Terrain drivability analysis in 3D laser range data for autonomous robot navigation in unstructured environments

Abstract

Three-dimensional laser range finders provide autonomous systems with vast amounts of information. However, autonomous robots navigating in unstructured environments are usually not interested in every geometric detail of their surroundings. Instead, they require realtime information about the location of obstacles and the condition of drivable areas.In this paper, we first present grid-based algorithms for classifying regions as either drivable or not. In a subsequent step, drivable regions are further examined using a novel algorithm which determines the local terrain roughness. This information can be used by a path planning algorithm to decide whether to prefer a rough, muddy area, or a plain street, which would not be possible using binary drivability information only.

Probabilistic terrain classification in unstructured environments

“…The roughness can be obtained by calculating the local height disturbance, as in Neuhaus et al [1]. The roughness feature offers valuable clues about how uneven a terrain cell is.…”

“…Therefore, as a first step, a reduction of the large point cloud is necessary and an efficient data structure is essential. Our work was motivated by the terrain analysis performed by Neuhaus et al [1], where a two dimensional grid structure was introduced to provide fast access to negotiability estimates.…”

Integrating visual and range data for road detection

“…In these studies, 3D points are either classified directly [6], or divided into grids for analysis. In the latter method, features such as elevation mean and variance [7], roughness [8], and point distribution [9] within each grid are extracted to identify grid-wise traversable regions.…”

A terrain description method for traversability analysis based on elevation grid map

“…It can be found that the eigenvectors of matrix S must be mutually orthogonal, because S is a real symmetrical matrix. In this case, the PCA of the covariance matrix S of Q f is in fact equivalent to eigenvalue decomposition for S. 9 The eigenvectors and eigenvalues of S can be solved by singular value decomposition. S can be transformed into a diagonal matrix L with an invertible matrix V generated by these orthogonal eigenvectors, which is shown in the following…”

“…More terrain analysis is not mentioned. Neuhaus et al 9 presented a grid-based algorithm for classifying regions as either drivable or not. The local terrain roughness feature based on PCA is estimated for terrain drivability analysis.…”