Neural network system for detecting and neutralizing remote unauthorized interference into the components of the Internet of Things
In this work the authors have analyzed the neural network system for detecting and neutralizing remote and unauthorized interference with components of the Internet of Things. The main focus is on considering the neural network approach to detecting intrusions into the Internet of Things network, its monitoring and countering suspicious activity on the host. Features of development of model of artificial neural networks for application of apparatus of neural network in this direction have been considered. This allows you to reflect the successful identification of various types of attacks in terms of true and false positive results. However, the problems of obtaining data on overload and critical modes of the system remain unresolved. The use of a neural network system for detecting and neutralizing remote and unauthorized interference with components of the Internet of Things allows you to implement a module for detecting anomalies in the network, based on the Voltaire series, which considers the theoretical prerequisites of the method of dynamically building an artificial neural network. The main types of attacks, types of intrusion detection systems, interpretations of the obtained data, a brief study of works in the field of neural network solutions have been analyzed. An effective solution has been offered to protect workstations in the Internet of Things network from unauthorized access, and to configure security for all component modules. In conclusion, recommendations have been given for implementing the construction of a neural network module that detects deviations in the operation of the Internet of Things from normal modes.