This paper studies relationships between the polynomial ILQ design method for robust servo systems and two-degree-of-freedom optimal servo systems. It turns out that the former method uses the same state feedback as the latter one, but attains a specified response by letting another tuning parameter large enough. The paper also shows that, because of model matching specification, the polynomial ILQ method needs some additional consideration for guaranteeing optimality. Numerical simulation is carried out to illustrate the results.
Recent study, the extended Kalman filter (EKF) is applied to the learning algorithm for a feedforward neural network and it is shown that the EKF based learning algorithm has good convergence performances. However, the feedforward neural network could not perform a dynamical signal processing such as time series pattern recognition. On the other hand, the recurrent neural network (RNN) could have dynamical characteristics because the RNN has feedback connections with time delay in the network. The EKF based learning algorithms for the RNN are also reported and some training properties became clear. In this paper, RNN training based on the EKF are applied to the time series signal processing. The connection weights of the RNN can be modified by using the filtering algorithm based on the EKF. Simulation results show that the RNN based on the EKF has good performances of the training for time series patterns generated from sine functions and for some pulse patterns.
We consider the infinite-horizon linear quadratic control problem for a descriptor system (DLQCP) based on the theory of dissipative system. First, we derive a dissipative inequality for the descriptor system satisfied by the optimal cost of DLQCP. This implies that the optimal cost is characterized by the solution of a linear matrix iniquality (LMI). Secondly, we show that the solution of LMI characterizing the optimal cost also satisfies a related generalized algebraic Riccati equation (GARE) if DLQCP is regular. Finally, we derive the optimal solution to DLQCP with fixed terminal condition.
Algebraic reconstruction techniques (ART) represent a set of iterative techniques that seek to obtain an estimate of material parameter distribution in which the predicted values of projection agree with the observed values. In this paper we apply a generalized ART to Neurofuzzy Geotomography to reduce the learning time or iteration times. We show the effect of ART by computer simulations. Futhermore, the proposed method is applied to the experimental data collected at a dam site by crossborehole seismic probing.
This paper describes a method of the position measurement of ceiling fluorescent lamps that are reference points to measure vehicle positions for autonomous navigation. The principle of the present method is basically the motion stereo that makes use of the vehicle itself and has an advantage that it works even when orientation of a camera is unknown. We can also obtain the vehicle traveling routes while positions of fluorescent lamps are measured. Because our final purpose is the measurement of the vehicle positions, we have examined the accuracy of the vehicle position measurement using the measured coordinates of the reference fluorescent lamps by both experiments and numerical simulations and confirmed the validity of the present method.
In this paper, we proposed a liquid container transfer system with the robustness for the change of static liquid level. First, nominal model is reasonably decided to suppress the sloshing (liquid vibration) for the change of liquid level, and a reference trajectory is also determined by an optimization method of Fletcher Reeves combined with a clipping-off technique to do so. Secondly, H∞ control theory has been applied to this system to obtain the robustness for the change of liquid level. Through simulations and experiments, the usefulness of the present control system has been demonstrated. Finally, the effectiveness has been also shown by comparing such a conventional control method of LQI control with Kalman filter.