Performance of Digital Implementations of an Adaptive Processor

The performance of a digital implementation of an Applebaum-Howells type adaptive processor is analyzed for both a limiter and nonlimiter configuration. The performance is evaluated in terms of steady-state residue power, using either a single-pole filter or a perfect integrator to smooth the output of the correlation mixer. The latter filter is the more commonly used for digital implementations. It is shown that when using the perfect integrator filter for both the limiter and linear digital implementations, the steady-state average weight vector equals the optimum weight vector. Thus, for this filter, the steady-state residue power is the minimum possible for either implementation. When using the single-pole filter, neither implementation achieves the minimum possible steady-state residue power. The relative performance of the two implementations depends upon the relative gain settings. When the gains are adjusted to give comparable servo stability for the design maximum jammer power, a reasonable criterion for digital implementations because of analog to digital saturation, the limiter configuration always has smaller steady-state residue power.     

Stable mid-latitude red arcs: Observations and theory

Summary The observational features of the arc are fairly well established. At present, the thermal conduction model appears to explain the red arc features most consistently, but it must be noted that a soft electron flux would give very similar results. Ion temperature measurements in the vicinity of an arc, which should be forthcoming in the very near future, can establish conclusively whether transverse electric fields play any important role in the formation of the arcs. Accepting the assumption that the arcs are the result of energy flowing down from the plasmasphere, the major remaining question is: where does the energy come from and how does it get into the plasmasphere? The various proposed mechanisms discussed in the previous chapter appear feasible, but much work needs to be done before this problem is completely resolved.On leave from the Department of Electrical Engineering, The University of Michigan, Ann Arbor.The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

A Note on Dynamic Range of Quantizers

Based on the relationship between the inputand output correlation functions for a symmetric2N step optimum quantizer withGaussian noise inputs a computer programwas generated to obtain plots of the inputnumber of levels versus dynamic range of thequantizer. For it to be physically useful thedynamic range should be related to the distortionin the fundamental band and thedesired harmonic suppression. A definitionbased upon these considerations is suggestedand curves are plotted for two different fundamentaldistortion levels and three differentlevels of harmonic suppression. It is foundthat, for low levels of harmonic suppression(0-20 dB range), the dynamic range can beincreased by tolerating a higher level of fundamentaldistortion. In the medium range ofharmonic suppression (20-35 dB), any increasein the dynamic range due to higherfundamental distortion levels disappearsfor high level quantizers (typically 4 bitsor more). For still higher harmonic suppression(40 dB or higher), the dynamic range ofthe quantizer is independent of the acceptablefundamental distortion.     

One-kVA, Three-Phase dc-ac Inverter with Digital Control

This paper describes a method of applying digital techniques to the control of a 1-kVA three-phase dc-ac inverter to generate a sinusoidal 400-Hz output, using high-frequency bridge-chopper techniques. The model which was constructed used predominantly off-theshelf digital microcircuits and resulted in a device with an overall efficiency of 85 percent, in a 0.52 cubic-foot package which weighed 19.5 lbs. Sinusoidal output with less than 2 percent harmonic distortion at 115 volts line to neutral was obtained with 28 volts dc input.     

A Performance Measure for Evaluating Aircraft Landing Trajectories

A general performance index is developed for evaluating aircraft landing trajectories. The primary term in the index is the effect of noise on people residing near the air terminal. Other terms included are passenger comfort, fuel consumed, and the time spent in the near-terminal area. Models are developed for aircraft engine noise, passenger comfort, the population distribution about a specific airport, and the aircraft flight behavior. While this performance index may be used in computing optimal trajectories, it is also useful for comparing nonoptimal trajectories which, for one reason or another, may be worthy of consideration. Some examples of such comparisons are included through simulations of landing. The aircraft considered is a Boeing 737.     

The Electric Antennas for the STEREO/WAVES Experiment

The STEREO/WAVES experiment is designed to measure the electric component of radio emission from interplanetary radio bursts and in situ plasma waves and fluctuations in the solar wind. Interplanetary radio bursts are generated from electron beams at interplanetary shocks and solar flares and are observed from near the Sun to 1 AU, corresponding to frequencies of approximately 16 MHz to 10 kHz. In situ plasma waves occur in a range of wavelengths larger than the Debye length in the solar wind plasma λ _D ≈10 m and appear Doppler-shifted into the frequency regime down to a fraction of a Hertz. These phenomena are measured by STEREO/WAVES with a set of three orthogonal electric monopole antennas. This paper describes the electrical and mechanical design of the antenna system and discusses efforts to model the antenna pattern and response and methods for in-flight calibration.     

Missions to Mercury

Mercury is a very difficult planet to observe from the Earth, and space missions that target Mercury are essential for a comprehensive understanding of the planet. At the same time, it is also difficult to orbit because it is deep inside the Sun’s gravitational well. Only one mission has visited Mercury; that was Mariner 10 in the 1970s. This paper provides a brief history of Mariner 10 and the numerous imaginative but unsuccessful mission proposals since the 1970s for another Mercury mission. In the late 1990s, two missions—MESSENGER and BepiColombo—received the go-ahead; MESSENGER is on its way to its first encounter with Mercury in January 2008. The history, scientific objectives, mission designs, and payloads of both these missions are described in detail.     

A method for estimating parameters of K-distributed clutter

A method for estimating the parameters of K-distributed clutter when the available sample size of the data is limited is proposed. In this method, the arithmetic mean and geometric mean of the given data are used to estimate the model parameters. Expressions characterizing the performance of the proposed estimator are presented, along with some simulation results. For spiky clutter, simulations show that parameter estimates obtained from the arithmetic and geometric mean are approximately equal to the numerically evaluated maximum-likelihood (ML) estimates. The method is also used to estimate the parameter of the Weibull density     

Chaotic zero dynamics in kinematically redundant robots

The authors attempt to integrate two areas in robotics that are maturing rapidly: the control of kinematically redundant rigid robots and feedback linearization. Kinematically redundant robots have several advantages over nonredundant robot arms. The extra degrees of freedom are used to reach around or between obstacles. Unlike nonredundant rigid link robots that do not exhibit zero dynamics, the self motions of kinematically redundant robots can be used for torque optimization or avoiding obstacles. It is shown that induced self motions can be chaotic. The main contribution is to demonstrate chaos in cases of feedback linearization when the decoupling matrix is nonsquare and there are less outputs than inputs     

An optimum filter for direct A/D conversion

The author describes a design procedure to find the optimum finite-impulse-response (FIR) filter coefficients when the characteristics of the bandpass filter preceding the analog-to-digital (A/D) converter are known. A simulation of direct A/D conversion performance shows the improvement over the filter coefficients designed to minimize the maximum error over a specified band of frequencies. DC cancellation may be added to the process with very little change in the coefficients of the optimum FIR filter or its performance.<>