This paper describes the design, fabrication and characteristics of a new type of near-infrared (NIR) photodiodes. The photodiode consists of a lattice-mismatched GaInAs light-absorbing layer, a very thin InP/InAsP cap layer and a InAsP strained superlattice (SSL) buffer layer. The very wide wavelength spectral response over a wavelength range from 0.6 to 2.1μm and low dark current have been achieved for the first time using this specially designed photodiode structure. The newly developed InAsP SSL structure were successfully used to realize the about fivefold reduction in the thickness of the buffer layer. This also demonstrates the feasibility of fabricating state-of-art lattice-mismatched GaInAs photodiodes grown by metalorganic vapor-phase epitaxy. A shallow pn junction through the very thin cap layer was formed using the new Zn diffusion technique.
The mean detectivity D^* at 300K was higher than 10¹¹cmHz^-1/2W^-1 in the wavelength range from 1 to 2μm. The photodiode mounted on a cooling unit in a hermetically-sealed package with a glass window can operate with D^*>10¹²cmHz^-1/2W^-1 from 1.3 to 2.1μm. These photodiodes are the preferred devices in a wide range of NIR system, including remote sensing, gas analysis, and pollution detection. They are also ideal for NIR spectroscopic applications (such as FT-NIR system) requiring wide-wavelengthspectral data to perform chemometric analysis.

View full abstract

This paper proposes the Vision Based Active Antenna (VBAA) that can detect the contact force, the stiffness of the environment, and the contact location between an insensitive elastic antenna and an environment, through the observation of the antenna's shape by a camera. We show that both the contact location and the contact force can be estimated by measuring two arbitrary points on the antenna after a pushing motion, even though the exact contact point is hidden by occlusion. By considering the geometrical relationship between the virtual (without environment) and the real (with environment) displacements of the contact point, the stiffness of the environment can also be estimated, while our conventional Active Antenna can not do. We present the basic working principle of the VBAA and give experimental verification.

View full abstract

In the spectral analysis of temporally varying signals such as human voice and sound from musical instruments, the short time spectral analysis is known to be effective because with that method one can change the frame length depending on the frequency involved and choose arbitrarily the analysis interval in the frequency domain. Among the methods, the analysis with the complex first order system by Aoshima and its extension by Chinone are well-known and currently being used. In order to optimize and develop further the methods, it is necessary to investigate their stability because of its recursive nature. In this paper we first discuss the stability of the Chinone's method and by extending the discussion, next propose a new method of IIR short time spectral analysis which utilizes data window in order to analyze even lower level spectral components of signals. Furthermore, the effectiveness of the proposed method for feature extracting the tone and consonants in the process of transcription of music is experimentally demonstrated.

View full abstract

The adaptive control system has been generally constructed by using an observer-controller with adjustable parameters and by setting an appropriate adaptive law. Recently, Masuda et al. has proposed a new adaptive control system with a free parameter via the coprime factorization method. They proved the boundedness of all signals in the closed-loop system and the convergence of the output error using an adaptive control scheme for a static error system. In this paper, we derive a dynamic error system and a non-minimal realization for SISO plants with a free parameter via the coprime factorization method. However, the dynamic error system is not strictly-positive real except for the plants with the relative degree 1. Therefore, we propose the parameter adjustment law for the plant whose dynamic error system satisfies a Riccati equation instead of the Lurie equation. The present adaptive control law guarantees the boundedness of all states of the error system and of all signals in the adaptive system.

View full abstract

This paper presents a solution to two disk mixed sensitivity problems on the basis of pole assignment. The sensitivity function and the complementary one are decomposed to the product of two functions respectively. One of them is chosen to satisfy the sensitivity specification or the robust stability one. If the gain of the other function are less than one, the problem is solved. It is shown that the decomposition and the gain adjustment can be easily accomplished by two parameters on a graphics. The procedure and numerical examples are presented. We get lower sensitivity solution than H_∞ control approach to one disk problem.

View full abstract

In this paper, a linear model for the induction motor is derived from a view point of magnetic energy. In the derived model, the induction motor is described a two-input and two-output controlled object by selecting an energy stored in the rotating field as a new state variable. Magnetic energy stored in the field and converted to the torque are considered to be controlled inputs. These magnetic energies are decoupled. This suggests that torque-producing component current is decoupled from flux-producing component current. And torque response isn't affected by rotor resistance variation. Therefore, a state of so-called field orientated control is achieved in the derived model. We confirm the availability of a derived model by comparing it with an approximated linearization model in simulation results.

View full abstract

Arm musclo-skeletal systems, driven by pairs of agonist and antagonist muscles, have redundancy in motor commands. The redundancy is caused by inherent arrangement of the antagonistic muscles and its synergistic coactivation, and is known to serve for flexible manipulation tasks, where positional movement and visco-elastic property of the arm should be controlled simultaneously. Additional pair of double-joint muscles, expands control range of visco-elastic property, increases more redundancy. Thus reduction of such motor command redundancy is a primary subject in the control of muscloskeletal systems.
In this paper, we propose a method for reducing the redundancy by explicitly specifying visco-elastic property and for realizing simultaneous control of positional movement and visco-elastic property. First, a musclo-skeletal system of the arm is modeled by a planar two-link manipulator with four single-joint muscles and two double-joint muscles. Then, motor commands to track the desired trajectories, specified with movement and visco-elastic matrices, are calculated by a relaxation method that uses a forward system model of the arm. Motor command redundancy is dissolved by optimizing a criterion function, composed of quadratic errors between the desired and the computed trajectories by the forward system model. Simulation results show that coactivation of double-joint muscles is important to control elastic property, and that the arm is controlled to execute the same positional movement with different elastic matrices..

View full abstract

To eliminate position-dependent disturbances in a control system for a constant-speed of rotation, we have proposed a very effective approach by introducing a new concept called the “position domain”. However, because the position domain model of a linear plant is in fact nonlinear, a linear control strategy may not meet the design requirements. To solve this problem, we propose a nonlinear compensation approach in this paper. The procedure for designing a control system comprises two steps. First, an input-output exact linearization approach is used to linearize the nonlinear plant in the position domain. Then, a repetitive controller is designed for the linearized plant to ensure good performance.

View full abstract

Branch and bound algorithms are general purpose intelligent enumeration techniques for solving combinatorial optimization problems. It is considered to be well suited to parallel processing. Typically, an application of parallel processing cannot enlarge the solvable size of combinatorial optimization problems. However, parallel branch and bound algorithms can achieve super-linear speedup versus increasing processing elements. In this case, a problem of larger size can be solved in a practical amount of computing time.
In this paper, we propose the hybrid selection rule for parallel branch and bond algorithms. Typical selection rules for sequential branch and bound algorithms can be naturally enhanced to parallel one. Using six networked workstations, experimental effectiveness comparisons among several selection rules are presented. The results show that the hybrid selection rule leads to super-linear speedup.

View full abstract

Recently, the importance of computational experiments has been indicated to realize complex phenomena. We propose a framework to describe a computational model with transitive functions for complex behaviors. A concrete example of this framework was described, based on a time series of BZ reaction, which has temporal intermittency chaotic behavior. In our framework, the approximate model of BZ reaction contains several properties of the time series of BZ reaction. The approximate efficiency of our framework is evaluated from Shannon's information theoretical point of view, with temporal mutual information. This result suggests that our framework has potential to describe the complex phenomena of not only a single system, but also an assembly system.

View full abstract