We tackle the challenges of decentralized multi-robot navigation in environments with nonconvex obstacles, where complete environmental knowledge is unavailable. While reactive methods like Artificial Potential Field (APF) offer simplicity and efficiency, they suffer from local minima, causing robots to become trapped due to their lack of global environmental awareness. Other existing solutions either rely on inter-robot communication, are limited to single-robot scenarios, or struggle to overcome nonconvex obstacles effectively. Our proposed methods enable collision-free navigation using only local sensor and state information without a map. By incorporating a wall-following (WF) behavior into the APF approach, our method allows robots to escape local minima, even in the presence of nonconvex and dynamic obstacles including other robots. We introduce two algorithms for switching between APF and WF: a rule-based system and an encoder network trained on expert demonstrations. Experimental results show that our approach achieves substantially higher success rates compared to state-of-the-art methods, highlighting its ability to overcome the limitations of local minima in complex environments.

Indoor parking lots are the GPS-denied spaces to which vision-based localization approaches have usually been applied to solve localization problems. However, due to the repetitiveness and symmetry of the spaces, visual localization methods commonly confront difficulties in estimating precise 3D poses. In this study, we propose four novel modules that improve localization precision by imposing the existing methods with the spatial discerning ability. The first module constructs a key text graph that represents the topology of key texts in the space and becomes the basis for discerning repetitiveness and symmetry. Next, the orientation filtering module estimates the unknown 3D orientation of the query image and resolves spatial symmetric ambiguity. The similarity scoring module sorts out the top-scored database images, discerning the spatial repetitiveness based on detected key text bounding boxes. Our pose verification module evaluates the pose confidence of top-scored candidates and determines the most reliable pose. Our method has been validated in two real indoor parking lots, achieving new state-of-the-art performance levels.

Indoor visual localization is significant for various applications such as autonomous robots, augmented reality, and mixed reality. Recent advances in visual localization have demonstrated their feasibility in large-scale indoor spaces through coarse-to-fine methods that typically employ three steps: image retrieval, pose estimation, and pose selection. However, further research is needed to improve the accuracy of large-scale indoor visual localization. We demonstrate that the limitations in the previous methods can be attributed to the sparsity of image positions in the database, which causes view-differences between a query and a retrieved image from the database. In this paper, to address this problem, we propose a novel module, named pose correction, that enables re-estimation of the pose with local feature matching in a similar view by reorganizing the local features. This module enhances the accuracy of the initially estimated pose and assigns more reliable ranks. Furthermore, the proposed method achieves a new state-of-the-art performance with an accuracy of more than 90 % within 1.0 m in the challenging indoor benchmark dataset InLoc for the first time.1

Planarity of checkerboards is a widely used feature for extrinsic calibration of camera and LiDAR. In this study, we propose two analytically derived covariances of (i) plane parameters and (ii) plane measurement, for precise extrinsic calibration of camera and LiDAR. These covariances allow the graded approach in planar feature correspondences by exploiting the uncertainty of a set of given features in calibration. To construct plane parameter covariance, we employ the error model of 3D corner points and the analytically formulated plane parameter errors. Next, plane measurement covariance is directly derived from planar regions of point clouds using the out-of-plane errors. In simulation validation, our method is compared to an existing uncertainty-excluding method using the different number of target poses and the different levels of noise. In field experiment, we validated the applicability of the proposed analytic plane covariances for precise calibration using the basic planarity-based method and the latest planarity-and-linearity-based method.

In this paper, we propose a dependable visual kidnap recovery (KR) framework that pinpoints a unique pose in a given 3D map when a device is turned on. For this framework, we first develop indoor-GeM (i-GeM), which is an extension of GeM [1] but considerably more robust than other global descriptors [2]-[4], including GeM itself. Then, we propose a convolutional neural network (CNN)-based system called KR-Net, which is based on a coarse-to-fine paradigm as in [5] and [6]. To our knowledge, KR-Net is the first network that can pinpoint a wake-up pose with a confidence level near 100% within a 1.0 m translational error boundary. This dependable success rate is enabled not only by i-GeM, but also by a combinatorial pooling approach that uses multiple images around the wake-up spot, whereas previous implementations [5], [6] were constrained to a single image. Experiments were conducted in two challenging datasets: a large-scale (12,557 m2) area with frequent featureless or repetitive places and a place with significant view changes due to a one-year gap between prior modeling and query acquisition. Given 59 test query sets (eight images per pose), KR-Net successfully found all wake-up poses, with average and maximum errors of 0.246 m and 0.983 m, respectively.

This paper proposes a realistic indoor modeling framework for large-scale indoor spaces. The proposed framework reduces the geometric complexity of an indoor model to efficiently represent large-scale environments for image-based rendering (IBR) approaches. For this purpose, the proposed framework removes geometrically excluded objects (GEOs) in point cloud and images, which represent the primary factors in high geometric complexity. In particular, GEOs are coherently removed from all images using a global geometry model. Then, the remaining holes are inpainted using globally consistent guidelines, to achieve accurate image blending in IBR approaches. The experimental results verify that the proposed GEO removal framework provides efficient point clouds and images for realistic indoor modeling in large-scale indoor spaces.

This paper proposes a calibration method that accurately estimates six parameters between the two centers of 2-D light detection and ranging (LIDAR) and a rotating platform. This method uses a simple plane, and to the best of our knowledge, it is the first to enable full-degree-of-freedom (DOF) estimation without additional hardware. The key concept behind this method is a decoupling property, in which the direction of a line on a plane does not contain 3-DOF translation terms. Based on this, a cost function for rotation is constructed, and 3-DOF rotation parameters are estimated. With this rotation, the remaining 3-DOF translation parameters are calculated in a manner that minimizes the cost function for translation only. In other words, an original 6-DOF problem is decoupled into two 3-DOF estimation problems. Given these cost functions, degenerate cases are mathematically analyzed for known cases (incomplete), and the robustness is numerically tested for all possible cases (complete). The performance of the method is validated by extensive simulations and experimentations, and the estimated parameters from the proposed method demonstrate better accuracy than previous methods.

In indoor spaces, plane features are general geometric primitives that are robust to noise. To estimate the relative displacement and orientation between two consecutive 3D poses, at least three-pair correspondences of planes with different normals are needed at both poses. Otherwise, the estimated state falls into degeneracy, which is usually caused by the narrow field-of-view (FoV) of sensors as well as environmental structures. In the case where there are changes in degeneracy, with the exception of rotation, the state can be resolved using existing methods. To compensate for rotation in degeneracy for one-pair correspondences, the main approach is to project the initial prediction of an inertial measurement unit's (IMU) orientation to the normal of the corresponding plane such that it satisfies the shortest arc length. We also propose an algorithm that detects various degeneracy cases through which a seamless implementation becomes feasible, regardless of the number of plane pairs. We show the usefulness of the proposed method by performing a 3D mapping experiment where several degenerate cases arise due to limited FoV.

In this paper, a realtime pseudo-structure extraction algorithm for 3D indoor point cloud data (PCD) is proposed. This algorithm is called Convex Cut (CC) because of its two main steps: cutting the PCD with arbitrary planes, and extracting convex parts. CC can be used as a preprocessing module for other existing algorithms to extract static parts in dynamic environments or to represent a principal 3D model of a given PCD. Its calculation time is 24 milliseconds for 50k PCD on a consumer PC, and it yields a precision value of 0.90 and a recall value of 0.99 on average in highly dynamic and cluttered environments. Some possible applications are explained such as simultaneous localization and mapping in dynamic environments, efficient dense map representation, robust 3D scan matching with plane features, and natural motion planning.

This paper proposes a path planner for a humanoid robot to enhance its performance in terms of the human-robot interaction perspective. From the human point of view, the proposed method uses the time index that can generate a path that humans feel to be natural. In terms of the robot, the proposed method yields a waypoint-based path, the simplicity of which enables accurate tracking even for humanoid robots with complex dynamics. From an environmental perspective through which interactions occur, the proposed method can be easily expanded to a wide area. Overall, the proposed method can be described as a scalable path planner via waypoints with a time index for humanoid robots. Experiments have been conducted in test beds where the robot encounters unexpected exceptional situations. Throughout these trials, the robot successfully reached the goal location while iteratively replanning the path.

This paper deals with a path planning problem for turning motion of a car-like vehicle. We propose a turning method which finds a curvature continuous optimal path between two positions for a car-like vehicle.

In this paper, we propose a new design of a flexible enveloping grasper for pick and place tasks with the low complexity in manipulation and task planning for the purpose of practical use in the near future. Flexible material for the grasper has many advantageous characteristics inherently including robustness against manipulation errors and the ability to increase contact area with a grasped object and the grasping force. Compliance of the grasper material also contributes to reduction in complexity of the processes such as the force control, sensor-motor coordination, and manipulation by self-adaptation. Two properties, flexibility and compliance, mentioned above help the proposed grasper minimize the internal forces in a passive manner and achieve the successful force distribution with self-adaptivity when performing enveloping grasping. In order to demonstrate our work, we have constructed 2 different prototypes of flexible enveloping grasper. Experimental results validate robust performances of the proposed grasper.

In this paper, we propose an algorithmic compass that yields the heading information of a mobile robot using the vanishing point in indoor environments: VPass. With the VPass, a loop-closing effect (which is a significant reduction of errors by revisiting a known place through a loop) can be achieved even for a loop-less environment. From the implementation point of view, the VPass is useful because it can be appended upon any existing navigation algorithms. Experimental results show that the VPass yields accurate angle information in indoor environments for paths with lengths of around 200 m.

This paper presents a practical algorithm for topological navigation in corridor environment using cheap sonar sensors. Topological navigation consists of topological mapping and obstacle avoidance navigation. In this environment, it has two major problems which are 1) identifying nodes under globally similar but locally slightly different shape of corridor environment for topological mapping and 2) obstacle avoidance navigation using cheap sensors. In order to detect nodes robustly, we define the eigenvalue ratio (EVR) which converts geometrical shape of the environment to a quantitative value based on the principal component analysis (PCA). This method is demonstrated by simulation and experiments to be robust, with the notable detail that at a given corridor intersection the method can detect a cluster of nodes defined in this manner. For safe obstacle avoidance navigation, we introduce circle following (CF) algorithm which uses geometry of environment and characteristics of sonar sensors. Experiments and simulation validate reliability of the CF algorithm

Edges in topological maps have rich information such as shape, numbers of obstacles, locations of obstacles which can be used for a data association. However, no systematic approach on using the edge data for the data association has been reported. This paper proposes a systematic way of utilizing the edge data for data association. First, we explain a Local Generalized Voronoi Angle(LGA) to represent the edge data in 1-dimension. Second, we suggest a key factor extraction procedure from the LGA to reduce the number by 27-28 times for computational efficiency using the wavelet transformation. Finally we propose a way of data association using the key factors of the LGA. Simulations and experiments show that the proposed data association algorithm yields higher probability for similar edges in computationally efficient manner.

A localization algorithm for topological maps with dynamics is proposed in this paper. Especially, this algorithm considers a localization when some nodes are deleted by door closing and has two main features. First, lots of edge information are used in efficient way. Second, the algorithm calculates probability of a node to be the current location in a systematic and general way. Experiments in a topological map with dynamics are conducted to show the performance of the proposed algorithm.

All mobile robots suffer from odometry error. Relative localization from odometry has both the systematic and the non-systematic errors. However, once a precise system error model and its parameters are given, the accuracy of odometry can be remarkably improved. Most previous works on this effort focused on the differential drive robots with little attention to the other types of mobile bases. In this paper, we analyze sources of odometry error and propose an error model for the synchro drive robot. We then describe a novel procedure to accurately estimate the error parameters of the derived error model and the covariance matrix of the synchro drive robot. However, this procedure is general for all mobile bases, so we also apply our method for the differential drive robots and show experiments. This new process uses the shape of the path, as opposed to just end points, to estimate the error parameters and covariance matrix. We happen to use the generalized Voronoi graph to generate this path. Experimental results validate the error model of the synchro drive robot and precise estimation ability of the proposed method for the synchro and the differential drive robots.

GVG has been effectively used as a sensor based navigation tool using 360/spl deg/ sensor data. For mobile guide robot applications, however, we can only use 180/spl deg/ sensor data and the robustness of the navigation algorithm is critical for successful applications. For that purpose, the robot should be equipped with three capabilities. Those are 1) incremental GVG construction, 2) robust GVG navigation and 3) navigation strategy that just uses half of the sensor scan, i.e. 180/spl deg/. In this paper, we propose a GVG navigation algorithm that has above 3 capabilities. We firstly propose a method that can estimate the invisible 180/spl deg/ range from previous range data. Moreover, we suggest a way of robust GVG navigation algorithm by using a sensor data matching technique. The simulation result validates that the proposed algorithm can incrementally and robustly navigate the semi-unstructured map by using 180/spl deg/ sensor scan.

In this paper, an algorithm for hyper symmetric environment simultaneous localization and mapping (SLAM) is developed based on the generalized Voronoi graph. The hyper symmetric environment is a challenging environment for the SLAM and the most difficult problem on this is a data association. To redeem this problem, we propose three techniques. The first one is a breadth first node navigation algorithm which highly reduces the complexity of the data association. The second scheme is an odometry calibration method which converts untrustable odometry into reliable one. The third method is a multi layered data association scheme. We integrate above three techniques into one framework and propose a SLAM algorithm for the hyper symmetric environment. The simulation result shows that the proposed algorithm can successfully cover a large hyper symmetric environment under 20% odometry uncertainty.

Present the development of a human-mobile robot communication system using electrooculogram (EOG) signals. An ideal velocity shape signal processing algorithm is proposed to extract position data where the eyes are focusing on from the noise and drift included in EOG signals. Additionally, an efficient algorithm for the detection of various eye-lip movements such as blink and wink is suggested. Two experiments were performed for the validation of the human-mobile communication system. One is point stabilization of a mobile robot, using extracted eye focusing position data. The other is a moving target following experiment using various eye-lip movements as mobile robot commands.

In this paper, we consider hard contact transition such as hammering when the manipulator's approaching velocity is high or the environment is very stiff. In contact transition, phase transition can be divided into two definitely different phases, "pre-transition phase" and "transition phase". Here we define the "pre-transition phase" and we propose three control methods. First, we propose a novel controller named as "suppression controller" which is not only stable but also simple to implement. Second, we present passive damper named as "flexible-damped joint" which is a good solution to circumvent the pre-transition phase. Third, we suggest a stable and simple controller which can maximize joint damping and minimize recontact velocity in the flexible-damped joint. It is named as "joint damping controller". The effectiveness of these three approaches are verified via experiments using the POSTECH 1-DOF DD arm.