BPNN Localization Method for A ISAC System under LOS/NLOS Scenario

Aiming at the problem of high-precision target localization by the system of perception-assisted communication, a joint target localization scenario of Unmanned Aerial Vehicle (UAV) sensing assisting Base Station (BS) communication is considered. The UAV adopts the Frequency Modulated Continuous Wave (FMCW) sensing signal, and the BS adopts the Binary Phase Shift Keying (BPSK) communication signal. And the target location is achieved through the correlation operation. To solve the problem of the low accuracy of traditional positioning algorithms, two target positioning methods combined with Back-Propagation Neural Network (BPNN) algorithm are proposed in this paper. One is to use the estimated distance value to enter the BPNN for training and positioning calculation, and the other is to use the correlation curve value to enter the BPNN for training and positioning calculation. Simulation results show that these two methods have good robustness in Line-of-Sight/Non-Line-of-Sight (LOS/NLOS) scenarios, which significantly improves the performance of the classical Least Square (LS) and Two-Step Weighted Least Square (TSWLS) localization algorithms.