Beom-Hee Lee

A survey of map merging techniques for cooperative-SLAM

This paper presents a survey of map merging techniques for cooperative-SLAM. The recently proposed map merging techniques are classified into two categories: direct map merging and indirect map merging. In each category, several techniques are briefly described. Then, their advantages and disadvantages are discussed in the context of accuracy and computation time. The description and discussion can contribute to realizing and improving cooperative-SLAM.

A KNOWLEDGE BASE FOR DYNAMIC PATH PLANNING OF MULTI-AGENTS

Abstract A fuzzy rule base is proposed to navigate multi-agents from initial positions to target positions in unknown environments. The proposed fuzzy rule base determines the highest priority of nine possible heading directions. The fuzzy rule base has been developed employing genetic algorithms as an approach to dynamic path planning of autonomous multi-agents in unknown environments. Paths which satisfy some optimization criteria with respect to moving distance, smoothness, and clearance of obstacles was obtained from the fuzzy rule base. The fuzzy rule base was obtained from off-line navigation with precise sensor modeling and applied to various simulated on-line navigation. The performance of the fuzzy rule base in different unknown environments is acceptable and shown in simulation results.

Robust scan matching with curvature-based matching region selection

This paper presents a novel scan matching algorithm which uses not whole scan region but only salient scan region selected around curvature-based features. A curvature function computed by the relative coordinates of neighbor scan points is used to extract salient features which are invariant to translation and rotation. Because the scan matching regions are selected around the salient features, the presented algorithm can be robustly performed even in noisy environments. The robustness of the presented algorithm was tested by datasets obtained from various noisy environments and was verified by consistently showing smaller errors than other scan matching algorithms. Moreover, the presented algorithm was successfully applied to SLAM.

A visual landmark design method for high speed multiple robots

The visual structured landmarks used in robot localization are popularly utilized in the mobile robot navigation field. When the localization of the robot implies fast and accurate performances, such landmarks can be used in visual surveillance systems to get very efficient results. This paper suggests a visual landmark designing methodology for high speed multiple mobile robots. The conventional visual landmark models are examined according to instructions of the proposed methodology. The experimental results are presented to evaluate the conventional landmarks and verify the feasibility of the proposed methodology in an indoor visual surveillance system.

Advances in Sea Coverage Methods Using Autonomous Underwater Vehicles (AUVs)

Due to rapid changes in modern technological development over the last several years, researches on small Autonomous Underwater Vehicles (AUVs) and Unmanned Underwater Vehicles (UUVs) appeared as important issues for various possibilities of application in the ocean. Military applications together with commercial need require practical details which are robust and cheap for realization as a product. In addition, works on coverage have been motivated by a wide range of real world applications that include non-humanitarian demining, deep-sea development sweeping and robotic spray-painting. Currently, many coverage applications utilize on-line coverage algorithms (Gabriely et al., 2003), where robots do not rely on a priori knowledge of a work-area, and thus must construct their motion trajectories step-by-step, addressing discovered obstacles as they move. This is the main contrast from the conventional off-line coverage algorithms, where robots are given a map of a work-area, and can therefore plan their paths ahead of deployment (Ge et al., 2005). In this paper, we focus on on-line coverage for underwater environment by multiple AUVs. Relevant works have shown that one of advantages of adopting multiple robots for a coverage task is the potential ability for more efficient coverage (Hazon et al, 2005). However, another advantage is that they usually offer greater robustness. Unfortunately,this important capability has been neglected in previous works done on on-line algorithms. As far as the details of the on-line coverage methods, there are many kinds of covering motions that can be realized either by back-and-forth motions in vast cells, or by following a general Voronoi diagram in narrow cells (Acar et al., 2001). Exact cellular decomposition can also be achieved through the use of the boustrophedon decomposition (Choset, 2001) as well as through Morse functions (Acar, 2002). The boustrophedon approach has been extended in (Choset, 2001) to the multi-robot domain. Such form of coverage requires the coverage to be executed in formations, which may be accomplished in a variety of ways (Ge et al., 2005). Spanning Tree Coverage algorithms (Gabriely, 2003) have also been proposed for online coverage of an unknown and gridded environment by a robot. For coverage with multi-robot teams, a frontier-based exploration technique has been used to guide robots towards the region between known and unknown areas on the map (Yamauchi et al., 1998). The Mark and Cover (MAC) algorithm has been proposed in (Wagner et al., 2000) with the proved convergence and the bounds on cover time of the algorithm. In (Yang & Luo, 2004), a neural network model has been used for exploration of a O pe n A cc es s D at ab as e w w w .in te hw eb .c om

Development of an Open Architecture Controller for a Commercial AUV

AUVs are essential to reduce the difficulty in building up flexible exploring and working systems in the water. However it is hard to integrate sub-modules or control algorithms into an AUV because AUV makers have developed control architectures and communication protocols of their own. So we develop an open architecture for AUV controller that integrates ready-made sub-modules to a reconfigurable and fault tolerant AUV. And we apply this open architecture to the cruising AUV.

A sensor fusion system using enhanced extended Kalman filter with double fuzzy logics for autonomous robot guidance

This paper presents a sensor fusion system for autonomous guidance of a robot. The sensor fusion system is physically composed of a laser range finder and two vision sensors. Also, it is systematically designed to fuse the information obtained from sensors and to overcome those sensors drawbacks. To be specific, it utilizes double fuzzy logics for fusion and extended Kalman filter for estimation sequentially. In experimental setup, we compare the proposed sensor fusion system and systems using sensors independently by linking a wall-following algorithm for autonomous robot guidance. The result shows that the proposed system has robustness against environments with some difficult conditions.

An Improved Resampling Technique using Particle Density Information in FastSLAM

FastSLAM which uses the Rao-Blackwellized particle filter is one of the famous solutions to SLAM (Simultaneous Localization and Mapping) problem that estimates concurrently a robot`s pose and surrounding environment. However, the particle depletion problem arises from the loss of the particle diversity in the resampling process of FastSLAM. Then, the performance of FastSLAM degenerates over the time. In this work, DIR (Density Information-based Resampling) technique is proposed to solve the particle depletion problem. First, the cluster is constructed based on the density of each particle, and the density of each cluster is computed. After that, the number of particles to be reserved in each cluster is determined using a linear method based on the distance between the highest density cluster and each cluster. Finally, the resampling process is performed by rejecting the particles which are not selected to be reserved in each cluster. The performance of the DIR proposed to solve the particle depletion problem in FastSLAM was verified in computer simulations, which significantly reduced both the RMS position error and the feature error.