Tightly-coupled integration of acoustic signal and MEMS sensors on smartphones for indoor positioning

Based on the built-in sensors of a smartphone, a tightly-coupled integrated indoor positioning solution was developed based on the acoustic ranging signal and measurements from other built-in sensors. First, a two-step time of arrival (TOA) estimation method based on short-time Fourier transform and enhanced cross-correlation is performed to achieve an accurate TOA estimate. Having obtained the accurate TOA estimates, studies on a tightly-coupled integrated navigation algorithm based on TDOA and PDR is carried out. The algorithm takes the advantage of the complementarity of PDR and acoustic ranging observables to effectively improve positioning accuracy in a dynamic positioning scenario. In order to evaluate the performance of the proposed algorithm, field tests for static case and dynamic case were carried out. In the static test case, an average positioning accuracy of 0.238 m was achieved with an improvement of 38.66% compared to the least square solution purely based on acoustic ranging observables. For the dynamic test case, a method based on predicted state of Doppler compensation was applied for Doppler corrections of the TDOA observables. Furthermore, asynchronous TOA compensation, which caused by the fact that TOA measurements are not estimated in the same epoch, was also applied before feeding to the positioning algorithm. The test results demonstrated that positioning accuracy in dynamic case is 0.513 m. Compared to the solution without applying the Doppler corrections and the asynchronous TOA compensations, the performance is improved by 27.64%. Three mobile phones (Huawei Mate 20, OnePlus 6 and Google Pixel 3) were used in the field test, the performance of the positioning algorithm is consistence in all three different phones