This paper proposes a novel algorithm to control wheels of a rover in rough terrains with keeping traction and minimizing energy consumption on the basis of observing a change of a robot velocity from a 4WD Kinematic model. Proposed kinematic-based rough terrain control (KRTC) can satisfy a rover to meet an integrated performance of following a desired velocity while keeping maximum traction and minimizing energy use without an estimator of wheel sinkage and an observer of a dynamic parameter. KRTC consists of two main parts; a slip optimizer and a slip controller. For slip optimization, optimal slip values regarding maximum traction and tractive efficiency are derived from indoor experimental data during a wheel-terrain interaction. Optimized slip ratio is used as an input of a PID slip controller. KRTC only uses the desired velocity of the robot as an input value and gets the two feedback data of actual slip and actual velocity of each wheel. For verification, KRTC was applied to a 4WD rover and the performance was analyzed based on acquired experimental data on a variety of rough terrains.

In the paper, designing process of a walking and swimming legged underwater robot is proposed. The robot has been inspired by diving beetles. To maximize swimming ability of the robot, the structural analysis of the diving beetle's leg has been conducted. The final model of the robot designed by AutoCAD(2D and 3D computer aided design) is presented based on employing structural advantages which diving beetles have. Additionally, underwater experiments with the designed robot leg has been implemented. Bio-mimicking swimming locomotion was applied to the robot leg during the experiments.

On rough terrain, there are a variety of soil types having different soil strength. It means that it is needed for outdoor robots to change wheel control strategies since optimal slip and maximum traction levels on wheels differ depending on soil strength. Therefore this paper proposes an algorithm for acquiring optimal control parameters, such as maximum traction coefficient and optimal slip ratio to maximize traction or minimize energy consumption, based on estimating strength of soils. In this paper the optimal models of wheel traction and slip are derived through indoor experiments by a testbed for analysis of wheel-terrain interactions on three types of soil; grass, gravel and sand. For estimating soil strength, actual traction coefficient, including information of motion resistance, is observed by a state estimator related to wheeled robot dynamics. The actual traction coefficient and slip ratio on wheels are employed to estimate soil strength by a numerical method on the basis of derived optimal models. The proposed algorithm was verified through real driving experiments of a wheeled robot on various types of soil.

This paper mainly focused on proposing a multi-functional leg for a six-legged underwater robot inspired by diving beetles(D.BeeBot). In order to make the robot have abilities both walking and swimming, one of legs is designed in advance by adopting structural advantages that the diving beetle has for efficient swimming. Trajectories of the leg employ various biomimetic motions produced by SPG(swimming pattern generator), setting simple control parameters. Performances of the leg are verified by measuring drag force by a force/torque sensor through underwater experiments. In the paper, swimming mechanism is mainly concerned for this time. Consequently, we proved the validity of designing consideration for the multi-functional leg.

On rough terrain, there are a variety of soil types having different soil strength. It means that it is needed to change wheel control strategies since optimal wheel slip levels differ depending on soil strength. Therefore, this paper proposes a novel algorithm for optimal slip control of wheeled robots with a trade-off between traction and energy consumption on the basis of observing a change of a robot velocity on a variety of soil types. Proposed intelligent slip-optimization control algorithm is largely divided into three main parts; (1) slip optimization (2) slip controller (3) intelligent disturbance compensator. For slip optimization, an optimal slip model is derived as a function of slip ratios regarding maximum traction and tractive efficiency from indoor experimental data during a wheel-terrain interaction. Optimized slip ratio is used as an input of a PID slip controller with compensation of disturbance which occurs by a change of surface shapes. This algorithm is verified by controlling a real four wheeled robot on rough terrains.

This paper presents an analysis of inverse kinematics solutions for YH050 which is an industrial robot manipulator developed by Hyundai Heavy Industries Co. We propose three methods: numerical approach, geometrical approach and combined approach. These approaches are addressed through theoretical explanation and experimental verification.

We propose a new cooperative localization method based on CP (constraint propagation) technique for multiple mobile robots sharing position information with each others. In case of conventional probabilistic methods such as EKF, they give the most probable estimation rather than guarantee any deterministic values or boundary for true position of the robots. To comply with such cases as at least boundary of robot position should be guaranteed, we propose a CP-based method that can handle the cases and gives a boundary containing true robot position value. To be specific, we apply the constraint propagation technique to localizing a set of mobile robots equipped with wheel encoders, gyro meters, and GPS receivers. Fusing the data and applying CP, we can get improved accuracy for localizing the robots. Also, by introducing the correlation between GPS data in common workspace to the sensor fusion framework, we further improve the localization accuracy for multiple robot cases. Several simulation results showing that the proposed method improves localization performance considerably compared to conventional methods will be shown.

In order to utilize hydrodynamic drag force on articulated robot moving in an underwater environment, a repetitive motion planning method subject to drag optimization is proposed. Repetitive trajectories, a preplanned periodic motion, are obtained from general polynomial trajectory planning method by connecting the end point of trajectory to start point by enforcing continuous motion. After drag torque is described in terms of joint angles and velocities, directional drag force in task space is derived, which is used as objective function to be optimized during a time period. The genetic algorithm (GA) is adopted to find optimal parameters composing the coefficients of polynomials. To verify the correctness of proposed method, periodic motion planning examples of simple two-link planner robot are presented in this paper.

In this paper, dynamic constraints are considered for the analysis of manipulability of robotic systems comprised of multiple cooperating arms. Given bounds on the torques of joint actuators for each robot, the purpose of this study is to derive the bounds of task acceleration of object carried by the system. Bounds on each joint torque form a polytope in joint-torque space and the task space acceleration is obtained from this polytope through dynamics and kinematics of both object and robots. A series of mathematical processes are proposed to get the reachable acceleration bounds from given robot actuator torque constrains, which includes the process for finding minimum infinite norm solutions. Several examples including two robot systems as well as PUMA560 robot systems are shown with the assumptions of complete-constraint contact model (or 'firm grip').

In this paper, the general problem of mapping an n-dimensional polytope Q to an m-dimensional polytope P by P=AQ with n>m is studied. It is well-known that some vertices of Q are mapped to internal points of P, and it is very time consuming to detect them when their number is large. A computationally efficient recursive algorithm based on dimension-growing has been developed which is capable of successively removing the internal points of each intermediate polytope when they are first detected, hence eliminating their further growth into additional internal points in subsequently intermediate polytopes. The method can be applied to compute the robotic manipulability polytope based on the mapping x/spl dot/=Jq/spl dot/. It is shown that substantial reduction in computation time can be achieved for higher DOF redundant robot manipulators. Examples including the 8-DOF ARMII robot are presented and comparisons to the conventional method are made.

A low cost motion capture system for the human body is handled. This system is based on low cost accelerometers. By extracting the gravity-related term from sensor data, we relate the sensor data with body configuration in real-time. One sensor unit is composed of 3 accelerometers and is capable of identifying 2 rotating angles of joints with 2 degrees of freedom. A geometric fusion technique is applied to cope with the uncertainty of sensor data. A practical calibration technique is also proposed to handle the error in aligning the sensing axis with coordination axis. In the case where motion acceleration is not negligible compared with the gravity acceleration, a compensation technique to extract gravity acceleration from sensor data is proposed. Experimental results of human motion capturing are also described with graphics.

For redundant robot manipulators, linear motion and angular motion are handled separately in solving inverse kinematics and trajectory planning. We apply the concept of decomposed specified manipulability to solve the inverse kinematic equation. With given joint velocity limits and given task velocity, we solve the inverse kinematic equation so that the linear velocity of the solution has maximum magnitude while the angular velocity of the solution is kept at a predefined value. For a trajectory where only the geometric path of the end-effector position and the angular velocity during travelling are specified, we derive a joint trajectory that makes the robot reach the goal position with shorter travelling time than conventional methods.

Regarding the dexterousness measure of robot manipulators, two geometric tools, manipulability ellipsoids and manipulability polytopes, are examined and compared with each other. Even though the manipulability ellipsoid approach is the most widely used techniques, it is shown that the manipulability ellipsoid does not transform the exact joint velocity constraints into task space and so may fail to give exact dexterousness measure and optimal direction of motion in task space. After showing that the polytope approach can handle such problems, we propose a practical polytope method which can be applied to general 6-dimensional task space. The relation between manipulability ellipsoids and manipulability polytopes are also explored for a redundant case as well as nonredundant case.

A practically applicable collision-free trajectory planning technique for two robot systems is proposed. The robot trajectories considered in this work are composed of many segments between which the robots stop to assemble, weld, or do other jobs by end-effector. The proposed method is based on the planning-coordination decomposition approach in which planning finds the trajectory of each robot independently according to their tasks and coordination finds velocity modification to avoid collision with each other. To fully utilize the convenience of independent trajectory planning and to ensure no geometrical path change after coordination, a simple technique avoiding collision by adding least delay-time in the local sense before moving along the next path-segment is developed. The least delay time is determined by a graphical method on the space called co-ordination space where collision between two robots and coordination of two robots are easily visualized.

A numerical trajectory planning method is proposed which ensures collision-free and near time-optimal motions for two robotic manipulators with limited actuator torques and velocities. Constraints on accelerations of the robots that are derived from the limited torques of actuators are transformed to the constraints on the curvatures of the coordination curve in coordination space. A dynamic programming technique is then provided on the discretized coordination space to derive a collision-free and time-optimal trajectory pair for the robots. A numerical example is included.>

A multilayer feedforward neural network and a design method that takes the distribution of given training patterns into consideration are proposed. The size of the network, initial values of interconnection weights, and parameters defining the nonlinearities of processing elements are determined from a specially selected portion of the given training patterns, which is called a set of feature points. With these initial settings, the performance of the network is further improved by a modified error backpropagation learning process. It is shown in several examples that the proposed model and the design method are capable of rapidly learning the training patterns compared to conventional multilayer feedforward neural networks with random initialization techniques.

A multilayer neural network and its initialization method, which takes the distribution of given training patterns into consideration, are proposed. The network is composed of four layers, and the role of each layer in the network is specialized by analyzing the internal information processing of the network. The size of the network, initial values of weights, and parameters defining the characteristics of the nonlinearities of processing units in hidden layer are determined from a selected portion of the given training patterns. With these initial conditions, the performance of the network is further improved by the general error backpropagation learning process. The proposed model and method give any desired mapping performance with smaller network size and faster learning speed than obtained with conventional multilayer neural networks and a random initialization technique. To show the usefulness of the proposed network, an illustrative mapping is realized by the proposed algorithm.

To improve limitations of fuzzy PI controller especially when applied to high order systems, we propose two types of fuzzy logic controllers that take out appropriate amounts of accumulated control input according to fuzzily described situations in addition to the incremental control input calculated by conventional fuzzy PI controllers. The structures of the proposed controller were motivated by the problems of fuzzy PI controllers that they generally give inevitable overshoot when one tries to reduce rise time of response especially when a system of order higher than one is under consideration. Since the undesirable characteristics of the fuzzy PI controller are caused by integrating operation of the controller, even though the integrator itself is introduced to to overcome steady state error in response, we propose two fuzzy controllers that fuzzily clear out integrated quantities according to situation. The first controller determines the fuzzy resetting rate by situations described fuzzily by error and error rate, and the second one by error and control input. The two structures both give reduced rise time as well as small overshoot.>

A trajectory planning method that ensures collision-free and time-optimal motions for two robotic manipulators with limited actuator torques and velocities is proposed. In terms of parameters defining the positions along the paths of two robots, a two-dimensional coordination space and a maximum velocity curve are constructed to detect collision regions and to plan a time-optimal velocity curve, respectively. An algorithm is then provided for the case of zero initial velocities and a single collision region. A numerical example is included.>

A trajectory planning method ensuring collision-free motion and minimum traveling time for two robots is proposed. The robot system under consideration is given specified geometric paths for two robots: then, by using traveled lengths along the paths, a two-dimensional coordination space is constructed for detecting collision. Also the robot dynamics is transformed as a function of the traveled lengths, and is described in the phase plane containing bounds on acceleration and velocity limits derived from a set of constraints on torques and maximally allowable velocities of actuators. Then the main algorithm is provided which simultaneously considers collision avoidance and time optimality on the coordination space and the phase plane, respectively. The proof for optimality of the proposed algorithm is given, and a simulation result is included to show the usefulness of the proposed method.>