Acoustic Positioning Algorithm Based On Multi-sensor Data Fusion Research

The current positioning technology, for the interference from acoustic environment, the impact of electronic noise, etc., can not position target precisely. The multi-sensor data fusion theory is applied in acoustic positioning, by increasing the number of sensors to achieve positioning target several times, and inputting these position data into the data fusion system to obtain the results of the final positioning in order to improve position accuracy.In this paper, on the basis of accuracy analysis of 5 elements across position method, a new BP neural network-based data fusion algorithm of multi-sensor acoustic positioning algorithm is proposed. Positioning target repeatedly is achieved by using 5 elements across position method and permutation. By increasing the number of statistic samples, the algorithm can reduce interference from the electronics, environment, therefore, the accuracy of acoustic positioning is highly promoted. At the same time, the results obtained are integrated so as to accomplish positioning more accurately.BP neural network-based data fusion approach, with its advantages in robustness and convergence, has become a commonly used method for data fusion, while the traditional single-BP neural network has some limitations. To improve the traditional single-BP neural network, a data fusion algorithm based on double-BP neural network is proposed. The algorithm first removes outlier data network for training to identify the effectiveness of target data, and then points through the effective integration of network to obtain the ultimate localization coordinates.Finally, simulation analysis is implemented to verify the algorithm in this paper. Results show that in this paper, the proposed algorithm of multi-sensor acoustic based on fusion positioning and dual-BP neural network-based data fusion technology have some advantages in positioning accuracy and speed.