3D normal distributions transform occupancy maps: An efficient representation for mapping in dynamic environments

Abstract

In order to enable long-term operation of autonomous vehicles in industrial environments numerous challenges need to be addressed. A basic requirement for many applications is the creation and maintenance of consistent 3D world models. This article proposes a novel 3D spatial representation for online real-world mapping, building upon two known representations: normal distributions transform (NDT) maps and occupancy grid maps. The proposed normal distributions transform occupancy map (NDT-OM) combines the advantages of both representations; compactness of NDT maps and robustness of occupancy maps. One key contribution in this article is that we formulate an exact recursive updates for NDT-OMs. We show that the recursive update equations provide natural support for multi-resolution maps. Next, we describe a modification of the recursive update equations that allows adaptation in dynamic environments. As a second key contribution we introduce NDT-OMs and formulate the occupancy update equations that allow to build consistent maps in dynamic environments. The update of the occupancy values are based on an efficient probabilistic sensor model that is specially formulated for NDT-OMs. In several experiments with a total of 17 hours of data from a milk factory we demonstrate that NDT-OMs enable real-time performance in large-scale, long-term industrial setups.

RGB–D terrain perception and dense mapping for legged robots

“…The aim of the mapping method is to create an accurate model of the environment, which is suitable to be used for motion planning of a walking robot. We chose the elevation map representation of the environment because it allows us to efficiently query the map in the motion planning of the robot on a rough terrain (Belter and Skrzypczyński, 2011a;Belter et al, n.d.;Saarinen et al, 2013). The main problem with this mapping method for a walking robot is related to the perception system and geometrical properties of the obstacles.…”

Multi-scale conditional transition map: Modeling spatial-temporal dynamics of human movements with local and long-term correlations

“…With the resulting temporal occupancy grid (TOG), the dynamics of each cell is classified by the occupancy values over different timescales. Saarinen et al [7] proposed a 3D modeling approach combining the normal distributions transform (NDT) and the occupancy grid map. The method has the ability of adapting to a dynamic environment with a recencyweighted strategy.…”

“…The occupancy grid map [1] and its variants have been employed in modeling dynamics in the environment by many methods, for example [2], [3], [4], [5], [6], [7]. In the work [3], the occupancy grid map is generalized and the static environment assumption is relaxed by introducing the Hidden Markov Model (HMM).…”

Modeling motion patterns of dynamic objects by IOHMM

“…Each set is updated at a varied learning rate to adapt to different extent of dynamics in the environment according to the timescale. Saarinen et al [6] proposed a 3D modeling approach combining the normal distributions transform (NDT) and the occupancy grid map. The method has the ability of adapting to a dynamic environment with a recency-weighted strategy.…”

“…Occupancy grids [3] are one of the most popular representations of environments where mobile robots operate, and are also used for dynamic environment studies [4] [5] [6] [7]. In a very recent work, Meyer-Delius et al [8] generalized the occupancy grid map and released the static environment assumption by introducing the hidden Markov model (HMM).…”