부산대학교 도서관

In order to solve the problem of a large positioning error, low refresh rate and high-power consumption of UWB in an indoor high-density environment. This paper proposes a UWB indoor positioning method based on SVM (support vector machine), which transforms the positioning problem into the solution of the multi-classification problem. The SVM model based on TDOA (time difference of arrival) is established, and the one-to-one classification model is used to solve the coordinate value quickly. Experimental results show that the proposed algorithm has higher accuracy and less time in indoor positioning. The algorithm proposed in this paper has more application prospects.