Simultaneous Localization, Mapping and Moving Object Tracking

Abstract

Simultaneous localization, mapping and moving object tracking (SLAMMOT) involves both simultaneous localization and mapping (SLAM) in dynamic environments and detecting and tracking these dynamic objects. In this paper, we establish a mathematical framework to integrate SLAM and moving object tracking. We describe two solutions: SLAM with generalized objects, and SLAM with detection and tracking of moving objects (DATMO). SLAM with generalized objects calculates a joint posterior over all generalized objects and the robot. Such an approach is similar to existing SLAM algorithms, but with additional structure to allow for motion modeling of generalized objects. Unfortunately, it is computationally demanding and generally infeasible. SLAM with DATMO decomposes the estimation problem into two separate estimators. By maintaining separate posteriors for stationary objects and moving objects, the resulting estimation problems are much lower dimensional then SLAM with generalized objects. Both SLAM and moving object tracking from a moving vehicle in crowded urban areas are daunting tasks. Based on the SLAM with DATMO framework, we propose practical algorithms which deal with issues of perception modeling, data association, and moving object detection. The implementation of SLAM with DATMO was demonstrated using data collected from the CMU Navlab11 vehicle at high speeds in crowded urban environments. Ample experimental results shows the feasibility of the proposed theory and algorithms.

Model-Instance Object Mapping

“…The Robot Object Mapping Algorithm [8] detects moveable objects by detecting differences in the maps built by SLAM at different times. Detection and Tracking of Moving Objects [9] is an approach that seeks to detect and track moving objects while performing SLAM. Relational Object Maps [10] reasons about spatial relationships between objects in a map.…”

Model based vehicle detection and tracking for autonomous urban driving

“…A number of approaches focused on the use of vision exclusively (Zielke et al, 1993;Dickmanns, 1998;Dellaert and Thorpe, 1998), whereas others utilized laser range finders (Zhao and Thorpe, 1998;Streller et al, 2002;Wang et al, 2007) sometimes in combination with vision (Wender and Dietmayer, 2008). We give an overview of prior art in Sect.…”

“…Zhao and Thorpe (1998); Streller et al (2002); Wang (2004); Wender and Dietmayer (2008)) including most recent developments by the UGC participants (Darms et al, 2008;Leonard et al, 2008). Typically these approaches proceed in three stages: data segmentation, data association, and Bayesian filter update.…”

Towards hierarchical blackboard mapping on a whiskered robot

“…We do not consider the full SLAM problem here, but instead work in a simulation of CrunchBot having zero odometry noise to avoid the localisation problem and focus on mapping only. Related object-based mapping models have recently appeared [51,23,46,40] using laser sensors to recognise and learn complex but nonhierarchical spatial models. However as data available through whiskers to CrunchBot is much sparser than that from laser scanners, the required level of sensor detail is unavailable, therefore we compensate with the new mapping technique of fusing contact reports into hierarchical models.…”