Signal-to-Noise Ratio of Arbitrary Power-Series Nonlinear Amplifier

A general analysis of the effect of an arbitrary power-series nonlinear amplifier followed by a coherent mixing device on signal-to-noise ratio (SNR) is performed. An expression is derived for the improvement factor which is defined as the logarithm of the ratio of the output SNR to the input SNR. This expression is applicable to the coherent amplitude detector and phase locked loop as well as noncoherent amplifier by appropriate selections of the detection angle. Moreover, the improvement factor can be obtained for noise with an arbitrary amplitude distribution. To demonstrate the applicability of this analysis, the improvement factors of the nonlinear amplifiers such as a power-law amplifier and a power-series amplifier with positive and negative discriminations are numerically calculated for the cases where the input noise amplitude distributions are Rician and triangular.     

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.     

Clutter Covariance Matrix Form Effect on Adaptive MTI Performance

The performance of an adaptive moving target indicator (MTI), which employs a Wiener predictor by means of a transversal filter, is discussed, taking into consideration the effect of the form of the clutter covariance matrix on the MTI performance. It is emphasized that the main tap position in the transversal filter is an important factor which provides degrees of freedom in the clutter covariance matrix to improve the MTI performance. Calculation results show that by exploiting these degrees of freedom, excellent performance is feasible, in particular shorter transient response.     

A Fast Beamforming Algorithm for Large Arrays

This beamforming algorithm is written specifically for array radars in which the number of array elements K is very large compared with the number of jammers L the radar is designed to suppress. It uses a set of M noise vectors to construct a basis for the jammer component of the antenna output vectors. The component of the quiescent weight vector orthogonal to each basis vector is calculated, renormalized to unit length, and identified as the adapted weight vector. This algorithm is effective in the suppression of many types of jammers. The number of noise samples M required in the construction of the adapted weight vector is approximately equal to L. In the special case of L narrowband noise jammers, for example, a choice of M = L usually reduces the receiver output jammer power to a few dBs above the white noise background. It is permissible to have M相似文献   

A Detection Algorithm for Optical Targets in Clutter

There is active interest in the development of algorithms for detecting weak stationary optical and IR targets in a heavy opticalclutter background. Often only poor detectability of low signal-to-noise ratio (SNR) targets is achieved when the direct correlation method is used. In many cases, this is partly obviated by using detection with correlated reference scenes [1, 2].This paper uses the experimentally justified assumption that most optical clutter can be modeled as a whitened Gaussian randomprocess with a rapidly space-varying mean and a more slowlyvarying covariance [2]. With this assumption, a new constant falsealarm rate (CFAR) detector is developed as an application of the classical generalized maximum likelihood ratio test of Neyman and Pearson. The final CFAR test is a dimensionless ratio. This test exhibits the desirable property that its probability of a false alarm(PFA) is independent of the covariance matrix of the actual noiseencountered. When the underlying noise processes are complex intime, similar considerations can yield a sidelobe canceler CFARdetection criterion for radar and communications. Performance analyses based on the probability of detection (PD)versus signal-to-noise ratio for several given fixed false alarm probabilities are presented. Finally these performance curves are validated by computer simulations of the detection process which use real image data with artificially implanted signals.     

Third-Order Loop Filter Design for Acceleration-Rate Tracking

A perfect third-order loop filter design that can be implemented as a digital filter is obtained which minimizes the noiseless steady-state acceleration rate (jerk) error for a fixed loop noise bandwidth. Simulations were performed to obtain transient responses of the third-order loop plus a sample fourth-order loop under a jerk input. The results enable one to obtain a loop design that minimizes the loop noise bandwidth required for a given steady-state jerk error and thus obtain better noise jitter performance.     

The Generation of Correlated Weibull Clutter for Signal Detection Problems

A method is discussed for generating Weibull vectors with a desired correlation matrix and specified parameters. Such vectors may represent samples of a correlated clutter signal. The presented method makes use of a suitable nonlinear transformation of random Gaussian vectors with correlated components. Computational aspects of the method are also discussed.     

In memoriam Julius Bartels

Ohne Zusammenfassung

---

In memoriam Julius Bartels

     

A New Mathematical Formulation for Strapdown Inertial Navigation

A differential equation is developed for the orientation vector relating the body frame to a chosen reference frame. The time derivative of this vector is the sum of the inertially measurable angular velocity vector and of the inertially nonmeasurable noncommutativity rate vector. It is precisely this noncommutativity rate vector that causes the computational problems when numerically integrating the direction cosine matrix. The orientation vector formulation allows the noncommutativity contribution to be isolated and, therefore, treated separately and advantageously. An orientation vector mechanization is presented for a strap down inertial system. Further, an example is given of the applica tion of this formulation to a typical rigid body rotation problem.