Tracking with biased measurements of signal strength sensors

Sensors that measure received signal strength from moving targets may have bias that need to be accounted for if accurate tracking of targets in time is needed. When the bias is unknown, it has to be estimated together with the other unknowns of the system model. If the applied methodology for tracking is particle filtering and if the number of sensors is large, the performance of the used particle filtering algorithm may degrade considerably, In the paper we show how the tracking can be performed by marginalizing the biases through the use of Rao-Blackwellization and how the number of used Kalman filters for marginalization can be reduced to only one. We demonstrate the performance of the proposed algorithm with computer simulations.