Object tracking in mmWave radar networks

Abstract

Location-aware devices enable new services such as localization and tracking of objects within existing wireless communication networks like cellular mobile, Wi-Fi, and radio. To ensure these services are also available in the evolving millimeter wave (mmWave) communication infrastructure, it is important to develop algorithms that enable mmWave devices, like radars and 5G nodes, to localize and track objects. The main challenges that these algorithms must address is localizing objects that are not carrying sensing equipment, synchronizing devices exclusively via the mmWave band, and solving a data association uncertainty problem to reliably track objects of interest. Our development of the Multistatic Networking with mmWave Radar Arrays for Positioning (MiNiMAP) system solved these challenges by implementing mmWave processing in a multistatic network, scheduling, and radar synchronization algorithms. Through the use of these three algorithms in addition to Bayesian filtering, MiNiMAP is capable of tracking a single object with a network of mmWave radars. Indoor localization experiments validate MiNiMAP's overall system performance and the impact of each algorithm.