Road Obstacle Detection via Wheel Speed Signal Segmentation

The patent segments the wheel speed signal into distinct phases (approach, impact, rebound) to extract specific characteristics. By dividing the continuous signal into meaningful segments corresponding to different obstacle interaction stages, the system can identify both the presence and location of obstacles while determining their severity levels through phase-specific parameter analysis.

The patent transforms the one-dimensional wheel speed signal into multiple dimensional information by computing derived parameters such as acceleration, jerk, and frequency content. This dimensional expansion allows the system to extract location information from signal timing, severity from amplitude characteristics, and obstacle type from frequency patterns, thereby recovering the lost spatial and qualitative information.

The patent performs preliminary processing of wheel speed data by pre-computing acceleration and jerk parameters, and establishing reference profiles for different obstacle types. This preliminary action creates a structured data foundation that enables detailed obstacle characterization without requiring complex real-time analysis, thus maintaining system simplicity while providing comprehensive obstacle information for maintenance prioritization.

The patent replaces complex mechanical sensing systems with computational analysis of existing wheel speed data. By using signal processing algorithms to extract obstacle characteristics from routine wheel speed measurements, the system achieves detailed obstacle characterization without adding mechanical sensors or increasing physical system complexity.

The patent applies partial action by focusing computational resources only on detecting and analyzing wheel speed variations that exceed predefined thresholds indicative of obstacle impacts. Rather than processing all wheel speed data equally, the system selectively analyzes only the portions of the signal containing obstacle-related information, thereby maintaining detection completeness while significantly reducing processing time and computational burden.

The patent implements periodic action through continuous monitoring at fixed sampling intervals and periodic computation of derived parameters. This structured periodic processing approach ensures that obstacle information is captured systematically without requiring continuous intensive computation, optimizing the balance between detection completeness and processing efficiency.