Adaptive Detection Probability for mmWave 5G SLAM

Abstract

In 5G simultaneous localization and mapping (SLAM), estimates of angles and delays of mm Wave channels are used to localize the user equipment and map the environment. The interface from the channel estimator to the SLAM method, which was previously limited to the channel parameters estimates and their uncertainties, is here augmented to include the detection probabilities of hypothesized landmarks, given certain a user location. These detection probabilities are used during data association and measurement update, which are important steps in any SLAM method. Due to the nature of mm Wave communication, these detection probabilities depend on the physical layer signal parameters, including beamforming, precoding, bandwidth, observation time, etc. In this paper, we derive these detection probabilities for different deterministic and stochastic channel models and highlight the importance of beamforming.