This paper proposes a new method for dynamic control of redundant manipulators via artificial potential field approach (APFA). The proposed method is based on the APFA with a combination of a time scale transformation and a time base generator which works as a time scale compressor. The dynamic behavior of the manipulator can be controlled by using the proposed method without any change of the form of the designed controller itself. The effectiveness of the proposed method is verified by computer simulations for redundant manipulators.

View full abstract

A shadow appears on a surface when rays from a source of light are cut off by an interposed opaque object. The object is always accompanied with the shadow. Both regions of the object and its shadow are often detected as one region in the image plane. As an estimation of shape (height) of the object is influenced by the region of the shadow, it is necessary to separate the two regions of the object and its shadow. This paper introduces a separation method of objects and their shadows in the image. The effectiveness of the proposed method was examined by processing still pictures.

View full abstract

One of the important issues in control of distributed autonomous systems (DAS) with dynamic movement is to obtain a decentralized control scheme by which each subsystem can self-organize its behavior to be suitable for the assigned tasks. Using the wheel-shaped multi-actuator system we proposed before, this paper describes a new GA-based co-evolving scheme which allows each subsystem to adapt its control law to the subsystems' common characteristics in real-time. Some of the difficulties in implementing a real-time GA search algorithm are overcome by distributing GA-related operations among the subsystems, making use of the system's symmetrical characteristics. The subsystem evaluates system performance from locally obtained information: -time interval between two consecutive signals notifying the system movement. This time-based evaluating technique enables recovery from badly-performing control laws. The proposed scheme is validated by simulation and experimental results which reveal that the subsystems successfully improve themselves by discovering, in real-time, control laws which performs better with respect to the given evaluating function.

View full abstract

Semiconductor ring lasers have many capabilities of realizing new functional devices. In this paper, we propose a novel optical inertial rotation sensor using a semiconductor ring laser. If a semiconductor ring laser operates as an optical inertial rotation sensor, a very small and simple optical gyroscope can be realized. To our knowledge this is the first demonstration of a semiconductor ring laser gyroscope (S-RLG). Experimental results are as follows. (i) The Sagnac frequency shift can be detected as a beat note by the terminal voltage change of the semiconductor ring laser without branching the circulating optical power. Therefore, the S-RLG system can be constructed very simply as compared with already proposed optical gyroscopes. (ii) The detected beat frequency between two counter propagating lasers in the S-RLG is directly proportional to the applied rotation rate. (iii) And furthermore, we present data demonstrating the injection locking phenomenon around low rotation rate.
These results verify that the proposed S-RLG operates as an optical inertial rotation sensor based on Sagnac effect.

View full abstract

The position information based disturbance observer estimates both a disturbance torque and a motor speed by using the information of a rotor position and a torque current. The speed control system using the position information based disturbance observer has already been realized and used in industrial robotics applications. It is expected to be useful and practical for several industrial motor drive applications. However, the conventional position information based disturbance observer has influence of overflow and underflow of encoder pulse counter directly. When the encoder pulse counter has an overflow or underflow phenomenon, the speed control system using the conventional observer has a wrong or oscillated speed response.
In order to overcome this problem, this paper proposes a new rotor position information based disturbance and speed observer having no influence of the overflow and underflow of encoder pulse counter. When the encoder pulse counter has an overflow or underflow phenomenon, the proposed observer well estimates a motor speed and a disturbance torque stably and quickly, and the speed control system using the proposed observer has a stable and robust speed response.

View full abstract

In this paper, a design method of simple adaptive control (SAC) scheme is proposed for discrete-time deterministic systems. Most of discrete SAC schemes require a direct control feedthrough or a parallel feedforward compensator in order that the controlled object may satisfy the ASPR(Almost Strictly Positive Real) condition. The existence of additive elements, however, makes the SAC schemes complicated. Whereas, according to the proposed control scheme whose adaptation rule is modified, it enables us to construct further simple adaptive control systems. The proposed discrete SAC scheme is experimentally evaluated on a 2×2 level plus water flow rate control system.

View full abstract

In this paper, a new control method for decentralized systems is proposed by using the concept of Nash equilibrium points of game theory. It is supposed that the system stated in this article is composed of many subsystems including their controllers, so each subsystem can be recognized as each player in game. From the above assumption it is pointed out that the Nash equilibrium points can be calculated by the commonly-used back propagation algorithm if the criterion function of each subsystem is given.
From tank-network simulations, it is shown that the controller obtained by the Nash equilibrium points can be used when the criterion function of each subsystem is independently given.

View full abstract

In this paper, an algorithm to recognize facial expressions considering the size difference between the individual faces is proposed. The size differences between the individual faces becomes a problem for extracting effective features. On the contrary, in the recognition of facial expressions, the differences between the individuals must be considered and the judgments should be achieved considering each specific area. From these points of view, we propose a facial expression recognition system considering the size differences between the individual faces. Computer simulation using 59 expressive facial images shows more than 86.7% accuracy. From the experimental results, the prospects of using this method are very encouraging.

View full abstract

It is predicted that the demands of robots in home increase in the future. The behavior of the robots must be changed by the structure of houses or the arrangement of furnitures there. Therefore, the robots will face with the unknown environments necessarily. In these situation, the cooperation tasks by plural robots is necessary when we consider the various kind of works and its efficiencies. In this paper, we consider about a cooperation task of robots in unknown environment. We take the problem of unknown environment exploration by mobile robots as an example here. The aim of this problem lets mobile robots explore unknown environment and make the map. There are two important points, the efficiency of work and independence on environment, to raise here. We propose algorithm for path planning and collision avoidance between mobile robots in order to form an action pattern to meet these conditions. In case for forming an action pattern using these algorithms, we use Genetic Programming (GP). Finary, we let mobile robots explore 5 unknown environments in the simulation with an action pattern established in this way.

View full abstract

The Genetic Algorithm (GA) is well known as an effective technique to solve the optimization problem recently. But, only one solution is obtained through one searching by the GA. In contrast with it, the Immune Algorithm (IA) has the mechanisms of the effective searching for a set of solutions. So, plural optimal solutions of the problem can be obtained in one lot. But the IA has many parameters of the algorithm and it is difficult to decide these parameters. Then we propose a new method as that the IA parameters are decided dynamically by using the GA in this paper. The proposed method was applied to the problem of solving simultaneous non-linear equations and then good performance of it was shown.

View full abstract

In order to improve the response property of the closed-loop control system constructed by generalized minimum variance control (GMVC), in general, the control-weighting of cost function is given a positive value. But this may cause the offset in the output of the system. Also, the offset may occur because of input disturbance. To eliminate these offsets, a pre-compensator is added to the conventional controller. But in this method, it is recognized that the derived control law cannot optimize a cost function completely. In this paper, we propose a new control law optimizing the cost function completely.

View full abstract

A set of programs was developed for analyzing transition of EEG (Electro-encephalo-gram) active sources during visual stimulus-finger reaction process. First, positions of the active sources were estimated by applying the Least Square Method to each mapping data of 16 channels EEG. Transitions of the sources (chains) were discriminated according to a criterion of norm between pairs of sources in sequent EEG maps. Important chains (longer than 100ms) were displayed as an animation, while features of these chains were coded as EEG chart compactly in one or two sheet of papers. This chart was useful to express the transition of EEG activity in usual language.

View full abstract

In the measurement of actual random phenomenon, the observed data often contain the fuzziness due to the existence of confidence interval in measuring instruments, permissible error in experimental data and a quantized error in digitized observation.
In this study, by focussing on the complicated sound environmental system which cannot be identified from a standard classical approach, a system identification method based on the above fuzzy observation is theoretically proposed. The effectiveness of the proposed method is experimentally confirmed through an evaluation for the output response probability in an actual acoustic environment.

View full abstract

Although simulated annealing is powerful to obtain good approximation solutions for multi-modal function optimization, it requires large amount of computation time. In order to speed up simulated annealing, it is worth while providing a good initial solution using high speed greedy heuristics, and then starting simulated annealing. One of the difficulties of this approach is the determination of the starting temperature of simulated annealing. The starting temperature should be such temperature that simulated annealing does not destroy some good structures of the initial solution produced by heuristic and ensures further optimization. We propose a solution to this problem, a new temperature estimation method for simulated annealing using neural networks. In the proposed method, the neural network learns some relationship between temperature and probability distribution of the difference of cost function between two neighboring solutions in the solution space. The neural network learns to estimate proper annealing temperature of solutions which are produced by greedy heuristics. We demonstrate the efficiency of the proposed method by applying it to VLSI cell placement problems. Computational experiments show that the proposed method can obtain the same quality placement results as those of simulated annealing within less computation time.

View full abstract

This paper presents a method for the automatic evaluation of the appearance of seam puckers on suits. We process the evaluation as pattern recognition. As the features of seam puckers we use the fractal dimensions.
Since gray levels of seam puckers are often confused with those of material's texture, we distinguish seam puckers from the texture using the concept of variance. Then, fractal dimensions are calculated using the knowledge of the seam puckers. By these processes, the features are extracted efficiently. Twenty unknown suits were used for the evaluation experiment and we could obtain a result close to the one assessed by human beings.

View full abstract