Visual-Inertial Odometry With Online Calibration of Velocity-Control Based Kinematic Motion Models

Visual-inertial odometry (VIO) is an important technology for autonomous robots with power and payload constraints. In this paper, we propose a novel approach VIO stereo cameras which integrates calibrates the velocity-control based kinematic motion model of wheeled mobile online. Including such can help to improve accuracy VIO. Compared several previous approaches proposed integrate wheel odometer measurements purpose, our method does not require encoders be applied when robot modeled model. We use radial basis function (RBF) kernels compensate time delay deviations between control commands actual motion. The calibrated online by system used as forward planning. evaluate data obtained in variously sized indoor environments, demonstrate improvements over pure method, prediction