一种面向智能终端的视觉陀螺仪%2fPDR%2fGNSS组合导航方法.pdfVIP

Visual-Aided Pedestrian/GNSS Coupled Navigation for Smart Terminals Gong Xuping, Wei Dongyan, Li Xianghong, Xu Ying, Yuan Hong (Academy of Opto-electronics, Chinese Academy of Sciences, Beijing, China, 10094) Abstract: Smart terminals are attractive platforms for the navigation systems due to their small size, low cost and the openness. Pedestrian navigation is well used for it can keep a long period accurate relative position accuracy. Apart from system solving error, obtaining a reliable yaw angle from a digital compass is one of the most challenging tasks. Visual-aiding methods have been considered to solve this problem, but they are sensitive to line detection methods. Considering GNSS has a good accuracy in global coordinate system and vision-aided PDR system gives a good structure in local coordinate system, this paper presents a visual-aided pedestrian/GNSS coupled navigation algorithm by fitting global position to local measured system. First we detect heading change by vanishing points from lines in consecutive images. Then we embedded KVISPDR and GNSS results into transform structure. We conduct a series of experiments in open areas, and results from fields test demonstrate that the performance of the KVISPDR algorithm in positioning can mitigate 30% localization error in comparison to embed GNSS positioning system. KVISPDR/GNSS uses merely mobile phone and requires no infrastructure or additional reference information. As an effective and light-weight augmentation to GNSS positioning, KVISPDR/GNSS holds promise in real-world feasibility.