Robust Sensor Registration Based on Bounded Variables Least Squares

The ill conditioning problem of sensor registration is considered. We analyze the ill conditioning in the dense-target scenario and the dense-sensor scenario, respectively, and present a robust registration method based on the bounded variables least squares (BVLS). The proposed approach can reduce the influence of ill conditioning by means of inserting prior constraints on the desired solution. Compared with the traditional least squares (LS) method and the recently proposed azimuth-LS method, the BVLS approach can accommodate the ill conditioning in densetarget and (or) dense-sensor scenarios. In the meantime, it is consistent with the LS estimator when the registration problem is good-conditioned. Simulation results demonstrate the superior performance of the proposed method. Keywords—Sensor registration, bounded variables least squares (BVLS), ill conditioning, dense-target scenario, dense-sensor scenario.