| Title |
A Stage-wise Dynamic Weight Map Fusion Method with Real-time Battlefield Recognition and Mission Conditions |
| Authors |
조은지(Eunji Cho) ; 이태영(Tae-Young Lee) ; 안종식(Jongsik Ahn) ; 조호진(Hojin Cho) ; 김준원(Joonwon Kim) |
| Keywords |
Grid map; Grid map fusion; Grid map generation; Unmanned ground vehicle; Map fusion |
| Abstract |
In modern battlefield environments, unmanned ground vehicle(UGV) must achieve both real-time responsiveness and survivability during mission execution. Grid maps that represent battlefield information such as traversability and threat levels are widely used for autonomous path planning; however, conventional approaches based on a single grid map or fixed fusion of multiple maps have limitations in reflecting diverse mission conditions and dynamically changing situations. In this paper, we propose a priority-based dynamic weight map fusion method that effectively incorporates real-time operational situations. Four analysis maps?an FDB-based grid analysis map, a stuff-based traversability analysis map, an obstacle analysis map, and a threat analysis map?are generated using a FDB and drone image. These maps are then used as inputs to a two-stage weighting process to produce the final fusion map. First, grid-level weights are adaptively adjusted according to situation conditions. Second, map-level fusion weights are applied based on mission conditions to reflect both traversability and survivability in the final fusion map. The proposed method was validated using analysis maps constructed from real drone image and panoptic segmentation, with fusion results compared under different mission conditions as well as default and situation-applied cases. Experimental results demonstrate that the proposed approach can effectively adjust fusion weights in response to real-time battlefield situation changes and mission condition variations, enabling more accurate and reliable generation of traversability-reflected fusion maps. |