Joint estimation of emitter power and location in cognitive radio networks

We propose an iterative numerical solution to the primary emitter power and location estimation problem in cognitive radio networks. The proposed solution is based on generating samples from the joint probability distribution of location and power parameters using the Metropolis-Hastings (M-H) algorithm. Our proposed estimate is superior than the ML estimate based on the grid search technique especially when the sensor density is low and/or the shadowing spread is high. Additionally, we introduce the concept of the credible primary exclusive region (CPER), which is calculated using the credible samples generated by the M-H algorithm. Because emitter location and power estimates always contain some amount of uncertainty in practical settings, adoption of CPERs may provide a better suited tool for defining protected regions in cognitive radio networks.