Informacionnye Tehnologii

A specialized methodology was developed, computer modeling was conducted, and a comparative analysis was performed for autonomous control methods for small spacecraft, differing in how the nonlinearity of the initial stabilization problem is taken into account:

— control constructed using a system of linear differential equations (SLDE);

— control constructed using SLDE applied to a more general system of differential equations in which nonlinear terms are taken into account in trajectory calculations;

— control using the SDRE method for a nonlinear system of differential equations. To evaluate the effectiveness of the three control methods under consideration under identical initial conditions, stabilization problems using the quadratic quality criterion (LQR) were considered.

A computational experiment showed that, in most cases, the SDRE method yields the lowest quality functional of the three control methods considered. However, the final quality functional values for control constructed using SLDE with nonlinear terms remain consistently higher than those for SDRE. This is due to the fact that the nonlinear term significantly contributes to the increase in the quality functional in the initial equations. Unlike linear optimal control, the SDRE method used does not provide the necessary condition for a minimum performance functional, but it does not require the linearity of the initial differential equation system, which is a significant advantage. Despite the lack of a mathematical proof of the minimal performance functional for all cases, numerical experiments nevertheless demonstrate the advantages of the SDRE method for calculating control. However, the lack of a proven minimum performance functional allows this method to be considered a suboptimal control method. The described approach, using a mathematical model of suboptimal control for small spacecraft, allows it to be used to solve energy-intensive problems that largely determine the operational life of such spacecraft

The article presents an innovative method for detecting and classifying multivector attacks in decentralized Internet of Things (IoT) networks, significantly enhancing information security. The proposed approach combines Generative Adversarial Networks (GAN) to create realistic synthetic anomalous data and Reinforcement Learning (RL) algorithms for adaptive real-time updating of the attack detection model. This integration effectively addresses the shortage of high-quality, balanced data on rare and emerging attack types, as well as the dynamic nature of threats in IoT environments, which traditional methods struggle to handle adequately.

The method comprises two key interconnected components: a synthetic anomaly generator that expands the training dataset, and an RL classifier capable of continuous learning on hybrid data that includes both real and generated events. To improve decision-making reliability, a weighted voting mechanism is employed among agents, taking into account the trust levels of each network node. The article provides a detailed description of the RL agent’s neural network architecture, featuring three hidden layers with Batch Normalization and Dropout, along with its training algorithm using Deep Q-Network (DQN) and double Q-learning.

Experimental evaluation was conducted on a synthetic dataset simulating real IoT scenarios with various multivector attacks such as trust undermining, behavior manipulation, data leakage, and cooperation disruption. The results demonstrate significant improvements in recall and F1-score metrics compared to classical machine learning methods, including Random Forest and Support Vector Machine, especially when trained on the extended dataset with GAN-generated synthetic data. The method offers high attack detection accuracy while reducing false positives and maintains adaptability to novel, previously unknown threats. This makes the proposed approach a promising solution for next-generation IoT security systems, capable of effectively operating under resource-constrained devices and dynamically changing environments.