Kalman Filter based NLOS Identification and Mitigation for M2M Communications over Cellular Networks

Machine-to-machine (M2M) communication is a crucial technique to improve the spectrum efficiency (SE) and energy efficiency (EE) by identifying the location of the unknown machine (UM) in fifth-generation (5G) cellular networks. This paper proposes a hybrid time difference of arrival (TDoA)/angle of arrival (AoA) based localization architecture using the Kalman filter (KF) algorithm for M2M communications over cellular networks. The proposed architecture incorporates the KF for identifying the non-line of sight (NLOS) propagation error which utilizes the time of arrival (ToA) measurement model. Furthermore, KF provides a promising solution to minimize the NLOS propagation error. We also adopt an extended Kalman filter (EKF) algorithm for estimating the location of UM by processing the formulated TDoA as well as integrating the AoA measurement model for mitigating the inaccurate AoA information. The proposed architecture is evaluated in terms of root mean square error (RMSE). Simulation results show that the proposed hybrid scheme provides better localization accuracy in the proposed architecture.