This research studies the use of a water drop to pick up a minute object for manufacturing small electric devices. The method can align the center of an object with the center of the nozzle even if there is a deflection between them initially. This is called a self-centering effect. The relationship between the size of the water drop and the parameters of storing water in the nozzle is studied experimentally. The time for drawing water into the nozzle and the pressure provided to the nozzle are investigated as parameters, and the conditions in which the formed drops become hemispherical are clarified. To prove the feasibility and efficacy of the proposed process using a hemispherical drop, experiments on picking up a resistor chip and spherical and cylindrical objects are carried out. The self-centering effect is investigated by varying the deflection of an object from the center of the nozzle. It is also confirmed that the self-centering effect leads to virtually zero deflection remaining after the object is picked up.
In this paper, optimization of diesel engine control parameters using a modified multi-objective particle swarm optimization (MOPSO) method is considered. This problem is formulized as a multi-objective optimization problem involving three optimization objectives: brake specific fuel consumption (BSFC), exhaust gas emission, and soot. A modified MOPSO is proposed with integration of particle swarm optimization (PSO) and a crossover approach. Several benchmark functions are tested, and results reveal that the modified MOPSO is more efficient than the typical MOPSO. Engine control parameter optimization with an extended PSO and the modified MOPSO is simulated, respectively. It proved the potential of the modified MOPSO for the engine control parameter optimization problem.
This paper proposes a continuous-time model estimation method by using the Unscented Kalman Filter (UKF) from sampled I/O data, in which plant parameters as well as the initial state are estimated. The initial state is estimated based on a backward system of the plant, and the parameters are estimated by using an iterated UKF, which repeats the estimation of the forward system and the backward system alternately. In order to demonstrate the effectiveness of the proposed method, a rotary pendulum is considered to estimate the parameters of a continuous-time nonlinear system.
The authors previously proposed a method to measure the diameter of the deformed reinforcing bars in concrete structures nondestructively using an electromagnetic wave radar. The method estimates the periodicity of the knots of the inspected bar and utilizes the standard relationship between the knot's pitch and the diameter of the bar to measure the diameter indirectly. The effectiveness of the method was verified using test specimens where the bars were placed parallel to each other. However, in practical case, where other reinforcing bars cross the inspected bar perpendicularly, the stronger reflections from the cross bars influence the reflection from the inspected bar. This paper thus proposes a general method which eliminates such unwanted influences from the cross bars and measures the diameter accurately even in practical environment.
Association rule mining is one of the most exploited areas in data mining, which includes applications from business to data classification and summarization. Thus, several association rule-based classification methods have been proposed. Most of them often produce too many rules for humans to read over, that is, the generated rules are usually complex and hardly understandable for the users. Only some of the rules extracted are of real interest. Most of the rules are either redundant, irrelevant, or obvious. In this paper, a new post-processing method for pruning class association rules is proposed by a combination of statistics and an evolutionary method named Genetic Relation Algorithm (GRA). The algorithm is carried out in two phases. In the first phase, the rules are pruned depending on their matching degree with data, and in the second phase, GRA selects the most interesting rules using the distance between them. The two-phase method has the following properties: 1) efficient since it reduces dramatically the pruning processing time. 2) reliable because a small rule set is produced which is accurate (it keeps at least the same prediction accuracy as the original large rule set), comprehensible (it is more understandable for the users since the number of attributes involved in the rule is also small) and interesting because of the diversity of rules. The advantages of the proposed method is demonstrated using several real datasets and it is compared with other conventional methods including GNP-based mining in terms of prediction accuracy and time consumption.
Parallel electrostatic suspension is proposed and studied analytically. In this suspension system, a single power amplifier is operated to suspend multiple floators or to control the multiple-degree-of-freedom motions of a floator. Parallel suspension systems are classified into two types according to the connection of electrodes for suspension: parallel-connected and series-connected. The controllability of the former type is discussed both for direct and indirect voltage controlled systems. The conditions for each system to be controllable are clarified. It is also shown that the latter type becomes uncontrollable.
This paper investigates an adaptive control problem for a class of cascaded nonlinear systems with discontinuity. The uncertainty is taken into account as an unknown parameter and bounded unknown functions in the differential equation with discontinuous right hand side. The Filippov-framework is used to analyze the behavior of the discontinuous dynamical systems. First, a state feedback design with parameter adaptation is presented, where the designed controller ensures the convergence responses of of the closed loop system to any given bounded set. Then, the design approach is extended to achieve asymptotical stability under a stronger condition on the driven subsystem. Finally, a numerical example is given to demonstrate the proposed design approach.
This paper is concerned with a new feedback control design methodology for multi-fingered robot hands applicable to multiple contact situations. As a first step, we especially consider the situations where all the fingers are in contact or not in contact with an object, considering the tasks of catching and releasing the object preceding to or followed by grasping/manipulating the object. Main features of the proposed method are: (1) the direction of the fingertip motion in the non-contact situation is selected to be directly linked to the direction of the object motion and the internal force in the contact situation; (2) by introducing a unified system description for multiple contact situations, a linearizing compensator applicable to multiple contact situations is designed. The controller can handle the tasks with the multiple contact situations by choosing appropriate desired trajectories for the linearizing compensator without switching control architecture. In addition, owing to the selection of the motion in the non-contact situation, all the fingers can approach to the object synchronously along the directions of the object motion and the internal force in the contact situation. A numerical example is shown to prove effectiveness of the proposed method.
This paper considers a stability problem for stochastic and discrete-time switched systems with arbitrary switching. Several stability conditions based on Lyapunov theory are presented. In particular, a common Lyapunov function approach and a multiple Lyapunov function approach are discussed. For linear switched systems, a tractable stability condition is given in terms of linear matrix inequalities.
An asymptotically exact approach is presented for robust semidefinite programming, where coefficient matrices polynomially depend on uncertain parameters. Since a robust semidefinite programming problem is difficult to solve directly, an approximate problem is constructed based on a division of the parameter region. The optimal value of the approximate problem converges to that of the original problem as the resolution of the division becomes higher. An advantage of this approach is that an upper bound on the approximation error is available before solving the approximate problem. This bound shows how the approximation error depends on the resolution of the division. Furthermore, it leads to construction of an efficient division that attains small approximation error with low computational cost. Numerical examples show efficacy of the present approach. In particular, the exact optimal value is often found with a division of a finite resolution.
This paper presents a model predictive controller for an electromechanical valve. The proposed method has an important feature that it is a deadbeat controller with a designed traveling time of the valve armature. Therefore, the armature reaches a target position and velocity. The controller calculates the optimum control force for the system in each sampling instant. A two-point boundary problem is solved analytically for calculating the optimal force for the system. A disturbance observer is used for estimating the armature velocity and for improving the controlled system robustness against parameter uncertainty. The proposed method is demonstrated in computer simulation and on an experimental testbed.
This paper proposes a fault-tolerant control (FTC) synthesis against partial actuator faults through particle swarm optimization (PSO). It is shown how to design a reconfigurable state feedback controller in such a way that the performance deterioration stays within a prescribed tolerance when the actuator faults occur. The proposed FTC method is applicable to a variety of linear systems with input constraints. Numerical examples subject to input constraints are given to demonstrate its effectiveness.