Indoor Radio Map Construction Based on Position Adjustment and Equipment Calibration

Abstract

The crowdsourcing-based wireless local area network (WLAN) indoor localization system has been widely promoted for the effective reduction of the workload from the offline phase data collection while constructing radio maps. Aiming at the problem of the diverse terminal devices and the inaccurate location annotation of the crowdsourced samples, which will result in the construction of the wrong radio map, an effective indoor radio map construction scheme (RMPAEC) is proposed based on position adjustment and equipment calibration. The RMPAEC consists of three main modules: terminal equipment calibration, pedestrian dead reckoning (PDR) estimated position adjustment, and fingerprint amendment. A position adjustment algorithm based on selective particle filtering is used by RMPAEC to reduce the cumulative error in PDR tracking. Moreover, an inter-device calibration algorithm is put forward based on receiver pattern analysis to obtain a device-independent grid fingerprint. The experimental results demonstrate that the proposed solution achieves higher localization accuracy than the peer schemes, and it possesses good effectiveness at the same time.

Localization with Transfer Learning Based on Fine-Grained Subcarrier Information for Dynamic Indoor Environments

“…[ 5 ]. Due to the ubiquitous Wi-Fi devices in office and home environments, Wi-Fi-based device-free localization has attracted a lot of research interest [ 6 , 7 ]. Manikanta Kotaru et al [ 8 ] proved that Wi-Fi localization can achieve decimeter-level positioning accuracy.…”