Digital Simulation Results Research Articles

Bang-bang aspect of human learning and its modeling

In manual tracking task with a linear control wheel, human tracking processes could be distinguished into two different control mode, a bang-bang and gradual mode. This paper discusses a bang-bang control mode, and proposes a model to explain human adaptation in this mode. When human operator finds that some lead compensation is necessary for good tracking, the tendency of a bang-bang mode is observed remarkably.His bang-bang switching motion is characterized by the predictive manoeuvre lines in the phase plane between the plant output and the rate: Namely, his control reversals turn up whenever the predictive line crosses a target position in the phase plane. The proposed model is based on this control law. Digital simulation results of the model show close agreements with experimental results obtained by human operators.

Monte Carlo study of the optimal non-linear estimator: linear systems with non-gaussian initial states †

The optimal, in the mean-square sense, estimate of state vector of a linear discrete system that is excited by white zero mean gaussian noise and that has non-gaussian initial state vector is presented. Both the optimal estimate and the corresponding error covariance matrix are given. It is shown that the optimal estimator consists of two parts : a linear estimator which is a Kalman filter and a non-linear part which is a parameter estimator. In addition, the a posteriori probability density function, p(x(k)λk), is also given. Finally, a suboptimal procedure that reduces the computational requirements is presented. The results of extensive digital computer simulations including Monte Carlo study have been presented to establish that the non-linear filter presented here is far superior to the best linear Kalman filter. A practical filter design criterion for utilizing this non-linear filter with reduced data processing requirements is also given.

Dynamics of piano mechanisms

Analysis of the structure of the modern upright and grand piano keyboard mechanism is given, and the linear model used in previous simulations by other authors is described. The comparison of the experimental results obtained for a real keyboard mechanism with those which were calculated from known equations of motion (linear model) showed, however, not sufficient coincidence, thus a necessity for improvement of the model arose. A new, non-linear model of the keyboard mechanism is presented, and some results of that model's digital computer simulations are given.

Spatial densities for specular points on a gaussian surface

Specular point densities for several models of randomly rough surfaces with a Gaussian height distribution are derived. Both area densities and volume densities conditioned by the surface height variate and surface curvatures are obtained. Several shadowing functions are compared with digital computer simulation results, and the application of the slope conditional shadowing function as a modifying factor for specular point densities is described. These densities, as modified by the shadowing function, are required in determining the statistics of fields reflected from random surfaces inducing deep-phase modulation whenever discrete scatterer models are employed in the analysis.

Theory and Flight Verification of the TIFS Model-Following System

The content of this paper describes the theoretical development and flight-test results of the model-following control system of the Air Force Total In-Flight Simulator (TIFS). A discussion of the conceptual design and detailed development of the system configuration is given. The manner in which the feedforward, gust compensator and lateral-directional feedback gains are obtained is developed. The feedforward and gust compensator gains are obtained by simple matrix algebra calculations. A sensitivity minimization approach using modern control theory is used to obtain the lateral-directional feedback gains. Digital simulation results are included to show the improvement in model-followi ng achieved with the feedback gains determined by this approach. Time histories of the model and TIFS responses from flight test are also included to show the quality of modelfollowing obtained with the system for both the lateral-directional and longitudinal modes of operation. These results verify the theory and design procedure used to obtain the TIFS model-following control system.

Filtering of Defects in Integrated Circuits with Orientation Independence

Defect enhancement in semiconductor wafers is shown to be possible by spatial frequency filtering without the necessity for filter alignment. The technique depends on the wafer being sufficiently periodic. Coherent optical and digital computer simulation results are presented.

An Algorithm Providing All-Attitude Capability for Three-Gimballed Inertial Systems

This paper describes an algorithm which protects a three-gimballed inertial system against gimbal lock while preserving the inertial reference. One sample of the gimbal angles and gimbal angle rates for a torque-free dynamic vehicle provides the information needed to determine whether or not gimbal lock will occur at some later time. When gimbal lock is imminent, the inertial platform is commanded to a new safe orientation, and knowledge of the inertial reference is updated accordingly. By iterating this process gimbal lock protection is extended to a vehicle subjected to torques. The results of digital simulations using the Apollo Primary Guidance, Navigation and Control System as a model are presented for various dynamic situations.

Crosscorrelation identification in the presence of drift

In this paper three related methods are proposed for eliminating the effect of low frequency polynomial type drift in the crosscorrelation identification of linear time-invariant discrete systems using pseudorandom sequences. The methods are similar to the one suggested by Macleod [8] in that a convolution summation using alternating sign binomial coefficients is performed on the measured system output. However, the proposed methods are computationally simpler than Macleod's method since no deconvolution process is required. Digital computer simulation results are presented to show the effectiveness of these methods in the presence of both low and high frequency disturbances.

Image detection through bipolar correlation

The purpose of this paper is to describe the bipolar reference concept. A bipolar reference is simply a reference that can take on a continuous range of both positive and negative numerical values. It will be shown that for map matching via the convolution operation, references that yield maximum signal-to-noise ratios, in general, will be bipolar references. Algorithms for generating bipolar references and digital simulation results are given.

Decision Processes in the Adaptive Behavior of Human Controllers

A decision algorithm which simulates the rapid adaptive behavior of human controllers following sudden changes in plant dynamics is developed. The control of a VTOL aircraft in hover following failure of the stability augmentation system is used as a specific example. The decision algorithm is based on the assumption that the human controller recognizes certain pattern features in the error/error-rate phase plane. Experimental data, obtained from pilots facing four possible alternatives following the time of failure, are presented. The proposed decision algorithm is developed, and digital simulation results are discussed. A theoretical justification for the algorithm, based on statistical decision theory, is presented in the Appendix.

Recursive generalised-least-squares procedure for online identification of process parameters

A recursive differential correction-estimation technique is presented for online estimation of process parameters in the presence of autocorrelated noise. The technique is shown to allow adaptive estimation of changing system parameters. The experimental results of digital simulations indicate that the proposed recursive-estimation algorithm is valuable for online-control applications, and for the estimation of parameters in the presence of large amounts of data by avoiding the need for excessive data storage.

A simplified explicit guidance scheme for orbital injection.

Onboard computer equations for rocket vehicle guidance to elliptical orbit, discussing iteration absence from computations and digital simulation results

The application of the MIDAS digital simulator to the study of kinetic alternatives in a chemical reaction system

The frequency response of the platinum-catalyzed sulphur dioxide oxidation system was simulated for four alterna tive rate expressions suggested in the literature for the chemical kinetics in a nonflow stirred reactor. By compar ing the results of MIDAS digital simulation with experi mentally obtained frequency responses, it was established that the dynamic behavior of this reaction system is not affected significantly by the proposed reaction mecha nisms.

Identification and adaptation in control loops with time varying gain

AbstractA method is presented to apply the gain identification procedures developed in a companion paper to a control system. The method employs a small identification tank which follows the control tank; the identification tank is perturbed to estimate the gain of the process (controlled tank system). The effects of load changes on the identification system are minimized by this approach.The results of analog and digital computer simulations of this adaptive process are given. Both a general system with linear change in gain and a pH control system with a step change in concentration of the buffer species are studied. Process gain changes up to 20:1 are introduced.It is concluded that an adaptive control system of this type can be designed to maintain good control characteristics in a process experiencing wide gain variation. Criteria are presented to aid in the design of an adequate identifier.

A GYRO MOMENTUM EXCHANGE DEVICE FOR SPACE VEHICLE ATTITUDE CONTROL

A gyro control device is described which may be used in attitude control systems requiring momentum exchange devices. It has been shown that the use of two single-degree-of-freedom gyros per vehicle axis permits nearly independent control of the three components of vehicle angular rate. The new controller will employ only two gyro wheels and three gimbals to effect control of the three components of vehicle angular rate. A moderate degradation of transient response is experienced, due to increased cross-coupling between vehicle axes; but, on the other hand, significant weight and volume reductions are achieved. It is shown that the gyro controller is an extremely efficient device when its energy requirements are compared to those for a flywheel controller. Digital simulation results are presented to illustrate the time response of an attitude control system employing the advanced gyro controller.

Time-optimal control of higher-order systems

Practical extension of time-optimal control to systems of higher order than three has been limited primarily by difficulties in physically representing surfaces in a phase space of these higher dimensions. A method is presented here for obtaining the forcing function as a function of the state variables without requiring use of the phase space concept. On line solution of a set of transcendental equations is required. Results of a digital simulation of a fourth-order, real-root, single-degree-of-freedom system are presented. In a digital solution the system operates as a series of short open-loop control intervals. The effect of including derivatives of the input for prediction is shown for second-order model inputs.