AI Edge Processing - A Review of Distributed Embedded Systems

The age of cybernetics is underway, with a lot of systems being computerized leading to the generation of a lot of data. Sending this data to remote systems and the cloud for analysis and/or processing introduces latency, reliability, and privacy issues. These issues are generally not acceptable to systems that need to respond in real-time or near real-time. Local processing of data is an alternative. Hence there is a need for high performance embedded computing systems to be employed at the edge. This paper proposes and reviews distributed embedded computing systems as a possible architecture for edge AI processing.