Lego voxelization method for arbitrary building information models to support path planning

Building information models are widely used as data sources for indoor path planning, with many approaches employing grid cells or voxels to represent the indoor navigation environment. Among these, the LEGO voxelization method is particularly suitable for creating path-planning environments, as the generated LEGO voxels fit the components better than other voxelization methods. However, current LEGO-based voxelization methods, which rely on triangle normals, can be unreliable when processing models containing triangles with inconsistent vertex sequences. To address this limitation, this study proposes an improved LEGO voxelization method that eliminates the need for triangle normal data. Experimental results demonstrate that the method achieves an accuracy ratio of 95.33% to 99.96% across 12 test models and outperforms the current LEGO-voxelization method in various model formats. Additionally, it can voxelize large-scale models within a reasonable time. By streamlining the creation of cell-based environments, a traditionally time-consuming and tedious task, this method allows researchers and practitioners to focus more on path-planning research and applications.