Spatiotemporal Registration for Multi-sensor Fusion Systems

Abstract

A method of online sensor fusion with spatially and temporally misaligned sensors is considered in this paper. A spatiotemporal registration mode is developed for sensor alignment, and augment the target state vector with spatiotemporal bias. An unscented Kalman filter (UKF) is used to fuse and register these sensors, then the target state estimation as well as the spatial and temporal bias estimation can be obtained simultaneously. Simulations show that the proposed algorithm not only can align these sensors properly with both spatial and temporal bias, but can also obtain accurate fused tracks.