Abstract
Modern special-purpose ground robotic complexes (GRCs) operate under conditions of intense interference from heterogeneous sources, including electromagnetic, optical, and acoustic. The combined effect of these interferences results in a significant deterioration of the information quality from sensor systems, which threatens the loss of controllability, disorientation, and a decrease in task performance efficiency. Existing approaches that consider the influence of interferences in isolation do not accurately reflect real operating conditions, necessitating the development of new, comprehensive solutions. The study proposes a mathematical model for assessing the impact of combined interference on multisensor NRC systems and develops an effective signal processing method based on it to enhance the stability and reliability of the control system. The mathematical model accounts for the simultaneous impact of various types of additive and multiplicative interference on sensor signals. To compensate for their impact, a method has been developed that combines: assessment of the signal quality of each sensor by calculating the signal-to-noise ratio (SNR) and the probability of correct detection; synthesis of the informative feature space using the principal component analysis (PCA) method to reduce the dimensionality of the data and level the impact of interference; classification of NRC states using machine learning algorithms in the synthesized feature space. The effectiveness of the proposed approach has been experimentally demonstrated using a multisensor system with 60 sensors. It is demonstrated that utilizing the synthesized feature space enables achieving high classification accuracy of NRC states (up to 99%) even under conditions of intense combined interference. It was found that even when using 40-50% of the most informative features, sufficient accuracy (70-95%) is ensured, indicating the stability of the method and allowing for a reduction in computational load. The scientific novelty lies in the complex model of the influence of combined interference, which takes into account the correlations between different types of interference, and in the method of interference compensation based on the synthesis of the feature space, invariant to changes in the nature of the interference influence. The implementation of the proposed approach enables an increase in the autonomy and reliability of the NRC's functioning in conditions of electronic warfare and other complex environments. The results can be used in the design of new generations of robotic complexes, the modernization of existing systems, and the training of operators.