Adding link quantity information to redundant RF signal strength estimates for improved indoor positioning

Abstract

A common approach of range-based indoor posi tioning methods use the received signal strength (RSS) of RF packets from several anchor nodes to estimate their distances to a mobile emitter with unknown position. The range-based weighted centroid localization (WCL) approach with link quality information (LQI) is known to be more accurate than range free centroid localization (CL) methods using only a binary radio link information according to the cell of origin (COO) principle. With the combining of redundant RF channels using spatial and frequency diversity the WCL position estimation is proved to be even more reliable, although the RSS-based distance estimations of a single RF channel are known to be error-prone in multi path indoor environments. A novel range free approach using the exact number of available diversity channels -the link quantity information (LQnI) -is proposed. It needs no more infrastructural effort or processing power and can easily be applied to the range-based WCL estimation technique with redundant sensor information. Especially the combining of the redundant RSS-based distance estimations together with the LQnI approach leads to a more accurate position estimation. Experimental results in an office building and in a real-life tracking application for maintenance staff in the underground coal mining show the improvements of the additional range-free approach.

MineLoc — Personnel Tracking System for Longwall Coal Mining Sites

“…The range-free centroid approach (rCL), proposed in Fink and Beikirch (2012a) uses the exact number of available diversity channels, the link quantity information (LQnI), for the weighting according to…”

“…In Fink and Beikirch (2012a), we propose a combined weighting using LQI and LQnI information according to…”

“…10, the position estimation error can be reduced by more than 50 % from e x = 0.79 m to e x = 0.38 m. A further analysis of the accuracy improvement of the enhanced WCL algorithm is given in Fink and Beikirch (2012a), using experiment results on a motion test track. The regarding cumulative error distribution functions are shown in Fig.…”

Radio-based human tracking for large indoor environments using distributed centroid location estimation

“…using IMU (inertial measurement unit)-based platforms, even can be run without any reference infrastructur and thus, the coverage is theoretically unlimited. A robust onedimensional indoor local positioning system (ILPS) for the localization and tracking of maintenance staff in the underground longwall coal mining is proposed in [6]. The underground mining ILPS uses a set of reference nodes which are connected to a data concentrator PC using a wired bus system.…”

“…The maximum bus load (cable length, data rate and connected nodes) is given by the network specifications like cable attenuation and the electrical load of a single CAN transceiver. For the localization application in [6], a maximum number of 20 blind nodes (BNs) can be localized simultaneously along a line of up to 32 reference nodes (RNs). To break up these limitations, we propose distributed location estimation and data compression strategies.…”