Indoor Acoustic Localization with IMU for UGV

Indoor acoustic localization systems have been widely concerned for their high positioning accuracy. However, the current systems cannot perform satisfactorily in noisy and dynamic environments. To address this problem, a localization system fused with acoustic and Inertial Measurement Unit (IMU) to avoid the multi-path effect and dynamic error is proposed in this paper. Firstly, an acoustic ranging constraint is built by a novel peak search algorithm to lessen the multipath effect. Secondly, a key frame association constraint which utilizes IMU pre-integration is introduced to achieve a tightly coupled localization. Finally, based on the above-mentioned constraints, the complex indoor Unmanned Ground Vehicle (UGV) positioning is transformed into an optimization problem which can be solved easily. Experimental results demonstrate that our designed system is accurate with a dynamic positioning error of 0.51 m at 95% with a five-anchors system.