Daulet Akhmedov

The structural and positional analysis of planar linkage mechanisms containing higher-class Assur groups requires the generation and solution of systems of algebraic kinematic equations. This paper presents algorithms for the automated analysis of such mechanisms based on an information matrix model. The proposed representation describes a mechanism as a system of material points connected by geometric constraints and provides a unified framework for algorithmic processing of mechanism structure and parameters. A square-matrix information model is developed to store structural and geometric information of a mechanism. Based on this model, three algorithms are proposed: (i) an algorithm for structural analysis and determination of mechanism mobility, (ii) an algorithm for automatic generation of kinematic equations, and (iii) an algorithm for automatic construction of a positional analysis procedure for solving the generated system of algebraic equations. For mechanisms containing higher-class Assur groups, the developed approach automatically generates an ordered sequence of two-link kinematic chains. When direct decomposition is not possible, conditional driving links are introduced by selecting conditionally disconnected links and defining corresponding driving coordinates. The proposed algorithms have been implemented in the SAPRM system. The applicability of the approach is demonstrated through positional analysis of three planar linkage mechanisms containing higher-class Assur groups.

The increasing need for accurate prediction of geochemical anomalies requires methods capable of capturing complex spatial patterns that traditional approaches often fail to represent adequately. For N datasets of the form (X_i,Y_i) representing the geographic coordinates of sampling points and C_i denoting the geochemical measurement, training multilayer perceptrons (MLPs) presents a challenge. The low informativeness of the input features and their weak correlation with the target variable result in excessively simplified predictions. Analysis of a baseline model trained only on geographic coordinates showed that, while the loss function converges rapidly, the resulting values become overly “compressed” and fail to reflect the actual concentration range. To address this, a preprocessing method based on anisotropy was developed to enhance the correlation between input and output variables. This approach constructs, for each prediction point, a structured informational model that incorporates the direction and magnitude of spatial variability through sectoral and radial partitioning of the nearest sampling data. The transformed features are then used in a dual-MLP architecture, where the first network produces sectoral estimates, and the second aggregates them into the final prediction. The results show that anisotropic feature transformation significantly improves neural network prediction capabilities in geochemical analysis.