform or with corresponding external
components. Applications can access the platform and utilize
its capabilities through publicly exposed open API endpoints.
Instances of the proposed architecture currently support AIaugmented software deployment in real automotive edge devices
and standard cloud services.