A Fast Database Correlation Algorithm for Localization of Wireless Network Mobile Nodes using Coverage Prediction and Round Trip Delay

A priori selection of high accuracy mobile station position estimates

“…Specificity k = TN k TN k + FN k (4) Table III shows that the proposed ANN classifier achieves high accuracy, precision and specificity for the three classes. However, in this work, whose objective is to identify and select high accuracy MS position estimates that shall be forwarded to critical LBS applications, the most relevant metric is the precision of the high accuracy class.…”

“…Therefore, RF-FING+RTD-PRED searches for matches only in the correlation space. The correlation space is the subset of all Rfings in the CDB that are compared to the Tfing, and it is obtained by a process comprising four successive filtering steps, called deterministic filtering [4]. After defining the correlation space, the Euclidean distance between the Tfing and each Rfing in the N dimensional RSS space is calculated.…”

“…The best server areas of each cell -which are used in the first filtering step to reduce the correlation space [4] and the CDBs used by RF-FING+TA-PRED were built using propagation modeling. To do so, the Okumura-Hata propagation model [13] was applied upon DEMs (Digital Elevation Models) of the test regions terrain, using the RF parameters of the RAN (antennas coordinates and azimuths, antennas radiation patterns, output power, transmission frequency, etc.).…”

“…Using the information received from the RAN (Radio Access Network) and queried at the BSA, the SMLC mounts both the RF fingerprint (F j ) -with the parameters listed in the NMRs (Network Measurement Reports) sent by the target MS, plus the RTD (Round-Trip Delay) measured by the serving cell -and the classifier input vector (V j ). The RF fingerprint F j is used by RF-FING+RTD-PRED, which is a network-based positioning method originally proposed by the authors in [4]. After carrying out a principal components analysis (PCA) to reduce the input vector V j dimension, it is presented to the backpropagation ANN (Artificial Neural Network) classifier, which works in parallel with the network-based positioning method, as shown in Figure 1.…”

Positioning in Bluetooth and Uwb Networks

“…It is crucial when the radio map is too large -which is the case for location systems in metropolitan areas (Campos and Lovisolo, 2009) or even in large multi-floor buildings (Campos et al, 2014). In MS-based Bluetooth indoor fingerprinting, search space reduction might also help reduce energy consumption -both of the MSs and, most importantly, of the fixed beacons -which is in keeping with one of the key features of BLE -i.e., enhancing battery lifetime.…”