Fingerprinting

A positioning technique creating a database mapping signal characteristics to specific locations, then matching observed signals against this database. Involves offline training phase (surveying and building database) and online positioning phase (comparing real-time signals). Can compensate for complex RF environments where simple models don't work. Labor-intensive but works in challenging environments using lower-cost hardware.

Positioning technique creating a database of location-specific signal characteristics (fingerprints) during training phase, then matching real-time measurements against database to determine position. Wi-Fi and BLE fingerprinting process: (1) Site survey - systematically measuring signal strengths from multiple access points at known reference positions throughout facility (typically 2-5 meter grid spacing), recording RSSI values and creating fingerprint database linking signal patterns to positions. (2) Positioning - mobile device measures current RSSI from visible access points, compares against fingerprint database using pattern matching algorithms (k-nearest neighbors, neural networks, probabilistic methods), and returns position of best-matching fingerprint. Fingerprinting advantages: achieves 3-8 meter accuracy using standard Wi-Fi/BLE infrastructure without precise infrastructure positioning or time synchronization, and adapts to site-specific RF environment (naturally accounting for obstacles and multipath). Limitations: requires labor-intensive site survey (typically 2-8 hours per 1000 m²), fingerprint database becomes outdated if environment changes (new equipment, furniture rearrangement), positioning accuracy lower than ranging-based methods, and database size grows with facility size (limiting scalability). Wi-Fi fingerprinting accuracy degrades 20-40% if survey is 6+ months old in dynamic industrial environments. Modern systems use automatic fingerprint updates during operation (crowdsourced from device measurements with known positions) to maintain database currency.