Abstract: A METHOD AND APPARATUS FOR CONTROLLING INTEGRATION TIME ON MULTIPLEXING STARING ARRAYS for selecting all or any portion of each row integration scan for charging integration capacitors from a scanned staring array by providing a row ON selection clock signal, providing a signal corresponding to all or a portion of the time the row is to be connected to the capacitors, and nanding the row ON signal and the corresponding signal to select the time allotted for charging said capacitors.

Abstract: A Lofargram Enhancement System having a processor having 1 to i rows and 1 to j columns of processing elements or data cells. Each data cell has a location corresponding to an index, a computational capability and a memory means for storing a a pixel intensity value. The memory cells of all processing elements are addressed concurrently. A FIFO receives and stores a scan line comprising a sequence of pixel intensity signal values. A program provides a means for calculating row and column total values for the total of the pixel intensity signal values stored in a predetrmined relationally related location for each respective processing element for each respective processing element in the systolic processor. An auxiliary processor reads the respective row total and column totals for each processing element in the jth column of the systolic processor and loads the corresponding first column processing element with a corrected pixel value.

Abstract: The invention comprises a gallium arsenide driver with float capability for logic high and logic low signals including logic means for receiving an input signal and providing logic high and low signals therefrom. Level shift means responsive to the high and low signals downshift the voltage levels thereof. Common gate means responsive to the downshifted signal voltage levels amplify these signals, and output source-follower means responsive to the amplified signals provide output signals to a pad.

Abstract: An alarm for detecting transient senic events determines an estimate of background environmental acoustic events or noise by computing a short time correlation function for individual segments of data and then subtracting an estimated mean of the correlation function. Acoustical events which represent major excursions from the measured background are detected as transient events. The background is continuously updated within selected processing periods chosen to maintain an accurate differentiation for transient events. In further aspects, the residual correlation function is generally detected using a .chi..sup.2 statistic which is computed using an estimated covariance of the correlation function and the residual correlation function for a specific window. If a given residual does not represent a detected transient event, that residual is used to estimate the mean and covariance of the background using an exponential integrator.

Abstract: A SOUND RECOGNITION PROCESS comprises training a sound recognition system to a shape Hidden Markov Model (HMM) from a shape codebook of shape codewords, and to a separate gain HMM from a gain codebook of gain codewords. An unknown token is assigned to the sound category associated with the shape HMM and gain HMM with the greatest probability product. Shape and gain codewords replacements for incoming Linear Predictive Coded (LPC) vectors are selected by choosing the codeword with minimum distortion between the codeword and a combination of the vector and some of the preceding codewords, exponentially weighted by time.

Abstract: The present invention provides a method and apparatus for performing finite impulse response (FIR) type filtering of 2N points of real-only input data using fewer multiplication elements than traditionally required. A filter multiplexer segregates or divides the input data into even and odd data values which are transferred to the "real" and "imaginary" inputs of a fast Fourier Transformation (FFT) or DFT element which selectively performs N-point transformations according to a preselected FFT algorithm. The transformed data output by the FFT transformer is divided into even and odd transformed, frequency domain, data. The even and odd transformed data is transferred to a multiplier where it is multiplied by a preselected weighting matrix of coefficients and the DFT of the desired filter response. The resulting data is divided into even and odd transformed data functions which are recombined to provide the transform of the multiplier input data function.

Abstract: Static electric charges are removed from the proof mass 10 and support 12 of an accelerometer by covering the proof mass 10 and support 12 with a thin metallic film 36, such as chromium oxide.

Abstract: A passive uniplanar double-balanced star RF mixer, comprising a substantially planar support substrate with a conductive layer of material disposed on one side and a slotline and first, second, and third coplanar waveguides formed therein which are configured to accommodate different frequencies. The first waveguide is bifurcated on one end into two waveguide branches which form first and second waveguide terminations each physically coupled to one end of the slotline. The second waveguide is electrically coupled to the first waveguide at a location spaced apart from the branches. The third waveguide is electrically coupled to the slotline at a position located approximately equi-distant from the slotline ends. A first grounding element is connected to the grounds of the first and third waveguides, and is positioned adjacent to the first branch termination. A conductive surface is enclosed by the waveguide branches and the slotline.

Abstract: The present invention provides a highly efficient, high speed, and low power method and apparatus for extraction of a single signal from a set of frequency division multiplexed (FDM) signals. The apparatus operates on an analytical digital FDM-type group of signals to decompose the overall signal spectrum by octaves until a single signal or channel is selected. The signal selection is accomplished by transferring received FDM signals into a bandsplitting filter where the spectrum is divided into an upper and lower half-band spectrum. The upper half-band is complex shifted into the lower haIf-band and one half-band eliminated from further consideration. The selected half-band is decimated by a factor of two and recirculated through the filter where it is repetitively half-band filtered, complex shifted, and decimated until a desired single channel remains. The signal data from one processing pass is interleaved with data from previous passes to increase processing efficiency and speed.

Abstract: The invention is a Self Propelled Underwater Device With Steerable Fin Stabilizer having a unique hull of generally tubular configuration and having a substantially conical shaped rear portion. The hull comprises a plurality of hollow interlocking tubular sections with one section including the rear portion; adjacent tubular hull sections include peripherally spaced apart threaded portions respectively disposed internally and externally thereof for mating relationship upon relative circumferential motion of the adjacent sections, the adjacent sections are further configured to provide an abutting joint when fully mated with watertight sealing means between the mated sections. The vehicle includes a plurality of steerable stabilization fins mounted on the hull about the rear conical portion at spaced apart locations with hull slots for receiving the respective fins when stowed.

Abstract: Using classic Fast Fourier Transform (FFT) rules, a radix-12 FFT is composed of a first tier of 2 multiplierless radix-6 transformers followed by multiplierless radix-2 transformers, or by its transpose configuration. Complex data are represented in a 1, W.sub.3 coordinate system rather than in a classic 1,j coordinate system. The only multiplicative scaler in the complex twiddle factors is the reciprocal of the square root of 3 which appears six times and which by conversion to canonical signed digit code, can be accurately expressed by 5 adds. As a consequence the complex twiddle factor multipliers and ancillary address reduce to a total of 144 real adds required to perform the entire complex 12-point FFT.

Abstract: A laser gyroscope signal generator for increasing the resolution of the gyroscope, the signal generator being responsive to a sequence of pairs of phase shifted first and second laser gyroscope output pulses each being characterized to have a first and second laser gyroscope output pulses each being characterized to have a first and second logic state, said first and second output pulses in each pair having a leading or lagging phase relationship determined by the direction of rotation the laser gyroscope sourcing said pairs of phase shifted output pulses, the laser gyroscope output signal generator also being responsive to a system clock signal source for providing a clock signal having a pulse recurrent frequency greater than the pulse recurrent frequency of said phase shifted first and second laser gyroscope output pulses, the output signal generator comprising: a memory means responsive to said phase shifted first and second LGO (laser gyroscope output) pulses and said clock signal for transferring the lo

Abstract: A plurality of heat sinks assembled to form a rack in a lightweight housing accommodate printed circuit cards between the heat sinks and circulate cooling oil through the heat sinks to drain the test heat from the boards quickly. Each heat sink uses a porous metal foamed core, configured in zigzag shape, to provide a tortuous path for the fluid flow which extends into lateral fins comprising the side plates of the porous metal foam for receiving the cards or ground plates therefore. These plates are clamped against the fins for close contact to the cooling fluid thereby shortening the heat flow path. Brazed on aluminum top and bottom cover plates make a fluid tight connection with the porous foamed metal. The brazed on top and bottom plates are also configured in the zigzag pattern of the porous foamed metal.

Abstract: The present invention is a high efficiency bi-directional spatial power combiner for extremely high frequency signals. In one embodiment, orthogonally polarized electro-magnetic waves are used to isolate signals going into and out of a planar array of amplifiers. A dual mode horn transmits vertically polarized fields to a MMIC planar array of amplifiers. Vertically polarized array elements receive the signals, amplify them, and retransmit back to the horn using horizontally polarized array elements. An ortho-mode transducer on the horn provides isolation for the two polarized signals. The backside of the array can be used for a heat sink, and to provide access for DC biasing of the array elements. In a second embodiment, monolithic grid oscillators are used in lieu of power amplifiers with orthogonal polarization array elements. The input signal and the output signal are of the same polarization. A circulator is used in lieu of an ortho-mode transducer to isolate the input signal from the output signal.

Abstract: Method and apparatus for using deep ultraviolet excimer laser beams to dice semiconductor substrates, such as sapphire, with or without integrated circuitry, by establishing guided relative motion between the beam and the substrate to achieve ablative photodecomposition with the angle between the beam and the substrate being approximately five (5) degrees out of normal.

Abstract: The centralized bus arbitration circuit 20 shown here allows processor 12-18 access to a bus 10 for a period of time determined by either the processor 12-18 or the circuit 20, as appropriate. Indefinite access and immediate return of control to the bus arbitration circuit 20 are also provided for.

Abstract: Method for organic free radical, including aliphatic radical, transport of dopant species for precise, predetermined, and reproducible doping concentrations to control electrical properties for chemical vapor deposition grown materials.

Abstract: Extended laser life may be obtained by measuring, with an ammeter AM, the discharge current applied to a laser (either linear or a ring laser gyro RLG), measuring the intensity of the resulting light beam with an intensity monitor IM, calculating the noise with an integrated circuit IC, and maintaining the discharge current at the lowest level possible which is consistent both with maintaining the intensity above the intensity service limit and the noise below the noise service limit.

Abstract: An apparatus for housing a focal plane array includes a tubular metal stem for housing a cryostat inside of it, and a hollow conical metal housing having its narrow end brazed onto the end of the tubular metal stem. A ceramic electrical interconnection assembly is rigidly sealed onto the wide end of the housing, having embedded electrical connections for providing electrical connections between the interior and exterior of the assembly. A metal ring-shaped mounting structure is rigidly attached to the ceramic interconnection assembly. A ceramic coldplate is sealed over the other end of the rigid metal tube in the interior of the structure for holding the focal plane array. The rigidity of the structure provides a known positional relationship between the coldplate and the mounting ring, to facilitate positioning the focal plane array during dewar installation. A metal cap is sealed to the mounting ring to provide a reworkable, hermetic closure.

Abstract: The tuner for ring laser gyro dither mechanism shown herein is a disk 32 which is bolted to the top end 48 of a base mounted vane 24 of the gyro. All such base mounted vane top ends 48 are thus bolted together, thereby stiffening the spring constant of the azimuthal vibration 52 of the top ends 48, thus increasing the frequency of the dither mechanism.