GVIM: GNSS/Visual/IMU/Map Integration Via Sliding Window Factor Graph Optimization in Urban Canyons

Globally referenced and accurate positioning is of great significance for the realization of fully autonomous systems. The visual and inertial measurement unit (IMU) integrated navigation system (VINS) can provide accurate positioning in a short period but is subject to drift over time. Meanwhile, the performance of the VINS is significantly degraded in urban canyons due to the numerous outlier visual features caused by moving objects and unstable illuminations. The global navigation satellite system (GNSS) can provide reliable and globally referenced positioning in open areas, but it is challenged in urban canyons due to the signal reflections and blockages from tall buildings. To exploit the complementariness of the GNSS and VINS, this paper proposed a sliding window factor graph optimization (FGO) based GNSS/Visual/IMU/Map Integration. First, the window carrier phase (WCP) and the Doppler measurements are explored to constrain the relative motion of the system within consecutive epochs. Second, a novel sliding window (SW) based map matching model is proposed to correct the states using the lightweight OpenStreetMap (OSM). Different from conventional filtering-based map matching, the states within the sliding window of the FGO are associated with the lane information from the OSM which effectively exploited the measurement redundancy arising from the factor graph model. The effectiveness of the proposed method is validated using the challenging dataset collected in the urban canyons of Hong Kong. The results showed that lane-level positioning can be achieved even in dense urban scenarios, with poor satellite visibilities and numerous visual feature outliers.