2024/08/29 by Maximiliano Rivera Figueroa, Figueroa, Maximiliano Rivera, Pradyumna Kumar Bishoyi +3
Computer Science · Engineering · #FOS: Electrical engineering #Face and Expression Recognition #Face recognition and analysis #Indoor and Outdoor Localization Technologies #Signal Processing (eess.SP) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2408.16464
openalex publication_date 2024/08/29 · openalex created_date 2024/09/22 · openalex updated_date 2026/07/28
Cooperative multi-monostatic sensing enables accurate positioning of passive targets by combining the sensed environment of multiple base stations (BS). In this work, we propose a novel fusion algorithm that optimally finds the weight to combine the time-of-arrival (ToA) and angle-of-arrival (AoA) likelihood probability density function (PDF) of multiple BSs. In particular, we employ a log-linear pooling function that fuses all BSs' PDFs using a weighted geometric average. We formulated an optimization problem that minimizes the Reverse Kullback Leibler Divergence (RKLD) and proposed an iterative algorithm based on the Monte Carlo importance sampling (MCIS) approach to obtain the optimal fusion weights. Numerical results verify that our proposed fusion scheme with optimal weights outperforms the existing benchmark in terms of positioning accuracy in both unbiased (line-of-sight only) and biased (multipath-rich environment) scenarios.