In this paper, we propose an indoor self-localization algorithm for mono microphones, which achieves efficient localization in a limited computational resource environment by generating two pairs of dipole sound fields and combining them with orthogonal detection techniques. Compared with the traditional time-driven mode, this algorithm introduces the frequency-division multiplexing technique, which dramatically improves the localization speed and shows excellent robustness at the same time. Numerical simulation experiments show that under the same environmental conditions, the improved algorithm in this paper is able to accurately estimate the azimuth and elevation angles by driving two dipoles synchronously, and the average error is no more than 0.5° compared with the original algorithm, while the positioning time is shortened by 3 to 4 seconds. The algorithm not only significantly outperforms the existing methods in terms of positioning timeliness, but also effectively ensures the accuracy of positioning, which has significant application value in the field of indoor positioning.