Akademik Veri Yönetim Sistemi
Applied Sciences (Switzerland), cilt.16, sa.7, 2026 (SCI-Expanded, Scopus)
This study examines a cost-shaping method that considers distance information to obstacles in a line-of-sight (LOS) any-angle path-planning approach on grid-based maps. In the proposed approach, the safety distance to obstacles is added to the cost in a controlled manner via a single adjustable and interpretable parameter; thus, the balance between safety and efficiency becomes practically adjustable. Node selection in the planning process is performed while maintaining the classical search rule; the additional penalty related to the safety distance is only included in the transit cost. This design strengthens consistency between method definition and implementation and eliminates the risk of the same safety term being considered multiple times. The experimental evaluation was conducted on a three-by-three scenario set encompassing map type and difficulty level dimensions. Starting and ending points were selected in a layered and matched manner as easy/medium/difficult; the safety parameter was scanned at different values, following a repeatable protocol under all conditions. Outputs were evaluated using efficiency metrics such as path length and number of turns, as well as minimum safety distance, safety distance violation rate, and a curvature indicator representing the smoothness of the path geometry. In addition, practical costs such as planning time, an expanded number of nodes, and memory footprint were reported. The results show that exposure to low safety distance zones decreases and the path geometry becomes more regular with increasing safety parameters. Furthermore, it was observed that the success rate increased in pooled analyses while memory usage remained constant; paired statistical tests and effect size measurements confirmed that the improvements were strong and consistent. These findings reveal that safety distance-sensitive cost-shaping offers a lean control mechanism that enhances safety and maintains practical applicability within line-of-sight-based any-angle planning.