Ground Target Tracking Using a Monocular Camera and IMU in a Nonlinear Observer SLAM Framework

This paper discusses moving ground target estimation using a recently proposed, uniformly globally asymptotically stable observer for simultaneous localization and mapping (SLAM), which only requires a monocular camera and an inertial measurement unit (IMU). The observer operates under a persistent excitation constraint and requires the inertial origin to remain in view throughout its operation. The contributions of this paper are an iterative procedure to increase convergence speed of nonlinear observers and a process for keeping the inertial origin within view. Simulation results demonstrate fast observer convergence and tracking of a maneuvering ground target relative to the observer's landmark estimates.