Monte Carlo Localization in Outdoor Terrains Using Multi-Level Surface Maps

Abstract

We propose a novel combination of techniques for robustly estimating the position of a mobile robot in outdoor environments using range data. Our approach applies a particle filter to estimate the full six-dimensional state of the robot and utilizes multilevel surface maps which, in contrast to standard elevation maps, allow the robot to represent vertical structures and multiple levels in the environment. We describe probabilistic motion and sensor models to calculate the proposal distribution and to evaluate the likelihood of observations. We furthermore describe an active localization approach which actively selects the sensor orientation of the 2D laser range scanner to improve the localization results. To efficiently calculate the appropriate orientation we apply a clustering operation on the particles and only evaluate potential orientations based on these clusters. Experimental results obtained with a mobile robot in large-scale outdoor environments indicate that our approach yields robust and accurate position estimates. The experiments also demonstrate that multilevel surface maps lead to a significantly better localization performance than standard elevation maps. They additionally show that further accuracy is obtained from the active sensing approach.

Multi-resolution surfel mapping and real-time pose tracking using a continuously rotating 2D laser scanner

“…Khoshelham pro poses using solely planar objects for localization in 3D within indoor environments [14]. Kuemmerle et al [15] apply Monte Carlo localization in Multi-Level Surface maps [16] which represent occupied height intervals on a 2D grid. Klaess et al [17] model the environment in surfel maps in a fixed resolution, similar to the 3D-NDT [9].…”

An effective exploration approach to simultaneous mapping and surface material–type identification of complex three‐dimensional environments

“…This volume is a function of the sensor's position and orientation and is proportional to n u . The prediction of V new ( Q) can be determined by the established technique of ray casting (Kummerle, Pfaff, Triebel, & Burgard, 2008). Figure 2 shows that for each of the laser scanner's rays, a ray is cast from the proposed sensor viewpoint, 0 T s ( Q), through the partially known map to the extent of the sensing range.…”

Autonomous Exploration for 3D Map Learning

“…Hence, we perform several 3D scans along the way, which has the benefit, that it allows us to optimize the localization of the robot with the pose returned by the scan matcher. We apply a 6D Monte Carlo localization proposed by Kümmerle et al [9]. After each 3D scan, we replan the path to the selected viewpoint.…”