An Improved TDOA-based DPD Method via Multiple-frequency Function Fusion

The direct position determination (DPD) method is superior to traditional two-step localization methods which have error accumulation problem caused by intermediate parameters estimations. However, the DPD method is still faced with the problems of high dimensional matrix computation and decomposition, which lead to high complexity in the fusion center. In this paper, a new method is proposed based on the fusion of the cost functions in multiple frequency bands, in which the orthogonality is initially established to realize multiple signal classification (MUSIC) algorithm. Data segmentation is used to reduce the amount of calculation, and Lagrange multiplier method is exploited to avoid the influence of unknown attenuation coefficients. Compared to current DPD methods, our method achieves better positioning performance, while it requires lower complexity. Effectiveness of the approach is demonstrated by both simulation and real-world tests.