The development of intelligent systems for assessing the functional state of vehicles is an urgent task, as it enables wear detection, environmental monitoring, and operational optimisation. Acoustic noise and vibration signals generated by mechanical components contain valuable information about their technical condition. However, their diversity, nonlinearity, and multidimensionality limit the effectiveness of traditional analysis methods. This study presents the results of experimental research and computational modelling of acoustic signals from gasoline engines under normal operating conditions and in the presence of defects. A parametric and 3D topological visualisation method of signals in the coordinate space (x, dx/dt, d2x/dt2) is proposed, allowing the identification of characteristic patterns corresponding to different engine states. It is demonstrated that the compaction or expansion of phase trajectories indicates defects and increased wear. The proposed approach enables the formation of an integral wear index based on the area of phase trajectories, the energy of the second derivative, and the fraction of high-frequency components.