Appendix D
Research
Ongoing Aircraft-Pilot Coupling Simulation Technique Research
A critical technical tool in the design and development of new or modified aircraft is the piloted simulator. These include in-flight simulators and ground-based simulators with varying degrees of sophistication, from simple piloted combat stations ("work stations") to high fidelity, wide field-of-view facilities with large-amplitude motion base capability. (For a discussion of the necessary considerations for using these facilities, see Chapter 5.) In general, the current simulation techniques have been inadequate for exposing APC characteristics prior to flight tests. The purpose of this appendix is to describe current research to improve piloted simulator capabilities in APC evaluations.
T-33 Have PIO Simulation Technique
One current research project is an evaluation of the relative effectiveness of ground-based piloted simulator equipment with varying degrees of sophistication. Using "best engineering practices" to model the T-33 APC flight test experiment (called "Have PIO"), the research project focuses on variations of parameters in the longitudinal aircraft axis that influence PIO events. Piloted evaluations are then conducted on each type of facility: the Large Amplitude Multimode Aerospace Research Simulator (LAMARS) with and without motion; the MS-1 (40-foot visual dome); and the Piloted Combat Station. Pilot ratings, both Cooper-Harper Pilot Ratings and PIO ratings, are recorded for each configuration and compared with flight tests. Subsequent changes can be made to the simulation model, facility, task, or pilot stress
level to try to improve upon comparisons to flight tests. These experiments will identify the most reliable adjusted simulation techniques for exposing linear APC tendencies during landing for fighter applications. This project is scheduled to be completed in the second quarter of fiscal year 1997.
C-17 Comparison between In-Flight and Ground-Based Simulation and Flight Test
Another more sophisticated research study on the development of simulation techniques has just been initiated to evaluate and compare piloted simulation capabilities for exposing APC tendencies on transport aircraft. First, the C-17 high fidelity, motion-based simulator will be used to evaluate APC tendencies with degraded flight control changes. The specific flying tasks, emergency procedures, and APC events will be systematically defined and recorded. Then, the C-17 development test aircraft, T1, with its Change-A-Gain system, will be used to force APC events to occur in a safe manner during the APC flight test tasks identified on the simulator. The results will be compared and changes made to the simulation process, as required, to maximize agreement between flight tests and the ground-based simulations. Finally, the simulation experiment will be repeated using the Total In-Fight Simulator (TIFS) aircraft to evaluate the effectiveness of in-flight simulation in exposing APC characteristics. This project will require a minimum of 20 months to complete; T1 flight testing is planned for the first quarter of fiscal year 1997, and TIFS flight testing is planned for the second quarter of fiscal year 1997.
Development of APC Simulation Techniques for Fighter Aircraft Using the Variable Stability In-Flight Simulator Test Aircraft
A research experiment similar to the C-17 project is planned for fighter-type aircraft in fiscal year 1998. A similar approach to the one outlined above is planned using the F-16 Variable Stability In-Flight Simulator Test Aircraft (VISTA). Unique APC simulation techniques for maximizing simulator effectiveness will be developed.
APC Compensation And Detection Research
Compensation System Research
In the past, several attempts aimed solely at APC attenuation have been tried, with varying degrees of success. The general application of the proposed
approaches has not been thoroughly investigated, and the inherent limitations resulting from limiting phase lag and frequency bandwidth parameters to keep APC events from occurring must still be studied. Rather than attempting to detect APC tendencies, these studies focus on preventing the buildup of phase lag or filtering high frequency control inputs. The filter approach is currently used on the space shuttle.58 Phase lag compensation has shown some promise in preventing rate-limit-induced APC events.1,10,41 The extent to which this approach may limit desired maneuverability for otherwise APC-free maneuvers has not been adequately investigated. An approach like this has been incorporated into the Swedish Gripen (JAS 39) control system.60 Parameter compensation techniques coupled with a highly reliable detection system may provide a good integrated approach that would not limit maneuverability until an APC event is detected or extreme values of key APC-related parameters have been exceeded without APC detection.
Development of Theory-Based Detection Algorithm
As part of an integrated research effort to develop APC-resistant design criteria and development processes, a set of theory-based engineering algorithms is being developed that will provide on-board early detection of incipient APC events. Subsequent efforts will result in a complete system design that defines the detailed components, from sensor signal to warning device or compensation. A complete system will be developed and validated on current aircraft. Emphasis will be placed on minimizing unnecessary warnings and compensation while minimizing the occurrence of APC events.
Neural Network Empirically-Based Detection System
A completely different approach to on-board detection and compensation is being investigated using data from current APC events to train a neural network. The resulting algorithms will not require that the theory be developed in advance of an effective solution. The trained neural network will recognize and distinguish APC events before they become unmanageable. Initial trials of a relatively crude, single-axis neural network were very promising on the limited number of APC events tested. It should be noted that these early test cases included neural network identification of a nonlinear APC event on an F-18 aircraft even though the network was trained on F-16 linear APC events. Much work still needs to be done to train the network on a sufficient number of APC events to make this approach effective for all types of aircraft and APC events. After detection algorithms have been developed, substantial efforts will be required to verify and validate that the network is effective and