The author examines the instrument hardware and programming environments and the role that recent advances in instrumentation can play in helping to achieve test system commonality. He considers the aspects of commonality that are affected by test equipment, the capabilities and limitations of the instrument environment, and the criteria affecting these abilities, including the differences and similarities between integrated diagnostics and common testers approaches.<>
The authors present novel circuits for residue arithmetic, which have been configured to form a 3*3 finite impulse response (FIR) filter with programmable coefficients. The filter has a pipelined architecture and includes testability in the form of scan path. Area efficient circuits for residue adders, subtractors, and binary-to-residue converters have been designed. An encoding scheme has been used to reduce the residue multiplier area. A tree architecture for residue-to-binary conversion has been developed. The filter is timed with a two-phase clock, which has an estimated frequency of 15 MHz.< >
A fast test generation algorithm called FPODEM (Fast PODEM) is developed to improve the original PODEM algorithm. FPODEM can generate test patterns more quickly to detect all single-stuck faults (including fanout faults) in combinational circuits. In order to accelerate the algorithm of test generation it is necessary to reduce the total number of PI (primary input) assignments which are examined to generate and propagate the D-cubes for all single-stuck faults. The idea of backtracking in combination with the input-sequence-reordering technique is developed to increase the speed of test pattern generation. No self-masking problem needs to be considered in this algorithm, i.e. all detectable faults can be detected. It is shown that the FPODEM algorithm is faster and more efficient than the PODEM algorithm. The algorithm and data structure used in FPODEM are described. The results of simulation on combinational circuits demonstrate that FPODEM performs test generation quickly and efficiently, especially when the size of the circuit under test is increased.<>
A shared system in which a human positions the refueling arm in the vicinity of an aircraft and then allows the refueling port detection and nozzle insertion to be done automatically by a combination of visual servoing and compliant control is discussed. Experimental concept evaluations have successfully demonstrated algorithms for port detection, visual servoing, and compliant control of nozzle motion and insertion. A project overview is provided.< >
The authors examine the theory of topological spaces that has proved to be useful for analysis, design, and pattern recognition of multidimensional large-scale systems. Those spaces are finite or locally finite. It is proved that any TW-complex, TCW-complex, or simplicial complex K is a T/sub 1/-discrete space.<>
Several critical cognitive dimensions which limit human decision making in reactive environments such as air-to-air combat are investigated. It is noted that future air battles will challenge pilots' abilities to cope with sophisticated, numerically superior adversaries. Aircrews will have to perform critical tasks while meeting the demands of a rapidly evolving and uncertain tactical situation. Human cognitive processes are limited by processing bandwidth, data fidelity, and situation-dependent stresses. Rigidly structured human-machine interfaces may diminish aircrew performance because they do not account for individual differences in cognitive approach. It is argued that automation and new concepts in display design are not the answer. However, results from human factors and cognitive psychology suggest a solution: intelligent intent-driven systems whose response characteristics match knowledge-based schemata compatible with pilots' mental models of their tasks
Different types of neural networks (NNs) are surveyed and their suitability as elements in flight control and flight management systems is analyzed. Advantages of neural networks over conventional digital avionic systems include speed (especially when implemented in special hardware, taking advantage of massive parallel processing), robustness, fault tolerance, and the ability to adapt to new situations by learning. An example shows how an artificial NN can be used as a gain adjuster in a stability augmentation system. A three-layer NN receives elevator commands and the sensed resulting longitudinal aircraft motion as input. The NN recognizes certain patterns in this response which are an indication that the gain is too high and that the control system is dangerously close to the stability boundary. Another example shows how a NN can solve a complex combinatorial problem which arises in search planning. It has similarities to the traveling salesman problem
The requirements modeling techniques that were successfully applied to the development of a facility to perform software development and integration testing of aircraft embedded software are described. This facility mimics the dynamic environment experienced by the software executing within actual avionics hardware. The requirements for each software component in the system are modeled from three perspectives: processing, information, and behavior. The process view treats the system as a planned response system. Events that require a system response are identified and assigned a process. Processes are decomposed to minimize interfaces, and are grouped to preserve the stimulus-response structure. The information view identifies data retained by the system and the system interfaces. Retained data attributes are defined in entity relationship diagrams. Interfaces between components are identified by grouping data into data structures. The behavioral view identifies system modes of operation and control. State transition diagrams are used to identify each system state and the events in which transitions occur. The behavioral view is linked to the process model by control specifications. Application of this approach improves communication to development groups by reducing complexity. It is amenable to both functional and object-oriented design approaches.< >
A prototype software system has been built to evaluate the usefulness of simple artificial intelligence systems in aircraft cockpits. Laboratory evaluations of the system were conducted using a simulated aircraft in a full-mission scenario. Pilot task performance and workload were measured under two conditions: baseline and enhanced with the prototype knowledge-based decision-aiding system; some of the results achieved statistical significance. On the average, the knowledge-based system improved pilot performance in locating surface objects by 36%. Performance was improved an average of 42% for tasks related to successful response to mission emergencies. Workload was decreased by 13%. Overall, 83% of the pilots perceived the prototype system to be as effective as, or more effective than, the baseline system. In direct comparisons 81% of the pilots preferred the enhanced to the baseline system
An efficient real-time algorithm which maps three-dimensional features into photographic pixel coordinates is dealt with. The algorithm is optimized for use on parallel processors such as transputers. It efficiently processes polygonal terrain data based on DMA (Defense Mapping Agency) DTED (digital terrain elevation data). The same algorithm also maps cultural feature polygons, such as buildings or models. The algorithm's output structure feeds a unique hardware architecture which converts the output structures into world addresses at the pixel rate. Nonlinear mapping of polygon surfaces is included. Range information is also calculated to resolve occulting priorities. Transputers, with their excellent parallel processing capability, are well suited to process the algorithm. A transputer architecture allows near linear expansion of polygonal mapping capability as a function of the quantity of transputers used in the processing. Physical and virtual links between transputers are combined to create an efficient architecture. The algorithm and efficient hardware processing can produce a real-time photo-based image generator. The use of transputers in the implementation of an elevation processor is discussed in detail. The division of the elevation processor task into individual processes is considered
HITBAT is a personal computer (PC) database. It not only prints out four data pages of the reliability and maintainability (R&M) Action Plan and the R&M Hitlist, but also displays to the screen interacting input forms which act like spreadsheets to compute and display information. The tool installs itself on a PC and brings up the first menu. There are four menus which guide the user through the process. The output can be either collected and stored on disketts, left on the hard drive of the computer, or stored on a VAX supporting Kermit protocol. All of the operations mentioned are accomplished by menu selection. The databases stored on disks can be combined to make a larger database. There are input forms for adding information into the database. They consist of a data sheet which contains such information as work unit codes (WUC), master stock number (MSN), mean time between failure (MTBF), goal mean time between failure (GMTBF), and many other data items. The second is a sheet which ranks the items according to the R&M 2000 goals. The third is a corrective action and interactive funding profile. It allows the required funding to be entered by fund type, year, and funding status. The total funded and unfunded amounts are continuously displayed by year and by fund
The avionics reliability-cost (ARC) program is configured as a design aid for advanced technology avionics. It is intended to be used early in the design process to aid the designer in searching for cost-effective avionics implementations. The focus of the model is on technology advances promised by the next generation of integrated circuits typified by phase-one and, ultimately, phase-two VHSIC (very high-speed integrated circuits). The model is designed to allow the study of trade-offs involving all phases of life-cycle cost. ARC facilitates design-cost trade-off analysis by modeling interactions among selected device characteristics and packaging, environmental, and operational variables. ARC carefully models device power dissipation; module thermal loads; heat conductivity as a function of module size, heat sink, and material choices; and consequent device junction temperatures and failure rates. The same device and module characteristics that are used to deduce failure rates are used to estimate module costs through calibrated engineering buildup relationships. ARC runs interactively, allowing the user to create data sets, calculate, view, and print results, and save and edit old data sets as different parameter combinations are evaluated during a single ARC session. All actions are menu-driven, and context-sensitive help screens are provided to assist the new user
The advanced launch system (ALS), is a launch vehicle that is designed to be cost-effective, highly reliable, and operationally efficient with a goal of reducing the cost per pound to orbit. An electromechanical actuation (EMA) system is being developed as an attractive alternative to the hydraulic systems. The controller will integrate 20 kHz resonant link power management and distribution (PMAD) technology and pulse population modulation (PPM) techniques to implement field-oriented vector control (FOVC) of a new advanced induction motor. The driver and the FOVC will be microprocessor controlled. For increased system reliability a built-in test (BITE) capability will be included. This involves introducing testability into the design of a system such that testing is calibrated and exercised during the design, manufacturing, maintenance, and prelaunch activities. An actuator will be integrated with the motor controller for performance testing of the EMA thrust vector control (TVC) system. The EMA system and work proposed for the future are discussed. >
A digital pulse compression processor for an airborne radar system is described. The processor implements a general-purpose convolution-based algorithm that enables pulse compression of any waveform subject to the radar real-time throughput requirements. The pulse compression processor uses very high speed integrated circuit (VHSIC) level technology and uses multichip packaging to achieve a small size compatible with airborne applications. It is developed in a modular structure to enable custom application of common modules to platforms with different requirements. The processor architecture, an analysis of throughput capability, the modular building blocks, and the integrated circuit and packaging technology are presented. Two specific designs are described, and the performance and physical characteristics of the designs for both the VLSI and modular VHSIC versions are compared. Future improvements that will enable a halving of the size and a doubling of the operation rate are presented.<>
Common characteristics of computer-aided software engineering (CASE) tools are reviewed from the viewpoint of real-time avionics software. Guidelines for optimum tools are discussed, and some technology programs that can affect future tool development are outlined. It is argued that avionics software engineering tools are extremely important because they impose an organizational discipline on software development that is generally lacking for large-scale, unprecedented systems, where teams of technical workers must cooperate. This discipline forces controls that aid in the management of software projects
Control Data's development and use of durability analysis, which is part of the Avionics/Electronics Integrity Program (AVIP), in the Advanced Tactical Air Reconnaissance System (ATARS) is described. The durability requirements of the ATARS program, the derivation of the durability profile, the methodology for analysis, plans for verification, and durability testing are discussed. The use of the AVIP concept of durability to put into place a practical, cost-effective program which will result in an improved, more durable, and more available tactical air reconnaissance system for the US Air Force, Navy, and Marine Corps is described. The AVIP for ATARS has resulted in major design improvements, and has had a significant impact on its expected durability. On the ATARS program, the equipment was designed to known environmental criteria, and durability has been designed in; costly redesign and field repairs are unlikely. The AVIP has lowered the life-cycle cost and has increased mission availability of the ATARS hardware
One of the major requirements for the F-15 STOL (short take off and landing) and maneuver technology demonstrator (S/MTD) was to develop and demonstrate an integrated flight and propulsion control (IFPC) system. The IFPC is a fly-by-wire (FBW) system that replaced the F-15 mechanical flight and engine controls. It was a USAF requirement that the IFPC be fault-tolerant and that loss-of-control not exceed once per 100000 flights. A program goal was that the IFPC meet or exceed the reliability and maintainability of the basic F-15. The authors describe the architecture and redundancy management of the IFPC and the associated aircraft system modifications that created a robust, fault-tolerant, and maintainable flight/propulsion control system. It is concluded that the techniques used to provide the required failure management characteristics for the S/MTD resulted in a system which could withstand a significant number of failures in various parts of the system without loss of aircraft control.< >
The supportability investment decision analysis center (SIDAC) is an initiative jointly sponsored by the Air Force logistics command (AFLC) and the Air Force systems command (AFSC). This initiative was conceived to help improve the full spectrum of weapon system supportability, beginning with the earliest stages of technology development and progressing through the final phases of life cycle. SIDAC's interest in enhancing weapon system supportability is focused on six key services: analysis methods, models and simulations, data access and processing techniques, communication networks, information retrieval and repository, and special studies and tasks. Each of these services is studied as part of the SIDAC concept definition program. A case study conducted during the concept definition phase is included
The results of a study of the sensitivity of digital electronic warfare (EW) receivers are reported. It is demonstrated with actual digitized data that the sensitivity of a digital receiver can be improved if enough signal processing is applied. The approach presented can be considered as a brute force one. This approach does not provide the best sensitivity that can be accomplished, but it will provide an improvement over conventional EW receivers. In this approach, the digital Fourier transform (DFT) is used to find the frequencies of the input signals. Several DFTs with fixed lengths are used. This approach will approximate the performance of a superheterodyne receiver with a fixed bandwidth, and is equivalent to several channelized receivers, each having a different resolution bandwidth. The technique improves the sensitivity of the receiver with reasonable complexity
An attempt was made to determine if artificial neural networks (ANNs) can be trained to classify the correlation signatures of direct-sequence and frequency-hopped spread-spectrum signals. Secondary goals were to determine if network classification performance can be modeled with a conditional probability matrix; if the symmetry of the matrices can be controlled; and if using a majority vote rule over independently trained networks improves classification performance. Correlation signatures of the spread-spectrum signals were obtained from US Army Harry Diamond Laboratories. The signatures were preprocessed and separated into various training and testing data sets. Thirty samples of network responses for several sets of training conditions were gathered using a neural network simulator. ANNs trained directly on correlation signature data yielded classification accuracies on test data at or near 80%. The probability matrices were stationary with regard to test sets, and the ability to shift the symmetry of the matrices was demonstrated. Improvement of classification accuracy via majority vote was possible if the nets were trained on different data sets. An average improvement of 1.8% was found to be statistically significant for α=0.05. A metric was developed to estimate the similarity of the solutions found by networks in a given training run
