Real-time underground localization using graph pruning-augmented graph optimization for directional drilling

Abstract

Underground localization is widely used for the directional drilling due to advances in the unconventional resources. To mitigate the accumulated errors, a novel underground localization algorithm using re-measurement of the magnetic field and graph SLAM (simultaneous localization and mapping) was proposed. However, since the penetration length of the drilling device is very long, the real-time localization is not assured due to huge number of nodes. To reduce the computation time, an underground localization method using graph optimization augmented by graph pruning is proposed in this paper. The graph pruning algorithm can remove the nodes and constraints that have low priority. In this paper, in order to reduce the constraints, minimum spanning tree algorithm is used.