MPNN-Loc, a novel Graph Neural Network (GNN)-based localization algorithm aims to address the collaborative localization problem in sensor networks based on distance measurements. Specifically, it focuses on layout-agnostic sensor networks where the number and positions of sensor nodes to be localized can vary dynamically. MPNN-Loc integrates a message-passing mechanism with a correlation-based refinement strategy. The extensive evaluations exhibits its excellent generalization ability and strong robustness in dynamic and noisy environments.
JingjunXu/MPNN-Loc
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|