Simultaneous localization and mapping: part I

Abstract

T he simultaneous localization and mapping (SLAM) problem asks if it is possible for a mobile robot to be placed at an unknown location in an unknown environment and for the robot to incrementally build a consistent map of this environment while simultaneously determining its location within this map. A solution to the SLAM problem has been seen as a "holy grail" for the mobile robotics community as it would provide the means to make a robot truly autonomous.The "solution" of the SLAM problem has been one of the notable successes of the robotics community over the past decade. SLAM has been formulated and solved as a theoretical problem in a number of different forms. SLAM has also been implemented in a number of different domains from indoor robots to outdoor, underwater, and airborne systems. At a theoretical and conceptual level, SLAM can now be considered a solved problem. However, substantial issues remain in practically realizing more general SLAM solutions and notably in building and using perceptually rich maps as part of a SLAM algorithm.This two-part tutorial and survey of SLAM aims to provide a broad introduction to this rapidly growing field. Part I (this article) begins by providing a brief history of early developments in SLAM. The formulation section introduces the structure the SLAM problem in now standard Bayesian form, and explains the evolution of the SLAM process. The solution section describes the two key computational solutions to the SLAM problem through the use of the extended Kalman filter (EKF-SLAM) and through the use of Rao-Blackwellized particle filters (FastSLAM). Other recent solutions to the SLAM problem are discussed in Part II of this tutorial. The application section describes a number of important real-world implementations of SLAM and also highlights implementations where the sensor data and software are freely down-loadable for other researchers to study. Part II of this tutorial describes major issues in computation, convergence, and data association in SLAM. These are subjects that have been the main focus of the SLAM research community over the past five years. History of the SLAM ProblemThe genesis of the probabilistic SLAM problem occurred at the 1986 IEEE Robotics and Automation Conference held in San Francisco, California. This was a time when probabilistic methods were only just beginning to be introduced into both robotics and artificial intelligence (AI). A number of researchers had been looking at applying estimation-theoretic methods to mapping and localization problems; these included Peter Cheeseman, Jim Crowley, and Hugh DurrantWhyte. Over the course of the conference, many paper table cloths and napkins were filled with long discussions about consistent mapping. Along the way, Raja Chatila, Oliver Faugeras, Randal Smith, and others also made useful contributions to the conversation.The result of this conversation was a recognition that consistent probabilistic mapping was a fundamental problem in robotics with major conceptual and computational issues that ne...

Simultaneous Localization and Mapping in Marine Environments

“…Moutarlier and Chatila provided the first implementation of this type of algorithm with real data [54], using data from a scanning laser range finder mounted on a wheeled mobile robot operating indoors. The authors noted that the size of the state vector would need to grow linearly with the number of landmarks and that it was necessary to maintain the full correlation between all the variables being estimated; thus, the algorithm scales quadratically with the number of landmarks [11].…”

AUV Benthic Habitat Mapping in South Eastern Tasmania

“…Simultaneous Localisation and Mapping (SLAM) is the process of concurrently building a feature based map of the environment and using this map to obtain estimates of the location of the vehicle. The SLAM algorithm has seen a considerable amount of interest from the mobile robotics community as a tool to enable fully autonomous navigation [3,4]. Our current work has concentrated on efficient, stereo based Simultaneous Localisation and Mapping and dense scene reconstruction suitable for creating detailed maps of seafloor survey sites [8,9].…”

Vision based persistent localization of a humanoid robot for locomotion tasks

“…This process selects reliable image features to be incorporated to the sparse map representing the environment. Filtering-based methods also solve the problem as reported in the literature (Durrant-Whyte and Bailey, 2006). Paradoxically, while the success of classical * Corresponding author approaches is far from being directly extrapolated to walking machines, humanoid robots mainly rely on vision systems in order to perceive the environment and resemble human capabilities.…”