Since it first appeared, differential evolution (DE), one of the most successful evolutionary algorithms, has been studied by many researchers. Theoretical and empirical studies of the parameters and strategies have been conducted, and numerous variants have been proposed. Opposition-based DE (ODE), one of such variants, combines DE with opposition-based learning (OBL) to obtain a high-quality solution with low-computational effort. In this paper, we propose a novel OBL using a beta distribution with partial dimensional change and selection switching and combine it with DE to enhance the convergence speed and searchability. Our proposed algorithm is tested on various test functions and compared with standard DE and other ODE variants. The results indicate that the proposed algorithm outperforms the comparison group, especially in terms of solution accuracy.

Optimization-based methods have been recently proposed to solve motion planning problems with complex constraints. Previous methods have used optimization methods that may converge to a local minimum. In this study, particle swarm optimization (PSO) is proposed for trajectory optimization. PSO is a population-based stochastic global optimization method inspired by group behaviors in wildlife, and has the advantages of simplicity and fast convergence. Trajectory modifications are encoded in particles that are optimized with PSO. The normalized step cost (NSC) concept is used for the initialization of the particles in PSO. A method for reusing previously optimized parameters is also developed. The optimized parameters are stored together with corresponding NSC vectors, and when a constraint violation occurs, the parameters associated with an NSC vector that is similar to the query NSC vector are selected. The selected vector is used for initializing the particles. The reuse of the previously optimized parameters improves the convergence of the PSO in motion planning. The effectiveness of these methods is shown with simulations and an experiment using a three-dimensional problem with constraints. The proposed algorithm successfully optimized a trajectory while satisfying the constraints and is less likely to converge to a local minimum.

Sound localization has been utilized in many fields. For good performance, sound signal obtained from the microphone has to be well filtered and amplified. Although filtering and amplification can be implemented in digital, analog circuit is needed because of sampling aliasing problem and limitation of computational resources in real-time application. The analog circuit consists of several components. Because all of possible combinations cannot be investigated, evolutionary algorithm (EA) is chosen to find promising solution candidates. In the paper, we built a multi-objective covariance adaptation evolutionary strategy based on decomposition to obtain the component values. Through the experimental results, our proposed algorithm showed good performance.

An evolutionary computational approach for a gait generation of a quadruped robot autonomously generates a gait that adapts in an environment. In this approach, a fitness function that measures a performance of the gait is defined and parameters are optimized by maximizing or minimizing the function with evolutionary computation algorithms. However the previous research only has considered the optimization on an environment. In this paper, we suggest a gait adaptation method for a quadruped robot using a terrain classification and a gait optimization for an adaptation on various surfaces. The surfaces for the adaptation are learnt with a classification algorithm and a gait parameter on each surface is optimized with Particle Swarm Optimization (PSO). After the learning and the optimization, the classifier is used for classifying a surface that a robot is located and an optimized gait parameter is selected based on the classification result for the adaptation. The adaptation framework, a feature design and a filtering method for a classifier and a gait design for a quadruped robot are proposed in this paper. The proposed method was verified in a realistic 3D simulator and it successfully classified surfaces and selected optimized gaits for adaptations.

In this paper, a novel actor-critic method using an incrementally constructed radial basis function network is developed to deal with continuous state and action problems. There exists one local model for each basis function, and the number of local models is increased as the basis function network grows. The normalized weighted sum of their outputs is used to estimate the value function for the critic, and the models are updated with the local temporal difference error in the receptive field of the corresponding basis function. A Gaussian policy is used for continuous action, and it is parameterized by the mean and the standard deviation. The parameters are determined by the normalized weighed sum of the corresponding sub-parameters assigned to the basis functions, and the regular policy gradient method is used for their update process. A new error is introduced for the online shape adaptation of the basis functions. Reducing this error prevents some of the basis functions from dominating the value function approximation and the policy, and improves the performance when the incrementally constructed basis function network is used. Simulation results for three benchmark problems show the performance and effectiveness of the proposed method in comparison to other methods.

Various meta-heuristic search methods have been employed to resolve the sensor arrangement problem, which is a type of NP-hard, combinational problem. However, the difficulty of weight tuning, when formulating a single objective function, is the chief obstacle to the use of the single-objective optimization methods. Although multiobjective optimization methods have been applied recently to avoid the difficulty involved in weight design, the original multiobjective optimization method still requires a greater number of generations for the solutions to converge to the optimal Pareto front. Moreover, unlike in previous works, we deal with four unknowns to define the sensor arrangement problem more practically: 1) The number of sensors is unknown, 2) no candidate is given for installation, 3) the coverage radii of sensors are variable, and 4) sensors cover a wide area in which obstacles exist in complicated arrangements. To improve the search approach for a sensor arrangement with these requirements, we first propose a representation scheme to encode the sensor arrangement problem as a set of chromosomes. Genetic operators and a repair scheme are also properly employed in the proposed encoding method. In addition, two strategies, i.e., the hierarchical fitness assignment strategy and the hybrid optimization strategy, are proposed to improve convergence. We also perform experiments with two commercial sensors to verify the proposed multiobjective optimization approach for sensor arrangement (MOASA). The results show that the proposed MOASA gives better performance than conventional search methods. The effects of the proposed strategies are investigated with additional experiments in terms of the quality of Pareto solutions.

The Efficient-Energy Coverage (EEC) problem is an important issue when implementing Wireless Sensor Networks (WSNs) because of the need to limit energy use. In this paper, we propose a new approach to solving the EEC problem using a novel Ant Colony Optimization (ACO) algorithm. The proposed ACO algorithm has a unique characteristic that conventional ACO algorithms do not have. The proposed ACO algorithm (Three Pheromones ACO, TPACO) uses three types of pheromones to find the solution efficiently, whereas conventional ACO algorithms use only one type of pheromone. One of the three pheromones is the local pheromone, which helps an ant organize its coverage set with fewer sensors. The other two pheromones are global pheromones, one of which is used to optimize the number of required active sensors per Point of Interest (PoI), and the other is used to form a sensor set that has as many sensors as an ant has selected the number of active sensors by using the former pheromone. The TPACO algorithm has another advantage in that the two user parameters of ACO algorithms are not used. We also introduce some techniques that lead to a more realistic approach to solving the EEC problem. The first technique is to utilize the probabilistic sensor detection model. The second method is to use different kinds of sensors, i.e., heterogeneous sensors in continuous space, not a grid-based discrete space. Simulation results show the effectiveness of our algorithm over other algorithms, in terms of the whole network lifetime.

This paper presents the sensor deployment method to design a RFID sensor network for the mobile robot localization using evolutionary approach. For this purpose, we employ the differential evolution (DE), which is well-known for promising performance. We propose two variation methods, the direct optimization strategy for the maximum usage of initial information intuitively and the full coverage optimization strategy for the dense coverage for the surveillance and the security. In that case, the proper tuning of parameters of DE is essential. We experiment sensor deployment in two maps for providing guidance about parameter tuning. The experimental results show better sensor deployment result according to guided parameter setting. The full coverage optimization strategy also shows proper result using guided parameters from the standard DE case.

The safety of humans working with robots is an important issue. Many studies have addressed related methods, but fundamental limits to meet required safety have been met owing to the absence of compliance in the robot actuators. Pneumatic muscle is considered to be a basic actuator and offers the advantage of intrinsic elasticity to achieve the joint compliance. Here, the joint compliance actuated by pneumatic muscle is actively utilized to enhance human safety during collisions. To this end, the authors present a novel approach to control compliance and position independently without affecting on the each other's performance using pneumatic muscles. The presented method is verified by experiments using a physical robot.

Localization of indoor environment is a fundamental issue for mobile robot. In this paper, we proposed the localization scheme to fusion the RFID localization system, ultrasonic sensor and wheel encoder. The uncertainty is factor caused by each localization system, and the estimation error is affected by this uncertainty. The sensor system for mobile robot localization is technical limitation such as sensing range, operation feature. Therefore we focus on sensor fusion scheme. When the mobile robot moves, certain data combination set to fuse is selected according to the environmental factor. The performance and simplicity of the approach is demonstrated with the result produced by experiments using mobile robot.

An adaptive control using multiple incremental fuzzy neural networks (FNNs) is proposed for robot manipulators. The structure and parameters of the FNNs are determined dynamically by using an incremental FNN. By incorporating incremental learning and adaptive control with multiple models, the proposed method not only reduces complexity and computation induced by the use of multiple models, but also provides favorable transient and tracking performance. The multiple FNNs are switched or blended to improve the transient response when manipulating objects are changed. The parameters are refined adaptively to compensate for system uncertainties. The resulting closed-loop system with a switching or blending law is proven to be asymptotically stable. The proposed scheme is applied to control a two-link robot manipulator in conjunction with varying payloads.

This paper proposes a robust method to estimate sound direction for mobile robot interface. In contrast to other methods, the proposed method estimates sound direction at intervals of short time to deal with deterioration by movement of sound source or robot. The two methods, discrete Kalman filter and time delay of arrival (TDOA) using generalized cross-correlation (GCC), are combined by quantifying reliability of TDOA. The combination prevents the deterioration by low power of signals, noise, and reverberation because discrete Kalman filter efficiently filters out such effects. The result of experiments, in which mobile robot estimates azimuth angle with three microphones, shows that the proposed method provides robustness on estimating sound direction even if the sound source or robot is moving during estimation procedure.

The variable projection (VP) method for separable nonlinear least squares (SNLLS) is presented and incorporated into the Levenberg-Marquardt optimization algorithm for training two-layered feedforward neural networks. It is shown that the Jacobian of variable projected networks can be computed by simple modification of the backpropagation algorithm. The suggested algorithm is efficient compared to conventional techniques such as conventional Levenberg-Marquardt algorithm (LMA), hybrid gradient algorithm (HGA), and extreme learning machine (ELM).

A new result of global synchronization for more general delay complex dynamical networks with time-varying delay is obtained in this paper. In this paper, based on an appropriate Lyapunov function, the sufficient condition for the global synchronization is given here by using the linear matrix inequality and introducing a reference state, and this condition is simply given based on the maximum nonzero eigenvalue of the network coupling matrix. Moreover, we give how to construct the coupling matrix to guarantee global synchronization of network, which is very convenient to use. A two-dimension system with time-varying delay as a dynamical node in network with global coupling is finally given, which is used to verify the theoretical results of the proposed global synchronization scheme.

Pneumatic muscle has many advantages such as elasticity, high power and structural similarity to a living thing's muscle. There has been many researches to control robot actuated by pneumatic muscles, but conventional theories are hard to apply on real robot plants because of their assumptions and disregards of pneumatic muscle's physical aspects like size of pneumatic muscle and its controller. Here, the new method for saving space which is occupied by many controllers to operate robot actuated by pneumatic muscles is proposed. Actually there is easy way to control pneumatic muscle using the commercial proportional pressure regulator, but its size is not suitable to be embedded on stand alone robot. So, new method using the pressure switches of compact size and encoders is suggested. This new method is tested on a robot link with ball joints, actuated by four pneumatic muscles

Developing Takagi-Sugeno fuzzy models by evolutionary algorithms mainly requires three factors: an encoding scheme, an evaluation method, and appropriate evolutionary operations. At the same time, these three factors should be designed so that they can consider three important aspects of fuzzy modeling: modeling accuracy, compactness, and interpretability. This paper proposes a new evolutionary algorithm that fulfills such requirements and solves fuzzy modeling problems. Two major ideas proposed in this paper lie in a new encoding scheme and a new fitness function, respectively. The proposed encoding scheme consists of three chromosomes, one of which uses unique chained possibilistic representation of rule structure. The proposed encoding scheme can achieve simultaneous optimization of parameters of antecedent membership functions and rule structures with the new fitness function developed in this paper. The proposed fitness function consists of five functions that consider three evaluation criteria in fuzzy modeling problems. The proposed fitness function guides evolutionary search direction so that the proposed algorithm can find more accurate compact fuzzy models with interpretable antecedent membership functions. Several evolutionary operators that are appropriate for the proposed encoding scheme are carefully designed. Simulation results on three modeling problems show that the proposed encoding scheme and the proposed fitness functions are effective in finding accurate, compact, and interpretable Takagi-Sugeno fuzzy models. From the simulation results, it is shown that the proposed algorithm can successfully find fuzzy models that approximate the given unknown function accurately with a compact number of fuzzy rules and membership functions. At the same time, the fuzzy models use interpretable antecedent membership functions, which are helpful in understanding the underlying behavior of the obtained fuzzy models.

This paper deals with the emotional model of the software-robot. The software-robot requires several capabilities such as sensing, perceiving, acting, communicating, surviving and so on. There are already many studies about the emotional model like KISMET, AIBO. Though many emotional models are implemented, their human response architectures are invariant as time passes. Conventional emotional models make a robot like and obey a human during robot operation. This is natural for robots which are used in industrial, service and military areas but for pet robots this property lets robot look like real robots not real pets. Actually, a robot's emotional model is studied to implement robot intelligence more easily but emotion expression has to be exaggerated in a pet robot to make it interesting to humans. Pet robots must show dynamical emotion variations. The emotional model with the modified friendship is studied in this paper to overcome the conventional emotional model's limits to apply to pet robots. Friendship is distinguished as positive, negative and static schemes and an emotional model with the modified friendship is implemented and simulated on a software-robot.

A neural network based controller is derived for a mobile manipulator to track the given trajectories in the workspace. The dynamics of the mobile manipulator is assumed to be unknown completely, and is learned on-line by the radial basis function network (RBFN) with weight adaptation rule derived from the Lyapunov function. Generally, a RBFN can be used to properly approximate a nonlinear function. However, there remains some approximation error inevitably in real application. An additional control input to suppress this kind of error source is also used. The proposed algorithm does not need a priori knowledge about the exact system dynamic parameters. Simulation results for a two-link manipulator on a differential-drive mobile platform are presented to show the effectiveness for uncertain system.

This paper presents a mobile gait training system with body weight support (BWS), which relieves the patient of body weight of a prescribed amount by supporting the subject with a harness. The system is based on electrical motors to achieve system mobility according to the subject's motion of walking while it maintains a constant force on the body based on pneumatic actuators. The mobility of the system seems to be important for a correct gait pattern since the subject is able to walk on the ground actually. The system has' a simple architecture to construct, and is practical to the gait rehabilitation.

This paper discusses the safety assurance, capability features, and user-friendliness of Internet-based telerobotics, in the sense that how these features could be incorporated in order to enhance teleoperator performance. A comprehensive set of safety features are incorporated to the interface to screen and accept only the safe commands from the remote operator while discarding risky commands. A laser pointer is used to help remote operator in perceiving self-location and navigation, whereas orientation control has been completely automated and synchronized to the position commands of the teleoperator. The effectiveness of these features has been demonstrated by playing telerobotic mini-golf.

Motion tracking and object segmentation are the most fundamental and critical problems in vision tasks such as motion analysis. An active contour model, snake, was developed as a useful segmenting and tracking tool for rigid or non-rigid objects. Snake is designed on the basis of snake energies. Segmenting and tracking can be executed successfully by energy minimization. In this research, two new paradigms for segmentation and tracking are suggested. First, because the conventional method uses only intensity information, it is difficult to separate an object from its complex background. Therefore, a new energy and design schemes should be proposed for the better segmentation of objects. Second, conventional snake can be applied in situations where the change between images is small. If a fast moving object exists in successive images, conventional snake would not operate well because the moving object may have large differences in its position or shape, between successive images. Snake's nodes may also fall into the local minima in their motion to the new positions of the target object in the succeeding image. For robust tracking, the condensation algorithm was adopted to control the parameters of the proposed snake model called "adaptive color snake model (ACSM)". The effectiveness of the ACSM is verified by appropriate simulations and experiments.

The generation of the optimal walking pattern is an important question for a biped robot to keep walking stably. This paper is proposed for generating the walking patterns resulted in best performance of the biped robot using multiobjective evolutionary algorithm. We formulate a trajectory generation problem as a multi-objective optimal problem. We obtain all Pareto-optimal solutions on the feasible solution region for various walking pattern generation of a biped robot in EA simulation.

Gear-position-decision (GPD) tactics strongly affect the performances of automatic transmissions (AT) and, therefore, the performance of the vehicle. Since the electronic control methods were introduced into ATs, many advanced techniques have been raised to make AT vehicles more human friendly and better in fuel economy and dynamic behaviors. As a type of emerging AT, the automated manual transmissions (AMT) are being researched and developed in all relevant technologies. In this paper, we proposed a driving knowledge-based GPD (KGPD) method for AMTs. The KGPD algorithm is composed of a driving environments and driver's intentions estimator, the shift schedules for each typical driving environment and driver's intention situations, and an inference logic to determine the most proper gear position for the present situation. The estimator identifies the driving environments and features of driver's intentions, which are divided into some typical patterns. Based on the identified results, the gear-position inference algorithm calculates the best gear position at the moment. In fact, the method just simulates the course of a driver's making gear-position decision when driving an automobile with manual transmission. The test results show that the AMT with the method gives less unnecessary shifting, conducts more proper gear positions, and behaves better in subjective assessment than that with the method that is directly based only on automotive state parameters.

A new adaptive multiple neural network controller (AMNNC) with a supervisory controller for a class of uncertain nonlinear dynamic systems was developed in this paper. The AMNNC is a kind of adaptive feedback linearizing controller where nonlinearity terms are approximated with multiple neural networks. The weighted sum of the multiple neural networks was used to approximate system nonlinearity for the given task. Each neural network represents the system dynamics for each task. For a job where some tasks are repeated but information on the load is not defined and unknown or varying, the proposed controller is effective because of its capability to memorize control skill for each task with each neural network. For a new task, most similar existing control skills may be used as a starting point of adaptation. With the help of a supervisory controller, the resulting closed-loop system is globally stable in the sense that all signals involved are uniformly bounded. Simulation results on a cartpole system for the changing mass of the pole were illustrated to show the effectiveness of the proposed control scheme for the comparison with the conventional adaptive neural network controller (ANNC).

A new adaptive controller based on multiple neural networks for uncertain robot manipulator system was developed in this paper. The proposed multiple neuro-adaptive controller (MNAC) switches to a memorized control skill or blends multiple skills by using visual information on the given job to improve transient response at the time of task variation like change of manipulating object. MNAC is a kind of adaptive feedback controller where system nonlinearity terms are approximated with multiple neural networks. The proposed controller is effective for a job where some tasks are repeated but information on the load cannot be scheduled before the operation. During learning phase, MNAC memorizes a control skill for each load with each neural network. For a new task, most similar existing control skill may be used as a starting point of adaptation, which improves the performance of learning. Lyapunov function based design of MNAC guarantees the stability of the closed-loop system to be independent of switching or blending law. Simulation results on two-link manipulator or changing mass of the given load were illustrated to show the effectiveness of the proposed control scheme by the comparison with the conventional neuro-adaptive controller (NAC).

This paper deals with handling the unknown factors, such as the external disturbance and the unknown dynamics, for a mobile robot control. We propose a RBF network based controller for compensate for them. The stability of the proposed controller is proven by using Lyapunov function. To show the effectiveness of the proposed controller, several simulation results are presented. Through the simulations, we show that the proposed controller can overcome the modelling uncertainty and the disturbances. Also the proposed RBF controller outperforms the previous works from the viewpoint of computation time, which is a crucial fact for a real-applications.

Previous studies of evolutionary algorithms (EA) focus on stationary environments where the evaluation function and constraints of the problem are fixed over time of evolution. This study introduces a new algorithm for optimization in nonstationary environments. In this study, we explore the use of a hierarchically distributed evolution model which enables us to utilize fast local search operators while maintaining the sufficient diversity to adapt the environmental changes. The results are compared with previous techniques with several test functions.

A power assisted gait rehabilitation robot system by which a patient who has difficulty in walking can exercise is proposed. Our developed system consists of a robotic manipulator which can adjust the height of the system according to the user and a mobile base with two-driving wheels. It is similar to the architecture of a mobile manipulator. The neural network based control algorithm for a mobile manipulator used to compensate for dynamic interactions, unmodeled dynamics, and disturbances by the user on the system, is derived and experimental results are shown.

In this paper, a new evolutionary scheme for designing fuzzy logic system is proposed. We have developed a new fitness function for optimizing membership functions of antecedent parts of FLS. The proposed fitness function considers not only mean squared error, but also calculates some penalty based on the overlapping degree of adjacent membership functions for the same input. Consequent parts of the FLS are calculated using pseudo-inverse method. To show the effectiveness of the proposed algorithm, we have applied the proposed algorithm to various problems: Iris classification problem, Mackay-Glass time-series prediction problem and vision-based mobile robot control problem. It has been proven that, without much modification, the proposed algorithm can be used in various problem fields efficiently.

An active contour model, Snake, as a useful segmenting and tracking tool for rigid or non-rigid (deformable) objects, was developed by Kass (1987). Snake was designed on the basis of Snake energies. Segmenting and tracking can be executed successfully by the process of energy minimization. Kass' Snake can be applied to the case of small changes between images because its solutions can be achieved on the basis of variational approach. If a somewhat fast moving object exists in successive images, Kass' Snake will operates not well because the moving object may have large differences in its position or form between successive images. Snake's nodes may fall into the local minima in their motion to the new positions of the target object in next image. When the motion is too large to apply image flow energy for tracking, a jump mode is proposed for solving the problem. The vector used to make Snake's nodes jump to a new location can be obtained by processing the image flow. The effectiveness of the proposed Snake is confirmed by simulations.

Parallel genetic algorithms are particularly easy to implement and promise substantial gains in performance. Its basic idea is to keep several sub-populations that are processed by genetic algorithms. Furthermore, a migration mechanism produces a chromosome exchange between sub-population. In this paper, a new selection method based on nonlinear fitness assignment is presented. The use of the proposed ranking selection permits higher local exploitation search, where the diversity of population is maintained by a parallel sub-population structure. Experimental results show the relation between the local-global search balance and probabilities of reaching the desired solutions using test functions and nonstationary route-planning problems.

In the design of a controller for mobile robot, there are only few results on the problem of integrating the nonholonomic kinematic controller and the dynamic controller for mobile robots. Also there are only few literatures on the robustness of the controller in the presence of uncertainties or external disturbances in the dynamical model of mobile robot. In this paper, a robust adaptive controller which can achieve velocity tracking while considering not only kinematic model but also dynamic model of mobile robot is proposed. The proposed controller can overcome model uncertainty or external disturbances by the robust adaptive technique. The stability of the dynamic system is shown through the Lyapunov method.

This paper presents a robust adaptive control scheme of an underactuated manipulator. The proposed control schemes are a joint position-to-position (PTP) control scheme and a Cartesian end-point control scheme tracking a continuous path in Cartesian space. The proposed joint and Cartesian controllers are verified by experimental results for an actual three-link SCARA robot manipulator with a passive joint. In this robot experiment, the third joint of this robot is passive.

This paper presents a fault detection strategy and a robust fault recovery control scheme for robot manipulators to overcome the actuator failures. The joint (or actuator) failure considered is the free-swinging joint failure that causes the loss of torque on a joint. The proposed fault-tolerant control framework includes a normal control with normal (non-failed) operation, a fault detection and a fault recovery control to achieve task completion. After the detection and identification of joint failures, the robot manipulator becomes an underactuated manipulator with failed actuators. Then the underactuated manipulator can be controlled by the presented robust controller. To show the feasibility of the proposed fault-tolerant control scheme, simulation results for a three-link planar robot arm with a failed joint are presented.

One of the most important issues in the teleoperation is to provide the sense of telepresence so as to conduct the task more reliably. In particular, teleoperated mobile robots need to have some kind of backup system when the operator is blind to the remote situation owing to failure of the vision system. In Hong S-G et al. (see Mechatronics), the idea of artificial force reflection was proposed to enhance the reliability of operation when the mobile robot travels over plain ground. We extend these results to help the teleoperator even when the robot climbs stairs. The extended artificial force reflection method has two modes: traveling on plain ground and climbing stairs. When traveling over plain ground, the force information is artificially generated using range data from the environment while generating the impulse force when climbing stairs. To verify the validity of our algorithm, we develop a simulator which consists of a joystick and visual display system. Through experiments using this system, we confirm the validity and effectiveness of our new artificial force reflection in teleoperated mobile robots.

An active contour model was developed as a useful segmenting and tracking tool for rigid or nonrigid objects. Snake, one of the active contour models (ACMs), is designed on the basis of Snake energies. Segmenting and tracking can be executed successfully by the process of energy minimization. The ability to contract is a important process for segmenting objects from images, but the contraction forces of Kass' Snake are dependent on the object's form. In this research, new contraction energy, independent of the object's form, is proposed for the segmentation. Kass' Snake can be applied to the case of small changes between two images because its solutions can be obtained from the basis of variational approach. Furthermore considering the case that the motion is too large to apply image flow energy to tracking. A jump mode was proposed for solving the problem.

Q-learning is a kind of reinforcement learning where the agent solves the given task based on rewards received from the environment. Most research done in the field of reinforcement learning has focused on discrete domains. But the environment with which the agent must interact is continuous. Thus a method that is able to make Q-learning applicable to the continuous problem domain is needed. In the paper, the basic fuzzy rule is extended so that it can incorporate Q-learning. The interpolation technique which is widely used in memory-based learning is adopted to represent the appropriate Q-value for current state and action pair. The resulting structure based on the fuzzy inference system has the capability of solving the continuous state and action problem in Q-learning and generating fuzzy rules via interacting with the environment without a priori knowledge about the environment. The effectiveness of the proposed structure is shown through simulation on cart-pole system.

In this paper, a robust fault-tolerant control scheme including fault detection, identification and accommodation for robot manipulators is presented to overcome the actuator failures. The joint fault considered in this paper is the free-swinging joint failure that causes the loss of torque on a joint. The presented fault-tolerant Cartesian control framework includes a normal control with normal (nonfailed) operation, a fault detection and a fault recovery control to achieve task completion and fault accommodation. After the detection and identification of joint failures, the robot manipulator with failed actuators can be considered as the underactuated robot manipulator. To show the feasibility of the proposed fault-tolerant control scheme, simulation results for a three-link planar robot arm are presented.

In this paper, a variable structure model following control system (VSMFCS) for robot manipulators is proposed. By using a function augmented sliding hyperplane, the proposed controller guarantees that the system is always in the sliding mode. Therefore, although the conventional VSMFCS methods have been able to ensure the asymptotic stability of the overall system, the proposed scheme can guarantee the globally exponential stability even though the controller does not use the model state (x/sub m/).

A feedback linearizing decoupling dynamic control scheme of nonholonomic underactuated manipulators with free-swinging passive joints is proposed in Cartesian space where the task is planned. The presented control scheme does not require any braking process of passive joints. The presented dynamic control scheme has so-called "dynamic singularities" in the controller. Therefore, the singularities must be avoided an the path planning so as to guarantee the availability of the control scheme. Singularity-free desired paths in Cartesian space avoiding the dynamic singularities are obtained through a computer simulation. Simulation results for a three-link planar manipulator with one passive joint are presented to show the feasibility of the proposed control scheme with the singularity-free desired trajectory.

A study on the control of robot manipulators with free-swinging passive joints overcoming actuator failures is performed. In the case that the passive joints have neither actuators nor brakes, a robust adaptive control scheme overcoming the uncertainties is proposed in Cartesian space where tasks are planned, for underactuated manipulators. The proposed control scheme does not need a priori knowledge of the accurate dynamic parameters and the exact uncertainty bounds. The availability of the proposed control scheme is guaranteed by the assumption that the joint configuration remains in the inside of nonsingular regions during the total control process. To overcome this dynamic singularity problem for a nominal decoupling matrix in the controller, a singularity-free desired Cartesian path-planning in the outside of singular regions is achieved via a computer simulation. To illustrate the feasibility and validity of the proposed control scheme, simulation results for a three-link planar robot manipulator with one passive joint are presented.

This paper deals with the problem of motion planning for a unicycle-like robot. A simple local planner for unicycle model, based on an approximation of the desired configuration generated by local holonomic planner that ignores motion constraints, is presented. To guarantee a collision avoidance, we propose an inequality constraint, based on the motion analysis with the constant control input and time interval. Consequently, we formulate the problem as a constrained optimization problem, and a feedback scheme based on local sensor information is established by simply solving this problem. Through simulations, we confirm the validity and effectiveness of our algorithm.

Modeled from human neurons, various types of artificial neurons are developed and applied to control algorithm. In this paper, the weights and structure of feedforward neural network controller are updated using new skill learning paradigm which consists of supervisory controller, chaotic neuron filter and associative memory. The pattern of system nonlinearity along the desired path is extracted while supervisory controller guarantees stability in the sense of the boundedness of tracking error. Next the pattern is divided into small segments and encoded to bipolar codes depending on the existence of critical points. Comparing the encoded pattern with pre-stored neural parameters and pattern pairs through associative memory, the most similar one is obtained. Also, chaotic neuron filter is used to add perturbation to neural parameters when the training of feedforward neural network is not successful with the pre-stored parameters. Finally the memory is updated with new successful parameters and pattern pairs. Simulation is performed for simple two-link robot in case of the slight modification of desired trajectory.

This paper describes a dynamical local path-planning algorithm of an autonomous mobile robot available for stationary obstacle avoidance using nonlinear friction. Dynamical path-planning algorithm is considered to accommodate the mobile robot to the dynamic situation of the path-planning nature. Together with the previous virtual force field method, the path of the mobile robot is a solution of a path-planning equation. Local minima problems in stationary environments are solved by introducing nonlinear friction into the chaotic neuron. Because of the nonlinear friction, the proposed path-planner reveals chaotic dynamics in some parameter regions. This new path-planner is feasible to guide, in real-time, the mobile robot to avoid stationary obstacles and to reach the goal. Computer simulations are presented to show the effectiveness of the proposed algorithm.

Using a neural network to identify a function in the dynamic equation brings about additional difficulties which are not generic in other function approximation problems. First, training samples can not be arbitrarily chosen due to hard nonlinearity, so are apt to be nonuniform over the region of interest. Second, the system may become unstable while attempting to obtain the samples. This paper deals with these problems in continuous-time systems and suggests an effective solution, which provides stability and uniform sampling by the virtue of a supervisory controller. The supervisory control algorithm can be applied to robot system dynamics. The algorithm can be applied to an n-th order robot system, a simulation result is given for a simple two link robot.

This paper proposes an adaptive robust control scheme for free-flying space robots. In order to simply and effectively control free-flying space robots subject to parameter uncertainty and disturbances to avoid the measurement of the base (space vehicle) acceleration, and to eliminate the assumption that the uncertainty bounds must be a priori known in the conventional robust control approach, we develop a continuous adaptive robust control scheme using the norm-bounded property of uncertainty based on the Lyapunov direct method. Simulation results are presented to show the feasibility and robustness of the proposed control scheme.

When a flexible arm is rotated by a motor about an axis through the arm's fixed end, transverse vibration may occur. The motor torque should be controlled in such a way that the motor rotates by a specified angle, while simultaneously stabilizing vibration of the flexible arm so that it is arrested at the end of rotation. In this paper, the authors propose deterministic and adaptive control laws for one-link flexible arm. Using the fact that matrix dD(/spl delta/)/dt-2C(/spl theta/, /spl theta//spl dot/, /spl delta/, /spl delta//spl dot/) is skew symmetric, controllers which have a simplified structure with less computational burden are proposed by using Lyapunov stability theory.

An inverse kinematics solution that utilizes the differential relationship between the joint space and Cartesian space of redundant manipulators is proposed. Until now, this solution could be obtained using the pseudo inverse of the Jacobian matrix. However the computation of the pseudo inverse is complex so that it is not easy to implement it on a digital computer for online motion planning of the robot. As an alternative to the pseudo inverse solution, an inverse kinematics solution using fuzzy logic is derived. We find the rough solution based on the gradient method, and then refine it by introducing the fuzzy logic through the extension principle. As a result, a simple and fast inverse kinematics solution is obtained. The simulation results verify the efficiency of the proposed solution.

Design of a sliding mode control algorithm with efficient online compensation for robot manipulator is studied for robust and accurate tracking of the desired trajectories. The sliding mode tracking control uses a disturbance observer to eliminate the chattering problem that exists in conventional sliding mode control under modeling error and payload uncertainty. The proposed sliding mode control is presented in two theorems. The bounded stability of the proposed control method is proved. The efficiency of the control algorithm has been demonstrated by simulations for position tracking control of a robot subject to parameter and payload uncertainties.>

The authors discuss adaptive control of a space robot system with an attitude-controlled base on which the robot is attached. An adaptive control scheme in joint space is proposed. Since most tasks are specified in inertia space, instead of joint space, the authors discuss the issues associated to adaptive control in inertia space and identify two potential problems, unavailability of the joint trajectory (since mapping from inertia space trajectory is dynamics-dependent and subject to uncertainty), and nonlinear parameterization in inertia space. For a planar system, the linear parameterization problem is investigated, the design procedure of the controller is illustrated, and the validity and effectiveness of the proposed control scheme are demonstrated.>