Multi-Slope Path Loss and Position Estimation With Grid Search and Experimental Results

A transmitting device can be localized based only on the received signal strength (RSS) measured by receivers. Typical RSS-based algorithms assume the a priori knowledge of the path loss (PL) the signal undergoes in the search area. Recently, algorithms removing this assumption have been introduced based on statistical and nonlinear methods to solve the optimization problem required to estimate the transmitter’s position. However, such nonlinear methods could not converge to the optimal solution, especially in scenarios characterized by multi-slope or angular dependent PL. This paper considers an exhaustive search algorithm (ESA) for network localization based on the weighted least square (WLS) minimization. The algorithm’s effectiveness is assessed by simulation and compared with the derived Cramér-Rao lower bound (CRLB). From experiments with very noisy measurements in indoor and outdoor, the localization error with 200 LoRa PL measurements is 5 m, whereas algorithms assuming a single slope incur errors between 12 m and 47 m.