Preliminary Study on Dynamic Updates of Spatial Digital Twins via 3D Mobile Crowdsensing
Kenta Hasegawa, Tatsuya Kase, Kaito Watanabe, Takumi Miyoshi, Taku Yamazaki · 2025
A city-scale spatial digital twin is an emerging technology for creating smart cities by replicating urban spaces in the virtual environment. However, building such digital twins requires vast amounts of spatial data to accurately represent the physical space and enable city services. To address this issue, we developed a 3D mobile crowdsensing system that the participants use light detection and ranging (LiDAR)-equipped mobile devices to capture nearby point clouds. However, it mainly focused on examining the impact of mobile crowdsensing on spatial information sensing, overlooking the details of constructing and maintaining spatial digital twins. This study introduces a dynamic update mechanism of digital twins based on global and local registration algorithms. The proposed mechanism contains three steps to register collected point clouds into a global point cloud as a digital twin: preprocessing, feature point matching, and registration steps. Finally, we evaluate the performance of the proposed mechanism using a city-scale point cloud dataset, assessing registration success rate, accuracy, and processing time.