Design security architecture for unmanned aerial vehicles by 5G cloud network based implementation of SDN with NFV and AI

Abstract

A recent progression of unmanned aerial vehicles (UAV) augmentation its employments for different applications. It’s also vulnerable to being, stolen, lost, stray, or destroyed at status of a security infringements for the UAV network. The proposed strategy is defending against of different attacks through using artificial intelligence by implements five steps: RGSK, GCSCS, SEDC, HSSC, and FVNF. UAV authentication is happened in the first step through the Curve448. We performance deep reinforcement learning to run with GCS for packet assignment as it implemented for switch current state identification before updating. In our work we ability to alleviate for attack of flow table overloading by assigned of packets as an under loaded or idle switches. Then, selected the least loaded switch by applied 5 tuples. Hence, we divided SDN to SEDCs and HSSC forms. First in the SEDC we using Shannon entropy to achieve classified of input packet in to regular and suspicious packets. Last will forwarded regular packets to cloud layer. By growing multiple self-organizing maps for maintained in NFV that used to classify suspicious packets as classes normal or malicious packet. The proposed performance work evaluates using NS3.26 show up the better strategy to secure UAV for different attacks.