As computer vision algorithms are developed on a continuous basis, the visual information from vision sensors has been widely used in the context of simultaneous localization and mapping (SLAM), called visual SLAM, which utilizes relative motion information between images. This research addresses a visual SLAM framework for online localization and mapping in an unstructured seabed environment that can be applied to a low-cost unmanned underwater vehicle equipped with a single monocular camera as a major measurement sensor. Typically, an image motion model with a predefined dimensionality can be corrupted by errors due to the violation of the model assumptions, which may lead to performance degradation of the visual SLAM estimation. To deal with the erroneous image motion model, this study employs a local bundle optimization (LBO) scheme when a closed loop is detected. The results of comparison between visual SLAM estimation with LBO and the other case are presented to validate the effectiveness of the proposed methodology.

Acoustic signal is crucial for the autonomous navigation of underwater vehicles. For this purpose, this paper presents a method of acoustic source localization. The proposed method is based on the probabilistic estimation of time delay of acoustic signals received by two hydrophones. Using Bayesian update process, the proposed method can provide reliable estimation of direction angle of the acoustic source. The acquired direction information is used to estimate the location of the acoustic source. By accumulating direction information from various vehicle locations, the acoustic source localization is achieved using extended Kalman filter. The proposed method can provide a reliable estimation of the direction and location of the acoustic source, even under for a noisy acoustic signal. Experimental results demonstrate the performance of the proposed acoustic source localization method in a real sea environment.

This paper compares methods for attitude estimation of a UUV(Unmanned Underwater Vehicle). Attitude estimation plays a key role in underwater navigation using DVL(Doppler Velocity Log). The paper proposes attitude estimation methods using EKF(Extended Kalman Filter), UKF(Unscented Kalman Filter), and CF(Complementary Filter). It derives methods using the measurements from MEMS-AHRS(Microelectromechanical Systems-Attitude Heading Reference System) and DVL. The methods are used for navigation in a test pool and their navigation performance is compared. The results suggest that even if there is no measurement relative to some absolute landmarks, DVL-only navigation can be useful for navigation in a limited time and range.

Fish generates thrust with a compliant fin which is known to increase the efficiency. In this paper, the performance of a robotic dolphin, the velocity and the stability, was improved using an optimal compliant caudal fin under certain oscillating frequency. Optimal compliance of the caudal fin exists that maximizes the thrust at a certain oscillating frequency. Four different compliant fins were used to find the optimal compliance of the caudal fin at a certain frequency using the half-pi phase delay condition. The swimming results show that the optimal compliant fin increases the velocity of the robotic fish. The compliance of the caudal fin was also shown to improve the stability of the robotic fish. A reactive motion at the head of the robotic dolphin causes fluctuation of the caudal fin. This phenomenon increases with the oscillating frequency. However, compliant fin reduced this fluctuation and increased the stability.

This paper proposes an underwater localization algorithm using probabilistic object recognition. It is organized as follows; 1) recognizing artificial objects using imaging sonar, and 2) localizing the recognized objects and the vehicle using EKF(Extended Kalman Filter) based SLAM. For this purpose, we develop artificial landmarks to be recognized even under the unstable sonar images induced by noise. Moreover, a probabilistic recognition framework is proposed. In this way, the distance and bearing of the recognized artificial landmarks are acquired to perform the localization of the underwater vehicle. Using the recognized objects, EKF-based SLAM is carried out and results in a path of the underwater vehicle and the location of landmarks. The proposed localization algorithm is verified by experiments in a basin.

By a SLAM (simultaneous localization and mapping) method, we get a map of an environment for autonomous navigation of a robot. In this case, we want to know how accurate the map is. Or we want to know which map is more accurate when different maps can be obtained by different SLAM methods. So, several methods for map comparison have been studied, but they have their own drawbacks. In this paper, we propose a new method which compares the accuracy or error of maps relatively and quantitatively. This method sets many corresponding points on both reference map and SLAM map, and computes the translational and rotational values of all corresponding points using least-squares solution. Analyzing the standard deviations of all translational and rotational values, we can know the error of two maps. This method can consider both local and global errors while other methods can deal with one of them, and this is verified by a series of simulations and real world experiments.

In this paper, a new mobile robot, so called a rollerbot, is presented, which has single body and rugby-ball shaped roller wheel. A rollerbot has single point contact on ground and low energy consumption in motion because of the reduced friction. By changing center of mass using a balancing weight, a rollerbot is able to get steering force. The vertical position of mass center of the rollerbot in this paper is designed to lie inside radius of the roller wheel, so that to have stable equilibrium position. Thus, the posture and the steering control of the rollerbot can be easily done by changing the center of mass. Kinematics of the rollerbot is derived by transformation of differential motion in this paper.

An adaptive robot hand (AR-Hand) has a stable grasp of different objects in unstructured environments. In this study, we propose an AR-Hand based on a tendon-driven mechanism which consists of 4 fingers and 12 DOFs. It weighs 0.5 kg and can grasp an object up to 1 kg. This hand based on the adaptive grasp mechanism is able to provide a stable grasp without a complex control algorithm or sensor system. The fingers are driven by simple tendon structures with each finger capable of adaptively grasping the objects. This paper presents a method to decide the joint stiffness. The adaptive grasping is verified by various grasping experiments involving objects with different shapes and sizes.

In this study, we have developed the humanoid joint modules which provide a variety of service while living with people in the future home life. The most important requirement is ensuring the safety for humans of the robot system for collaboration with people and providing physical service in dynamic changing environment. Therefore we should construct the mechanism and control system that each joint of the robot should response sensitively and rapidly to fulfill that. In this study, we have analyzed the characteristic of the joint which based on the target constituting the humanoid motion, developed the optimal actuator system which can be controlled based on each joint characteristic, and developed the control system which can control an multi-joint system at a high speed. In particular, in the design of the joint, we have defined back-drivability at the safety perspective and developed an actuator unit to maximize. Therefore we establish a foundation element technology for future commercialization of intelligent service robots.