It is extremely difficult to investigate cogging torque of a rotating electric machine precisely. Two dimensional finite element based analytical approaches have been developed and implemented for the magnetic force in general salient-pole motors. However, for Lundell alternators which are widely used for factory automation, automobiles and so on, it is necessary to use three dimensional analysis. In this paper, a new and precise cogging torque measurement method and the data processing technique are proposed for Lundell alternator. By using this method, practical test data for a Lundell alternator are presented and the characteristics of cogging torque are clarified and evaluated.
A fast method of Newton-Raphson-based algorithm is presented for maximum likelihood estimation of the time-delay as well as the attenuation parameter of received signals in observation data corrupted by white Gaussian noise. Based on the mathematical method described by an Ito stochastic differential equation for the noisy observation data, the likelihood-ratio function is derived instead of the usual likelihood function, and the Cramér-Rao lower bounds for the time-delay and attenuation estimates are evaluated in a modified form using the likelihood-ratio function to show the uncertainty relation between these two estimates. Unbiasedness of these two estimates are theoretically demonstrated. Simulation results that confirm the effectiveness of the proposed method are given.
In this paper, we consider strictly proper digital filters with hypercomplex coefficients, and we also consider reduction of multiplications in hypercomplex algebra. As for the former problem, the first-order digital filter with hypercomplex coefficients can realize arbitrary-order strictly proper digital filters with real coefficients less than 4, whose order of numerator is smaller than that of denominator by 1. As for the latter problem, if we apply Quadratic Residue Number System to hypercomplex algebra, we can reduce the number of multiplications to 1/4 of the original one. And the most important thing of this result is that we only need 4 multiplications per multiplier and that these 4 multiplications can be done at the same time.
A learning control based on repetitive operations of robotic manipulators is one of the most promising method to realize high speed and high precision control for robotic manipulators. The joint servo system usually has an unstable zero in discrete-time domain. This fact leads to unstable learning, when a sampling period is small. In previous paper, we have proposed a new learning algorithm based on so-called 2-delay input method, By using the 2-delay input method we can stabilize the unstable zero and realize the stable learning even if the sampling period is small. However, concrete algorithm for parameter design and experimental results have not been shown yet. In this paper, we propose new design algorithm of two parameters in 2- delay input system especially to suppress ripple of the trajectory and show some experimental results to verify effectiveness of the proposed method.
This paper considers decentralizedH∞control problems for large-scale interconnected systems, the information structure constraints of which are conformable to the subsystems. Both output feedback and state feedback are dealt with. The design of localH∞controllers involves solving two algebraic Riccati equations of the subsystem dimensions. Sufficient conditions on the interconnections among subsystems are presented for desiredH∞ attenuation bounds to be achieved from the overall disturbance inputs to the overall controlled outputs by localH∞controllers.