Roadway Bump Detection Using Sensor Change Probability Maps

The patent divides the road surface into discrete locations and processes sensor data in subsets corresponding to different road segments. By segmenting the detection task into location-specific analyses, the system achieves comprehensive coverage while managing computational load through distributed processing of smaller data subsets rather than analyzing all sensor data simultaneously.

The system performs preliminary sensor data collection during vehicle traversal before conducting detailed bump and depression analysis. Sensor data is gathered and pre-processed during normal vehicle operation, allowing the computationally intensive detection algorithms to work with preorganized data subsets, thereby reducing real-time processing requirements while maintaining detection accuracy.

The patent combines sensor data from multiple vehicle traversals of the same roadway into unified data subsets for each location. By merging data across multiple passes, the system accumulates sufficient evidence to reliably identify bumps and depressions while processing aggregated results in batch operations rather than analyzing each traversal separately, thus improving reliability without proportionally increasing total processing time.

The system creates multiple data subsets representing the same physical locations from different traversals. These copied data representations allow parallel processing and comparison across traversals, enabling reliable identification of consistent road features while distributing computational work across multiple processed copies rather than requiring sequential analysis of raw traversal data.