Intelligent tour guide is a comprehensive service based on tourist's location, which is closely related to Geographic Information System (GIS), mobile positioning technology and Location-Based Service (LBS). But the intelligent tour guide field urgently needs the integrated positioning and navigation technology inside and outside the room. IR-UWB technology is suitable for positioning, tracking, navigation and communication in complex indoor environment, and is considered as the most potential indoor positioning technology to realize seamless connection between indoor and outdoor with outdoor positioning technologies such as GPS. However, one of the main problems facing IR-UWB positioning is Non-Line-Of-Sight (NLOS) error. Based on the advantages of BP neural network and support vector machine, this paper proposes a multi-model fusion algorithm to mitigate the NLOS propagation error of the time difference of arrival (TDOA) and the angle of arrival (AOA) of IR-UWB signal, and then uses TDOA/AOA hybrid positioning that mitigates the NLOS error. Simulation results show that the combined algorithm has stronger NLOS resistance and higher positioning accuracy than the single machine learning algorithm in mitigation NLOS errors.
Song, Bo; Li, Shenglin; Ren, Qinghui; and Chen, Chen, "Research on Impulse Radio Ultra - wideband Positioning Method Based on Combined BP Neural Network and SVM" (2018). ICEB 2018 Proceedings (Guilin, China). 48.