Abstract: An RLG OPTICAL NOISE INJECTOR places a random signal on a PZT driving one mirror controlling the optical path length (OPL) of a ring laser gyro (RLG) and the negative of the signal on a PZT driving a second mirror controlling the OPL. The OPL is therefore maintained constant while injecting random noise into the phase of the counterpropagating beams to avoid dynamic lock-in. Static lock-in may also be avoided by superposing an oscillatory signal on the random signal fed to the first PZT and the negative of the oscillatory signal on the random signal fed to the second PZT.

Abstract: A first-order filter apparatus 48 (FIG. 4) includes an integrate-and-dump (I&D) circuit 50 and an output loop 52. The I&D circuit 50 is driven by a cyclic scaling element 54, which multiplies N consecutive input signals 56, x(m+1) to x(m+N), by a cycle of N scaling factors 58, c(1) to c(N). The I&D summer 60 drives a double-throw switch 62. The double-throw switch 62 applies the output of the I&D summer 60 to an I&D delay element 64 for N-1 input clock cycles, and for an Nth input clock cycle to the input of a non-cyclic scaling element 66. The non-cyclic scaling element 66 scales its input by a non-cyclic scaling factor, and applies its output to one input of a feedback summer 68. The output of the feedback summer 68 is y(k), the output of the apparatus. The index k advances by one every time that the index m advances by N. The output of the feedback summer 68 is also applied to an output feedback multiplying element 70, which multiplies it by B.sup.N.

Abstract: This invention relates to fiber optic networks for the two way communication of light signals. One optic fiber may carry bidirectional signals to many nodes connected to the optic fiber by bidirectional passive optic couplers. Passive unidirectional to bidirectional passive optic couplers can bridge between bidirectional optic fiber networks and unidirectional fiber optic networks. Thus optical signals can be transported to and from each node to all other nodes such that each node is in contact with all the other nodes on the network.

Abstract: Planar antennas 10a, 10b require impedance matching with their associated transceivers 20a, 20b. Conventionally, an inductance coil is placed between the transceiver 20a, 20b and the antenna 10. Such coils add loss, require space within the transciever, and increase costs. This invention replaces the inductance coil with a planar transmission line 18a, 18b within the planar antenna 10a, 10b, such as a co-planar line, slotline, or microstrip line. If desired, active circuits 30 may be applied across the transmission line 18a, 18b, with an RF choke 42 being used to allow a dc bias to drive the active circuits 30 while preventing interference with RF operation.

Abstract: A vehicle radar emits four groups of single-frequency stepped radar pulses. In group A, the frequency of each pulse is a fixed amount higher than that of the preceding pulse. In group B, the frequency of each pulse is a fixed amount lower than that of the preceding pulse. In group C, the frequency of each pulse is the same as that of the preceding pulse. In group D, the frequency of each pulse depends on a modulo algorithm. The signals reflected from other vehicles may readily be processed with inexpensive equipment to discriminate among such other vehicles, and to determine the distance to, and relative speed of, each such vehicle.

Abstract: The in-phase channel 28 of a complex demodulated resonator data output signal 12 should contain all of the sensed information, and the quadrature-phase channel 32 should contain none of it. This will not happen if the phase of the reference signal 14 is incorrect. The phase may be adjusted by first filtering each demodulated channel with a respective low-pass dc-blocked filter 34, 38 which passes only the frequencies of the sensed information. If the sensed information gets through on both channels, then there will be a non-zero cross-correlation between the channels. This cross-correlation can be servoed to a minimum by use of a feedback signal 22. Doing so will cause all of the sensed information to be in one channel 28, and diagnostic information to be in the other channel 32.

Abstract: Noise may be reduced or eliminated from a digital sawtooth signal representing the phase of a periodic signal. This may be done precisely, even when inexpensive fixed-point arithmetic is used. In one aspect of the invention, the input signal (noise plus true signal) 12 is filtered to produce, in succession: (a) mod one differentiated noise plus slope of true phase signal 28; (b) mod one differentiated noise plus slope of residual phase signal (true phase signal minus estimated slope of true phase signal) 36; (c) mod one differentiated noise 46; (d) estimated noise 62; and (e) smoothed phase signal 72. In a second aspect, a noisy phase signal 12 is extracted from a first arbitrary periodic signal and the above steps are used to generate a noise-reduced phase signal 72. The noise-reduced phase signal 72 is then used to generate a second arbitrary periodic signal of the same frequency.

Abstract: A focal plane array (FPA) including a layered composite structure having several layers of materials with differing thermal expansion coefficients (TECs) and thicknesses to stabilize and maintain FPA performance during thermal cycling and excursions. An optical substrate is coupled to a multiplexer through a network of indium bumps. The typical TEC mismatch of the layered materials is offset by particularly selecting the types of materials used and the thicknesses of the individual layers of the FPA. Consequently, undesirable deflections and distortion of the multiplexer and the indium bumps are minimized, thereby substantially improving FPA reliability.

Abstract: The inventive laser radar uses a multiple frequency processing technique for the purpose of extracting relative range information from a received signal by transmitting and receiving two frequencies that are derived from a reference signal source modulated by two modulators that are driven by a different frequency, the reference frequency source also serving as a local oscillator.

Abstract: A radar digital IF (intermediate-frequency) signal is to be complex demodulated, then processed through a two-channel real lowpass filter, prior to Doppler correction, clutter filtering, and other usual processing steps. A particularly efficient form of lowpass filtering for such applications is the so-called CIC (cascaded integrator/comb) filter. The first stage of such a filter is always an integrator. A conventional mechanization would therefore feature a conventional complex demodulator followed by an accumulator on each of the two demodulator outputs, followed by more processing. This invention more efficiently mechanizes this function by replacing the usual pair of multipliers and pair of accumulators (or integrators) with a single two-delay (54, 58) accumulator having an add/subtract control (48) on the input adder (42), and a pair of multiplexer switches (62, 66).

Abstract: A cylindrical microwave cavity 12 has a pinched center 28, 30. Two apertures 28, 30, one in the center. of each face of the cylinder 10, provide access for a gas/powder mixture 124 having a dielectric coefficient depending on its density or some other measurement of interest. The faces 20, 22 of the cylinder 10 are much farther apart at the unpinched periphery, which comprises most of the volume of the cavity. The wider spacing allows an antenna 34 to be placed in the cavity 12, and provides a high Q-factor to maximize the sensitivity of the cavity 12. At the same time, certain resonance modes are greatly affected by a change in the dielectric coefficient of the gas/powder mixture 124 in the pinched portion 28, 30. The density of the powder or any other convenient variable may thus be servoed to any desired value. The powder 124 may be blown through the cavity 10, encased in tube 120 affixed to the cavity 12.

Abstract: The blade flash signature in a radar return is enhanced and used to detect and classify helicopter targets. The radar return of a target is fast Fourier transformed 34, and the frequencies corresponding to DC and the body of the target are filtered out 38. The filtered signal is inverse transformed 44, moving averaged 52, and auto-correlated 54. Auto-correlation 54 may take place in the frequency domain. If there are significant peaks in the auto-correlation output (other than at zero time delay), then a helicopter is present. The time delay between peaks is the period of the blade. The height of the peak indicates whether it is the main blade or the tail blade which has been detected, and allows a signal indicating both periods to be output. The main blade period and the ratio of the period to that of the tail blade can be compared with a data base to classify the helicopter.

Abstract: A radar lens 14 is made from a conventional Fresnel lens 10, but replaces the conventional curved surface 32 with a stepped approximation thereto 22, 24, 26, preferably of three steps. The thickness of the stepped lens 14, at each step, is a half-wavelength or a multiple half-wavelength of the radar operating frequency in the medium of the lens 14. The half-wavelength or multiple half-wavelength separation of the steps 22, 24, 26 causes reflections from the front 16 and rear 18 surfaces to cancel, thereby minimizing the (undesirable) standing wave between the lens 14 and the feed horn or feed horns 46, 48, 50. This avoids the necessity of reducing the standing wave by presenting the curved or stepped side 18 forward. The planar side 16 of the lens 14 (unlike the stepped side 18) doesn't need to be protected from road debris. The lens 14 can therefore be molded as an integral unit of a radome, desirable in the automotive setting.

Abstract: Sparse sampling of an analog signal is desirable because it allows a large number of analog signals to be monitored using a minimum of electronics. The analog signals may be multiplexed and the multiplexed signal applied to a single analog-to-digital converter (ADC) driving a single digital processor. However, such sparse sampling undesirably lengthens the time it takes to get an accurate measurement of the signal. This delay is minimized by Hilbert transforming the digitized signal 16 to produce an in-phase signal I and a quadrature-phase signal Q, and by measuring I.sup.2 +Q.sup.2. I.sup.2 +Q.sup.2 is the square of the amplitude of the input signal. The input signal 12 is sampled at about four times the signal's frequency. DC bias is eliminated in a two-delay delay-subtracter dc blocker 18B. The de-blocked signal 20 is applied to a two-delay Hilbert transformer 22. Only five samples are required to accurately measure the amplitude of the input signal 12.

Abstract: The seeker of a guided missile may be tested by a simulator. The simulator mimics the reflection of a pulsed laser beam from a target. The simulator has a low power IR emitting LED 18. This IR emitting LED 18 is pulsed by an energy storing capacitor 22. The pulse is controlled by an SCR 16 driven by a pulse repetition oscillator 12. A collimating lens 26 collects light from the IR emitting LED 18 and directs it to the seeker. When being self tested, IR from the IR emitting LED 18 is reflected by a mirror 40 onto a detector 24. The output of the detector 24 is passed through a high pass filter 30. This removes signals from IR noise sources (fluorescent bulbs, the sun, incandescent bulbs), and passes only signals from the IR emitting LED 18. A first buffer 32, preferably an amplifier, provides power between the high pass filter 30 and the gate of a second SCR 34. This second SCR 34 drives a visible light LED 36 through a second buffer 38. The visible light LED 36 acts as a positive self test indication.

Abstract: A vehicle navigation system uses a one-multiplier Gray-Markel filter. The sign parameter of each stage of the filter is selected by an algorithm which limits the maximum signal passing through the filter, thereby preventing overflow.

Abstract: The STREAMER FOR RISC DIGITAL SIGNAL PROCESSOR shown herein allows a CPU 46 to interface with a memory 60 via data registers 50. Pre-fetch and post-store of the correct address is determined by an address generator 58 according to a rule determined by a context register 52. An index indicative of this address is stored in an index register 54. The data, context, and index registers together form a streamer 56, streaming data between the CPU 46 and data memory 60. The rule of the context register 52 also drives a converter 62 for converting data between memory format and register format. The speed and flexibility of a RISC device is combined with the intensive memory access of a digital signal processor.

Abstract: An accurate estimate of the amplitude of a sparsely sampled sinusoidal signal is obtained by filtering the squares of the sampled values of the sinusoidal signal with an adjustable notch filter in order to remove a double-frequency component at twice the frequency of the sinusoidal signal. This amplitude estimate, for example, is used for automatic gain control of the amplitude of the sinusoidal signal. Preferably, the notch filter is a digital filter for computing an output signal (v) from successive samples (x.sub.n, x.sub.n-1, x.sub.n-2) of an input signal (x) according to: v.sub.n =x.sub.n -.beta.x.sub.n-1 +x.sub.n-2. The frequency control parameter (.beta.) is computed, so as to automatically track the frequency of the sinusoidal signal, by integrating a product of a derivative (x.sub.n-1 -x.sub.n-2) of the input signal (x) and a derivative (v.sub.n -v.sub.n-1) of the filtered signal (v).

Abstract: For use with a quartz angular rate sensor, a frequency and phase-locked synthesizer recovers a reference signal virtually free of phase noise, and generates a quadrature-phase reference signal for complex demodulation of the angular rate signal. The synthesizer also ensures a precisely adjusted phase shift of approximately zero across the drive tines of the sensor. Moreover, the digital synthesizer provides a precise numerical indication of the drive frequency, which can be used for compensation and automatic tuning of filters, such as a tracking filter, a filter in an automatic gain control, and notch filters in the phase and/or frequency detectors in the digital synthesizer. The tracking filter is used as a pre-filter for the synthesizer, and is responsive to a passband-width control signal generated from the magnitude of the frequency and phase error signal controlling the frequency generated by the synthesizer.

Abstract: A single bit bandpass analog-to-digital converter has an analog summer, an analog bandpass filter, a single bit quantizer, and a single bit digital-to-analog converter connected in a loop. An input signal to the single bit filter is applied to a plus input terminal of the summer, and the output of the digital-to-analog converter is applied to a minus input terminal of the summer. The output signal from the single bit filter is taken from the output of the quantizer. The bandpass filter is preferably driven by a digital clock running at the same frequency as the quantizer and the digital-to-analog converter. This architecture reduces quantization noise within the passband at the possible expense of increasing it outside the passband. The passband is centered precisely on one-quarter of the clock frequency.