Abstract: A novel microelectromechanical system artificial neural network (MEMS ANN) device performs the function of a conventional artificial neural network node element. Micro-machined polysilicon or high aspect ratio composite beam micro-resonators replace as computational elements the silicon transistors and software simulations of prior-art ANNs. The basic MEMSANN device forms a non-linear (e.g., sigmoid) function of a sum of products. Products of the magnitudes of sine waves, applied to the input drive comb and shuttle magnitudes, are formed in the frequency domain and summed by coupling a plurality of resonators with a mechanical coupling frame, or by integrating them into one resonator. A sigmoid function is applied to the sum of products by shaping the overlap capacitance of the output comb fingers of the resonator. Methods of building and using various single MEMS ANN devices and multi-layered arrays of MEMS ANN circuits are also described.

Abstract: An Aircraft Ground Power Unit (“AGPU”) is a mobile vehicle, preferably self-propelled, on which is mounted an enclosure. Within the enclosure are a number of subsystems, each of which provides a service required by a stationary aircraft for servicing and diagnostic testing prior to flight. Such services include electrical power, hydraulic power, engine-start capability (bleed-air), conditioned air (hot or cold) for the aircraft's equipment recesses or interior, and nitrogen to inflate struts and operate power tools. The control, status, configuration, and automation of all of the subsystems contained within the enclosure are consolidated under computer control. The enclosure and the subsystems contained therein can be readily mounted on and dismounted from the vehicle.

Abstract: A device for tracking a light pointer on a display. The device processes analog interlaced field video images from a video camera, seeking the point of brightest intensity, which is the current location of the pointer on the display. Synchronously operating counters and a latch count the lines and pixels of the video data until a high-speed comparator detects the brightest point of the video data. The comparator then sends a detect signal, causing control logic to freeze the counters, as well as freezing the latch, effectively storing in the counters and latch the line and pixel location of the pointer. The same detect signal from the comparator instructs a microcontroller to output the data from the counters in computer readable, digital data format. The device operates in substantially real time in either two or three-dimensional space, with greater speed and significantly reduced cost over the prior art, and an adjustable resolution.

Abstract: Apparatus and method to reduce the duration of timeout periods in fault-tolerant distributed computer systems. When nodes execute a task redundantly and communicate their results over a network for further processing, it is customary to calculate timeouts on a worst-case basis, thereby prolonging their duration unnecessarily. By applying Tchebychev's inequality, which holds for any statistical distribution, to adaptively determine the distribution of the arrival times of the results at the point where further processing of those results takes place, the duration of timeouts is reduced. Successively refining the statistical distribution of the arrival times leads to an improved forecast of future arrivals. Thus timeouts are kept to a minimum without compromising the reliability of the system.

Abstract: The present invention demodulates a FM, WM or AM optical input beam, amplifies the signal impressed on the beam, and then modulates without intervening electronics the signal amplitude of an output beam. The apparatus can be made from any semiconductor laser, including edge-emitting lasers and VCSELs. Light transmitted through the waveguide (the “control beam”) interacts with the carrier population of the laser, reducing the available gain and thereby the output intensity. The present invention has three key advantages: (i) a FM control beam produces an amplitude-modulated output beam from the main-laser, (ii) the amplitude of the AM beam depends directly on the differential gain ∂G/∂&ohgr;, and (iii) the apparatus and method provides gain for all LOGiC devices. The present invention works with available and future semiconductor lasers, including those emitting in the blue as well as those operating at 670, 850, 980, 1300, and 1500 nm.

Abstract: Apparatus and method for high-speed analog-to-digital conversion employs photonic circuits with variable light-absorption. A coherent light source is clocked to produce a stream of coherent light pulses onto which is modulated an analog signal to be converted to digital form. An optical splitter divides the stream of pulses into a number of streams that is proportional to the resolution sought. A passive photonic quantizer absorbs, to a predetermined degree, the light energy of the incident pulse streams. The output of a fully absorbed pulse does not trigger a comparator. Pulse streams of sufficient intensity pass through the light absorbers, and their output does trigger the comparator. The output state of the comparators may be read into a digital memory to form a digital word that represents the analog signal during any discrete clock interval. Most of the elements of the apparatus can be fabricated on a substrate of photonic semiconducting material.

Abstract: Apparatus and method to measure the characteristics of the substantially periodic motion of a mechanical system. A non-intrusive technique measures such motion by analyzing images of a movable component of the system without requiring a motion sensor built into the system. If the system includes a motion sensor, the present invention can calibrate it. The present invention applies to an object of any length scale if it can be imaged. The amplitude of a component's motion is obtained from a single, time-exposed image while the system is in periodic motion and a reference image made with the component at rest. The technique, implemented as one component of an automated test-bed apparatus, is significantly faster than the prior art. The present invention is especially efficient in characterizing the mechanical performance of MEMS. Speed facilitates both hands-on testing of prototypes and testing in production environments.

Abstract: Apparatus and method to detect corrosion in metal junctions. Corroded metal junctions are usually discovered by visual inspection. The present invention detects corrosion in metal junctions when it is not visually apparent. A corroded metal junction acts as a nonlinear device. It generates harmonics and other nonlinear products (such as intermodulation) of any signals applied to the junction. The presence of relatively high level harmonics and/or intermodulation products indicates directly that corrosion has occurred. To detect corrosion in a metal junction, one couples a fundamental frequency signal (f0) into the junction and tests for harmonics of that frequency, especially the third harmonic. Harmonic frequency signals that are relatively large (i.e., above the harmonics generated by the testing system) indicate the presence of corrosion. Measurements to determine if a metal junction is corroded are performed without disturbing the junction.

Abstract: Apparatus and method to identify chafing of a conduit, thereby reducing the failure of any system which would be damaged or whose function would be impaired by abrasion of the conduit. Such a system may carry electrical power, fuel, other fluid, hydraulics, pneumatics, optical, or electromagnetic signals. Wear caused by rubbing against external structures is detected by wrapping the conduit with a sensing element, which may be a conductive wire, waveguide, fiber optic cable, or a tube (wound around the conduit or enclosing it) that holds a fluid under pressure. The sensing element is positioned so that chafing on the conduit electrically contacts, breaks, or punctures the sensing element well before the conduit fails.

Abstract: A single Printed Circuit Board (PCB) designed to acquire data from a multiplicity of heterogeneous sources and convert the data to a high performance protocol suitable for transmission over long distances via fiber optic lines. A specific embodiment uses the Fiber Channel protocol on fiber optic cables to carry information between sensors and high performance computers (HPC). The High Performance Parallel Interface (HiPPI) is used as the protocol to connect to the HPC. Simplex (unidirectional) and full duplex communications are supported.

Abstract: Apparatus and method for improving the detection of signals obscured by either correlated Gaussian or non-Gaussian noise plus additive white Gaussian noise using Estimates from multi-channel data of model parameters that describe the noise disturbance correlation are obtained from data that contain signal-free data vectors, referred to as “secondary” or “reference” cell data. These parameters form the coefficients of a multi-channel whitening filter. A data vector to be tested for the presence of a signal passes through the multi-channel whitening filter. The filter's output is then processed to form a test statistic. The test statistic is compared to a threshold value to decide whether a signal is “present” or “absent”. Embodiments of the apparatus and method include estimating the signal amplitude both implicitly and explicitly and calculating test statistics for signal detection in both Gaussian and non-Gaussian noise.

Abstract: Spatial-temporal waveform diversity methods vary or modulate the far-field radiated waveform of a radar and/or communication antenna as a function of look direction (sidelobe structure). Radar detection performance and angle of arrival estimation are enhanced to deny a coherent reference to non-cooperative bistatic radars and coherent repeater jammer systems. In the most general case, the spatial-temporal modulated waveform varies as a function of angle, based upon the principles of multi-dimensional Fourier synthesis. Spatial-temporal denial is achieved with as few as two auxiliary antennas bracketing a main antenna. The same methods for spatial-temporal waveform diversity can also embed communications signals into the transmitted radar waveform for one-way simulcast of both waveform types. These same directionally dependent simulcast waveforms can incorporate navigation signals for enhanced precision engagement.

Abstract: Variable locations on a suitably coated light reactive semiconductor sheet can be illuminated by a pattern of diffracted light to form discrete conductive pathways between antenna radiating elements and an antenna groundplane. Varying the diffracted light pattern temporally and/or spatially changes the conductive pathways and the antenna's beam pattern. Similar variations modify the characteristics of an antenna's radiating element or reflective groundplane, thereby providing frequency control or limited directional control of the beam pattern. Methods for controlling the diffracted light permit an antenna beam pattern to form, redirect, and scan rapidly.

Abstract: A method and system for power management of components within a system by introducing low-level instructions, such as NOPS, delay loops, and sleep modes, into the instruction sequences within logical components of the system, thereby allowing the system to maintain desirable power dissipation over a given time interval. Through employment of a control mechanism, a system can interrupt itself periodically to determine if the power dissipation must be limited and to what extent. A system can determine if and by how much it must limit its power by polling external components. The method may be implemented by a controller that intelligently manages power dissipation of components through selective manipulation of computer operations by selected components. Controll of power dissipation for the selected components is intelligently based on overall system performance requirements.

Abstract: A digital image is "watermarked", that is, authenticated by an embedded pattern. The pattern is created by hashing the image and adding a signature element. Manipulating this result by the seed for a random number generator leads to an initial two dimensional random black-and-white pattern. This pattern is manipulated by a cellular automaton and smoothed before being added to the original image. To determine whether the image is authentic, one retrieves the watermark by subtracting the watermarked image from the original to obtain the difference. The value of the correlation between the difference thus obtained and the smoothed pattern determines the presence or absence of the watermark.

Abstract: Apparatus and method to improve the speed of electronic file transfer between remote computers by parallel processing. The most common transfer protocol is the File Transfer Protocol (FTP) Though the Internet is the most popular means by which users electronically transport data, the Internet's Transport Control Protocol/Internet Protocol ("TCP/IP") model, upon which FTP operates, does not adequately support the transfer of large data sets over long distances. Typical transfers between distant sites linked via the Internet have throughputs of about 20 Kilobytes per second ("Kb/sec") or less, and they are prone to dropping packets or losing connections. Making FTP parallel makes transfers up to five times faster, that is, it offers transfer rates of up to 100 Kb/sec. FTP is operated in parallel by segmenting a file into discrete packets, simultaneously transmitting these packets to the receiving computer, and reassembling the packets into the original file.

Abstract: Chip-like stacks of thinned chips are mounted in wells etched into a substrate. A "chip-like" stack is a stack of chips, which in the aggregate have a height approximately equal to that of a single conventional chip. These chip-like stacks are mounted in a variety of packages. In a preferred embodiment, the stacks are mounted in wells within the substrate of an integrated circuit and the stack is provided with a patterned overlay so that all the circuit connections can be made from the upper surface of the stack. The patterned overlay is protected by a planar insulator. A plurality of substrates may be stacked, one upon the other.

Abstract: A compact, continuous-wave blue laser is developed from a fiber made from heavy metal fluorides ("ZBLAN") doped with a rare-earth ion. The footprint required to create blue laser light is reduced because the fiber can be wound into spools of radius <25 mm and stacked one atop the other without cross talk. IR diodes (.lambda..about.790 nm and .lambda..about.1050 nm) are fiber-pigtailed to silica fiber in a conventional way. The light from the IR diodes is coupled to a single fiber through a 2.times.1 fiber coupler that has silica inputs and a ZBLAN output. The IR light optically excites the electrons of the rare-earth ions in the ZBLAN fiber host. This excitation causes the electrons to emit light at 480 nm (in the blue region of the visible spectrum) as they relax to the ground state. Dielectric mirrors feed back the emitted light. A high-reflector, high-transmitter ("HRHT") is the input coupler of the pumping light; a partial reflector, the output coupler.

Abstract: Apparatus and method for estimating the angle-modulation imposed on a transmitted Radio Frequency (RF) or Intermediate Frequency (IF) carrier. The method estimates angle-modulation when communications channels add distortions such as noise to transmitted signals. The system provides reconstituted frequency modulation from a feedforward demodulator. The goal is to reduce the modulation index of a desired signal and to employ a narrower band-pass filter that passes this signal while rejecting the distortion that accompanies it. A plurality of stages, 1 through M, exist within the system. Stages 2 through M contain the same components, although filter coefficients and alignment delays may differ from stage to stage. Stage 1 of the demodulator differs slightly from the remaining stages, since the only input required by Stage 1 is a complex envelope signal that contains both the desired signal and the distortion added by the channel.

Abstract: An imaging spectrometer provides a three-dimensional (two (2) spatial and one (1) spectral) image cube of a target. In this apparatus, a diffractive optical element (DOE) performs the imaging and provides the dispersion necessary to separate a multi-spectral target into separate spectral images that are detected by a photodetector array. A lens (or set of lenses) relays the image formed by the DOE to the photodetector array. This lens allows the DOE and photodetector array to be mounted with a fixed separation and the spectral images to have a constant magnification. The lens is stepped or scanned along the optical axis, and each position of the lens corresponds to a particular wavelength being imaged onto the photodetector array. At each position of the relay lens the photodetector array records a spectral image by a process called diffractive spectral sectioning.