Artificial Intelligence/ Machine Learning in IoT for Authentication and Authorization of Edge Devices

Internet of Things (IoT) is progressing at a fast pace. Issues of security and privacy, emerged with introduction of IoT in late nineties, are still amongst the main challenges. In security issues, authentication and authorization of edge devices are main concerns due to resource constrained nature of edge devices. Various solutions have been proposed in the past to address said concerns but most of the solutions are based on increasing the computational capacity, storage and power in edge devices. However, said solutions are not practical since these solutions are either not possible due to small size of edge devices of IoT or not economical for their wide spread adoption. Some of the solutions also suggest the use of light weight cryptographic primitives. However, same are also not practical since all edge devices do not have requisite resources to implement these solutions. This paper proposes use of Artificial Intelligence (AI)/ machine learning in addressing the issues of authentication and authorization in edge devices. Proposed solution is based on fog computing model within a framework of a smart house but without reliance on computational capacity, storage or power of edge devices.