Engineering topics
Publications and source records attributed to Curry, R. E..
A simulation task is reported which incorporates both kinds of variables, informational and psychological, to successfully study pilot decision making behavior in the laboratory. Preliminary experiments in the measurement of decisions and the inducement of stress in simulated low visibility approaches are described.
A model of the pilot as a monitor of instrument failures during automatic landing is proposed. The failure detection model consists of two stages: a linear estimator (Kalman Filter) and a decision mechanism which is based on sequential analysis. The filter equations are derived from a simplified version of the linearized dynamics of the airplane and the control loop. The perceptual observation noise is modelled to include the effects of the partition of attention among the several instruments. The final result is a simple model consisting of a high pass filter to produce the observation residuals, and a decision function which is a pure integration of the residuals minus a bias term. The dynamics of a Boeing 707 were used to simulate the fully coupled final approach in a fixed base simulator which also included failures in the airspeed, glideslope, and localizer indicators. Subjects monitored the approaches and detected the failures; their performance was compared with the predictions of the model with good agreement between the experimental data and the model.
Mathematical models of the human operator have been concerned primarily with his input/output characteristics and his adaptive behavior to sudden changes in the controlled element dynamics. Newer models have examined the ability of the human to detect failures when acting as a monitor. However, models for simultaneous monitoring and control (e.g., an aircraft pilot flying a split axis approach) are necessary for performing pilot task allocations and for coordinated design of display and control subsystems. Flight test results of simulated instrument helicopter approaches conducted have shown the following: (1) constant speed approaches can be made quite comfortably by the pilots; (2) pilots cannot hover on situation displays alone; and (3) pilots can hover with a flight director display, but feel uncomfortable because they do not have enough time to monitor the situation displays.
A multinomial maximum likelihood program (MUNOML) for signal detection and for behavior models is discussed. It is found to be useful in day to day operation since it provides maximum flexibility with minimum duplicated effort. It has excellent convergence qualities and rarely goes beyond 10 iterations. A library of subroutines is being collected for use with MUNOML, including subroutines for a successive categories model and for signal detectability models.
The small laboratory computer is ideal for experimental control and data acquisition. Postexperimental data processing is often performed on large computers because of the availability of sophisticated programs, but costs and data compatibility are negative factors. Parameter optimization can be accomplished on the small computer, offering ease of programming, data compatibility, and low cost. A previously proposed random-search algorithm ('random creep') was found to be very slow in convergence. A method is proposed (the 'random leap' algorithm) which starts in a global search mode and automatically adjusts step size to speed convergence. A FORTRAN executive program for the random-leap algorithm is presented which calls a user-supplied function subroutine. An example of a function subroutine is given which calculates maximum-likelihood estimates of receiver operating-characteristic parameters from binary response data. Other applications in parameter estimation, generalized least squares, and matrix inversion are discussed.
Two experimental capabilities were developed on the PDP-12 computers. The first of these was an experimental monitor to perform experiments in the psychophysics of visual target motion prediction, discrimination, etc., and exercises the (limited) graphical capabilities of the PDP-12. The original monitor was modified to include the capability of handling more than 64 stimuli without reinitializing the program. The second experimental monitor was programmed and was used to investigate pilot decisions in low visibility approaches. The details of this experimental capability, which has great potential for exploring decisions and behavior, are described in detail.
Explore the source record for details and available documents.
The Air Traffic Control (ATC) system is evolving under a general plan specified by the Federal Aviation Administration. Among the developments being considered is the Discrete Address Beacon System (DABS). The use of this system, although relieving congestion on the communications frequencies, would eliminate information about other aircraft because the party line communications now in use would be lost. One alternative to restore this lost information is an Airborne Traffic Situation Display (TSD). Experienced airline and military pilots participated in a factorial design to evaluate two types of communication (discrete address, party line) and two types of displays (TSD, no TSD). A stop-action quiz was used to evaluate their knowledge of other aircrafts' position, altitude, speed, heading, rate of climb, identity, and landing sequence number. Significant differences between conditions were detected, primarily in the position variables. Workload, as measured by a spare capacity side-task, showed a main effect of displays and a significant interaction between displays and communications. The data are summarized by plotting each display/communication condition configuration in the plane defined by information and workload index. A limited number of blunders by other aircraft were included in the simulations with a significant, but not entirely satisfactory, improvement in blunder detection attributed to the TSD.
The expected traffic environment for an intercity STOL transportation system is examined, and operational procedures are discussed in order to identify problem areas which impact STOL avionics requirements. Factors considered include: traffic densities, STOL/CTOL/VTOL traffic mix, the expect ATC environment, aircraft noise models and community noise models and community noise impact, flight paths for noise abatement, wind considerations affecting landing, approach and landing considerations, STOLport site selection, runway capacity, and STOL operations at jetports, suburban airports, and separate STOLports.
The basic concept of remoting information to the cockpit is used to design and develop a computerized airborne traffic situation display device that automatically selects and presents segments of a controller's scope to the aircraft pilot via a narrow band digital data link. These data are integrated with aircraft heading and navigation information to provide a display useful in congested air space. The display can include alphanumerical symbols, air route maps, and controller instructions.
A model for visual space perception is proposed that contains desirable features in the theories of Gibson and Brunswik. This model is a Bayesian processor of proximal stimuli which contains three important elements: an internal model of the Markov process describing the knowledge of the distal world, the a priori distribution of the state of the Markov process, and an internal model relating state to proximal stimuli. The universality of the model is discussed and it is compared with signal detection theory models. Experimental results of Kinchla are used as a special case.
Display research for aircraft collision warning systems
Random-input describing function as linear approximation to instantaneous nonlinearity in terms of minimum mean-squared error
Monograph on estimation and optimal control with quantized measurements covering linear, nonlinear and time varying systems
M measurement optimal feedback control algorithm for stochastic discrete time systems, considering nonlinear plant, constrained controls, nonquadratic cost and simulations
Separation theorem for nonlinear measurements, discrete-time linear systems and quadratic cost to achieve stochastic control without dynamic programming
Separation theorem for arbitrary nonlinear measurements to find optimal stochastic control without dynamic programming
Nonlinear estimation with quantized measurements and application to pulse code modulation, data compression, and predictive quantization