Hitoshi Doki

Simultaneous optimal design of structural and control systems for cantilevered pipes conveying fluid

We consider a shape optimization problem for a cantilevered pipe conveying fluid. The outer diameter distribution of the pipe and the location of the sensor and the actuator are optimized such that the critical flow velocity of the closed-loop system is maximized. The outer diameter distribution is optimized such that the total volume of the pipe is unchanged with the initial diameter distribution. The critical flow velocity of the closed-loop system is defined as the flow velocity which cannot be stabilized by active control law with a predetermined energy quantity. By this definition, it is physically reasonable comparing the quality of several design candidates since all candidates are actively controlled with the same energy consumption. We propose a method for obtaining the critical flow velocity with consideration of the amount of computation for the optimal design. This optimal design problem results in a maximization problem with equality and inequality constraints. We adopt simulated annealing method which is known as one of discrete optimization techniques for obtaining the optimal design.

Task based approach on trajectory planning with redundant manipulators, and its application to wheelchair propulsion

We introduce an optimization method for the trajectory planning of redundant manipulators which achieve a given task with high efficiency, and apply the method to a wheelchair propulsion problem. A GA is used to optimize the redundant variables of a manipulator. Additionally, the procedure of the method does not use any forward dynamics computation. The effectiveness of the proposed algorithm has been shown by applying the method to wheelchair propulsion.

Global Optimal Design of Dynamic Parameters of Robot Manipulators

In this paper we deal with a global optimal design of dynamic parameters of robot manipulators. The dynamic characteristic of robot manipulators is uniquely determined by base parameters, which are functions of physical parameters, e.g., the mass, the moment of inertia and the damping of each link, etc.. Using the fact that the equation of motion of manipulators generally becomes an afflne function on the base parameters we propose an LMI based global optimal design of the base parameters of robot manipulators. In this paper we obtain the global optimal base parameters minimizing the required energy to achieve a given task (given by the reference trajectory of the end-effector). We adopt the computed torque method as the control law for the robot manipulators. The proposed design method guarantees the global optimality of the optimized base parameters. As a design example an optimal design of the base parameters of planar two link manipulator is presented.

Integrated optimal design of passive and active elements for hard disk servo systems

To achieve high density and high speed access of magnetic disk drives, integrated optimization method is applied for designing the head-suspension assembles. Approximation is used to the model of a head-suspension assembly, which makes it possible to obtain a nice analytical result on Riccati equations. Using this approach, we can obtain an explicit expression of the optimal controller in a state space form. Based on this setup, the simultaneous optimization problem both on active and passive parameters is reduced to a simpler one. A numerical example is given to show that a better performance of a magnetic disk drive can be obtained by the simple design procedure.

Estimation of Golf Club Head and Ball Motions by Neural Network

In sports that are one of the life science fields, the performance of tools such as wear, shoes, clubs and so on advances remarkably, and the record has improved spectacularly. However, it tends to the restriction of “tools that expand distance of a jump” in golf that is one of ball sports toward the target in the distance. The improvement of the sports skill will need not only improvement of tools performance but also improvement of humans sports skill in the future. In the previous paper, we related the golfers cock-roll motions in the golf driver swing and eight club head and/or ball motions (head velocity, ball velocity, upper angle, side angle, back spin, side spin, carry and traverse) by experiment and multiple regression analysis method. In this paper, the club performance is newly added to the explaining variable. Moreover, by using measured variables and/or factors, the estimated system at head and ball velocity based on the neural network is proposed. When the estimated value by the suggesting method is compared with the measurement value, the maximum error margin and the average error margin are smaller. When the presumption value is compared with the measurement value, the maximum and the average error of the suggesting method are smaller than that of the previous method. Using the sensitivity analysis, we found that the most important factors in improving the head velocity and ball velocity

B17 A Simultaneous Design Method of Mechanical/Dynamical Parameters and a Control System for Robot Manipulators

ロボットマニピュレータは,産業等の幅広い分 野で使われている.ロボットマニピュレータやそ のコントローラを設計する際には,定量的な性能 評価指標が必要になる.性能評価を考えた場合, 評価方法は大別して静的評価と動的評価の二つに なる. 静的評価の代表である機構評価は,性能評価に おいて最も基本的で重要なものである.現在まで に,マニピュレータの(主として静的な)操作性 の良否を表す代表的な指標として,可操作性楕円 体と操作力楕円体1)が提案されている.可操作性 楕円体と操作力楕円体は,それぞれロボットマニ ピュレータの動かしやすさと力の出しやすさを定 量的に評価することができる.また,可操作性楕 円体と操作力楕円体は,それぞれヤコビ行列の特 異値とその逆数を主軸に持つ楕円体である.この 性質を利用し,機構系の評価指標として可操作性 楕円体の体積が大きければ操作性がよく,小さけ れば操作性が悪いと考えた評価指標である可操作 度2)が提案されている.また,ヤコビ行列が正方 行列になる場合,ヤコビ行列の行列式と可操作性 楕円体の体積が比例することから,ヤコビ行列の 行列式を指標とした評価指標3)も提案されている. このような評価指標に基づいて機構の最適化を実 際に行った研究としては,ヤコビ行列式を指標と した内山,清水らの研究3)や精密はめ合い作業の 作業の達成条件から得られる許容関節角度誤差に 基づいて,その作業に適したマニピュレータを求 めた甲斐,原等の研究4)がある.これらの研究に おいて,評価指標を最適値にするための機構設計 法としては,設計する機構の全ての組み合わせに

Structural design for reduced-order H/sub /spl infin// controller

We consider an integrated design problem of structural and control systems. Using the reduced-order H/sub /spl infin// controller design method proposed by Mustafa and Glover (1991), we formulate the integrated design as a simultaneous design of the structural design parameter and the reduced order H/sub /spl infin// controller. The formulated problem does not need to take an iterative procedure (controller synthesis and structural modification) which is employed in most integrated design.

A Design Method of a Vibration Control System for Flexible Structures. Vibration Control with Control Energy Constraints.

This paper deals with a sensor/actuator placement problem in design of active vibration control system for flexible structures. This problem is formulated as a minimization problem of the total energy which is defined as a sum of a kinetic and strain energy in a colatrolled structure with a constraint of control effort. The inequality constraint on the variance of the closed loop control effort is adopted to represent the capacity of the actuator. Using a design algorithm which iteratively tunes the weighting matrix of the quadratic performance index in LQG problem, the controller which meets these specifications can be synthesized. The optimal location of the sensor/actuator is determined by calculating the total energy for each candidate of them under several energy constraints of the control effort. It is found that the optimal placement of the sensor/actuator depends on the input energy constraint. The simulation and a experiment for a cantilevered beam as a typical example of flexible structures are conducted. It is shown that these results of the optimization can be used as a guide to the design of active vibration control systems.

Active Control of Cantilevered Pipes Conveying Fluid. Stabilization with Constraint on Input Energy.

We presents an active control of cantilevered pipes conveying fluid with constraint on input energy. The controller used in this study is designed with constraints on input energy variances. LQG control law is used to obtain the feedback controller. This controller design method takes the performance of the actuator into consideration. The control object which consists of a pipe system and an actuator is modeled mathematically by the Galerkin method. The experimental setup consists of the pipe system, a laser sensor, an A/D converter, a personal computer, a D/A converter, and a DC servomotor which produces a transverse force for active control of the pipe. The experimental results show that the effectiveness of the controller in stabilizing unstable phenomena of cantilevered pipes conveying fluid varies according to the conditions of input variance constraints.

Design Method of Decentralized Control System via Stable Factorization - Approach from Simultaneous Stabilization -

Abstract Decentralized control problem with guaranteeing the stability both of overall system and the decoupled subsystems is considered Using parametrization of simultaneous stabilizing controller, optimization of the closed-loop characteristics both for the overall system and the decoupled subsystems is formulated as model-matching-type problem with constraint. Moreover, it is shown that a subclass of this problem can be cast as convex optimization problem.