Three-Dimensional Angle of Arrival Estimation in Dynamic Indoor Terahertz Channels Using a Forward-Backward Algorithm

A novel angle-of-arrival (AoA) estimation method for both azimuth and elevation based on Bayesian inference and statistical state transition probabilities is presented for the dynamic indoor terahertz (THz) channel. A precise AoA estimation is crucial for the deployment of a directive antenna, which can compensate for the high path loss and reduce the intersymbol interference. In many application scenarios, the user equipment is moved by the user during the data transmission, and the AoA is not constant. The novel algorithm exploits the fact that the AoA movement can be represented as a Markov process and that the Bayesian inference can be used to combine the likelihood and a priori information to provide a more precise estimate than using the likelihood alone. An indoor human movement model is developed to generate the realistic application scenario and obtain the statistical transition probabilities. The forward-backward algorithm is implemented to carry out the Bayesian inference. The algorithm performance is illustrated using the channel models generated by a ray launching simulator. The background log-likelihood is suggested to adapt the algorithm to the instant channel state change in a multipath environment.