Eco-Friendly Low Resource Security Surveillance Framework Toward Green AI Digital Twin
Hyunbum Kim, Jalel Ben‐Othman
Abstract
Most intelligent systems focused on how to improve performance including accuracy, processing speed with a massive number of data sets and those performance-biased intelligent systems, Red AI systems, have been applied to digital twin in smart cities. On the other hand, it is highly reasonable to consider Green AI features covering environmental, economic, social costs for advanced digital twin services. In this letter, we propose eco-friendly low resource security surveillance toward Green AI-enabled digital twin service, which provides eco-friendly security by the active participation of low resource devices. And, we formally define a problem whose objective is to maximize the participation of low source or reusable devices such that reusable surveillance borders are created within security district. Also, a dense sub-district with low resource devices priority completion scheme is proposed to resolve the problem. Then, the devised method is performed by expanded simulations and the achieved result is evaluated with demonstrated discussions.